American Journal of Engineering and Technology Research Volume 16, No. 1, 2016

Transcription

1 American Journal of Engineering and Technology Research Volume 16, No. 1, 2016 U.S. Library of Congress ISSN:

2 Editorial Board of this issue: Editor in Chief Dr. George Yang, Professor Department of Engineering Technology Missouri Western State University, St. Joseph, MO 64507, U.S.A. Editor in Charge Professor Wang, Zhengsheng Xi'an Technological University, China Associate Editors Dr. Long Qiao Assistant Professor, Department of Engineering Technology, Missouri Western State University, St. Joseph, MO 64507, U.S.A. Dr. Manik Mandal Research Associate Department of Chemistry 6 East Packer Ave Lehigh University Bethlehem, PA Dr. Chance M. Glenn, Sr. Professor and Associate Dean Office of Graduate Studies Rochester Institute of Technology Rochester, NY 14623, USA Dr. Omar Zia Professor and Director of Graduate Program Department of Electrical and Computer Engineering Technology Southern Polytechnic State University Marietta, Ga 30060, USA Dr. Seckin Gokaltun Research Scientist in Mechanical Engineering Applied Research Center Florida International University W Flagler ST, EC 2100, Miami, FL, USA Dr. Samir El-Omari Assistant Professor General Engineering University of Wisconsin, Platteville, WI 53818, USA 2

3 Dr. Allan Ayella Assistant Professor and Head Department of Natural Sciences McPherson College, McPherson, KS, USA Dr. M.M. Wang System Engineer Beth Israel Deaconess Medical Center, Harvard Medical School,. Boston, MA, USA Dr. Ahmed Nabih Zaki Rashed Professor, Electronics and Electrical Engineering, Menoufia University, Egypt Dr. Gururaj M. Neelgund Research Assistant Professor, Prairie View A&M University, Prairie View, TX, USA Dr. Rungun R Nathan Assistant Professor in the Division of Engineering, Business and Computing Penn State University - Berks, Reading, PA 19610, USA Dr. Haifa El-Sadi. Assistant professor Mechanical Engineering and Technology Wentworth Institute of Technology, Boston, MA, USA Professor Wang, Zhengsheng Xi'an Technological University, China Dr. Manvir Singh Kushwaha Research Scientist Department of Physics & Astronomy Rice University Houston, TX USA 3

4 American Journal of Engineering and Technology Research ISSN: Volume 16, No.1, 2016 Table of Contents Design Of Modify Wilson Current Mirror Circuit Based Level Shifters Using Stack Techniques 1-11 Design and Fabrication of Aluminium MeltingFurnace Using Locally Available Materials A Comparative Study Of Optimization Techniques For Capacitor Location In Electrical Distribution Systems Study of Sag by the Influence of Altitude Parameter Response Analysis of Phosphorus Removal to Operational Effects of Inoculum Concentration and Leaching Time during Biotreatment of Iron Oxide Performance Comparison Of Speech Recognition For Voice Enabling Applications- A Study The Effect Of Bit Error Rate On Hiperlan/2 Using 16 Qam, 64 Qam And 256 Qam Text Detection In Video Using Haar Wavelet Transformation And Morphological Operator Knowledge Management And Transfer In 2010 Vancouver Olympic Winter Games Implementation On Stm-16 Frame Termination Vlsi With High-Speed And Low- Power Gdi Techniques Using Sram 86-94

5 DESIGN OF MODIFY WILSON CURRENT MIRROR CIRCUIT BASED LEVEL SHIFTERS USING STACK TECHNIQUES M.Ragulkumar 1, Placement Officer of MikrosunTechnology, Namakkal, 1. Abstract Wide Range of level Shifters is fast energy efficient converters capable of converting the low voltages to high voltage. Level Shifters are mainly used to shifting their voltage from one level to other levels. Multi voltage systems utilize the advantage of level shifters. Multi voltage systems consist of low voltage as well as high voltage. Existing methods was implemented by using the LVT (Low Threshold Voltage Transistor) which produces the leakage power dissipations. Proposed method is implementing by using the forced PMOS method to reduce the leakage powers. Usually LS are inserted only while crossing low voltage domains to high voltage domain. In this paper modified the Wilson current mirror based level shifter is designed by using stack techniques. Measurement results were demonstrated by using cadence Tools. Keywords Level Shifters, Multi voltage systems, Widlar Wilson Circuits, Forced PMOS technique. I. INTRODUCTION The growing market of mobile batteries power electronic systems (e.g., cellular phone, personal digital assistant, etc.) Demand the designer of microelectronics circuits with low power dissipation [1]. More generally density, size and complexity of the chips continues to increase the difficulties in providing to adequate the cooling might either s add significant cost or limits the functionalities of the computing system which make use of those integrated circuit. In the past years, several techniques methodologies and tools for designing a low power circuits have been presented. However, only few of them have found their ways in current design flows [2]. The major three major sources of power dissipation in a CMOS circuits. These are switching short circuit power and leakage power. Switching power is mainly due to charging and discharging capacitors driven by the circuits. Short circuit powers are mainly caused by the short circuits current that arise when pairs of PMOS/NMOS transistors are conducting simultaneously [3]. Finally leakage power originated from substrate injections and sub threshold effect. One of the main reason causing from the leakage power increases is the increase of sub threshold leakage powers. When a technology feature size scales down suddenly supply voltage and threshold voltage also scale down. Sub threshold 1

6 leakage power increases exponentially as threshold voltage decreases. Stack method forced NMOS, Forced PMOS and sleepy keeper method are the some of the leakage current reduction methods [4]. II. LEVEL SHIFTERS A. CONVENTIONAL LEVEL SHIFTER A Sub threshold LSs are surveyed in their section. Conventional cross coupled LS are differentially cascade voltage switch logic (DCVSL) for raising a low voltage level, as shown in figure.1. The drive strength of NMOS transistors is enhanced to overcome the leakage of weakly conducting PMOS transistors. The operating range of CC LS depends on the transistor threshold voltages (v t ) and size; however, the operating range of CC LSs is very difficult to extend the sub threshold region (with respect to NMOS v t ) because of the NMOS drive strength decreases exponentially. For a converting sub threshold voltage, CC LS requires an exponentially increased in NMOS transistor size [5,6]. Figure.1.Conventional Level Shifter A. Current Mirror based Level Shifter Fig.2 shows conventional level shifters that use a basic current mirror circuit. The conventional current mirror level shifter can converts a deep sub threshold level because a high drain to source voltage of PMOS transistor facilitate the constructions of a stable current mirror, which offers an effective ON - OFF current comparison at the output nodes. However, a high amount of Quiescent current occurred when the input voltage is high threshold. These high power consumptions limit the use of the conventional Current Mirror Level Shifter [7]. 2

7 Figure.2. Current Mirror Based Level Shifter B. WILSON CURRENT MIRROR BASED LEVEL SHIFTER Fig.3 shown a CM - type LS that uses a Wilson Current Mirror (WCM), which clamps the quiescent power consumptions under a super threshold input. Figure.3. Wilson Current Mirror based level shifter A two stage CCLS of Pull up driving strength is reduced by their header NMOS, which expands the convertible input voltages. In shows that CC-type LS (CCPNR), which the output stage is a part of their NOR gate. The primary input is accelerating to the overall LS speed. It uses Logic Error Correction Circuit (LECC), it monitors input and output updating data during the pulse. The MWCMHB LS is a hybrid structure of comprising a modified Wilson current mirror and CMOS logic gates [7]. The input and output level ranges from a sub threshold voltage to their standard supplies voltage defined in a transistor technology. A Bidirectional level conversion is available; that is input and output levels can be scaled independently. 3

8 C. MODIFIED WILSON CURRENT MIRROR BASED LEVEL SHIFTER The MWCMHB LS structure is illustrated with three circuit blocks, as shown in Figure.4. A Modified Wilson Current Mirror Circuit (MWCM) is located in Block 1. When VDD1 is a sub threshold and VDD2 is high, the MWCM structures balanced the rising and falling delay at node A, without losing the original statics bias is favored in the WCMLS [4]. However, when the VDD1 and VDD2 levels are closed, the MWCM encountered the same problems as the WCM. The cascade PMOS has been insufficient drives currents and increasing the rising delay. Therefore in a block 3 delay path is designs with an adaptively to reducing the rising delay and maintained at moderate duty cycles. Output inverters offered with sufficient drives strength, which is required into a standard cell designs. Unlike the CCPNR level shifter, which has similar structures, the proposed level shifters uses a CM type of amplifiers, balancing delay path, and complementary OR gates in a block 2. The current mirror type s structure provides a wide operating range, and stacked PMOS transistors in their complementary OR gate limits the leakage current [6]. Level shifters should satisfied the following condition, 1. Small area for sub threshold level conversions. 2. Low power consumptions in super threshold operation. 3. Balancing rising and falling delay in the operating Range. 4. Bidirectional level conversions. Figure.4. Modified Wison Current Mirror based Level Shifter III. POWER REDUCTION METHOD Forced PMOS is main methods of leakage power reduction Techniques. In this method pull down networks is not modified. Pull up networks is only modified with the forced PMOS method. 4

9 An existing PMOS transistor is split into a two each transistor is having the Width/Length ratio of half compare with existing transistors. Consider a inverter as an example. By applying this forced PMOS methods to the inverter the power can be decreased from 2 pw to 1.75 pw. In these methods two PMOS transistors increased the delay in the current flow paths. So that power can be increased. Other leakage power reduction method is stacked method and a forced NMOS method. Stack approach is one of the leakage power reduction methods, which forces a stack effect by breaking down an existing transistor into two half size transistors [9]. When the two transistors are turned off together, induced reverse bias between the two transistors results in leakage current reduction. However, divided transistors increase delay significantly and could limit the usefulness of the approach [10]. In forced NMOS approach if input is given low as compared to threshold voltage., then at the same time PMOS turns on and NMOS turns off and if input is given high at the gate terminal as compared to threshold voltage, then at the same time PMOS turns off and NMOS turn on. Here, the two NMOS transistors which increase the delay in the flow of the current which ultimately decreases the leakage power in the circuit [11]. IV. PROPOSED LEVEL SHIFTERS In this section existing Level Shifters are modified by using the forced NMOS approach. Figure (5, 6, 7, 8) shows the conventional cross coupled method, current mirror based level shifters, Wilson current mirror based level shifters and modified. These level shifters were designed by using the Forced PMOS method. Power was reduced when comparing with existing level shifters. In forced PMOS method pull up networks is only modified. Here existing PMOS transistor W/L ratio is 0.6u/0.4u proposed circuit was designed by splitting the existing PMOS transistors into two each having the W/L ratio of 0.45u/0.6u. Fig.5. Modified Conventional cross coupled level shifter 5

10 Fig.6. Current mirror based level shifter using Forced PMOS Fig.7. Wilson Current Mirror based level shifter using Forced PMOS Fig.8. MWCM based level shifter using Forced PMOS V.SIMULTION AND RESULT Simulation results for proposed Level shifters is shown in Figure (9, 10, 11, 12) respectively. Waveform for modified conventional cross coupled level 6

11 shifters is shown fig.9. Input cross coupled voltage is 2.5v and output voltage is 0.22v. Current mirror based level shifters is also shift the voltage from 2.5v to 0.22v. Wilson current mirror based level shifters are used to shift the voltages from 2.5v to 0.13v. Modified Wilson Current Mirror (MWCM) based level shifters are a bidirectional level shifter which is used to shift the voltage from 1v to 5v. Waveform for MWCM is shown in Figure 12. Fig.9. Waveform for conventional level shifter Fig.10. Waveform for current mirror based level shifter Fig.11. Waveform for Wilson current mirror based level shifter 7

12 Simulation can be performed by using the Mentor Graphics. Mentor Graphics is one of the Electron Device Automation Tool. In this tool Eldo Calibre are mainly used for the circuit simulation, LVS and DVS check. Fig.12. MWCM based level shifter TABLE I. Power and Delay for Various Level Shifters without Forced PMOS Method Level shifters Power Dela y Conventional Level Shifter Current based shifter mirror level Wilson current mirror based level shifter Modified Wilson current mirror based level shifter (m w) (m w) (mw ) (n w) 0.1 ns 0.2n s 0.1n s 0.3n s Table. II. Power and Delay for Various Level Shifters with Forced PMOS Method 8

13 Level shifters Power Delay Conventional Level Shifter (mW) 0.2 ns Current mirror based level shifter (mW) 0.5ns Wilson current mirror based level shifter Modified Wilson current mirror based level shifter (mW) (nW) 0.2ns 0.5ns Table I and II shows the comparison of Normal level shifter design and level shifter design using Forced PMOS method. Comparing with existing and proposed method delay is slightly increased and power is decreased in proposed method. VI.CONCLUSION Conventional level shifter has the power dissipation of mw. The same circuits are designed by using a forced PMOS methods mean that time power is decreased from mw to mw. Likewise current mirror based level shifter, Wilson current mirror based level shifters, and Modified Wilson current mirror based level shifter is reducing the power consumptions from mw, mW and nW to mW, mW and nW. REFERENCES [1] Massimo Alioto Ultra- low power VLSI circuit design demystified and explained: A Tutorial, IEEE transaction on circuits and system- I: regular paper, vol.59, NO.1, january2012. [2] Y. Osaki, Tetsuya Hirose, Nobutaka Kuroki, and Masahiro Numa, A lowpower level shifter with logic error correction for extremely low-voltage digital CMOS LSIs, IEEE J. Solid-State Circuits, vol. 47, n0. 7, pp , Jul

14 [3] A vignesh Performance analysis of various Level shifters using LECC, International Journal of advanced research in electrical vol.2, issue 4, April 2013 [4] Sven lutkemeier and Ulrich ruckert A sub threshold to above threshold Level shifter comprising a Wilson current mirror, IEEE transaction on circuits and system-ii: Express briefs, vol.57, no.9, Sep [5] S. N. Wooters, B. H. Calhoun, and T. N. Blalock, Level Converter in 130- nm CMOS, IEEE Trans Circuits Syst. II, Exp. Briefs, vol. 57, no. 4, pp , Apr [6] S. Hsu, A. Agarwal, M. Anders, S. Mathew, H. Kaul, F. Sheikh, and R. Krishnamurthy, A 280 mv-to- 1.1 V 256 b reconfigurable SIMD vector permutation engine with 2-dimensional shuffle in 22nm CMOS, In Proc. IEEE Int. Solid-State Circuits Conf. Dig. Tech.Papers (ISSCC), pp , Feb [7] Y. Osaki, T. Hirose, N. Kuroki, and M. Numa, A low-power level shifter with logic error correction for extremely low-voltage digital CMOS LSIs, IEEE J. Solid-State Circuits, vol. 47, no. 7, July [8] Shien-Chun Luo A Wide-Range Level Shifter Using a Modified Wilson Current Mirror Hybrid Buffer, IEEE transactions on circuits and systems-i: regular papers, vol. 61, no. 6, June [9] Harshv ardhan upadhyay Comparison among different CMOS inverter with stack keeper approach in VLSI design, IJERA vol.2, issue 3, pp , May-Jun [10] Jun Zhou, Chao Wang, Xiu Liu, and at al., A fast and energy-efficient level shifter with wide shifting range from sub-threshold up to I/O voltage, in IEEE A-SSCC conf., 2013, [11] Sehunkim Sleepy keeper: a new approach to low-leakage power VLSI design. [12] A. Chavan and E. Macdonald, Ultra low voltage level shifter to interface sub and super threshold reconfigurable logic cells, in Proc. IEEE Aerospace. Conf, [13] B. Zhai, S. Pant, L. Nazhandali, and et al.., Energy-efficient sub threshold processor design, IEEE Trans. Very Large Scale Integr. (VLSI) Syst., vol. 17, no. 8, Aug [14] S. N. Wooters, B. H. Calhoun, and T. H. Blalcok, An energy-efficient sub threshold level converter in 130-nm CMOS, IEEE Trans. Circuits Syst. II, Exp. Briefs, vol. 57, no. 4, pp , Apr [15] H. Shao, and C. Tsui, A robust, input voltage adaptive and low energy consumption level converter for sub- threshold logic, in Proc. 33rd ESSCIRC, 2007, pp

15 [16] A. Hasanbegovic, and S. Aunet., Low-power sub threshold to above threshold level shifter in 90 nm process, Proc. NORCHIP Conf., Trondheim, Norway, 2009, pp [17] M. Lanuzza, P. Corsonello, and Stefania Perri., Low-power level shifter for multi-supply Voltage Designs, IEEE Trans. Circuits Syst. II, Exp. Briefs, vol.59, no. 12, pp , Dec

16 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 Design and Fabrication of Aluminium Melting Furnace Using Locally Available Materials 1 Ekpe E. E, 2 *Yahaya B. S, 3 Achema Felix, 4 Fabiyi M. O Engineering Materials Research Department, Nigerian Building and Road Research Institute, Km 10 Idiroko Road, Ota, Ogun State, Nigeria. *ybabatunde@gmail.com Abstract. The increasing demand for aluminium scrap (as waste constituent) in our environment cannot be over emphasized. Hence, an improved system of continuous aluminium recycling as a secondary aluminium production is designed and constructed. A gas-fired crucible furnace is designed and fabricated for melting scrap aluminium. This furnace is a modified model suitable for laboratories and workshops. This furnace is designed and fabricated using locally available materials and butane gas as the thermal energy source in heating up the system to the melting point of aluminium (660.4 o C). The molten aluminium is casted to a desired shape and size either as ingot or end product. Keywords: Aluminum, Furnace, Heat Efficiency, Materials. Introduction According to [1] secondary aluminium production (recycling) saves 95% energy needed to produce aluminium from its ore. Hence, it is more of economic importance that aluminium recycling becomes a major source of aluminium production [2]. Aluminium is primarily used to produce pistons, engine and body parts for cars. Beverages cans, doors, sliding door and aluminium foil. It may also be used as sheet metal, aluminium plate and foil, Rods, bars and wire, aircraft components, windows and door frames. The leading users of aluminium include the container and packaging industry, the transportation industry, building and construction industry [3]. Aluminium can either be produced from bauxite ore (primary aluminium refining) or from aluminium scrap (secondary aluminium refining). Refinement of aluminium ore is sufficiently expensive that the secondary production industry commands much of the market. It is recorded that about 40% of aluminium of aluminium in the US is recovered from secondary refining [4]. Hence, the interest of this project lies on the secondary aluminium production. According to [5] furnace is a space surrounded on all sides by walls and a roof for heating metal or glass to very high temperatures. The earliest furnace was excavated at Balakot, a site of the Indus valley civilization. Then, furnace was most used for the manufacturing of ceramics objects [6]. Furnaces operate in aggressive environment, where several components-molten metal, furnace lining, atmospheric gases, and products from combustion of fuels-coexist at extreme high temperature. Aluminium melting furnace derives its heat from solid fuel (coke and breeze), natural gas, electricity, or other source of energy. Furnaces vary in design, geometry, production capacity (melting rate), materials of construction, and mode of operation [7]. Other factors related to the energy source also affect the furnace design which includes how the energy is transferred to the molten material, how combustion gases are removed, and what refining and 12

17 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 treating equipment must enter the furnace, how long the holding periods are, and how the molten metal will be tapped. Several factors come into play besides the core ingredient of heat and metal as illustrated in fig. 1 below: Chimney Energy Source Output Melting Chamber Fig. 1 Schematic diagram of furnace operation The operating temperature required in the furnace depends on the melting and pouring temperature of the materials being melted. They can range from about C (650F) for zinc alloy to C (3100F) for alloy steels. But for aluminium, the operating temperature is from C and above. There are various types of furnace. They are mainly classified according to their energy source, and application in this case melting, or holding purpose. Different furnaces types are: Crucible Furnace, Induction Furnace, Dosing Furnace, Immersion Furnace, Reverberatory Furnace, Stack Furnace, etc. The furnaces of interest in this research work are the crucible furnace and reverberatory furnace [8-10]. Design Model The design of this research work is based on thermodynamic analysis of furnaces, materials availability, energy source and transfer. The following parameters guilded during the design and fabrication of this melting furnace Thermodynamic analysis/governing equation: This considers conservation of energy and Newton s law of cooling in determining heat transfer coefficient, heat flux, flow rate, and temperature distribution across the composite furnace wall. In other to account for energy balance, furnace efficiency, and capacity/performance, equation and formulations are called into play. These equations enable proper calculations and design functions of the system determined, and are based on the first law of thermodynamics. An energy balance can be expressed as H input = Q + H L. (1) Where, Q is the heat transferred to the melted metal, H input is the energy generated from the fuel combustion H L is the heat loss 13

