CHAPTER 2 LITERATURE SURVEY

Transcription

1 8 CHAPTER 2 LITERATURE SURVEY 2.1 REVIEW OF LITERATURE Human operators involved in co-ordinating the individual extrusion process, controller timers, counter, relays, individual temperature controllers, and speed controllers do not provide a satisfactory performance. Transition from idle state of the machine to the operation state causes significant variation in the temperature zone produces inhomogeneous melting temperatures and inconsistent product quality. The decoupling controller design for linear time-invariant square Multiple Input Multiple Output (MIMO) plants under the unityfeedback configuration is discussed and the design of decoupling controllers are achieved as pre assigned closed loop poles. (Lin and Hsieh 1991) The first paper on multi input multi output furnace control was published by Marzuki Khalid et. al. (1993) proposes a method for temperature control with the help of multilayered neural network with back-error-propagation algorithm for water bath temperature control system. The online learning of the emulators and controllers adapt to changes and to the environment, further improve performance in

2 temperature control. The prior training and judicious selection are of neural network models essential for its success. 9 Its steady state performance is improved by considering the integral error of the system. The Proportional Integral Derivative (PID) fuzzy controller for a four stage experimental system achieves a shorter rise time and smaller overshoot compared to PID control. Application of multivariable self tuning control by Kangling Fang et. al. (1996) can eliminate the couple influence of temperature of heating circles for the barrel effectively. The minimum phase, open loop unstable, variable time delay of the second order system is eliminated and provides good set point tracking performance to predict the emissivity changes for the compensation of measure temperature by neural fuzzy network proposed by Jiun-Hong La et. al. (1999). The trajectory of the temperature set point is achieved. The self-tuning technique proportional integral and proportional derivative type fuzzy logic control to conduct simulation analysis for a wide range of different linear and nonlinear second-order processes including a marginally stable provide an acceptable performance. Performances of the proposed self-tuning Fuzzy Logic Controller (FLC) are compared with those of their corresponding conventional FLCs in terms of several performance measures such as peak overshoot, settling time, rise time, integral absolute error and integral of time multiplied absolute error, in addition to the responses due to step set point change and load disturbance. The proposed scheme shows a remarkably improved performance over its conventional counterpart (Mudi and Pal 1999).

3 10 The intelligent control techniques used for the exit air temperature control of an air heat plant for the drying process. The schemes achieve the set-point tracking with high performance indices (Thyagarajan et. al. 1999). Industrial temperature regulation with adaptation mechanism to compensate the dynamic changes is developed by Keller (2000). The stability of the system is improved and exhibits robust performance for plants with significant variation in dynamics. Discrete time variable structure is developed by considering the control choice of dynamic sliding surface for a system with relative degree greater than one, computation of the discrete time dynamic sliding surface variable, and self-tuning of the switching control magnitude to reduce chattering. This simplified lumped parameter implemented for a temperature control of a MIMO system and experimental results demonstrate the effectiveness of the proposed method with a good set point trajectory proposed by Leehter Yao and Chin Chin Lin (2001). Fuzzy PID controllers are physically related to classical PID controller. The parameters settings of classical and fuzzy controllers are based on deep common physical background. The original newly introduced method considerably simplifies the setting and realization of fuzzy PI/PD/PID controllers (Pivonka 2002). Takagi Sugeno type recurrent fuzzy network with the direct inverse control configuration is proposed for the temperature control with reduced proposed rules and using field programmable gate array chip efficiently reduces the hardware design cost by Juang and Lin (1999).

4 11 The model for the three stage heaters with long delay time, large time constant and strong couple affects the model developed by Jaswinder Singh and Aman Ganesh (2008). The adapted PID neural network performs a perfect decoupling and control by training and selflearning process with short training time produces very good final result. Genetic algorithm based adaptive control which adequately adjusts controller gains according to the changes in plants. The controller gains are automatically tuned so as to minimize the error between the open loop frequency response of the reference model and that of the actual system at a few frequency points. The least square algorithm tuning is used for the controller (Saini 2005). The bigger overshoot and slower response of the conventional temperature control in the furnaces of the steel industry is eliminated by fuzzy and artificial intelligence. High temperature oxidations of billets are decreased by 50%, 12%, and 10% respectively over the current manual method (Yingxin Liao et. al. 2005). The conventional analog instruments with poor performance and reliability for power plants is eliminated by intelligent digital controller and upgrades the automatic control. The tuning of the PID parameters control strategy provides a good control performance (Yunhe Du et. al. 2005). The Takagi Sugeno model for plastic molding process utilizes the clustering method to generate the rule base of the fuzzy system. The back propagation algorithm is used to tune the adjustable parameter that has better performance than that of other methods.