18 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 Materials availability: Materials for the construction of the furnace are selected due to their availability and ability to withstand the desire temperature and purpose of the system. Energy source and transfer: Cooking gas (butane) is combusted in air (oxygen) and the heat produced is directly used as the thermal energy source. Energy transfer applicable in the design is multimode, which is a combination of radiation, convention and a little of conduction. Structural Design The structure of the furnace is cylindrically designed. This cylindrical structure is so desired, technically due to effectiveness in thermal distribution within the furnace chamber. The structure is explicitly detailed in the drawing below Fig. 2 Detailed Aluminium Furnace Drawing Materials Selection and Criterion The materials used in the course of this research work are sourced locally and selected based on their thermal properties and availability. The selected materials and there specifications are shown in the table below Table 1 Materials Components/Specifications S/N COMPONENTS/MATERIALS QUANTITIES SPECIFICATION 1 Metallic plate ½ sheet 2mm mild steel 2 Refractory bricks 10 pieces Ceramics 3 Insulating materials 1 Fibre glass 4 Burner 1 Locally fabricated 5 Gas cylinder 1 5kg 6 Gas line/pipe mm 7 Crucible pot 1 Ceramics 8 Gas 5kg Butane 9 Refractory ¼ bag White 10 Gas regulator 1 1.0kg/hr 14

19 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 Major Feature of the Furnace The design model is a modification having the following under listed features; i. The body and chamber: the body is formed from a cylindrical shaped of 2mm steel sheath. The body form a melting chamber enclosing the insulating materials (fibre glass), refractory bricks, and crucible. ii. The burner: the burner is externally attached to the body and tangentially positioned to burn the supplied gas using atmospheric oxygen as the heat source to the melting chamber. iii. The crucible: the crucible is a locally made ceramic pot placed in the melting chamber for the purpose of holding the molten aluminium to the time of casting. iv. Roof/cover: the use of furnace cover is critical to energy efficiency. This is constructed with the same materials as the body, from refractory or ceramics with low thermal conductivity, which reduces conduction and radiation heat loss. v. Vent/chimney: a controlled vent of 5mm diameter pipe is attached to the cover, for reduction of internal furnace pressure gradient built across the ceiling, and waste gas exit. Fabrication Procedure and Assembly 2mm mild steel sheet metal can was cut to dimension and shaped to form the cylindrical structure of the furnace. Fibre glass was used as insulating materials, cut and placed internally on the sheet-metal can floor and wall, before refractory lining set in. The refractory lining was built inside a sheet-metal can of 450mm high and 320mm in diameter. Drill and ream 80mm hole in diameter on the side of the cylinder. Then the firebricks are neatly cut to sizes and shaped for the furnace floor and wall. The refractory lining consists of ganister and pieces of firebricks. Ganister is a mixture of equal parts of pulverized firebricks and fire clay and the mixture having the consistency of rather stiff mortar. The bottom of the can is covered with ganister of about 30mm deep, and tamped down to eliminate air pocket. The pieces of firebricks are placed in the positions and pressed down into the ganister so that their top surface will be level 6mm below the hole in the side of the can. Ganister is choked into the spaced between the pieces of firebrick. After a 3-day drying-out period, the gas is ignited and allowed small flames to burn for 30 minutes to complete the curing of the lining. When the furnace is cooled the lining was inspected for cracks which are almost certain to develop. The cracks were filled with prepared refractory cement of fireclay and allowed to dry out before the next firing. Crack filling is repeated until no more crack develops. Furnace Design Calculations The various parameters, methods and techniques used to design the furnace, quantify the losses from the furnace and performance assessment of the furnace is as shown in the calculations below. The two major parameters used in determining furnace capacity are as follows: 1. Energy Consumption, Cap ec ; this describes the amount of energy or heat intake by the furnace. In other word, the actual quantity of heat used in melting a unit of aluminium. The unit for capacity is KJ/Kg or Btu/Ib (British unit/pound). It is determined by this equation; Cap ec = 5.5[KW] (from burner rating) = 5.5[KJ/Sec] = 5.5 x 3600 = 19800[KJ/hr] = 4729[kcal/hr] = 4729[Kcal/Kg] 15

20 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 Using the conversion factor, I kcal/kg = KJ/Kg = 1.8 Btu/Ib Hence, Cap ec, = 4729 x = 19800KJ/Kg or Btu/Ib 2. Stock Capacity, Cap sc ; this capacity refers to the quantity of stock (in this case, mass of aluminium) that can be melted at a certain period of time. This is determined by the size of the crucible by volume or mass (in Kg) Cap sc, = 5Kg (maximum) Heat Supplied to the Furnace Heat supplied to the furnace is the heat input to the furnace chamber as a result of gas combustion from the burner. This is determined by the ratings on the gas regulator which has 1.0Kg/hr. Hence, the heat input, H input = KJ/Kg x 1.0Kg/hr = KJ/hr Using the conversion factor; 1 KJ/hr = Kcal/hr KJ/hr x = Kcal/hr The heat output is calculated as follows: Q = m C p T. (2) Where, Q = quantity of heat stock in Kcal/hr M = weight of the stock melted per unit time Kg/hr C p = mean specific heat of stock in Kcal/Kg o C T = change in temperature from the initial temperature of stock before it enters the furnace to the final temperature of stock, o C Substituting the values; = 5 x 0.22 x (750 25) = 797.5Kcal/hr = KJ/hr = KW or 927.5W From equation 1 Heat loss, H L, = H input Q = = KJ/hr Furnace Performance/Efficiency From Newton s law of cooling equation; q = Q A = h c T. (3) Where, Q = heat flow rate A = area of cylinder h c = heat transfer coefficient T = temperature difference across wall 16

21 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 A = πd2 2 + πdh l = (320) = mm 2 = m 2 T = = C or 998K Substituting the values; Q h c = = = 1.52 A T W/m2 K Temperature Drop across the Composite Wall q 1 ln d1 d2 T a T b =. (4) 2πk 1 K, is the thermal conductivity for different wall materials. Substituting the values T b = ln(306/238) = 9 Hence, T b = = C For T c ; T b T c = q 1 ln d 3 d2. (5) 2πk T c = ln (318/306) T c = 55 0 C For T d ; 55 T d = q 1 ln d4 d3 2πk 3 T d = C Determination of Composite Wall Thickness Steady-state, one dimensional heat flow through insulation systems is governed by Fourier s law Q = K. A. (dt) t. (6) Where, Q = rate of heat flow, w A = cross sectional area normal to heat flow, m 2 T = thickness of insulating materials dt = temperature difference in material, o C From equation 6; t = K. A. (dt) Q. (7) Applying equation (7) for each wall; For t a ; t a = k 1. A. (T a T b ) Q For t b t b = k 2. A. (T b T c ) Q For t c = 1.4 x x ( )/927.5 = 0.008m = x x (741 55)/927.5 = 0.018m 17

22 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 t c = k 3. A. (T C T d ) Q Results and Discussions = 11.5 x x (55 114)/927.5 = 0.448m After a thorough analysis and calculations, the following parameters where used during the designing and fabrication of aluminium melting furnace using locally available materials. Table 2 Obtained Parameters S/N PARAMETER SYMBOL VALUES UNITS 1 Capacity Cap KJ/Kg Btu/lb 2 Heat Input H input Kcal/hr KJ/hr 3 Heat Output Q Kcal/hr KJ/hr 4 Initial temperature T 1 15 O C 5 Final Temperature T o C 6 Efficiency E % 7 Heat loss H L KJ/hr Mass of 8 Aluminium M 4.98 Kg Produced 9 Area of Furnace A m 2 10 Heat Flux Q w/m Heat Transfer Coefficient Temperature across furnace wall h c 1.52 w/m 2 K T a T b T c T d o C Furnace Performance and Evaluation The aluminum melting furnace was evaluated to ascertain its performance by melting aluminum scraps at a temperature of over a period of 300 seconds. The results obtained were tabulated as indicated in Table 4 and graphically in Fig. 4. It was observed that the temperature of the melting furnace was maintained at 720 degrees over a period exceeding 300 seconds and this shows that the furnace was designed to attain a maximum temperature of Table 4 Performance Analysis of the Melted Aluminum Scrap on the Furnace 18

23 Temperature (Degree) American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 Time (Seconds) Temperature (Degree) Time (Seconds) Fig. 4 Melting Furnace Performance The results obtained above from Table 4 shows that the melting furnace can achieve its main objective of recycling laboratory size aluminium (5Kg) in a single operation. A significant improvement in efficiency of 28% was recorded compare to the 4-6% in using the traditional furnace [12]; it also takes a shorter time to melt the aluminium with great reduction in the energy consumption. Hence, the success of this research work supersedes that of the traditional techniques. Conclusion This research work was undertaken to design and fabricate an aluminum melting furnace for laboratory and workshop use. The furnace was constructed putting into consideration; its temperature attainment, capacity of metals it can hold, operators safety, space to be occupied in the workshop floor/laboratory, cost restrictions, availability of the materials used, its maintainability and portability. This research has revealed that the locally fabricated aluminum melting furnace for laboratory and workshop operations have an efficiency of 28%, low energy consumption and shorter time of operation. Hence, a more effective and performing furnace were developed. Finally, the actualization and realization of this research work is a boost to the development of local manpower capacity in Nigeria and also to advance the reliability of engineering materials in service. 19

24 American Journal of Engineering and Technology Research Vol. 16, No. 1, 2016 References [1] Goldstick R. and Thumann A., Principle of waste recovery, 2 nd Ed, Fairmont press, 1986 [2] Joshi, Amit M, Aluminium Foundry Practice, 2 nd Ed, IIT, Bombay, India, 2002 [3] Kazantsev E.I, Industrial Furnaces, 2 nd Ed, Mir Publishers, Moscow, 1977 [4] September 26, 2014 [5] Whitehouse R.C, 1993, the Valve and Actuator User Manual, Mechanical Engineering Publication, ISBN [6] August 20, 2014 [7] Daviess Clive, Calculation in furnace technology, 1 st Ed, Pergamum press, 1970 [8] Charles H. Fink Jr., Modern Induction Melting Improves Process and Productivity, Foundry Management and Technology, 2005 [9] Gray W.A and Muller R., Engineering Calculations in Radiation Heat Transfer, 1 st Ed, Pergamum Press, 1974 [10] Abubakre, O. K., Development of Aluminium Based Metal Matrix Particulate Composites (MMPC) Reinforced with Alumina, Silica and Mill Scale, PhD Thesis, Mechanical Engineering Department, Federal University of Technology, Minna, Nigeria, [11] Bala, K. C., Design and Development of Sand Muller and Standard Sand Rammer., M. Eng. Thesis, Mechanical Engineering Department, Federal University of Technology, Minna, Nigeria, [12] Ilori, B. O., Design of Electric Arc Furnace. B. Eng Thesis; Mechanical Engineering Department; Federal University of Technology, Minna, Nigeria,

25 A Comparative Study Of Optimization Techniques For Capacitor Location In Electrical Distribution Systems Ganiyu A. Ajenikoko 1, Jimoh O. Ogunwuyi 2 1, Department of Electronic & Electrical Engineering, Ladoke Akintola University of Technology, P.M.B. 4000, Ogbomoso, Nigeria. 2, Department of Electrical & Electronics Engineering, Osun State Polytechnic, Iree, Osun State, Nigeria. Corresponding ajeedollar@gmail.com Abstract. The location of capacitors to be installed in electrical distribution system is an optimization problem that needs immediate attention. In this paper, an overview of various approaches for solving problems of optimal location of capacitors in electrical distribution system is carried out. The objective functions and constraints of the problem are formulated. The total energy losses are minimized; the cost of energy losses are kept at minimum, the sum of energy loss cost and the investment cost required for the capacitor location are minimized. The cost due to energy loss cost, peak power loss and the investment cost of the capacitors are also minimized. The optimal solution for the power flow equation and bus voltage limits are obtained to limit the capacitor KVAr to be located. The various approaches for solving capacitor location problems are discussed and compared. The Particle Swarm Optimization technique is the most popular of all the techniques due to its simple implementation and relatively smaller computation with a faster convergence even though; premature convergence of algorithm is the draw-back. The accuracy of the results can be improved upon by combining two or more of the optimization techniques to locate capacitors in electrical distribution system. Keywords: Optimization, Capacitors, Objective Functions, Heuristic Technique, Distribution System. 1. Introduction, In electrical distribution system, a portion of power loss can be reduced by adding shunt capacitors in order to supply a part of the reactive power demand (Kannan et al, 2007, Chin and Lin 1994).There is also the need to decrease the losses associated with the reactive power flow through the branches in the distribution systems. Location of capacitors in electrical distribution systems is accompanied with benefits such as bus voltage regulation, power quality improvement, feeders and system capacity release, energy loss reduction and power factor correction (Schmil 2005, Song et al 1997, Kaplan 1994, Neg et al 2000). The location and control of capacitors under different loading conditions depict the extent of the advantages of the capacitor placement 21

26 in distribution systems. In this case, the optimization problem needs to be formulated as well as the objective functions after which the solution approaches will be applied to determine optimal position to install the capacitors on electrical distribution systems(wang et al 2010, Mekhamer et al 2002, Salama et al 1995, Baran and Wu 2009). The power loss in a distribution system is significantly high because of lower voltage and hence high current compared to that in a high voltage transmission system. The pressure of improving the overall efficiency of power delivery has forced the power utilities to reduce the loss, especially at the distribution level (Grainger and Lee 2011). Reconfiguring the network can reduce the power loss in a distribution system. The reconfiguration process changes the path of power flow from the source to the loads. The loss can also be reduced by adding shunt capacitors to supply a part of the reactive power demands. Shunt capacitors not only reduce the loss but also improve the voltage profile, power factor and stability of the system (Salama et al 1995, Neg et al 2000). The active power demands at all nodes and losses must be supplied by the source at the root node, as distribution system is mainly radial. However, addition of shunt capacitors can generate the reactive power and therefore it is not necessary to supply all reactive power demands and losses by the source. Thus, there is a possibility to minimize the loss associated with the reactive power flow through the branches (Schmil 2005). Distribution system accounts for a major portion of power system losses. The power loss in a distribution system is significantly high because of lower voltage and hence high current, compared to that in a high voltage transmission system. The pressure of improving the overall efficiency of power delivery has forced the power utilities to reduce the loss, especially at the distribution level. The same technique can be applied to other types of feeders (Masoum et al 2004). Capacitors are used widely to reduce the distribution system loss. In addition, shunt capacitors could also accommodate voltage regulation and VAR supply. For capacitor optimal location, general considerations include (Song et al 1997): (i) The number and location; (ii) Type (fixed or switched); (iii) The size; When capacitors are placed, power loss is reduced and also energy loss is reduced. These factors contribute to increasing the profit. Cost of capacitors decreases this profit. So profit is weighted against the cost of capacitor installation. The whole saving can be given as follow: S=K p P+K e E K C C (1) Where, K P = Per unit cost of peak power loss reduction($/kw) K E = Per unit cost of energy loss reduction ($/KWh) K C = Per unit cost of capacitor ($/KVAr) p = Peak power loss reduction (KW) E= Energy loss reduction (KWh) C= Capacitor size (KVAr S= Saving in money per year ($/year) The voltage constraint must be satisfied i.e voltage (pu) should be between min (0.9) to max (1.1).i.e. V min V V max (2) Real Power Loss formulation (RPL) 22

27 If vi.. is the ith bus voltage and vj is jth bus voltage, then, the line current between ith and jth buses is given by (Sallam et al 1994): I ij = V i V j (3) Z ij Where, Z ij is the impedance between ith and jth buses, the transmission power between the ith and jth buses and vice versa can be expressed as (Kaplan 1994): S ij = V i I V ij (4) S ij = V i I V ji (5) The real active loss between ith and jth buses is defined as: S ij = V i I V ji (6) Total loss power in power system is defined by (Eajal and El-Hawary 2010, Masoum et al 2004): P Total N N i,j Loss = i=j j=i P loss (7) Where, N is the number buses of power system and the real power loss is given by: RPL = Total P Loss Pno min al (8) Loss no min Where, P al Loss is the real power loss. Voltage Profile Improvement (VPI) One of the components of optimal location of capacitor banks is the improvement in voltage profile. This index penalizes the size-location pair which gives higher voltage deviations from the nominal value (Vnom). In this way, closer the index to zero better is the network performance (Santoso and Tan 2013, Grainger and Lee 2011). 2. Materials and method. i. Formulation of the objective functions for the capacitor location problem. This is given as: Min F= Min F= s j=1 T j j P L s j=1 K e j T j P L j (9) (10) Min F= s (K j j=1 e T j P j L ) + C c (11) Min F= s j=1 (K j e T j P j p L ) + C c + K p P L (12) Where F = the value of the desired objective function S = number of load levels Tj = time duration for the j-th load level P j L = power loss at the j-th level k j e = per unit cost of energy loss at the j-th load level C C = investment cost of capacitor 23

28 K P = per unit cost peak power loss P L p = power loss at peak load level From equations (9) to (12) above, the following sub-tasks were performed: (a) Minimization of the total energy losses. (b) Keeping the cost of energy losses at minimum. (c) Minimizing the sum of energy loss cost and the investment cost required for capacitor location. (d) Minimizing the total cost due to energy loss cost, peak power loss cost and the investment cost of the capacitors. ii. Obtaining the optimal solution for the power flow equations and bus voltage limits. 3. Review of the capacitor location approaches: The capacitor location problems have been solved with a variety of techniques. Conventional analytical method is one of the methods recently used with some heuristics. Algorithms based on heuristic programming have been emphasized ln recent times. The features of heuristics algorithms include robustness and lesser computational effects. Some of the approaches for optimal location of capacitor on distribution feeders are as fellows: (a) Ant colony optimization; The behaviours of real ants are imitated in this approach. Without using visual cues, real ants can take the shortest path from two sources to the nest. They are capable of adapting to changes in the environment. The ants establish the Pheromone. The pheromone is the material deposited by the ants, which serves as critical communication information among ants, thus guiding the determination of the nest movement. (b) Harmony search algorithm: This is a meta-heuristic optimization algorithm inspired by playing music. This algorithm deals with meta- heuristic analysis which combines races and randomness to imitate natural phenomena. The best harmony between components that are involved in the operation process for optimal solution is motivated by the operation of orchestra music. I n this case, optimization design variables can be assumed to be discrete in nature as a result of computational intelligence and random processes involved. Musical instruments can be played with some discrete musical notes determined by player experience and random processes in improvisation. Based on aesthetics standards, music players improve their experience while improvement in the design variables of computer memory is determined by the objective function. (c) Graph: A set of nodes connected by arcs is called a graph. A possible combination of capacitor sizes and locations determine each node for a given number of capacitors to be placed. A separate graph is used for each possible number of capacitors to be installed. The nodes of a graph is usually found by the search process to determine the optimal solution. The minimum 24

29 value of the objective function is determined by a search process to locate the node in the graph with some minimum number of capacitors to be placed. The next graph is searched if the maximum number of capacitors to be installed is not reached and in this case, the number of capacitors to be placed is incremented. The process is then terminated with the addition of another capacitor which will increase the objective function. (d) Simulated Annealing: This approach has to do with annealing of metals or crystals, temperature that is controlled by cooling schedule at each iteration is employed by simulated annealing. The state converges to solution when the temperature cools down. The state moves freely to other state when the temperature increases and stops moving to other state at low temperature (e) Fuzzy Logic: A computational representation of heuristic knowledge about a particular problem is illustrated here. Solutions of control and optimization problems are some of the few application areas. Solving the problem using fuzzy logic involves identification of the main variables that have influence on the decision to be taken and quantify their values. The profile of this variable is established by a relationship function to express the degree of compatibility of each of them with information that is previously known. In this case, necessary actions are determined after the establishment of the rules. Rules are set up to find out the advantage of installing a capacitor bank in a particular bus. The fuzzy variables in this case are the bus voltages and the objective function values. They are a function of the bus that is accessed to verify the practicability of installing capacitor. (f) Genetic Algorithm: This is an optimization statistical method. The robustness feature of genetic algorithm makes it to have a flexible balance between efficiency and necessary features to survive in different environment. Coding is a main concept in genetic algorithm. It encodes problem parameters of variables and uses the code. The solution accuracy, range of parameter changes and the relation between variables determine the number of bits used for encoding procedure. String of bits in the encoded form of possible solutions of the problem constitutes chromosomes. Each set of genetic algorithm deals with the sets of chromosomes in the search spaces. To solve the problem of capacitor location, the chromosome is made up of n+1 bit of binary numbers where n is the number of probable capacitor location.the capacitor is usually located by the first bit and it is called location selector. The length of location selector bit is equal to the number of probable locations of capacitor and each location is represented by bit of the string. In the bit position, a 1(one) is used to select a location and a 0(zero) bit implies non selection of the locations. The KVAr bit for a particular location may have a non zero value when the selection bit for that location in the location selector bit is zero. The KVAr bit in this case is ignored and determines the fitness function value of the chromosome. (g) Particle Swarm Optimization: This is a meta-heuristic parallel search approach. The iteration of particles in swarm, using common evolutionary computation algorithm guide the direction of swarm towards the optimal region of several space. Some of the advantages of particle swarm optimization are: computational efficiency, simplicity in concept and implementation, less computation time and inexpensive memory for computer resources. (h) Tabu Search: This is used for solving a combinatorial optimization problem and it is a metaheuristic method. It is a hill- climbing method that evaluates the final solution through repeating the process of creating solution in the neighbourhood around the initial solution and selecting the 25

30 best solution in them. Tabu search is an extension of the hill- climbing method since it has the adaptive memory called the tabu lists which are just some attributes in order not to change. A new attribute that enters into the tabu list stays in the tabu list for a while. There after, the oldest attribute is released from the tabu list as soon as a new attribute enters into the tabu list. A new attributes is also assigned to the tabu list each time the iteration count is updated. Tabu search employs intensification and diversification methods to obtain optimal solution. The solution structure has a frequency counter which is used in the intensification or diversification approach. Whenever the frequency counter is less than three times the frequency counter threshold, then there is possibility of intensifying the search in the neighbourhood of the sub optimal region. Tabu search in this case now finds an optimal solution to the combinatorial optimization problem. (i) Analytical optimization technique: This technique involves computation of minimum losses and cost savings with the assumption that the distribution feeders do not have any sub-branch. The 2/3 rule is a peculiar feature of this method. (j) Plant Growth Simulation Algorithm: Plant growth process is the oasis upon which this technique rests. In this method, the trunk of a plant grows from its root, some branches will grow from the nodes and some new branches will grow from the nodes on the branches. The process continues until a plant is formed. (k) Body immune algorithm: This technique imitates the behaviour of human body against the external invasions. It is a computational tool in pattern recognition. It considers the objective function while satisfying the constraints. (l) Hybrid Artificial Intelligence Technique: This is a four-stage approach involving the simulated annealing, genetic algorithm, Tabu-search and the hybrid genetic algorithm-fuzzy logic algorithm for solving a combinatorial optimization problem with a non-differential function. 4. Comparative study of the optimization techniques. The most popular method is Particle Swarm Optimization technique. This technique is simple to implement, the computation load is relatively smaller with a faster convergence. Most difficult problems are easily solved with this method. Premature convergence of algorithm is found to be a draw-back to this approach The genetic algorithm optimization technique is next to particle swarm optimization technique in term of popularity even though its divergence nature makes it a disadvantage during application. Fuzzy approach is employed in Hybrid optimization technique to find capacitor location while appropriate sizes of capacitors are found using genetic algorithm, Tabu search, Harmony search algorithm and Simulated annealing. 5. Conclusion. A comparative study of optimization techniques for capacitor location on electrical distribution system has been presented. The optimal location of capacitors on electrical distribution systems is a challenging task as a result of its combinatorial nature. Technical and functional objective functions were formulated for an optimal location of capacitors. The total energy losses, cost of energy losses, sum of 26

31 energy loss cost and the investment cost required for capacitor location were minimized to obtain the optimal solution for the power flow equation and bus voltage limits. The analytical optimization technique is very simple to implement, even though, its convergence is relatively poor with a slow computation. The convergence of Heuristic technique is fast. Large and complex networks are easily solved using these optimization techniques. A high level of result accuracy can be achieved by combining two or more of the aforementioned techniques. References Baran M.E, and Wu F.F (2009): Optimal sizing of capacitors placed on a radial distribution system, IEEE Trans on power delivery, Vol.5, No.4, Pp Chin C and Lin W.M(1994): Capacitor placements for distribution systems with fuzzy algorithms, Procceedings of the 1994 Region 10 Ninth Annual International Conference, Pp Eajal A.A and El-Hawary M.E(2010): Optimal capacitor placement on radial distribution system and sizing in unbalanced distribution systems with harmonics consideration using particle swam optimization, IEEE Transactions on power delivery, Vol.25, No. 3, Pp Grainger J.J and Lee S.H (2011): Optimum size and location of shunt capacitors for reduction of losses on distribution feeders, IEEE Transactions on power delivery,vol.1, No.5, Pp Kannan S.M, Rathina G.S and Slochanal S.M.R(2007): Fuzzy logic based optimal capacitor placement on radial distribution feeders, IEEE Transactions on power Apparatus and Systems, Vol. 100, Pp Kaplan M (1994): Optimization of number, location, size, control type and control of shunt capacitor on radial distribution feeders, IEEE Trans. on power apparatus and system. Vol PAS- 103, Vol.9, Pp Masoum M.A.S, Jafarian A, Ladjevard M and Fuchs E.F (2004): Fuzzy approach for optimal placement and sizing of capacitor banks in the presence of harmonics, IEEE Transactions on power delivery, Vol. 19, No. 2, Pp Mekhamer S.F, Soliman S.A, Moustafa M.A and El-Hawary(2002): Load flow solution of radial distribution feeders: a new contribution, International Journal of Electrical power and Energy systems, Vol. 24, No. 9, Pp Neg H.N, Salama N.MA and Chikhani(2000): Capacitor allocation by approximate reasoning fuzzy capacitor placement, IEEE Transactions on power delivery, Vol. 15, Issue 1, Pp Salama M.M.A, Chikhani A.Y, Hackam R and Mansour E.A.A(1995): Control of reactive power in distribution systems with an end-load and fixed load condition, IEEE Transactions on power Apparatus and Systems, Vol. 104, No. 4, Pp s 27

32 Sallam A.A, Desouky M and Desouky H (1994): Shunt capacitor effect on electrical distribution system reliability, IEEE Transaction on Reliability, Vol. 43, No. 1, Pp Santoso I.I andtan O.T (2013): Neural-Net based real time control of capacitors installed on distribution systems, IEEE Trans. on power delivery, Vol.3, No. 9, Pp Schmil J.V(2005): Optimal size and location of shunt capacitors on distribution feeders, IEEE Transactions on Power Apparatus and Systems Vol. 84, Pp Song Y.H, Wang G.S, Johns A.T and Wang P.Y(1997): Distribution network reconfiguration for loss reduction using fuzzy controlled evolutionary programming, IEEE Trans. Gener., trans Distri., Vol. 144, No. 4 July Wang J. C, Chiang H. D, Miu K. N and Darling G(2010): Capacitors placement on real time control in large scale unbalanced distribution system: loss reduction formula, problem formulation, solution methodology and mathematical justification. IEEE Trans. On power delivery, Vol.12, No.2, Pp

33 Study of Sag by the Influence of Altitude Parameter Shamim Ahmed Martin-Luther-Universität Halle-Wittenberg Halle, Germany Abstract. In this paper, transmission line sag is studied by direct effect of altitude parameter. More precise result is achieved when the measurement is performed considering the height of the overhead transmission lines from ground. Barometrical data from American standard atmosphere is used to carry this study. The complete form of mathematical sag expression is considered here, in presence of ice coting and wind pressure. Study is performed for both cases, when the supports are equal and unequal levels. Numerical calculations are performed in Matlab. Keywords: Transmission lines, Sag, Quintic polynomial, Quartic polynomial, Altitude function, Matlab. Introduction In a simple definition, the vertical distance between the points of supports and the lowest point of the transmission line is referred as dip or sag. It is not possible to install the electric cables without any sag even though we wish to do so because of its acting downward weight. This sag is in practice, allowed to form to prevent the conductors from break due to the excessive tension. Thus the sag has become a frequently handled parameter in power systems. The conventional mathematical expressions of sag are involved with the ice coating and the wind pressure [1]. This pressure again varies by the altitude or height. Some useful barometric formulae are available with different variables including the altitude. These equations are generally used to find out the barometric pressure and hence the sag of the overhead power transmission lines. But in this paper, the sag is studied considering the altitude directly. Conventional Measurement of Sag It is considered that the supports are at unequal levels (Fig. 1). The transmission lines are assumed with ice coating and influenced by wind pressure. Then the mathematical forms [1] of sag can be expressed as by (1) and (2). 29

34 S l = W r 2T (l 2 Th W r l )2 (1) S u = W r 2T (l 2 + Th W r l )2 (2) Where, S l is sag at the support of lower level and S u is sag at the support of upper level. T is the tension in the conductor. The span length is l. Difference between the two supports is h. Resultant weight per unit length of the conductor is denoted by W r. The resultant weight is occurred because of the weight of conductor itself, ice coating and the wind flow. Thus W r is given by (3). W r = [(W c + W i ) 2 + W 2 w ] 1 2 (3) W c, W i and W w in Eq. (3) denote the weight per unit length of the conductor, weight of ice for unit length and wind force per unit length respectively. Weight of ice per unit length can again be written as: W i = πtρ i (D + t) (4) ρ i is the density of ice. Diameter of the conductor is D and t for the thickness of ice coating. Finally wind force per unit length is calculated using (5), where P is the wind pressure [1]. W w = P(D + 2t) (5) If the supports are exactly at equal level (h = 0) or almost at equal level (h 0) then Eq. (1) and Eq. (2) are reduced in a single formula. In that case the sag S becomes [1]: S = W rl 2 8T (6) A schematic diagram is shown in Fig. 1 to illustrate the involved parameters of sag formation. 30

35 Fig.1. Formation of sag at unequal supports These mathematical expressions [Eq. (1), Eq. (2) and Eq. (6)] are generally used for the conventional measurement of sag. The wind pressure P is in practice, calculated by some sort of barometric formulae. This is of course a correct process but the problem is that those barometric formulae may have many variables [2]. So it would be better if the sag expressions can have only the altitude parameter, for the barometric estimation. This altitude is the height of the overhead power transmission lines. In the next section, sag is studied in this direction. Measurement Using the Altitude Parameter Many theoretical and experimental works have been done on sag calculation and later sag was calculated from GPS measurements [3]. But up to this time no research had been conducted to calculate the sag directly by altitude measurement system. Let us consider the variation of wind pressure with respect to the height. According to the information provided by American Standard Atmosphere [2], this variation follows the Height (H) and Pressure vector (P), as below. H = [ ] in Km P = [ ] in KPa This data set is widely used to get the overview of the wind pressure for certain altitude levels. The variation of wind pressure with altitude and the standard deviation of wind pressure are observed in Matlab. For this purpose, the corresponding code is written in Matlab, which is shown below. 31

36 % observe barometric data and find standard deviation of wind pressure % % P and H are row vectors % H = [ ]; % height in Km P = [ ]; % wind pressure in KPa % figure; axis('normal') box 'on' grid 'off' zoom 'off' barh(h,p,1.5) % horizontal bar diagram xlabel('wind Pressure (in KPa)') ylabel('height (in Km)') xlim([0 110]) text(50,5,'pressure at Sea Level') legend('wind Pressure','Location','NorthEast'); % % find standard deviation of wind pressure % M = mean(p); % mean of wind pressure % N1=P(:,1); s1=round(abs(n1-m)); % standard deviation of P at zero level N2=P(:,2); s2=round(abs(n2-m)); N3=P(:,3); s3=round(abs(n3-m)); N4=P(:,4); s4=round(abs(n4-m)); N5=P(:,5); s5=round(abs(n5-m)); N6=P(:,6); s6=round(abs(n6-m)); N7=P(:,7); s7=round(abs(n7-m)); % standard deviation of P at 71 Km % S = [s1 s2 s3 s4 s5 s6 s7] % standard deviation for seven altitude levels % figure; bar(s) ylabel('standard deviation of wind pressure') % str1=['at H = 0; std = ',num2str(s1)]; str2=['at H = 11 Km; std = ',num2str(s2)]; str3=['at H = 20 Km; std = ',num2str(s3)]; str4=['at H = 32 Km; std = ',num2str(s4)]; str5=['at H = 47 Km; std = ',num2str(s5)]; str6=['at H = 51 Km; std = ',num2str(s6)]; str7=['at H = 71 Km; std = ',num2str(s7)]; % text(0.6,86,str1); 32

37 Height (in Km) text(2,25,str2,'rotation',90); text(3,25,str3,'rotation',90); text(4,25,str4,'rotation',90); text(5,25,str5,'rotation',90); text(6,25,str6,'rotation',90); text(7,25,str7,'rotation',90); % Wind pressure at different altitude levels and standard deviation of wind pressure are illustrated on bar diagrams in Fig. 2 and Fig. 3 respectively. 71 Wind Pressure Pressure at Sea Level Wind Pressure (in KPa) Fig. 2. Wind pressure variation with altitude 33

38 Standard deviation of wind pressure at H = 11 Km; std = 4 at H = 20 Km; std = 13 at H = 32 Km; std = 18 at H = 47 Km; std = 19 at H = 51 Km; std = 19 at H = 71 Km; std = at H = 0; std = Fig. 3. Standard deviation of wind pressure for seven altitude levels From figure 3, it is obvious that almost every cases, the standard deviations are too high. So it is not a good idea to consider the average of wind pressure at the desired level, where the overhead transmission lines are installed. This problem could be solved with a quintic polynomial, which is established by curve fitting method in Matlab. 34

39 Wind Pressure, P (in KPa) P = e-07*H *H *H *H *H P vs H Fitted Curve Height, H (in Km) Norm of Residuals = Residuals Fig. 4. Fitted wind pressure for fifth degree The residuals are found as below. Residuals = [ ] Hence the curve fitting leads to achieve (7). α(h) H H H H H 5 (7) Thus α(h) forms a quintic polynomial, where the leading coefficient is and the leading term is H 5. This polynomial implies that higher the power of altitude, its coefficient becomes smaller. This Altitude function α(h) can directly be used to measure the wind pressure P, only using the height from the ground to the overhead power transmission lines. Eq. (5) can now be given in terms of α(h) as: 35

40 W w = (D + 2t) α(h) (8) The resultant weight, using Eq. (8) stands as: Finally the mathematical expression of sag in Eq. (6) becomes: W r = [(W c + W i ) 2 + {(D + 2t) α(h)} 2 ] 1 2 (9) Eq. (10) allows to express Eq. (1) and Eq. (2) as below. S = l2 [(W 8T c + W i ) 2 + {(D + 2t) α(h)} 2 ] 1 2 (10) S 1 = [(W c+w i ) 2 +{(D+2t) α(h)} 2 ] 1 2 2T ( l 2 Th ) 2 l [(W c +W i ) 2 +{(D+2t) α(h)} 2 ] 1 2 (11) S 2 = [(W c+w i ) 2 +{(D+2t) α(h)} 2 ] 1 2 2T ( l 2 + Th ) 2 l [(W c +W i ) 2 +{(D+2t) α(h)} 2 ] 1 2 (12) Eq. (11) and Eq. (12) are the desired expressions of sag. Where S 1 the sag at support of is lower level and S 2 is the sag at support of upper level. These equations are perfectly able to calculate sag by the direct effect of altitude function α(h). Some Special Conditions For the minimum height that is at the sea level α(h) is to be reduced as: α(h) = Lim α(h) KPa (13) H 0 But it is important to note that wind pressure does not follow the conventional barometric formulae above 32 Km from the sea level [3]. This is true both for the case of standard temperature lapse rate as zero or nonzero. As a result, the study of measurement is limited up to this level. But in practice, it is not a problem at all. Because the installed transmission lines are much lower than that level, even in the hilly places. 36

41 However, the rate of change of wind pressure with respect to height becomes: dα(h) dh = H 0.05 H H H 4 (14) Thus Eq. (14) is a quartic polynomial when the rate of change of wind pressure with respect to the height of the transmission line is considered. It is observed from Fig. 5. dp/dh (KPa/Km) 200 Altitude, H (Km) Formation of Quartic Polynomial -400 (In presence of Lapse rate) Fig. 5. Formation of quartic polynomial from quintic form From Eq. (7), it is obvious that when we consider the height of the transmission lines at the sea level then H is to be considered zero. Thus α(h) becomes: α(h = 0) KPa α(h = 11 Km) = 22 KPa; Almost same as in vector P As a same manner, for desired level (where the transmission line is installed), the corresponding wind pressure can be obtained, for the altitude range 0 H <

42 Conclusions Conventional sag has been studied by influence of the height of overhead transmission lines. This study is more convenient for sag calculation for any types of sag formation. In special case, the quintic polynomial has been reduced on its limiting value same as the wind pressure at sea level. This allows to have a single constant of the altitude function. The quintic polynomial was turned into a quartic polynomial when the rate of change of wind pressure near to the overhead transmission lines is considered. This approach can be used when the altitude is below 32 kilometers. The measurement is valid for both conditions when the supports are at equal and unequal levels. References [1] V. K. Mehta and Rohit Mehta, Principles of power system, S. Chand & Company Ltd. New Delhi, 2003, pp [2] U.S. standard atmosphere: [3] K. Sangeeta, R. Dahiya and P. K. Chaturvedi, A method of getting sag of power transmission line from DSP of GPS measurements, International Journal of Engineering Science and Technology, vol. 2(8), pp ,

43 Response Analysis of Phosphorus Removal to Operational Effects of Inoculum Concentration and Leaching Time during Biotreatment of Iron Oxide E. M. Ameh C. A Mgbachi Department of Metallurgical and Materials & Electrical/Electronic Engineering, Enugu State University of Science and Technology, Enugu, Nigeria. ebere.monica@yahoo.com cysmart2007@yahoo.com. ABSTRACT Response analysis of phosphorus removal to operational effects of inoculum concentration and leaching time concentration was carried out during biotreatment of iron oxide ore using Acidithiobacillus Ferrooxidans. The model was expressed as; ξ = ϑ The validity of the empirical model was found to be rooted on the expression ξ = ϑ where both sides of the expression are correspondingly approximately equal. Statistical analysis of removed phosphorus concentration for each value of the inoculum concentration as obtained from experiment and derived model-predicted results show standard errors of and x 10-7 % respectively. Furthermore, removed phosphorus concentration per unit inoculum concentration as obtained from experiment and derived model-predicted results were and 21.0 % / ml respectively. Deviational analysis indicates that the maximum deviation of model-predicted removed phosphorus concentration (from experimental result) was less 8%. This translates into a derived model confidence level of above 92% and response coefficient of dependence of phosphorus removal on the inoculum concentration and leaching time. Keywords: Response Analysis, Phosphorus Removal, Inoculum Concentration, Leaching Time, Acidithiobacillus Ferrooxidans, Agbaja Iron Ore. INTRODUCTION During service, the catastrophic failures in steel structures containing high phosphorus content has raised the need for intensive research and development aimed at improving the methods and procedures for dephosphorizing iron ore. It is strongly believed that the development of an economical dephosphorisation process is critical factor for the survival and growth of iron and steelmaking industry. Research [1] has shown that during the effective dephosphorisation of Western Australian iron ore, sulphuric acid was chosen as the leachant on the basis of its availability and low cost. The iron ore sample used in this study typically contained 0.126% phosphorus. After roasting at 1250 C, lump ore (P mm), pellet 1 (grinding to 100%, -1.5 mm before pelletisation) and pellet 2 (grinding to 100%, mm before pelletisation) were leached in solutions with different sulphuric acid concentrations. After leaching for 5 hours at 60 C in 0.1 M sulphuric acid solution, 67.2%, 69.0% and 68. 7% of the phosphorus was leached from the above three samples, respectively. The phosphorus content was reduced from 0.126% to 0.044%, 0.055% and 0.042% respectively. It was observed that during this process, the dissolution of iron during leaching was negligible. Report [2] has revealed that during dephosphorization of phosphor-containing iron, the phosphor-containing iron ore is broken to the granularity of less than 0.074mm and mixed with iron pyrite pre-broken to the granularity of less than 0.074mm based on mass percent of 5%-20%. The mass concentration of ore slurry is adjusted to 10%-20% by the aphosphorosis 9K culture and ph of original ore slurry is in the range of 1.5 to 3.5. This method is suitable for direct-extracting and dephosphorizing from each phosphor-containing 39

44 iron ore microbe and the ore dephosphorizing yield can exceed 80%. The content of phosphorus satisfies the requirement of blast furnace iron making materials, and having comprehensive utilization of iron pyrite. It is pertinent to state that the process [2] provides a reliable technical support for high phosphorus iron ore in countries with low cost and good dephosphorization. Biological processes for phosphorus removal has been evaluated [3] based on the use of several types of fungi, some being oxalic acid producing. Results generated from this research where Aspergillus niger was used as the microbe show that phosphorus removal efficiencies at the end of the 49 days of the leaching process are 81, 63 and 68% for 5, 100 and 250 mesh grain sizes of Agbaja iron ore respectively. Results of experiment and associated derived models [4-6] have shown that dephosphorization of iron oxide ore using hydrometallurgy; using oxalic acid solution, is highly dependent on the final ph of the leaching solution which varies time and other factor such as initial solution ph, initial leaching temperature, mass-input of the iron oxide ore and ore mineralogy etc. These results [6] have indicated that the removal of phosphorus during leaching of iron oxide ore in oxalic acid solution is also achievable based the mass-input of iron oxide ore and the final solution ph. This is because varied mass-inputs of iron oxide ore results to varied final solution ph. An empirical model for predictive analysis of removed phosphorus concentration during leaching of iron oxide ore in sulphuric acid solution has been derived [7]. In this work, both experimental results and results from the predicted model indicate that phosphorus removal from the iron oxide ore is dependent on the initial and final ph of the leaching solution. This is because the final ph of the leaching solution is greatly determined by the initial ph which initially has a specified mole of hydrogen ion. Further studies [8] have shown that the concentration of removed phosphorus during leaching of iron oxide ore in oxalic acid could be dependent on the leaching temperature. The model indicates that at a leaching temperature range C, the maximum deviation of the model-predicted concentration of removed phosphorus from the corresponding experimental values is less than 29%. It has also been shown [7] that phosphorus removal (during leaching of the iron ore in oxalic acid solution) could be dependent on the final ph of the leaching solution and weight input of the iron oxide ore. The model is expressed as; P = 150.5/μ (1) indicates that the concentration of phosphorus removed is inversely proportional to the product of the weight input of the iron oxide ore and the final ph of the leaching solution. Process conditions considered during the formulation of the model [7] include: leaching temperature of 25 0 C, initial solution ph 5.5 and average ore grain size; 150μm). The aim of this work is to carry out a response analysis of phosphorus removal with respect to the operational effects of inoculum concentration and leaching time during biotreatment of iron oxide ore using acidithiobacillus ferrooxidans. The essence of this work is to ascertain the level of dephosphorization achievable at different inoculum concentration where all other input process parameters are kept constant. MATERIALS AND METHODS Agbaja (Nigeria) iron ore concentrate used for this work was obtained from Nigeria Metallurgical Development Centre (NMDC) Jos. The concentrate was dried in air (under atmospheric condition) and used in the as-received condition. Five samples of constant weight quantity of the dried iron ore concentrate of particle size 0.35 mesh size were each added to a culture of ATF in a conical flask of inoculum concentration 0.2 ml and the mixtures allowed to react for 2 weeks at a temperature of 25 0 C after which the mixtures were filtered and the concentration of removed phosphorus determined using atomic absorption 40

45 spectrometer (AAS). The average of the removed phosphorus concentration determined in each experiment set was taken as the precise result. The experiment was repeated with different inoculum concentrations; 0.5, 2, and 3 ml and the corresponding phosphorus removal also determined using AAS. Details of the experimental procedures and process conditions prevailing during the biooxidation process are as reported in the previous work [9]. Model Formulation Experimental data obtained from the highlighted research work were used for the model derivation. Computational analysis of these data shown in Table 1, gave rise to Table 3 which indicate that; ξ - S Nϑ + K (2) Introducing the values of S, N and K into equation (2) ξ = 21.0ϑ (3) ξ = 21.0ϑ (4) Where (ξ) = Conc. of removed phosphorus (%) (ϑ) = Inoculum concentration (ml) ( ) = Leaching time (weeks) S = , N = 21.0 and K = 75.0 are equalizing constant (determined using C- NIKBRAN [10]) Table 1: Variation of removed phosphorus concentration with inoculum constant leaching time [9] Boundary and Initial Condition ( ) (ϑ) (ξ) (%) Consider iron ore (in a furnace) mixed with a culture of ATF. The atmosphere was not contaminated i.e (free of unwanted gases and dusts). Initially, atmospheric levels of oxygen are assumed just before the reaction between the ore and the microbes. Mass of iron oxide ore: (50g), particle size: 0.35 mesh size, treatment time: 2 weeks, constant treatment temperature: 25 o C, range of inoculum concentrations used; The boundary conditions are: oxygen atmosphere at the top and bottom of the ore particles interacting with the microbes. At the bottom of the particles, a zero gradient for the gas scalar are assumed and also for the gas phase at the top of the particles. The reduced iron is stationary. The sides of the particles are taken to be symmetries. RESULTS AND DISCUSSIONS The result of the chemical analysis carried out on the beneficiated iron ore concentrate is presented in Table 2. The table shows that the percentage of phosphorus present in the as-beneficiated ore is 0.9%. 41

46 Table 2: Result of chemical analysis of iron ore used [9] Element/Compound Fe P SiO 2 Al 2 O 3 Unit (%) Model Validation The validity of the model is strongly rooted in equation (3) (core model equation) where both sides of the equation are correspondingly approximately equal. Table 3 also agrees with equation (3) following the values of ξ and 21.0ϑ evaluated from the experimental results in Table 1. Table 3: Variation of ξ with 21.0ϑ ξ ϑ Furthermore, the derived model was validated by comparing the removed phosphorus concentration predicted by the model and that obtained from the experiment. This was done using various evaluative techniques such as computational, statistical, graphical and deviational analysis. Computational Analysis Computational analysis of the experimental and model-predicted removed phosphorus concentration was carried out to ascertain the degree of validity of the derived model. This was done by comparing phosphorus removal per unit inoculum concentration valuated from model-predicted results with those from actual experimental results. Removed phosphorus concentration per unit inoculum concentration (%) was calculated from the equation; ξ I = ξ / ϑ (5) Therefore, a plot of the concentration of phosphorus removed against inoculum concentration as in Fig. 1 using experimental results in Table 1, gives a slope, S at points (0.2, ) and (3, ) following their substitution into the mathematical expression; ξ I = Δξ / Δϑ (6) Equation (6) is detailed as ξ I = ξ 2 ξ 1 / ϑ 2 - ϑ 1 (7) Where Δξ = Change in removed phosphorus concentrations of ξ 2, ξ 1 at two values of the inoculum concentrations ϑ 2, ϑ 1. Considering the points (0.2, ) and (30, 769) for (ϑ 1, ξ 1) and (ϑ 2, ξ 2) respectively, and substituting them into equation (7), gives the slope as % /ml which is the removed phosphorus concentration per unit inoculum concentration during the actual leaching process. 42

47 Conc. of Phosphorus (%) Conc. of Phosphorus (%) American Journal of Engineering and Technology Research Vol. 16, No.1, R 2 = Inoculum Conc. (ml) Fig. 1: Coefficient of determination between concentration of removed phosphorus and inoculum concentration as obtained from the experiment [9] Similarly a plot of the concentration of removed phosphorus against inoculum concentration (as in Fig. 2) using derived model-predicted results gives a slope: % / ml on substituting the points (0.2, ) and (30, ) for (ϑ 1, ξ 1) and (ϑ 2, ξ 2) respectively into equation (7). This is the model-predicted removed phosphorus concentration per unit inoculum concentration. It is very pertinent to state that the actual removed phosphorus concentration per unit inoculum concentration (as obtained from experiment and derived model) was just the magnitude of the signed value. The associated sign preceding these values signifies that the associated slope tilted to negative plane. Based on the foregoing, removed phosphorus concentration per unit inoculum concentration as obtained from experiment and derived model were and 21.0 % / ml respectively R 2 = Inoculum Conc. (ml) Fig. 2: Coefficient of determination between concentration of removed phosphorus and inoculum concentration as obtained from derived model 43

48 Conc. of Phosphorus (%) American Journal of Engineering and Technology Research Vol. 16, No.1, 2016 A comparison of this set of values for removed phosphorus concentration (per unit inoculum concentration) also shows proximate agreement and a high degree of validity of the derived model. Statistical Analysis The standard errors (STEYX) in predicting the removed phosphorus concentration (using results from experiment and derived model) for each value of the inoculum concentration are and x 10-7 % respectively. The standard error was evaluated using Microsoft Excel version Considering the coefficient of determination R 2 from Figs. 1 and 2, the correlations between removed phosphorus concentration and inoculum concentration as obtained from experiment and derived model predicted results, was calculated using the equation; R = R 2 (8) The evaluations show correlations and respectively. These evaluated results indicate that the derived model predictions are significantly reliable and hence valid considering its proximate agreement with results from actual experiment. Graphical Analysis Comparative graphical analysis of Fig. 3 shows very close alignment of the curves from model-predicted removed phosphorus concentration (MoD) and that of the experiment (ExD). The degree of alignment of these curves is indicative of the proximate agreement between both experimental and model-predicted phosphorus removed concentration ExD MoD Inoculum Conc. (ml) Fig. 3: Comparison of the concentrations of removed phosphorus (relative to inoculum concentration) as obtained from experiment [9] and derived model Deviational Analysis Analysis of removed phosphorus concentrations from the experiment and derived model revealed deviations on the part of the model-predicted values relative to values obtained from the experiment. This is attributed to the fact that the surface properties of the iron ore and the physiochemical interactions between the ore and the bacteria (ATF) which were found to have played vital roles during the process were not considered during the model formulation. This necessitated the introduction of correction factor, to bring the model-predicted removed phosphorus concentration to those of the corresponding experimental values. 44

49 Conc. of Phosphorus (%) Deviation (%) American Journal of Engineering and Technology Research Vol. 16, No.1, 2016 Deviation (Dn) of model-predicted removed phosphorus concentration from that of the experiment is given by Dn = Pv Ev x 100 (9) Ev Where Pv = Removed phosphorus concentration as predicted by derived model Ev = Removed phosphorus concentration as obtained from experiment Correction factor (Cr ) is the negative of the deviation i.e Cr = -Dn (10) Therefore Cr = - Pv Ev x 100 (11) Ev Introduction of the corresponding values of Cr from equation (11) into the derived model gives exactly the removed phosphorus concentration as obtained from experiment MoD Deviation Inoculum Conc. (ml) Fig. 4: Variation of model-predicted removed phosphorus concentration with associated deviation from experimental results (relative to inoculum concentration) Fig. 4 shows that the maximum deviation of the model-predicted removed phosphorus concentration from the corresponding experimental values is less than 8%. The figure shows that the least and highest magnitudes of deviation of the model-predicted removed phosphorus concentration (from the corresponding experimental values) are and % which corresponds to removed phosphorus concentrations: and %, as well as inoculum concentrations: 0.2 and 2 respectively. 45

50 Conc. of Phosphorus (%) Correction factor (%) American Journal of Engineering and Technology Research Vol. 16, No.1, MoD Corr.factor Inoculum Conc. (ml) ` Fig. 5: Variation of model-predicted removed phosphorus concentration with associated correction factor (relative to inoculum concentration) Comparative analysis of Figs. 4 and 5 indicates that the orientation of the curve in Fig. 5 is opposite that of the deviation of model-predicted removed phosphorus concentration (Fig. 4). This is because correction factor is the negative of the deviation as shown in equations (10) and (11). It is believed that the correction factor takes care of the effects of surface properties of the iron ore and the physiochemical interactions between the iron ore and bacteria which have played vital roles during the process, but were not considered during the model formulation. Fig. 5 indicates that the least and highest magnitudes of correction factor to the model-predicted removed phosphorus concentration are 2.27 and % which to removed phosphorus concentrations: and %, as well as inoculum concentrations: 0.2 and 2 respectively. It is important to state that the deviation of model predicted results from that of the experiment is just the magnitude of the value. The associated sign preceding the value signifies that the deviation is a deficit (negative sign) or surplus (positive sign). CONCLUSIONS Response analysis of phosphorus removal to operational effects of inoculum concentration and leaching time concentration was carried out during biotreatment of iron oxide ore using Acidithiobacillus Ferrooxidans. The validity of the empirical model derived, validated and used for the analysis was rooted on the expression ξ = ϑ where both sides of the expression are correspondingly approximately equal. Statistical analysis of removed phosphorus concentration for each value of the inoculum concentration as obtained from experiment and derived model-predicted results show standard errors of and x 10-7 % respectively. Furthermore, removed phosphorus concentration per unit inoculum concentration as obtained from experiment and derived model-predicted results were and 21.0 % / ml respectively. Deviational analysis indicates that the maximum deviation of model-predicted removed phosphorus concentration (from experimental result) was less 8%. This translates into a derived model confidence level of above 92% and response coefficient of dependence of phosphorus removal on the inoculum concentration and leaching time. 46

51 REFERENCES [1] Cheng, C. Y., Misra, V. N., Clough, J., Mun, R., (1999). Dephosphorisation of Western Australian iron ore by Hydrometallurgical process. Miner. Eng. 12, pp [2] Patent Publication No , [3] Anyakwo, C. N., and Obot, O.W. (2008), Phosphorus Removal from Nigeria`s Agbaja Iron Ore by Aspergillus niger, Inter. Res. J. Eng. Sc. Tech. 5(1), pp [4] Nwoye, C. I., C. N. Mbah., C. C. Nwakwuo., and A. I. Ogbonna. (2010). Model for Quantitative Analysis of Phosphorus Removed during Leaching of Iron Oxide Ore in Oxalic Acid Solution. J. Min. Mater. Charact. Eng. 9(4), pp [5] Nwoye, C. I. (2009). Model for Evaluation of the Concentration of Dissolved Phosphorus during Leaching of Iron Oxide Ore in Oxalic Acid Solution. J. Min. Mater. Charact. Eng. 8(3), pp [6] Nwoye, C. I., Agu, P. C., Mark, U., Ikele, U. S., Mbuka, I. E., and Anyakwo, C. N. (2008). Model for Predicting Phosphorus Removal in Relation to Weight of Iron Oxide Ore and ph during Leaching with Oxalic Acid. Inter. J. Nat. Appl. Sc. 4(3), pp [7] Nwoye, C. I. and Ndlu, S. (2009). Model for Predictive Analysis of the Concentration of Phosphorus Removed during Leaching of Iron Oxide Ore in Sulphuric Acid Solution J. Min. Mater. Charact. Eng. 8(4), pp [8] Nwoye, C. I.. (2009). Model for Predicting the Concentration of Phosphorus Removed during Leaching of Iron Oxide Ore in Oxalic Acid Solution. J. Eng. Appl. Sc. (in press) [9] Ameh, E. M. (2013). Dephosphorization of Agbaja Iron Ore By Sintering And Bioleaching Processes. Ph.D Thesis, Nnamdi Azikiwe University Awka, Anambra State, Nigeria. [10]Nwoye, C. I. (2008). C-NIKBRAN; Data Analytical Memory. 47

52 PERFORMANCE COMPARISON OF SPEECH RECOGNITION FOR VOICE ENABLING APPLICATIONS - A STUDY V. Karthikeyan 1 and V. J. Vijayalakshmi 2 1 Department of ECE, VCEW, Thiruchengode, Tamilnadu, India, Karthick77keyan@gmail.com 2 Department of EEE, KPRCET, Coimbatore Tamilnadu, India, vijik810@gmail.com Abstract - Performance Comparison of Speech Recognition for Voice Enabling Applications is presented. This paper focuses on the speaker independent system to provide all the mobile phone applications without touching the device especially for those visually challenged. The proposed system is evaluated through two different classification modeling such as Template generation using Dynamic time warping (DTW) and Hidden Markov Model (HMM) / Vector Quantization (VQ) with Mel Frequency Cepstral Coefficient (MFCC) features. The performance comparison between two classification modeling is made based on the average recognition accuracy Hidden Markov Model (HMM) / Vector Quantization (VQ) classification modeling with Mel Frequency Cepstral Coefficient (MFCC) features gives recognition rate of 82.77% for the utterances which is higher than the conventional methods. TMS320C2x DSP processor can be used to implement voice enabled mobile phone. Index Terms- Template generation, Dynamic time warping (DTW), Hidden Markov Model (HMM) / Vector Quantization (VQ) and Mel Frequency Cepstral Coefficient (MFCC) INTRODUCTION Speech recognition is a field in which the system recognizes the spoken words [1]. It enables the system to identify the words that a person speaks into microphone and recognizes them by converting them into written text. Automatic Speech Recognition (ASR) is the process of determining a sequence of words spoken by human using machines. The goal of ASR is to have speech as a medium of interaction between man and machine. This Automatic Speech Recognition (ASR) is commonly used in voice enabled mobile phones implemented using voice as an input. In these systems, it is necessary to perform recognition of voice input so that the appropriate action can be enabled. Voice enabled mobile phone refers to the enabling of mobile applications where a user may dial a contact in the device memory by saying the digit of the contact which is saved in speed dial or name of the contact[2], sending and receiving the message and making the phone to put on hold mode and loudspeaker mode. The first step in the process of voice enabling application is speech recognition. To do so, it is necessary to extract the features from the uttered word. Recognition of the uttered word is done by means of comparing both trained sample and test sample. The most commonly used feature extraction algorithm is Mel frequency Cepstral coefficient front end. The power frequency data is then filtered with filter that resembles the human ear s sensitivity curve called Mel scale Filter. While the human ear is sensitive to frequency variations in lower frequency values, this sensitivity is reduced for higher frequency 48

53 signal components. From the experimental results, it is well known that Mel Frequency Cepstral Coefficients (MFCC) is among the best acoustic features used in automatic speech recognition [3]. The Mel Frequency Cepstral Coefficients are robust, contain much information about the vocal tract configuration regardless the source of excitation, and can be used to represent all classes of speech sounds. This paper is organized as follows: In section 2, the general principles of ASR system have been discussed. Section 3 deals with the Mel Frequency Cepstral Coefficient (MFCC) feature extraction technique which is used in the proposed system and section 4 deals with two classification modeling. In section 5 results and conclusion are discussed which shows the accuracy of two classification modeling. AUTOMATIC SPEECH RECOGNITION SYSTEM In general, the Automatic Speech Recognition (ASR) system [4] consists of two modes such as training mode and testing mode. First the input speech signal is pre-processed and then features are extracted. From the extracted features reference samples are created from which the comparison and recognition is made. The general block diagram of Automatic Speech Recognition (ASR) system is shown in figure.1. Training Mode: In speaker dependent and speaker independent system, the system has to be trained by the speaker. It means that samples have to be collected from different speakers using microphone as an input device Accuracy is directly proportional to the no of samples Preprocessing methods are used to extract acoustic characteristics of the speech signal. Software analyses and generates patterns from the extracted feature vectors and stores it in matrix form as a reference pattern. These reference patterns are matched with the input speech during the recognition mode. TESTING MODE: In this mode, the test sample is analyzed for its acoustic characteristics and the important features are extracted from the input speech sample. This feature vectors are used to generate an input pattern using and stores it in matrix form. This unknown pattern is compared against known reference pattern, element by element. Once the best match is found, the appropriate action is enabled. Fig. 1.Block Diagram of Automatic Speech Recognition System FEATURE EXTRACTION Feature extraction is the key to the front-end process in speaker verification systems. The performance of a speaker verification system is highly dependent on the quality of the selected 49

54 speech features. The speech signal is a slowly varying signal and is often termed stationary. Therefore, short-time spectral analysis is the most common way to characterize speech signal. Before extracting the features the speech signal is preprocessed using following steps. i) Framing ii) Windowing, The speech signal is divided into short fixed length frames. The continuous speech signal is divided into frames where each frame consists of M samples [5]. Very often successive frames are overlapping with each other by M samples. For the proposed system, frame size of M = 256 with an overlap of 50% i.e. N=128 have been used. After frame segmentation, windowing is carried out to minimize the spectral distortion by using the window to taper the signal on both ends thus reducing the side effects caused by signal discontinuity at the beginning and at the end due to framing. Hamming window is used as spectral leakage is less. It is multiplied with each frame and the window function is as given in eqn. (1), (1) Where M is the number of samples in each frame There are different types of feature extraction techniques available such as Mel-Frequency Cepstral Coefficient (MFCC), Linear Prediction Cepstral Co-efficient (LPCC), Bark Frequency Cepstral Co-efficient (BFCC), Perceptual Linear Prediction (PLP) and Rasta Perceptual Linear Prediction. Among those Mel Frequency Cepstral Coefficient (MFCC) provides good recognition accuracy. Mel-Frequency Cepstral Coefficient (MFCC) vectors are used to provide an estimate of the vocal tract filter [6]. Background noise energy level is evaluated at the beginning and the end of speech signal and energy thresholds are applied to find speech beginning and end points. The pre emphasized speech signal is blocked into frames of N = 256 samples, with adjacent frames separated by M = 128 samples. Then windowing is done to minimize the speech signal discontinuities at the beginning and end of each analysis frame. The general block diagram of Mel Frequency Cepstral Coefficient (MFCC) is shown in figure.2 Fig. 2. Block Diagram of MFCC Technique After windowing, Fast Fourier Transform (FFT) is applied. Then a spectrum is passed through 20 Mel scale triangular filter bank. The Mel scale is a critical band frequency scale that takes into account the frequency perception in the human auditory system. Discrete Cosine Transform (DCT) is applied to the log Mel scale filter outputs and thus 12 Mel-frequency Cepstral coefficients (MFCC) are obtained. Then Cepstral filtering is performed with the help of eqn. (2), (2) 50

55 Where N is the number of filter bank channels. Mel-Frequency Cepstral Coefficients (MFCCs) are calculated from the log filter bank amplitudes {mj} using the eqn. (3) Where N is the number of filter bank channels. SPEAKER MODELLING As the extracted features requires more storage memory, it is necessary to convert those features into vectors so that storage memory requirement can be met which can be done by using classification modeling. The proposed system is verified through Template model and Hidden Markov Model (HMM) / Vector Quantization (VQ). The block diagram of the proposed system is as shown in figure.3. (3) Fig.3. Block Diagram of Proposed System Template generation modeling uses Dynamic Time Warping (DTW) for speech pattern matching [7] for speaker dependent system in which it expands or contracts the time axis non-linearly to match the input speech with the reference template. The reason here to use Dynamic Time Warping (DTW) algorithm is that Dynamic Time Warping (DTW) -based recognition engine has been widely embedded inside Qualcomm MSM (Mobile Station Modem) chips because of its less computational complexity for phone dialing. Template Generation using DTW Templates are reference patterns that are derived from features recorded over the length of the whole word rather than at particular points. The procedure is as follows: Choose one utterance from the training data for reference. Use Dynamic Time Warping (DTW) technique to align all the training data to match with the reference. Once the training data are aligned, compute the reference pattern vector as the Centroid of the feature vectors (of cepstral coefficients) corresponding to all the occurrences of the digit. Dynamic Time Warping is used to create reference templates and to find the best match between the reference template and the input template derived from the test input speech sample. The matching process needs to compensate for length differences and take account of the non-linear 51

56 nature of the length differences within the words. Dynamic Time Warping (DTW) grid is used to find the best match between input data and stored sequence. We can find a path through the grid which minimizes the total distance between them. The input data is either stretched or compressed in order to match with the input. Once an overall path has been found, the total distance between the input sequence and reference sequence can be calculated for this particular input template. In Dynamic Time Warping method (DTW), when comparing sequences with different length, the sequence length is modified by repeating or omitting some frames, so that both sequences will have the same length. This modification of sequences is called time warping. Hidden Markov Model with Vector Quantization (HMM/VQ) Here the proposed method, speaker independent isolated word recognition is implemented using Vector Quantization (VQ) and Hidden Markov model (HMM) which is suitable to provide higher accuracy rate with more no of samples. One of the most popular approaches to speaker independent [8] speech recognition, is the combination of Vector Quantization (VQ) for the encoding of segments of speech with a Hidden Markov Modelling (HMM) for the classification of sequences of segments [9] as in figure.4. After extracting the features, K-means clustering algorithm is used to iteratively create the vector quantizer codebook until the average distance falls below a preset threshold. A set of such vectors (corresponding to multiple utterances of the same word) is used to re-estimate the Hidden Markov Model [10] for that word. Fig.4. Block Diagram of Hidden Markov Model with Vector Quantization After extracting the features, K-means clustering algorithm is used to iteratively create the vector quantizer codebook until the average distance falls below a preset threshold. A set of such vectors (corresponding to multiple utterances of the same word) is used to re-estimate the Hidden Markov Model [10] for that word. This procedure is repeated for each word in the vocabulary. In the testing mode, the set of Mel Frequency Cepstral Coefficient (MFCC) vectors corresponding to the unknown word is quantized by the vector quantizer to give a vector of codebook indices. This is scored on each word Hidden Markov model (HMM) to give a probability score for each word model. The decision rule is used to choose the word whose model gives the highest probability. VQ is a process of mapping vectors from a large vector space to a finite number of regions in that space. Each region is called a cluster and can be represented by its center called a centroid. The collection of all code words is called a codebook [11]. In speech recognition, vector quantization can be used to train Hidden Markov model (HMMs). RESULTS AND DISCUSSION The proposed speaker independent isolated speech recognition system is implemented in 52

57 MATLAB software. Database containing 0-9 digits and 8 words Dial, Cut, On, Off, Hold, Read, Write and Loudspeaker are considered for voice dialing applications. The digits are used to dial a contact no assigned for speed dial. The words like Dial, Cut and On, Off and Hold are used to control voice dialing application and Mobile phone on and off operation respectively. The words like Read, Write and Loudspeaker are used to enable message read and write application and hands free mode respectively. Database is collected from 25 speakers pronouncing each word 10 times yielding 250 utterances per word and 180 utterances per speaker, totally 4500 samples. The first 8 utterance of each word is used for training and remaining is used for testing. In Training phase, the uttered digits are recorded using 8-bit Pluse Code Modulation (PCM) with a sampling rate of 8 KHz, which holds the information of the communication and converted as a wave file using Total audio converter software The performance of the speech recognition systems is given in terms of a word error rate (%) and is calculated from eqn.4, W= (M/N) x100% (4) Where: N = Total no of samples taken M = Total no of recognized samples In the confusion matrix, each uttered word is compared with all other words and it shows how many times each word was correctly recognized and from which the average recognition accuracy is calculated. The Table.1 shows the result of speech recognition system using Template generation modeling with Mel Frequency Cepstral Coefficient (MFCC) features. As shown in Table.1, digit 0 and 2 have the highest accuracy rate of 60% and digit 4 & 9 and the word cut have the least accuracy rate of 50%. The word off has the second highest accuracy rate of 58%. In Hidden Markov Model (HMM) / Vector Quantization (VQ) model, codebook is generated using Vector Quantization (VQ) method. In this codebook generation, some n no of clusters are generated after different no of iterations as shown in Table 2. As shown in Table.3, digit 0 has the highest accuracy rate of 88% and digit 2 has the least accuracy rate of 78%. Digit 9 and dial has the next highest accuracy rate of 87%. Words hold and off have the accuracy rate of 85%. Words Loudspeaker and on have the accuracy rate of 84%. Digit 5, 8 and word write have the accuracy rate of 83% Digit 3 and word cut have the accuracy rate of 82%. Digit 6 has the accuracy rate of 81% followed by the digit 4 and word write which have the accuracy rate of 80%. Comparison between the recognition rates of two different classification modeling with Mel Frequency Cepstral Coefficient (MFCC) features is shown in Table.4 which implies that Hidden Markov Model (HMM) / Vector Quantization (VQ) has higher recognition rate than Template generation modeling CONCLUSION The proposed system designed using Hidden Markov Model (HMM) / Vector Quantization (VQ) classification modeling with Mel Frequency Cepstral Coefficient (MFCC) features gives recognition rate of 82.77% for the utterances which is higher than Template generation modeling using Dynamic time warping (DTW) which is of 53.88%. Further it can be implemented using DSP processor for hardware implementation so that voice enabling applications can be made. TMS320C2x dsp processor can be used to implement voice enabled mobile phone. 53

58 54

59 References [1] Rabiner, L.R. and R.W. Schafer, Prentice-Hall Inc. Juha Iso-Sipila, Design and Implementation of a Speaker-Independent Voice Dialing System: A Multi- Lingual Approach Ph.D., 1978, Digital Processing of Speech Signals, April [2] Tilo Schurer, An Experimental comparison of different feature extraction for Telephone speech, in proc of 2 nd IEEE workshop on interactive voice technology for 55

60 Telecommunication applications, 1994 [3] Jason Chong and Roberto Togneri, Speaker Independent Recognition of Small Vocabulary, M.S.Thesis, Centre for Intelligent Information Processing Systems, the University of Western Australia [4] L. Rabiner and B. H. Jaung, Fundamentals of Speech recognition, Prentice Hall Englewood Cliffs, New Jersey, [5] S.B. Davis and P. Mermelstein, Comparison of Parametric representations for Monosyllabic word recognition in continuously spoken sentences, IEEE Transactions on Acoustics, Speech, Signal Processing, vol. ASSP-28(4), pp , August 1980 [6] Itakura F. (1975). Minimum prediction residual applied to speech recognition. IEEE Trans. Acoustics, Speech, Signal Proc., ASSP-23 (1), [7] Schafer, R.W. Scientific bases of human-machine communication by voice in Proceedings of the National Academy of Science. Vol. 92, [8] Hoshimi M., Miyata M., and Hiraoka S., Speaker independent speech recognition method using training from a small number of speakers, in proc IEEE international conference on acoustics, speech and signal processing, pp ,1992. [9] Nilsson M., Ejnarsson M., Speech Recognition Using HMM: Performance Evaluation in Noisy Environments, MS Thesis, Blekinge Institute of Technology, Department of Telecommunications and Signal Processing, 2002 [10] Ferrer M.A., Alonso I., Travieso C., Influence of initialization and Stop Criteria on MM based recognizers Electronics letters of IEEE, Vol. 36, pp , June [11] Rabiner L.R., Levinson S.E., Rosenberg A.E., Wilson J.G., "Speaker independent Recognition of isolated words using clustering techniques, IEEE Trans. Acoustic Speech Signal Process Vol.27, pp ,

61 THE EFFECT OF BIT ERROR RATE ON HIPERLAN/2 USING 16 QAM, 64 QAM AND 256 QAM 1 B.O.Omijeh and 2 S. A. Adegoke 1 Department of Electronic and Computer Engineering, 2 Centre for Information and Telecommunication Engineering, University of Port Harcourt, Choba, Rivers State, Nigeria bourdillon.omijeh@yuniport.edu.ng; omijehb@yahoo.com Abstracts With today rise in the demand for mobile communication services, tomorrow generations of mobile systems are expected to provide very high data rate transmissions as well as improved quality of service to users within commercial feasibility so as to avoid poor reception which the users are experiencing nowadays. This paper shows the effects of Bit error rate on hiperlan2 using 256 QAM modulation techniques.the developed model was simulated using MATLAB application package. OFDM-256QAM modulation format was employed to provide the parallel transmission and the received signal was compared withthe transmitted signal to analyse the bit error rate performance of the OFDM system.it was discovered that If the medium between the transmitter and receiver is good and the signal to noise ratio is high, then the bit error rate will be very small - possibly insignificant and having no noticeable effect on the overall system.the increase in bit error rate indicates performance degradation. Keywords: OFDM, BER, SNR, MATLAB, Eb/No. QAM, HIPERLAN/2 1.0 INTRODUCTION One of the developments that modern digital communications systems has brought to radio engineering is the need for end-to-end performance measurements. The measure of that performance is usually bit-error rate (BER), which quantifies the reliability of the entire radio system from bits in to bits out, including the electronics, antennas and signal path in between. Designing the first version of the standard named HiperLAN/1, started 1991, when designing was by now going on. The aim of the HiperLAN was the high data rate, higher than The standard was accepted in The functional specification is EN300652, the rest is in ETS The standard covers the Physical layer and the Media Access Control part of the Data link layer like There is a new sub layer named Channel Access and Control sub layer (CAC). This sub layer deals with the access requests to the channels. The achievement of the demand is dependent on the usage of the channel and the precedence of the request. CAC layer provides hierarchical independence with Elimination-Yield Non-Pre-emptive Multiple Access mechanism (EY-NPMA). EY-NPMA codes precedence choices and other functions into one variable length radio pulse preceding the packet data. EY-NPMA enables the network to function with few collisions even though there would be a large number of users. Multimedia applications work in HiperLAN because of EY-NPMA precedence mechanism. MAC layer defines protocols for routing, security and power saving and provides naturally data transfer to the upper layers. On the physical layer FSK and GMSK modulations are used in HiperLAN/1. 57

62 HiperLAN/2 functional specification was achieved February Version 2 is designed as a fast wireless connection for many types of networks. Those are UMTS back bone network, ATM and IP networks. Also it works as a network at home similar to HiperLAN/1. HiperLAN/2 uses the 5 GHz band and up to 54 Mbit/s data rate[1]. The physical layer of HiperLAN/2 is very similar to IEEE a wireless local area networks. However, the media access control (the multiple access protocol) is Dynamic TDMA in HiperLAN/2, while CSMA/CA is used in a. Basic services in HiperLAN/2 are data, sound, and video transmission. The accenting is in the quality of these services (QoS). The standard covers Physical, Data Link Control and Convergence layers. Convergence layer takes care of service dependent functionality between DLC and Network layer (OSI 3) [2]. 2.0 RELATED WORK A commonly used method to estimate the quality of a wireless link is to compute the BER using pilot symbols. Pilot symbols represent a predefined sequence of symbols, which are known at the transmitter and the receiver side. Therefore, the BER can be computed from these pilot symbols. For example, HiperLAN/2 as well as third generation telephony uses pilot symbols [2]. This approach has two disadvantages. First, the transmission of the pilot symbols introduces overhead. Second, the BER is only computed over a small amount of the total bits that are transmitted [3]. 2.1 Bit Error Rate (BER) Bit error rate is a key parameter that is used in assessing systems that transmit digital data from one location to another. BER is applicable to radio data links, Ethernet, as well as fibre optic data systems. When data is transmitted over a data link, here is a possibility of errors being introduced into the system. If this is so, the integrity of the system may be compromised. As a result, it is necessary to assess the performance of the system, and BER provides an ideal way in which this can be achieved. BER assesses the full end to endperformance of a system including the transmitter, receiver and the medium between the two. BER is defined as the ratio of the number of wrong bits over the number of total bits.in simple form, BER = Eqn.1 Using a Matlab simulation model for HIPERLAN/2 with ½ and ¾ code rates the BER performance for different modulation schemes is compared. It can be easily seen that the BER curves for all modulations with ½ code rates are less degraded as compared to that with ¾ code rate. Also the SNR is improved for ½ code rate that for ¾ code rate [4] 2.2 Factors affecting Bit error rate, BER Interference: The interference levels are set by external factors which are present in a system and these levels cannot be changed by the main system design. Whether it may be possible to set the system-bandwidth. If we reduce the bandwidth, then the level of interference will be reduced by good amount. 58

63 Reduce bandwidth: Another way that can be adopted to diminish the BER (bit error rate) is to reduce the bandwidth. Increase transmitter power: If the power level of the system is boosted, then the power per bit is improved. It would be balanced against interference levels factors. BER(Bit error rate) is a parameter which provides us an exceptional clue for the performance of a data link. Number of errors is one of the main parameters in any data link, But the BER(bit error rate) is a key parameter. One should have knowledge of the BER so that it gives information on other features like bandwidth and power. 3.0 Simulation Model The study of bit error rate effect on hiper LAN/2 was done by modifying an existing model in Simulink (Matlab ). This model was designed to show the effect of bit error rate on hiper LAN/2 using 64-QAM and 256-QAM modulation techniques with high power amplifier (HPA). Fig.1 and Fig.2 represent the whole system model for the proposed design. The Rapps model[5] and Linear Model of Travelling wave Tube was selected as HPA model in the system and it is connected after the OFDM transmitter. HPA s are used to amplify the power of the signal before sending it on the channel. This block of HIPERLAN/2 with RAPP S amplifier and LINEAR amplifier are shown in Fig.1 and fig.2 respectively. The input data (random binary data) were mapped unto 64-QAM and 256-QAM constellation. The network is generated using the data of OFDM transmitter and receiver[6-8]. Each subcarrier was assigned a particular number of bits and the subcarriers were transmitted in parallel. The received bits were compared with the transmitted bits in order to calculate the BER [9-11] The following parameters and system configurations were used: Modulation: 64, 256 QAM Data rate: 54Mbps and 72Mbps FFT size: 64 Number of subcarriers: 32, 64 and 128 OFDM Frame size: 80 Guard interval type: Cyclic prefix Number of paths: 2 Path delay vector: [ ] Distance: 2km Carrier frequency: 900MHz Amplifier Model: Rapp s Model, Linear Model. Linear Model:0-20db Table 1: Table of mode dependent physical layer parameters of HIPER LAN/2 [4] MODULATION CODING RATE BITS PER SYMBOL BPSK ½ 1 6 BPSK ¾ 9 NORMAL RATES(Mbps) BIT 59

64 QPSK ½ 1 12 QPSK ¾ QAM 9/ QAM ¾ QAM ¾ QAM ¾ 8 72 Fig.1: Simulation Block Diagram for 64-QAM with HPA (Mathworks, 2010) Fig.2: Simulation Block Diagram for 256-QAM with HPA (Mathswork,2010) The simulation was carried out by varying the number of subcarriers and mobile speed parameters and their BER performances at Eb/No of 0 to 20dB were obtained. 60

65 Addition of extra zero bits in the OFDM symbol is used to avoid aliasing. Cyclic prefixing can be implemented by adding the last few bits of a symbol at the beginning of the symbol and is used for both timing and frequency synchronization. On the receiver side, most of the functions are just the opposite of the equivalent transmitter blocks. The time domain signal is converted into the frequency domain by the FFT and symbols are extracted by a demodulator. Removal of pilot carriers, frame synchronization and elimination of cyclic prefixes are performed beforehand in the receiver block. After de-normalisation, frames are passed through a de-interleaving process. Viterbi algorithm is used to decode convolution ally encoded input data. With the Viterbi algorithm, the zero-valued dummy bit has no effect on the outcome of the decoder. Finally the received data bits are compared to the transmitted bits by a bit error calculator. The following steps were taken to analyse the Hiper Lan/2 model alongside various QAM outcomes: Variation of different QAM modulation schemes such as 64 and 256. Variation of simulation time to determine its effects on the simulation results. Variation of signal to noise ratio to determine its effects on simulation results. Introducing amplifier to the model to see its effect on the simulation result. Variation of different model of the high power amplifier and compare it result with when it was not introduced. 4.0 RESULTS AND DISCUSSION The main purpose of the simulation was to design a base-band HIPERLAN/2 model and observe its behaviour under different channel conditions. The aim of doing this simulation is to measure the performance of 64, 256-OFDM with BER and SNR calculations, under 64-QAM and 256- QAM modulation techniques. According to the standard, the highest data rate is 54 Mbit/s, which should be obtained using 64-QAM and a coding rate of 3/4 achieved using a puncturing rate of 2/3. However, 256-QAM was introduced into the simulation with a coding rate of ¾ using a puncturing rate of 3/4 to obtain a higher data rate of 72Mbit/s.Math lab/simulink modelling demonstrated that 64 QAM mode with coding rate ¾, a data rate of 54 Mbit/s is achieved with 15db SNR improvement compared to 256 QAM coding rate ¾ (with puncturing) as required for the maximum 72Mbit/s. The advantage of using QAM is that it is a higher order form of modulation and as a result it is able to carry more bits of information per symbol. By selecting a higher order format of QAM, the data rate of a link can be increased. 61

66 (a) Transmitted Signal (b) Received Signal Fig.3: Simulation result of HIPER LAN/2 with Rapp Amplifier Fig 3(a) and (b) shows the transmitted and received signals and (c) shows the spectrum plot of HIPER LAN/2 with Rapp model of HPA. After including amplifier in an effect of nonlinearity can be seen on the received constellation points. The received signal is more scattering. The Bit Error Rate Calculation block gave varying results as the value of the SNR were varied from 1dB to 20 db which was shown in Table 2. The first value obtained for SNR of 1dB gave a high error rate of , and for 16 QAM, 64 QAM and 256 QAM respectively. As the value of the SNR increases from 1dB to 15 db, there was a significant reduction in the rate at which error was occurring in the calculation. The number of errors was significantly low. When the SNR was further increased to 16dB down to 20 db for 16QAM the bit error rate was zero which implies that no errors occur in the transmission. Table 2:Simulation values of 16 QAM, 64 QAM & 256 QAM against SNR. S/N Simulation Time SNR 16 BER QAM 64 BER QAM BER QAM

67 BER BER American Journal of Engineering and Technology Research Vol. 16, No.1, E E BER VS SNR PLOT OF HIPER LAN 2 FOR 256 QAM WITHOUT AND WITH LINEAR MODEL SNR WITHOUT HPA HPA RAPP'S MODEL Fig.4 Comparative Analysis of 16 QAM, 64 QAM and 256 QAM. Fig.5. BER VS SNR PLOT OF HIPER LAN2 FOR 256 QAM WITHOUT AND WITH LINEAR MODEL BER VS SNR PLOT OF HIPER LAN 2 FOR 256 QAM WITHOUT AND WITH LINEAR MODEL SNR 63 WITHOUT HPA HPA RAPP'S MODEL

68 Fig.4 shows simulation time vs Bit Error Rate for 16-QAM, 64-QAM and 256-QAM. As can be seen from the graphof 64-QAM, it increases and then decreasesat 0.05 as the simulation time increases over a constant distance. While for 256-QAM the graph increases as the simulation time increases over a constant distance. Fig.5 shows BER vs SNR plot of Hiper/Lan 2 for 16-QAM, 64-QAM and 256-QAM.The BER for 32, 64 and 128 subcarriers were observed. Furthermore for 16dB down to 20dB for 64QAM and 256QAM the bit error rate was constant which implies that at higher SNR the BER will be constant. Bandwidth calculation for 256 QAM is 72 Mb/s. The simulation results obtained show that for certain Eb/No and QAM, the more the number of SNR increases, the more the BER also increases. The spectrum is much wider after the signal passes through nonlinear amplifier. Figure 4.3 shows that the curve for HIPERLAN/2 with Rapp s model is much more degraded than the curve without model. Figure 4.4 shows that the curve after passing through the amplifier is much more degraded than the curve without amplifier. 5.0 CONCLUSION In this paper, It was discovered that If the medium between the transmitter and receiver is good and the signal to noise ratio is high, then the bit error rate will be very small - possibly insignificant and having no noticeable effect on the overall system. However if noise can be detected, then there is chance that the bit error rate will need to be considered.lower order modulation schemes can be used, but this is at the expense of data throughput. Higher modulation schemes can be used, but this is at the expense of noise.the results obtained in this work will serve as a useful reference for mobile wireless communication systems designers and researchers working on HIPER LAN2 for high data rate transmissions.the increase in bit error rate indicates performance degradation. 64

69 REFERENCES 1.TORSNER J AND MALMGREN.G Radio Network Solutions for HIPERLAN/2. Proc. of VTC '99 Spring (Houston), pp [2] ETSI TS V1.1.1, Broadband Radio Access Networks (BRAN); HIPERLAN Type 2; Physical (PHY) layer, April [3] Bit Error Rate (2015): ACCESSED ON 10/08/ N.D. Jigisha and D.D. Upena, A Comparative Performance Analysis of OFDM using MATLAB and Simulink with M-PSK and M-QAM Mapping, International Conference on Computation Intelligence and Multimedia Applications, ] Rapp, Effects of HPA-nonlinearity on 4-DPSK-OFDM-signal for a digital sound broadcasting system, in Proc. 2nd European Conference on Satellite Communications, Liege, Belgium, Oct. 1991, ESA-SP-332. [6] T.S. Rappaport, Wireless Communication Principles and Practice, 2 nd Edition, 2002, Prentice Hall of India Private Limited, New Delhi. [7] H. ZHANG ET AL Research of DFT-OFDM and DWT-OFDM on Different Transmission Scenarios. Proceedings of the 2ndInternational Conference on Information Technology for Application (ICITA). [8] Heiskala, J., and Terry, J.: OFDM wireless LANS a theoretical and practical guide (Sams Publishing, 2002) [9] Chang, R.W.: Synthesis of band limited orthogonal signals for multichanneldata transmission, Bell Syst. Tech. J., 1996, 45, pp [10]Ahmed R. S. Bahai and Burton R. Saltzberg, Multi Carrier Digital Communication. Kluwer Academic Publishers, [11] Omijeh, B.O and Biebuma, J.J (2013): Path Loss Model using Geographic Information System (GIS), International Journal of Engineering and Technology Volume 3 No. 3, March,

70 TEXT DETECTION IN VIDEO USING HAAR WAVELET TRANSFORMATION AND MORPHOLOGICAL OPERATOR Dinesh Annaji Kene 1, Dr. D. J. Pete 2 Department of Electronics and Telecommunication Engineering. Datta Meghe College of Engineering, Airoli, Navi Mumbai University of Mumbai. India *dineshkene@gmail.com pethedj@rediffmail.com ABSTRACT: This paper presents simple and efficient method for text detection, extraction and localization from video or static images using Haar wavelet and Morphological operator. Haar wavelet transform have its coefficients either 1 or -1, so that the operation speed of Haar wavelet transformation is fastest among all wavelets. The sub bands contain both text edges and non-text edges however the intensity of text edges is different that of the non-text edges. Instead of using Canny operator we used Sobal operator for edge detection because Sobal operator detect more edges than Canny operator when there is text information. Morphological operators are applied to edit or smoothing out the text region. Then detected text regions are further decomposed into character level. Then using some refinement the final text region are obtained. Keywords : CC Based method, Texture based method, Edge based method, optical character recognition (OCR) I. INTRODUCTION A brief history of Video Text detection:- The origin of video text detection research may be traced to document image analysis because of its core techniques in optical character recognition (OCR). Patent on OCR for reading aids for the blind and for input to the telegraph was filed in the Nineteen century and working models were demonstrated by 1916.however the field of document image analysis itself is only about 40 years old, OCR on specifically designed printed digits (OCR fonts) was first used in business in the 1950 s. The first postal; address reader and the social security administration machine to read typewritten earnings reports were installed in Devices to read typeset material and simple hand printed forms came in to their own in 1980 s, when the prices of OCR systems dropped by a factor of a 10 due to the advent of microprocessors bit mapped display and solid state scanners. The study of automatic text detection started about three decade ago. In the , a lot of methods were proposed.[1] Video content can be seen as two major categories, Perceptual content and Semantic content. Perceptual content includes attributes such as colour intensity, shape, texture and their temporal changes, While semantic content includes object, events and their relations. Among them text in images / videos is of particular interest as 1) It is very useful for understanding the content of an image 2) It is very easy to extract compared to other semantic content and 3) It is useful for application such as text based image search and automatic video logging. The semantic gap refers to the different interpretation and meaning of the extracted low level feature (e.g. texture, colour, shape etc.) by the uses with respect to applications. Text in video can be used to fill this gap because in many videos there is considerable amount of text, Video text can thus help to improve the effectiveness of content based image and video retrieval system [1] 66

71 Compared with the text in a typical document the text in a video frame is in much smaller quality. However these texts in video often give crucial information about media contents. They usually appears in the form of names, location, products, brands, score of the match, date & time etc. which are helpful information to understand the video contents. The text in images and videos can be superimposed on arbitrary background or embedded on the surfaces of the object in the scene with varying fonts size, colour, alignment, movement and lighting condition. Hence text detection and extraction are extremely difficult. A large number of approaches such as Connected Components based, Texture based and other methods have been proposed. Some of these methods have already obtained impressive performances. A majority of text detection and extraction approaches in the literature are developed based on scanned document nature, although most of them can be adopted for video images detection and extraction. Text in video presents unique challenges over that in document images, due to many undesirable properties of video for text detection and extraction problem. Fortunately text in video usually persists for at least several seconds to give human viewer sufficient time to read it. This temporal nature of video is very valuable and can be well utilize for text detection, extraction and recognition in video. II. RELATED WORK In order to show that the proposed method is effective and robust in terms of metrics and multioriented text lines detection, we have studied variety of related works. In the research book by Palaiahnakote Shivkumara, Chew Lim Tan, Wenyin Liu and Tong Lu, Video Text Detection A Research book published by Springer- Verlag London 2014, shows the variety of methods, operators and models for image as well as video text detection. In that book he analyse the minute details of Text Detection in Video and also shows some application oriented methods.[1] P. Shivkumara, Trung Que Phan and Chew Lim Tan in their paper New Fourier Statistical Features in RGB Space for Video Text Detection proposed Fourier statistical method for color images, text to convert in gray and then again in color for detection and extraction [2]. As well Plaiahnakote Shivkumara, TrungQuyPhan et al. represented gradient vector flow and grouping based method. [3]. P. Shivkumara, T.Q. Phan and C.L. Tan [3] in their paper A Laplacian Approach to Multi Oriented Text Detection in Video represented a Laplacian method for multi-oriented text detection [4]. In which it shows that the performance of these methods is low because scene text poses many challenges compared to graphics text. Multi-oriented text has only been partially addressed where the algorithm is limited to caption text and a few selected directions. Recently, addressed this multi-oriented issue in based on Laplacian and skeletonization methods. However, this method still has room for improvement in the following respects: Recall, Precision and F- Measure. [4] Chung-Wei Liang et.al. used DWT based text localization method in his paper, DWT Based Text Localization. In which he used Haar wavelet transformation and Morphological operator along with AND operator.[5] Neha Gupta et. Al. also used Haar Wavelet and Morphological operator in their paper, Localization of Text in Complex Images using Haar wavlet transformation. For the experimental result they used complex background images as input.[6] M. Nagu et. Al. used Neural network along with wavelet and Morphological operator in their research paper, Morphological Text Localization using Wavelet and Neural network. Their 67

72 paper implemented Neural networks for recognition of text character from isolated text images and make it editable. [7] Narsimha Murthy K N et. Al. in his research paper, A Novel Method for Efficient Text Extraction from Real Time Images with Diversified Background using Haar Discrete Wavelet Transform and K-Means Clustering shows that they used real time images for data input and get efficient text output using K-means clustering[8]. III. PROPOSED SYSTEM In order to show that the proposed method is effective and robust in terms of metrics and multi-oriented text lines detection, we have considered a variety of datasets which include images and videos of sports news, low contrast, different fonts and size, different orientation etc. Methodology Many efforts have been made earlier to address the problems of text area detection, text segmentation and text recognition. Current text detection approaches can be classified into three categories: The first category is connected component-based method, which can locate text quickly but have difficulties when text is embedded in complex background or touches other graphical objects. The second category is texture-based, which is hard to find accurate boundaries of text areas and usually yields many false alarms in text-like background texture areas. The third category is edge-based method. Generally, analyzing the projection profiles of edge intensity maps can decompose text regions and can efficiently predict the text data from a given video image clip. Text region usually have a special texture because they consist of identical character components. These components contrast the background and have a periodic horizontal intensity variation due to the horizontal alignment of many characters. As a result, text regions can be segmented using texture feature. 3.1 Video Preprocessing Automatic video text detection is extremely difficult so we need video preprocessing to reduce the complexity of the succeeding steps probability consisting of video text detection, localization, extraction, tracking, reconstruction and recognition. 3.2 Image Segmentation In video image segmentation, after inputting an original videos frames image segmentation techniques are employed to extract all the pixels belonging to text. The aim of segmentation is to divide the pixels of each frame from a video clip in to two classes of regions, which do not contain text and region which contain text. A lot of techniques have been developed to perceive images, which in general consist of three layers (as shown in figure) of so called Image Engineering (I. E.), processing (low layer), analysis (middle layer) and understanding (high level). The objective of segmentation is to partition an image in to separated region which ideally composed to interested object and accordingly obtain a more meaningful representation of image. There are some general rules to be followed. 1. They should be uniform and homogenous with respect to specific characteristics. 2. Their interior should be simple enough and without too many small holes. 68

73 3. Adjacent regions should have significantly different values with respect to the characteristics on which they are uniform. 4. Boundaries of every segment should be simple, not ragged and must be spatially accurate. Each image can be segmented in to multiple separated regions using two strategies namely, Bottom-Up & Top-Down segmentation. 3.3 Pre-processing Operators The aim of preprocessing operations for video text detection is an improvement of the image data by suppressing undesired degradation and simultaneously enhancing specific text relevant features. There are several preprocessing operators that are either enhance the image feature or suppress the information which is not relevant to video text detection. a) Image Cropping and Local operator : These are useful for brightness adjustment & contrast enhancement. Commonly used local operators are Multiplication & Addition, Another commonly used operator is Linear Blend operator, which is designed to cross-dissolve two video frames. b) Neighborhood operator : These operators are used to remove noise and sharpness the image detail. The neighborhood operators are 1. Linear Filter Operator i) Average Filter ii) Median Filter 2. Derivative operator i) Robert operator ii) Sobel operator iii) Laplacian operator c) Morphological operator : Mathematical Morphology is a tool for extracting image component that are useful in the representation and descriptive of region shape, such as boundaries, skeleton and the convex hull. It is defined fundamental Morphological operation such as Dilation, Erosion, 69

74 Majority, Opening and Closing to perform such operations. We often need, first convert colorful video frame in to a binary image which has only two elements of either a real foreground or a background complexity with respect to the others. d) Color Based Preprocessing : A color image is built of several color channels, each of them indicating the value of the corresponding channel. For example, an RGB image consist of three primary color channels of Red, Green and Blue, An HIS model has three intuitive color characteristics as Hue, Saturation and Intensity. These color representation are designed for specific devices and have no accurate definition for a human observer. Therefore color transformation is required to convert the representation of color to gray scale representation. If the input image is a gray-level image, the image is processed directly starting at discrete wavelet transform. If the input image is colored, then its RGB components are combined to give an intensity image. The pictures are often in the Red-Green-Blue color space. Intensity image Y is given by: Y= 0.299R+0.587G+0.114B (1) Image Y is then processed with 2-D discrete wavelet transform. The Y is actually value component of the Hue-Saturation-Intensity (HSI) color space. In stage the color image RGB is converted in to HSI color space, then value component is extracted from HSI color space using above expression. The RGB to HSI transformation is H= arctan (β/α) (2) S= (3) I = (R+G+B)/3 (4) Where α = R-1/2 (G + B), β = / 2 (G - B) The noise of image is reduced by using a medium filtering that is applied on the gray scale image. After this filtering a major part of noise will be removed while the edges in the image are still preserved. IV. WAVELET TRANSFORMATION Digital image is represented as a two-dimensional array of coefficients, each coefficient representing the intensity level at that coordinate. Most natural images have smooth color variations, with the fine details being represented as sharp edges in between the smooth variations. Technically, the smooth variations in color can be termed as low frequency variations, and the sharp variations as high frequency variations. The low frequency components (smooth variations) constitute the base of an image, and the high frequency components (the edges which give the details) add upon them to refine the image, thereby giving a detailed image. Hence, the smooth variations are more important than the details. Separating the smooth variations and details of the image can be performed in many ways. One way is the decomposition of the image using the discrete wavelet transform. Digital 70

75 image compression is based on the ideas of sub-band decomposition or discrete wavelet transforms. Wavelets, which refer to a set of basic functions, are defined recursively from a set of scaling coefficients and scaling functions. The DWT is defined using these scaling functions and can be used to analyze digital images with superior performance than classical short-time Fourier-based techniques, such as the DCT. Shivkumar et.al. used 2-D Haar wavelet decomposition to directly detect video text based on three high frequency sub-band images of LH, HL and HH.after computing features for pixels, K means algorithm is applied to classify the feature vectors in to two cluster of background and text candidates. V. HAAR WAVELET The Haar Wavelet is a sequence of rescaled Square Shapes functions which together form a wavelet family or basis. Wavelet analysis is similar to Fourier analysis in that it allows a target function over an interval to be represented in terms of an orthogonal function basis. The Haar sequence is now recognized as the first known wavelet basis and extensively used as a teaching example. The Haar sequence was proposed in 1909 by Alfred Haar. Haar used these function to give an example of an orthogonal system for the space of square integrable function on the unit interval [0, 1] The Haar wavelet s mother wavelet function ψ (t) can be described as ψ (t) = 71

76 Its scaling function ф (t) can be described as ф (t) = A. Haar Transform : The Haar transform is the simplest of the wavelet transform. This transform cross multiplies a function against the Haar wavelet with various shift and stretches like the Fourier transform cross multiplies a function against a sine wave with two phases and many stretches. The Haar transform is derived from the Haar matrix. An example of a 4 x 4 Haar transformation matrix is shown below H4 = ½ Property The Haar transform has following properties 1) No need for multiplications. It require only addition and there are many element with zero value in the Haar matrix, so the composition time is short, it is faster than Walsh transform whose matrix is composed of + 1 and 1. 2) Input and output length are the same. however the length should be a power of 2 i.e. N = 2 K, K N 3) It can be used to analysis the localized feature of signals, Due to the orthogonal property of the Haar function the frequency components of input signal can be analyzed. VI. MORPHOLOGICAL OPERATION Mathematical morphology is a tool for extracting image components that are useful in the representation and descriptive of region shape, such as boundaries, skeletons and the convex hull. It is defined two fundamental morphological operations, dilation and erosion, in terms of the union or intersection of an image with a translated shape called as structuring element. To perform such operations we often need first convert a colorful image which has only two elements of either a real foreground or a background complementary with respect to the other. Morphological operations often take a binary image and structuring element as input and combine them using a set operator (Insertion, Union, Inclusion, Complement). They process object in the input image based on characteristics of its shape which are encoded in the structuring element. Usually the structuring element is sized 3 X 3 and has its origin at the center pixel. It is shifted over the image and at each pixel of the image its elements are compared with the set of 72

77 the underlying pixels. If the two sets of elements match the condition defined by the set operator, the pixels underneath the origin of the structuring element is set to a pre-defined value (0 or 1). A Morphological operator is therefore defined by its structuring element and the applied set operator. VII. MOTION ANALYSIS As a preprocessing technique motion vector analysis from temporal frames is believed helpful for video text detection. It produces time dependent information about the motion of both the foreground object and their background to provide hints for identifying text from consecutive video frames. More ever it potentially play a key role in detecting and tracking video text from multiple directions or with occlusions. The major goals of motion analysis are to refine the text detection in terms of removing false alarms in individual frames interpolating the location of accidently missed text character and words in individual frames and temporarily localizing the beginning and end of eachword as well as its spatial localization within each frame. Their method proposed for motion analysis is based on the comparison of region in successive frames and includes five typical steps. 1. Text Tracking: this step is responsible for tracking the already detected text along with the frames that constitute each text sequence targeting the formation of temporally related chains of characters. Each character chain here represents the same characters during its existence in a collection of similar regions occurring in several continuous frames. Every time a character region is detected for the first time, a position is stored and a signature is computed for that character region by using the features of luminance, size and shape. Each frame contributes with one region classified as a character chain. 2. Text Integration :- this step groups the character chains in order to form word based on the special and temporal analysis of each character chain. Three temporal elements are adopted a) Temporal coexistence b) Duration and c) Motion. The chain not include in words at this phase are considered as noise and are discarded. 3. Character Recovery ;- Video temporal redundancy is explored to complete the words with missing character as least for some frames e.g. due to noise or too textured background. In order to complete the words they are extended to the size of their biggest chain of character and the characters missing in the chains are recovered by means of temporal interpolation with motion composition. Thus by using temporal redundancy, the text detection for each frame is improved by completing the words with missing character for some frames 4. Elimination of overlapped words:- It is important to improve the performance for scene text usually, overlapping of words occurs when false words are detected e.g. due to shadow or three dimensional text. Every time two or more words overlap more precisely their bounding boxes overlap at least for one frame, the words with smaller area are discarded. 5. Text Rotation :- This steps perform the rotation of the text to the horizontal position in order to be recognized OCR system. VIII. RESULTS The results of the work of first stage Localizer are computed in the following manner: 1. Created dataset with various Text images and videos 2. Colour images and frames of videos are converted in to Gray scale 3. Texture of text area analysed by Haar wavelet transformation 73

78 4. Localized the text region or segments using Connected Component method ( Bottom Up approach) and Sobal operator. 5. Used Morphological operator to edit the localized text region. 6. Resize the image. 7. Adjust Gray colour at extracting place with neighbours colour. 8. Label the connected component region. 9. Apply the bounding boxes. 10. Convert the image again back from gray to RGB scale by comparing colours of the starting image. 11. Display the result. The results of the work of second stage Localizer are computed in the following manner: 1. After 1 st localizer output, the detected text region further decomposed in character level using Connected Component and Morphological operator. 2. Applied the bounding box. 3. Display the result. The results of the work of Third stage Localizer are computed in the following manner. 1. Eliminate the lines of bonding box. 2. Resize the text region using Connected Component major axis lenth& minor axis length, pixel index list. 3. Used again Morphological operator to edit the detected text. 4. Label the text region. 5. Display the result Result of Localizer 1 Result of image 1.1 Result of Video frame

79 Result of Localizer 2 Result of image 2.1 Result of Video frame

80 Result of Localizer 3 Result of image 3.1 Result of Video frame 3.1 IX. CONCLUSION & FUTURE SCOPE In this work, we present a relatively simple and effective algorithm for text detection and extraction in videos by applying DWT to the images. As it requires less processing time which is essential for real time application. Mostly all the previous methods fail when the character are not aligned well or when the characters are too small. They also results in some missing character when the character have very poor contrast with respect to the background. But this algorithm is not sensitive to image color or intensity, uneven illumination and reflection effects. This algorithm can be used in large variety of application fields such as vehicle license plate detection to detect number plate of vehicle [5], mobile robot navigation 76

81 to text based land marks, object identification of various parts in industrial automation, analysis of technical papers with the help of charts, maps, and electrical circuits etc. this algorithm can handle both scene text images and printed documents. In future work we plan to exploit the colour homogeneity of text in video, temporal text detection from frame to frame, Multi-frame integration for image enhancement. X. REFERENCES 1) Palaiahnakote Shivkumara, Chew Lim Tan, Wenyin Liu and Tong Lu, Video Text Detection A Research book published by Springer- Verlag London ) P. Shivakumara, Trung Quy Phan and Chew Lim Tan, New Fourier-Statistical Features in RGB Space for Video Text Detection, IEEE Transactions on CSVT, VOL. 20, NO. 11. pp , November ) Palaiahnakote Shivkumara and TrungQuyPhan, Gradient Vector Flow and Groupingbased Method for Arbitrarily Oriented Scene Text Detection in Video Images. IEEE Transaction on circuits and systems for video technology, Vol.23, No.10, pp october ) P. Shivakumara, T, Q. Phan and C. L. Tan, A Laplacian Approach to Multi-Oriented Text Detection in Video, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.33, No.2, pp February ) Chung-Wei Liang and Po-Yueh Chen, DWT Based Text Localization, International Journel of Applied Science and Engineering, pp ,2004 6) Neha Gupta and V. K. Banga, Localization of Text in Complex Images using Haar wavlet transformation International Journal of Innovative Technology and Exploring Engineering(IJITEE), Vol-1, Issue-6, pp , November ) M. Nagu,B. Raja Rao, Morphological Text Localization using Wavelet and Neural network. International Journal of Computer Science and Information Technology(IJCSIT), Vol-2(2), Issue-6, pp , ) Narsimha Murthy K N et. al., A Novel Method for Efficient Text Extraction from Real Time Images with Diversified Background using Haar Discrete Wavelet Transform and K-Means Clustering International Journal of Computer Science issues (IJCSI), Vol-8, Issue-5, pp ,

82 Knowledge Management And Transfer In 2010 Vancouver Olympic Winter Games Xi Xi 1,2 Richard Duncombe 3 1.&3. Institute for Development Policy and Management, The University of Manchester, Manchester, United Kingdom 2. Morgan Philips Executive Search, Shanghai, China Abstract : With the development of sports, Olympic Games are much more focused by people all over the world. Meanwhile, knowledge, knowledge management(km) and transfer in Olympic Games become more and more important. This paper gives some analysis of KM and transfer through 2010 Vancouver Olympic Winter Games. Key words: Knowledge Management, transfer, 2010 Vancouver Olympic Winter Games Introduction With the development of sports, Olympic Games are much more focused by people all over the world. Meanwhile, knowledge, knowledge management(km) and transfer in Olympic Games become more and more important. This paper gives some analysis of KM and transfer through 2010 Vancouver Olympic Winter Games. Knowledge can refer to a theoretical or practical understanding of a subject. Polanyi (1966) categorizes knowledge into explicit and tacit knowledge. Explicit knowledge is experience and skills that can be put into words. It is easy to be articulated. Tacit knowledge is capabilities and skills that sometimes can be observed from one s behavior but it is hard to describe. Knowledge exists everywhere but it needs managing. From the opinions of Delong & Beers (1998) and Jones & Leonard (2009), it can be concluded that knowledge management (KM) is a process, dealing with knowledge by sharing and recreating, and aiming at adding values to organization. By applying KM, several improvements can be achieved and bring benefits. Combining previous researches, benefits of knowledge management can be categorized as dimensions in Table 1. Table 1: Benefits of knowledge management Dimensions Organizational Benefits of knowledge management Gain competitive advantages by increasing effectiveness and 78

83 Benefits Managerial Benefits Resource Benefits Employment Benefits Customer Benefits efficiency. Better adaption to market change. Increase profit and decrease operating cost. Improve decision making accuracy and avoid reinventing the wheel according to previous record and experience. Increase effectiveness by taking the knowledge from expert performers and applying it in similar situations. Time savings in doing routine work and a better re-use of internal knowledge. Maximize utility of current knowledge assets. Catalyze innovation based on previous information and knowledge. Build better sensitivity to brain drain. Build mutual trust between employees. Motivate staffs by valuing them and their expertise. Improve skills and capability. Shorten training time and enabling less-trained brokers to achieve higher performance levels. Increase team work spirit and collaboration by knowledge sharing. Shorten respondent time for inquiry. Develop new product to meet customers needs. Customer retention enables responsive service. Source: adapted from Jarrar (2002), Mertins (2001), Ballal and Elhag (2006), North and Hornung (2003), Breedt & Rensburg (2012), Awad & Ghaziri (2004), Duncombe (2013b) KM and transfer involved in the Olympic Games The knowledge management process application in sport event management began from 1999 through the Olympic Games Knowledge Management (OGKM) program ( Parent, MacDonald and Goulet,2013). The main purpose of OGKM is to promote knowledge transfer between Organizing Committee of Olympic Games (OCOG). The knowledge includes services, personal experiences, and information that includes an observer program, workshops, technical manuals and other documentation, videos, extra-net access, debriefs, second program, and a cross-cultural awareness program. Although OGKM began from the late 1990s, one branch of OCOG, the 2010 Vancouver Olympic Winter Games Organizing Committee (VANOC) was firstly required to implement knowledge management from the bid phase. As the Olympic Games is one of top sport events, there are many stakeholders involved in and help to accomplish this task. The main stakeholders in 2010 Vancouver Olympic Winter Games can be analyzed by the CASOLO model from the literature of Duncombe (2013a),that is Clients, Actors, Sponsor, Owner, Legitimizers, and Opinion leaders. KM and transfer in 2010 VANOC 79

84 The McKinsey 7-S model developed by Waterman, Peter and Phillips in the late 1980s (Waterman, Peter & Phillips, 1980). They claim that productive organization change is not just about structure and strategy. Effective organizational change is really the relationship between structure, strategy, systems, style, skills, staff, and something we call super-ordinate goals. The 7-S model was designed to check various factors and interactions between them (see Figure 1) Figure 1: McKinsey 7-S model Source: Waterman, 1980 Based on the McKinsey s 7-S model, the implementation of knowledge management change strategy of VANOC can be analyzed from each factor as follows. Structure VANOC is an interdependent and complete organization, while it is also a branch of the International Olympic Committee (IOC). On one hand, formed by IOC and and NOC (the National Olympic Committee), the structure of VANOC is complex. Members that consist its executive body contain people from various fields and backgrounds. VANOC is also responsible for communication with various stakeholders including governments, sponsor companies, volunteers, delegations, athletes, media and audiences. As different people and forces seek different benefits, it is not easy to balance their needs. On the other hand, the structure is flexible to embrace the knowledge management change strategy. The purpose of OGKM is to help OCOG manage opportunities and risks to benefit from the long-lasting impact of the Games through knowledge creation and sharing. The multiplicity 80

85 stakeholder structure can provide information, experiences and knowledge in various aspects, which can contributes to knowledge identification, sharing, transfer and adoption. Moreover, it can be learned from part a that this change strategy have broad advantages in general. As organization, managers, employees and customers all can benefit from knowledge management, this structure enables their participation to give contributions and gain benefits. Strategy The OGKM program is crucial for Olympic Games and OCOGs. The task facing any OCOG is massive, and every resource available is valuable in achieving successful completion. Although the local context changes considerably with each edition of the Olympic Games, the IOC and other members of the Olympic Movement make a great deal of information available, which helps OCOGs and their partners to avoid re-inventing the wheel every four years. (Final Report of the IOC Coordination Commission) The strategy for this change is progressive implementation. Early in 1999, a predecessor to the OGKM program began. The 2000 Sydney Organizing Committee of the Olympic Games (SOCOG) was the first OCOG that transfer knowledge to the next edition of Games. Then it starts as a information management project. Gradually, IOC and OCOGs began to gain benefits and mastered tools and technologies in knowledge management. Therefore, the VANOC implement completes knowledge management strategy. System Many stakeholders, including VANOC s workforces, gain external knowledge from personal relationships. A network is a crucial platform for knowledge transfer. Information technology and media such as VANOC s intranet (ICE) and extra-net (SNOW) were built and used as a holistic, trans-disciplinary approach to coordinating, facilitating and leveraging intellectual assets within organisations and communities (Halbwirth, 2005). To support knowledge management strategy, VANOC provides other several tools and mechanisms. For example, in knowledge acquisition practices, s, past experiences/attending events records, presentations, technical documents, person/verbal reporting and/or debriefing, and document management systems are used to enable smooth operation. Style Openness is critical components in successfully implementing knowledge management (Maznevski & Athanassiou, 2007). Knowledge sharing is the basis of other knowledge management practices. This process requires highly interaction and collaboration between various stakeholder groups. Moreover, the openness style should be put into practice in leadership from main groups in VANOC. In this paper, a host city government representative revealed that we share reports and information that we have. 81

86 Openness does not only mean taking in knowledge. VANOC s knowledge management also allows other organizations to capture knowledge and information they require. More importantly, it offers great experience to the whole OCOG, especially the next Games holder. Staff Individuals play significant role in OGKM and VANOC. Firstly, they are holders of tacit knowledge from their past Games experiences. Secondly, they open the gate of external knowledge by building their social capital through communication with stakeholders and other knowledge resources. Thirdly, they offer internal sources of knowledge based on their previous Games experience or explicit knowledge from their colleagues. However, many staffs cannot distinguish between information and knowledge. Some of them use these two words as interchangeably or even considered them synonymous. Therefore, knowledge sharing becomes an exchange of information, knowledge, resources, problem solving and politic management. Moreover, many staffs cannot consider and treat knowledge management as a whole system. Instead, they just focus on one step, such as knowledge record and storage. These misunderstandings of knowledge and knowledge management can create constrains and weaken benefits of this change. Skills The skills in Olympic Games are all about knowledge. The first skill is knowledge creation. To share with others, knowledge should be identified from information and experience as well as transfer tacit knowledge to disseminated explicit knowledge. However, not every member in stakeholder groups has this skill, which influences the strength of knowledge management. A international delegation representative said I don t think you can create knowledge. Another critical skill is knowledge adoption, which means using the experience and information gained/ received and applying them in a more effective manner in the organization. As there are several similarities in each Game, some knowledge can be used directly while others need transfer processing. The adoption skill enables organization and other stakeholders take full advantages of knowledge instead of just recording and storing. Shared values The goal of OGKM is to ensure that future hosts could draw from the wealth of knowledge that is available about Games organization, while using it to make their own projects more efficient and effective in delivering the highest quality conditions for the athletes of the world (OGKM & The Vancouver 2010 Debrief). Based on this value, as every host country and host city are willing to complete a successful sports event and be a good example, there are less constrains for sharing knowledge and knowledge management. A member from OCOG said we think that we have an obligation to try to help other future hosts of the Olympic Games to do a good job. 82

87 Moreover, it is obvious that the strength that sponsors the sports events and unites all stakeholders together is the Olympic spirit. People believe in this spirit and are willing to spread the spirit all over the world by providing wonderful Olympic Games. This shared value enables smooth operation of Games organization and knowledge management implementation. Evaluation VANOC has some unique features that makes it special and not easy to evaluate. Firstly, OGKM program is a change strategy for IOC while it is actually the VANOC implemented knowledge management. VANOC is a new organization that is built because of 2010 Vancouver Winter Olympic Games. As itself is born, there are few formed working styles and regulations, therefore applying a new strategy would be much easier and receive less constrains. Secondly, both VANOC and IOC are not companies whose first and foremost goal is maximizing profits. Reputations of the game, country and city may be more important. Moreover, although each session of the game is compared between each other, the comparison is definitely weaker than competition in business world. Thirdly, there are multiple stakeholders in this case. As various stakeholder groups seek various results, it is difficult to evaluate achievements of this change by checking whether it enables them to reach their goals. Main constraints in VANOC Although knowledge management has plenty of benefits and seems welcomed by everyone, there are still several constrains in 2010 VANOC. Information: There are misunderstandings of knowledge and knowledge management. As some staffs do not have the awareness of correct concepts, inaccurate or useless knowledge and application may be result in. Process: VANOC is the first OCOG that implement knowledge management, so there are few examples. The process of identifying, sharing, transferring and adopting knowledge need to be carefully designed and be clear. Lacking of well-designed process, knowledge management will be in chaos. Staffing and skills: Lacking of skills of knowledge gaining, identification, transfer and adoption directly influence the effectiveness of knowledge management processing. Moreover, IOC staffs who are delegated to VANOC and are used to their previous working style may resist to change and additional tasks such as recording and sharing knowledge. Other resources: The culture and language gaps between members of IOC, NOC and OCOGs also influence implementation of knowledge management. The language of knowledge record, in the official language of host country or in official language of IOC, the accuracy of translation, and budget of multiple language copies are constrains in specific implement process. 83

88 Positive actions in VANOC Overall, the implementation of knowledge management in VANOC is a successful change to IOC and following OCOGs. As each session of Olympic Games is organized by different OCOGs, knowledge management is a successful strategy that can ensure the consistency of service standard and experience. Led by the IOC, the staffs of the VANOC shared their knowledge and experience with representatives from London 2012, Sochi 2014, Rio 2016 and the three applicant cities for 2018 (OGKM & The Vancouver 2010 Debrief). The VANOC absorbed experience from various aspects and became a successful example in knowledge management, which laid the foundation for future application for other OCOGs. The success of VANOC is resulted from the following factors. Shared values. The shared values play decisive role in the success. Each member is attracted by the value, believe in the value, and willing to spread the value by their work, the constrains therefore are significantly reduced. Supporting system. To achieve successful knowledge management, supporting systems, especially information system and reward system, should be well designed and built. Progressive implementation. The pace of change is also crucial and should be controlled. A change, especially the one that has wide impacts on other processes, should be implemented gradually. When most people in the system understand this change and can foresee their benefits from this change, the implementation will be much smoothly. Reference 1.Polanyi, M. (1966). The tacit dimension. New York: Doubleday 2.Davenport, T.H., Delong D.W., & Beers, M.C. (1998), Successful Knowledge Management Projects, Sloan Management Review, Vol. 39 (2), pp Jones, K. & Leonard, L.N.K., (2009), From Tacit Knowledge to Organizational Knowledge for Successful KM in Knowledge management and organizational learning: Annals of Information Systems, King,W. R. (ed), Germany: Springer Verlag 4.Jarrar, Y. F. (2002), Knowledge management: learning for organisational experience, Managerial Auditing Journal, Vol. 17 (6) 5.Mertins, K., Heisig, P. & Vorbeck, J. (2001), Knowledge management : best practices in Europe, Berlin ; London : Springer 6.Ballal TMA1 & Elhag TMS2 (2006), Knowledge Management Systems: their benefits and obstacles from an industry s perspective, 2ND Specialty Conference on Leadership and management in Construction, available at (accessed 20 th December, 2013) 7.North & Hornung (2003), The benefits of knowledge management: results of the German reward knowledge manager 2002, Journal of Universal Computer Science, vol. 9, no. 6 (2003),

89 8.Breedt, M., & Rensburg, A.V., (2012), Knowledge management principles, South African Journal of Industrial Engineering, Vol.10(1) 9.Awad, E. M. & Ghaziri, H. M. (2004), Knowledge management, Prentice Hall International, Upper Saddle River,Great Britain 10.Duncombe, R. (2013b),Knowledge Management - Handout, Manchester: IDPM, University of Manchester 11.Parent, M.M., MacDonald, D. & Goulet, G. (2013), The theory and practice of knowledge management and transfer: The case of the Olympic Games, Sport Management Review, Duncombe, R. (2013a), Understanding Organizational Change - Handout, Manchester: IDPM, University of Manchester 13.Waterman, R. H., Peters, T. J. & Phillips, J. R. (1980) Structure is not Organization, Business Horizons, Vol. 23(3), pp Halbwirth, S. & Toohey, K. (2005), Leveraging knowledge: Sport and Success. IASI 12 th world congress. 15.Maznevski, M., & Athanassiou, N. (2007). Bringing the out side in: Learning and knowledge management through external networks. In Kazuo & Nonaka (Eds.), Knowledge creation and management: New challenges for managers (pp ). New York, Oxford University Press 85

90 IMPLEMENTATION ON STM-16 FRAME TERMINATION VLSI WITH HIGH-SPEED AND LOW-POWER GDI TECHNIQUES USING SRAM G.GOWRI LAKSHMI PG Student [VLSI], Departement of electronics and communication engineering, Vivekanandha College of Engg for Women, Tiruchengode, Tamilnadu, India Abstract: Many of the current wireline networks are digitalized, In Japan, a synchronous digital hierarchy (SDH) system is installed in the public switched telephone network, and appliance data are transferred with a synchronous transfer module (STM). In this system to put together the advanced STM-16 using the CMOS technology in the technique. The STMs are constructed hierarchically with a nested structure. When commerce with the data for high-speed digital services, a nested level of four is required for each STM; that is, the lowest level of container, virtual container, administrative unit (AU), AU group, and STM. The VLSI has two kinds of I/O ports; one serial port operating at 2.5 Gb/s/pin and sixteen 8-b parallel ports operating at 19.5 Mb/s/pin. The high-speed port is connected to the set of connections side, while the low-speed ports are connected to the fatal sides including tool switches.using CMOS technique with STM- 16 frame termination VLSI, the power utilization during the sampled by is 34 mw, and that for 2.5Gb/s operation is 1.2w at 25 C.the proposed Gate Diffusion Technique(GDI) allows reducing power consumption, delay and area of digital circuits, while maintaining low difficulty of logic design.a multiplicity of logic gates have been implemented in 0.35 μm technology to compare the GDI technique with CMOS and PTL. A prototype test chip of STM-16 has been fabricated, based on GDI and CMOS cell libraries, showing up to 45% reduction in power-delay product in GDI. KEYWORDS: STM-16, CMOS, high speed, low power, GDI technique, PTL, frame termination, SDH. 1. INTRODUCTION INNOVATIONS in device technology including very large-scale integrated circuits (VLSIs), semiconductor lasers, Photodiodes, and optical fibers are having a great brunt on telecommunications. Many wire line networks are digitalized, and toll switches are connected to one another with optical fiber cables, which are mainly used to broadcast highly multiplexed digital data. In the fixed-line phone system, sending person s voice is sampled every 125 μs and converted to 8-b digital data by the sender s-side local switch. Using a public switched telephone network, the data are agreed to the receiver s-side local switch and reconverted to the original voice. In 1989, Nippon Telegraph and Telephone (NTT) Corporation adopted a synchronous digital hierarchy (SDH) for its public switched telephone network. An SDH is an internationally standardized system for synchronous multiplexed transmission for both current low-speed services including subscriber telephones and near future high-speed services. All of the equipment in the network is synchronized with a timing signal provided via a clock-signal tree. 86

91 A feature of the SDH system is that the broadcast rate in each multiplex level is configured to an integer multiple of the fundamental rate Mb/s (hereafter, 156 Mb/s). Also, with careful consideration of the actually required transmission capacity in Japan, one third of the fundamental rate (i.e.,52 Mb/s) is added to the Japanese national standard. Therefore, the network speeds in NTT are currently 52 Mb/s, 156 Mb/s,624 Mb/s, 2.5 Gb/s, and 10 Gb/s, excluding the conventional1.5 Mb/s (64 kb/s 24 channels) and 6.3 Mb/s rates. The SDH defines dedicated transmission frames, called synchronous transfer modules (STMs) for every individual multiplex level, where a frame is a transmission data package. Five kinds of frames have been defined: STM-0 (level 0; 52 Mb/s), STM-1 (level 1; 156 Mb/s), STM- 4(level 4; 624 Mb/s), STM-16 (level 16; 2.5 Gb/s), and STM-64 (level 64; 10 Gb/s). Every frame at each level has the same configuration in relation to two-dimensional byte data. With STM-1, the frame consists of a 9-row 9-column section overhead and a 9-row 261-column payload. The overhead area, which is 3.3% of the total frame data, is used to store the control data for frame synchronization and/or error checks. Moreover, a part of the overhead area is used to store the data that point the beginning of virtual containers in the payload. The frame frequency at each multiplex level is8 khz, and the frame data are transmitted bit by bit with an optical fiber cable. Major requirements for SDH transmission equipment are downsize and power consumption reduction.to reduce power consumption at the required operating speeds, employ a multi-vdd architecture using 2- and1-v power supplies. Also, the use of fully depleted silicon on insulator (FD-SOI) devices is considered from the view point of enhancing the operating speeds and/or reducing the dynamic power dissipation. The 2-Vmultiplexer/demultiplexer (MUX/DEMUX) macros provide a high operating speed of 2.5 Gb/s, while the dual-port static random access memory (SRAM) macro operates at a low supply voltage of 1 V. Moreover, 50-_-terminated input/output buffers (hereafter termed an I/O buffer) are designed to operate without devoted power supplies, and then a new direct-drive amplifier using a feedback reference scheme is devised for a 2-V 2.5-Gb/s input buffer use. In addition, multiplexing to STM-n (n: integer) is performed in the same way. Therefore, adopt a reconfigurable architecture to change the multiplicity with an outside selected mode signal. Our VLSI chip can handle five kinds of STMs: STM-1, 2, 4, 8, and 16, where STM- 2 and 8 are prepared only for the VLSI chip test. This function is implemented by 2-V 20- kilogate logic. To guarantee the required operating speed of 2.5 Gb/s, eliminate the delay flip-flop from the functional block located at the root node and also apply the double-rail type for every functional block located with a depth of one. The single-to-double transfer in the rail configuration is performed on the inside of the flip-flop in each functional block located with a depth of two. 87

92 That is, the slave latch in a flip-flop is replaced with a double-rail type. Hence, the timing slant between the complementary signals on a pair of wirings is minimized.stm frame section is given below.. Fig. 1. STM-16 frame termination VLSI. (a) Configuration. (b) Series of procedures in STM-1 terminator blocks 2. STM-16 FRAME TERMINATION VLSI The STMs are construct hierarchically with a nested structure. When dealing with the data for high-speed digitalservices, a nested level of four is required for each STM; that is, the lowest level of container, virtual container, administrative unit (AU), AU group, and STM. User data for each service are stored in a container and packed in an STM. When an STM is received, it is unpacked in reverse order to obtain the original container. The series of procedures needed for packing and/or unpacking is called frame termination. A. Design Concept The arrangement of our STM-16 frame termination VLSI is shown in Fig. 1(a). The VLSI has two kinds of I/O ports; one serial port operating at 2.5 Gb/s/pin and sixteen 8-bparallel ports operating at 19.5 Mb/s/pin. The high-speed port is connected to the network side, while the low-speed ports are connected to the terminal sides including tool switches. The configurations of STM-16 are defined with STM-1, and so adopt a line of attack that creates sixteen STM-1s as an transitional data form. The series of procedures required for each STM-1 frame termination is shown in Fig. 1(b). Here, the network side (that is, the core side of the VLSI chip) deals with STM-1 (19.5 Mb/s 8 b), while the terminal side deals with acontainer-4 (C-4). When packing a C-4 in an STM-1, a path overhead and a section overhead consisting of a pointer and control data are attached at a proper nested level. 88

93 The path overhead consists of the data for monitoring and/or error container and it is stored in the extended area in each payload. When unpacking a received STM-1, the section overhead, the pointer and the path in the clouds are removed at a proper nested level. B. Strategy For Reducing Power Consumption The simplest and most extensively used low power technique is power gating.cutting off the power supplies for stand by blocks reduces the wasted power dissipation caused by sub threshold leakage currents. The same concept was reported in1996, when multi threshold voltage CMOS (MTCMOS) was proposed for mobile applications. The other techniques for reducing active power are clock-signal gating, multiple supply voltages, and dynamic voltage and frequency scaling (DVFS). These techniques were developed for specific applications, and so it is important to employ proper low-power techniques to obtain a greater reduction of influence consumption. In CMOS digital VLSIs, a large part of the power consumption is engaged by dynamic components caused by the charge and/or discharge currents for parasitic capacitance. The dynamic power dissipation P is given by (1) P = C V2 f (1) Where C, V, and f are the parasitic capacitance, supply voltage, and operating frequency, respectively. The most effective way to reduce the power dissipation is to lower the supply voltage, but at the same time, the operating speed decreases. Hence, this methodology finds limited use. Another effective way is to reduce the parasitic capacitance. This is also effective in improving the degraded operating speed for low voltage operation. Multi-VDD Architecture: A feature of the STM-16 frame termination VLSI is that a high operating speed of 2.5 Gb/sis required only for the serial port that includes the peripheral logic. The required operating speed falls to 19.5 Mb/s as it approaches the parallel ports. To consider this situation, use two different kinds of power supplies VDD (typical2 V) and VLL (typical 1 V). With regards to I/O buffers, restrict the use of extra power supplies for terminators and/or references. From the view point of reducing power dissipation, use a 1-V power supply for many components, and the use of a 2-V power supply is restricted to the critical components for which a high operating speed is actually needed. That is, 2.5-Gb/s I/O buffers, a parallel-to-serial converter for byte multiplexing (hereafter termed a MUX macro), a serial-to-parallel converter for byte demultiplexing (i.e., a DEMUX macro) use a 2-V power supply, whereas19.5-mb/s I/O buffers and STM-1 terminators including dual-port SRAM macros use a 1-V power supply [Fig. 1(a)].Regarding the intermediate blocks operating at MHz(i.e., STM-1 to STM-16 MUX, scrambler, synchronizer and STM-16 to STM-1 DEMUX), use a multi-vdd architecture to reduce power consumption without degrading data throughput. FD-SOI Devices: SOI devices (MOSFETs) have a feature, whereby the parasitic capacitance in the source and/or drain is small because the p-n junction area is greatly reduced compared with bulk devices. Moreover, CMOS/SOI is latch up free. Owing to this feature, the shortened isolation distance between MOSFETs reduces the chip size, resulting in a higher operating speed and/or less dynamic power dissipation. 89

94 C. Key Component Design When customizing a standard cell library, a total of two kinds of power lines (VDD and VLL) are installed in every standard cell To suppress the increase in VLSI design cost, the use of macros is restricted to solely the 2-V2.5-Gb/s MUX/DEMUX and 1-V dual-port SRAM. Also, the configuration of 50-Ωterminated I/O buffers is carefully considered to avoid the need for dedicated power supplies. D.1-V Dual-Port SRAM Macro Design the SRAM macro in a way that allows us to reuse the design including the physical layout data. Therefore, both 1- and 2-V operations are targeted. That is, the macro operates over a wide supply-voltage range V, and it uses a configurable architecture with a metatile methodology. A unique idea of the methodology is that all the interconnection wires needed to generate an SRAM macro are installed in each leaf cell by way of preparation. The word count and bit lines is reduced to a near GND level to save energy consumption. In these techniques, the first targets the application with a wide bit width for I/O data. The latter two are more effective when the bitline have a large parasitic capacitance. In addition, conventional low power techniques including partitioning and/or sub bank configuration are also available to enlarge data storage capacity with less increase in power consumption. The configuration of our SRAM macro is shown in Fig. 11.Each bitline is directly connected to a data input circuit or a data output circuit without bitline selectors. In designing the physical layout, these peripheral circuits are integrated into data I/O circuitry and located on one side of the memory array. The dual clock signals for the write- and read-port are unrelated to each other, and so the macro can be used as a buffer memory for asynchronous data transfer. 2.5-Gb/s MUX/DEMUX Macros: To achieve the required high operating speed, adopt complete binary tree architecture. A functional block and the interconnection between functional blocks are assigned to a node and a path in the binary tree, respectively. With an8-to-1 MUX, the depth required for the binary tree is three. The functional block for the MUX macro consists of a 2-to-1selector and a cascaded delay flip-flop, whereas that for the DEMUX macro consists of two parallel-connected delay flip-flops. In the architecture, the higher operating frequency limit is provided by the MUX macro not by the DEMUX macro because the MUX macro has 2-to-1 selectors in the critical paths. To enhance the maximum operating incidence of delay flip-flops, consider the use of the double-rail type. The configuration for twostage delay flip-flops is shown in, compared with that for the conventional single-rail type. A feature of the double-rail type is that it has a pair of complementary inputs and a pair of harmonizing outputs. 90

95 Fig to-1 MUX with complete binary-tree architecture, (a) configuration.b) time chart. Fig.3.1-to-8 DEMUX with complete binary-tree. Architecture a) Configuration.(b) Time chart.. 8-to-1 MUX: The configuration with the complete binary-tree architecture is shown in Fig. 9(a), where D-F/F and TS-F/F are two-stage and three-stage delay flip-flop, respectively. A 1.25-GHz (= f /2) half-rate clock signal with a 50%-duty cycle is provided, while sub rate clock signals at frequencies of 624 MHz (= f/4) and 312 MHz (= f /8) are generated with frequency dividers. Each functional block consists of a 2-to-1 selector and a D-F/F, or a TS-F/F, and it operates synchronously with a half-rate or sub rate clock signal. That is, the data given by a pre stage D- F/F and TS-F/Fare selected in the order of the D-F/F data followed by the TS-F/F data. The data hold time required for the latter multiplexed data are yielded using the TS-F/F instead of a D-F/F. In this way, multiplexed data (i.e., 2-b serial data) are obtained at the output of the functional block. 1-to-8 DEMUX: The configuration with the complete binary-tree architecture,. In contrast with the MUX, the input is 8-b serial data, while the output is8-b parallel data. That is, the input data pass through the binary tree from the root node to the leaves. A 1.25-GHzhalf-rate clock signal with a 50%-duty cycle is provided, while the sub-rate clock signals at frequencies of 624 and312 MHz are generated with frequency dividers. Each functional block consists of a TS-F/F and a D- F/F, and these components operate synchronously with the same frequency clock signals. In detail, the phase of the clock signal for the D-F/F is delayed by a half clock cycle, compared with that for TS-F/F. In this way, the phases of 2-b demultiplexed data are focused, and they are output from the functional block. 91

96 3. THE GATE DIFFUSION INPUT TECHNIQUE This paper presents the Gate Diffusion Input (GDI) technique. To reduce the power consumption different logic design techniques like CMOS complementary logic, Pseudo nmos, Dynamic CMOS, Clocked CMOS logic (C2MOS), CMOS Domino logic, Cascade voltage switch logic (CVSL), Modified Domino logic, Pass Transistor Logic (PTL) have been proposed. Although Static CMOS Logic has been the most popular design approach for the past three decades, many attempts have been made to propose a better alternative to achieve lower power dissipation. Circuit designed with transmission gate (TG) solves the problem of low logic level swing by using PMOS as well as NMOS but this implementation needs true and complemented control signal and requires more area than pass transistor logic. Pseudo-NMOS is simple and fast but reduces noise margins and increases power consumption. Pass-transistor logic is good for certain classes of circuits (MUX/adders). On the other hand, PTL implementations of logic gates such as NANDs and NORs were found to be slower and consume more power than CMOS implementations mainly because of the reduced outputs wings due to the threshold drop across a single-channel pass transistor. Fig.4. Proposed GDI technique using logic gates GDI technique is one to minimize the power dissipation.gdi technique here one of the inputs are directly diffused into the gates of transistors of N-type and P-type devices so it is called as Gate Diffused Input technique. 4. GATE DIFFUSION INPUT TECHNIQUE USING STM FRAME A new technique of low power digital circuit design is described. This technique allows reducing power consumption, delay and area of digital circuits, while maintaining low complexity of logic design. Performance comparison with traditional CMOS and various PTL design techniques is presented, with respect to the layout area, number of devices, delay and power dissipation, showing advantages and drawbacks of GDI as compared to other methods. A variety of logic gates have been implemented in 0.35 μm technology to compare the GDI technique with CMOS and PTL. 92

97 A prototype test chip of STM-16 has been fabricated, based on GDI and CMOS cell libraries, showing up to 45% reduction in power-delay product in GDI. Table1.Example Performance of Full adders in MGDI, GDI, CMOS and PT logic GDI technique reduces the power dissipation, propogation delay, area of digital circuit and it maintain low complexity of logic design. 5.PROPOSED WORK The Gate diffusion input (GDI) is a novel technique for low power digital circuit design. This technique reduces the power dissipation, propagation delay, area of digital circuits and it maintains low complexity of logic design. To reduce power consumption at the required operating speeds, employ a multi-v architecture using 2- and 1-V power supplies.also,the use of GDI technique is considered from the viewpoint of enhancing the operating speeds and/or reducing the dynamic power dissipation.reduction of both threshold and gate leakage compared to basic CMOS logic gates. A variety of logic gates have been implemented in 0.35 μm technology to compare the GDI technique with CMOS and PTL. A prototype test chip of STM-16 has been fabricated, based on GDI and CMOS cell libraries, showing up to 45% reduction in power-delay product in GDI. 6.CONCLUSION In this paper, described the multi-vdd GDI techniques used in our STM-16 frame termination VLSI. Two kinds of powerlines (VDD and VLL) were installed in each standard cell without growing the cell size. To reduce higher performance from SOI devices, 2-V 2.5-Gb/s MUX/DEMUX and 1-Vdual-port SRAM were realized as hard macros. Also, the pact of 50-_- terminated I/O buffers was carefully considered to eradicate the need for dedicated power supplies. An STM-16 frame termination VLSI, fabricated with a0.3-μm quintuple-metal GDI process, successfully achieved 1 W, 2.5-Gb/s operation under typical conditions with 2- and 1-V power supplies at 25 C. An sophisticated experimental result with the latest VLSI fabrication technology will be reported a better way to suppress the impact of the drain-induced barrier lowering(dibl) is to curb the use of fine gate length as long as a required operating speed is consummate. 93

98 7.REFERNECES [1] Nobutaro Shibata, Yusuke Ohtomo An STM-16 Frame Termination VLSI With2.5- Gb/s/Pin Input/output Buffers: High-Speed andlow-power Multi-VDD CMOS/SIMOX Techniques IEEE transactions on very large scale integration (VLSI) systems, vol. 23, no. 6, june 2015 [2] S. Narasimha et al., 22 nm high-performance SOI technology featuring dual-embedded stressors, epi-plate high-k deep-trench embedded DRAM and self-aligned via 15LM BEOL, in IEEE Int. Electron Device Meeting (IEDM) Dig. Tech. Papers, Dec. 2012, pp [3] R. Tsuchiya et al., Silicon on thin BOX: A new paradigm of the CMOS FET for lowpower high-performance application featuring wide range back-bias control, in IEEE Electron Device Meeting (IEDM) Dig. Tech. Papers, Dec. 2004, pp [4] Y. Hirano et al., Impact of actively body-bias controlled (ABC)SOI SRAM by using direct body contact technology for low voltage application, in IEEE Int. Electron Device Meeting (IEDM) Dig. Tech.Papers, Dec. 2003, pp. 35. [5] M. M. Pelella, C. T. Chuang, C. Tretz, B. W. Curran, and M. G. Rosenfield, Hysteresis in floating-body PD/SOI CMOS circuits, in Symp. VLSI Technol., Syst., Appl. Dig. Tech. Papers,Jun. 1999, pp [6] K. Usami et al., Automated low-power technique exploiting multiplesupply voltages applied to a media processor, IEEE J. Solid- State Circuits, vol. 33, no. 3, pp , Mar

99 American Journal of Engineering and Technology Research Volume 16, no 1,