5 12 Adaptive control adequately adjusts controller gains according to the changes in plants. The controller gains are automatically tuned so as to minimize the error between the open loop frequency response of the reference model and that of the actual system at a few frequency points. The proposed method has good control performance, overcomes long time delay (Saravanakumar and Wahidhabanu 2006) Takagi Sugeno type quantum neural fuzzy network has better performance than other methods. The rule base of the fuzzy system is implemented by dividing the original input output data into several clusters according to the similarity of the data. The proposed fuzzy model provides accurate input output data in a large temperature range and has the capability of applications in precision process control in plastic molding process (Bhogle et. al. 2007). The online adaptive control of non linear plants using neural networks with the application of temperature control system provides a neural inverse model added to the control scheme and an online update of the weight is provided (Harkouss et. al. 2010). Simulations have been carried out to show the robustness of the control algorithm. The parameter variations and disturbances have no effect on the tracking performance, since they have been compensated online. Sub neural inverse model is added to the structure scheme and weights are updated online. Good tracking performances have been obtained in both simulations and experimental application, although parameter variations and disturbances are unknown to the neural controller (Hedjar et. al. 2007).

6 13 The strong coupling effects between zones are eliminated by treating each zone as an independent non linear time varying factor, second order system and compensation is achieved with a proposed self tuning PID controller separately. The recursive instrumental variable method is employed to avoid noise and keep the online continuously to identify the capability for time varying parameters and validate the controller as the best one with good stability, high accuracy and good adaptability by Yinghua Song et. al. (2007). The optimal performance over the entire operating range with dynamic crossover and mutation probability rate improves the output in water bath temperature controller by Melba Mary et. al. (2007). Syed Faiz Ahmed (2007) proposes a micro controller based embedded system to control the sequence of hydraulic actuators movement effectively. The performance and the productivity of injection molding machine is increased at reasonable cost. GA based temperature controller design for thermal power plant proposed by Ali Reza Mehrabian and Morteza Mohammad Zaheri (2007). GA search for determination of the optimal PI controller parameters for a de-superheater in thermal power plant is implemented and it is efficient compared to other controllers. The PID like fuzzy logic controller on field programmable gate array device utilizes 1394 slices of the target FPGA, and is able to produce an output at sec with maximum frequency of MHz. The plant responses smooth output for the controller in higher sampling time for the temperature control. The absolute mean of

7 differences between the responses, was less than 0.5% of the output range (Mohammed Hassan and Waleed Sharif 2007). 14 The 10 % overshoot is eliminated in fuzzy control in the injection molding machine by Hongfu Zhou (2008) for the first order and time delay system. Adaptive neural fuzzy controller is developed to adapt process changes continuously with the genetic algorithm. The identified adaptive neuro controller balances the need to adapt with the need to preserve generalisation, and constitutes a general tool for adapting neural controllers on-line. Results indicate that inclusion of fuzzy and neural controller reflects a remarkable improvement in the system response by reducing the overshoot and integral average error, and maintaining the system dynamics, even in the presence of noise. PIC 16F877A microcontroller improves the performance of the temperature control by making considerable improvement in rising and settling time, besides, reducing overshoot and steady state error compared to a conventional PID controller proposed by Sheroz Khan et. al. (2008). Combined strategy of feedback control, iterative learning feed forward strategy tightly around the set point during normal operation is tested. Ideal state of the machine and transition eliminate the inhomogeneous problem for the reciprocating screw injection molding machine by Ke Yao et. al. (2008) The controller is robust against changes in the system parameters and has good set point tracking and disturbance rejection properties compared to conventionally proposed PI controller. The heater temperature control is complex because of non linear, parameter time variation, control variables (Hongfu Zhou 2008). The performance

8 15 of the temperature control of the plastic extruder is achieved with the help of genetic algorithm optimization method (Javier Causa et. al. 2008).The automatic genetic algorithm is optimized, solving with global optimization and makes the system in the optimized state throughout the entire operation process. The identification problem for time-delay nonlinear system uses gradient algorithm for updating the weights of the delayed neural networks and achieve the system to a stable for the bounded uncertainties (Talel Korkobi et. al. 2008). The controller performance was validated using a quadruped robot for the navigation problem and the results show high-quality performance with the help of intelligent control methods based on fuzzy logic, artificial neural networks, genetic algorithms and neuro fuzzy techniques (Pedro Ponce Cruz 2008). Bioreactor PID tuning is improved by GA and the results shows the controllers improves the performance of the process in terms of time domain specifications, set point tracking, regulatory changes and also provides an optimum stability compared to other tuning techniques (Giriraj Kumar et. al. 2008). Adaptive neuro fuzzy inference system based controller generates the membership function for fuzzy system, using water temperature control with the experimental results for various operating conditions and set points, with the better performance. The average error between the target and actual time response specifications with ANFIS tuning was less for the pilot scale jacketed batch reactor (Alipoor et. al. 2009).

9 16 The conventional fuzzy controlling is with oscillation, but the fuzzy genetic arithmetic control system has good adaptability proposed by Liu Yucheng et. al. (2009). Artificial Intelligence techniques can learn the accurate models and control the temperature of the continuous stirred tank reactor (Suja Mani Malar and Thyagarajan 2009). This paper addresses the problem of designing a fuzzy logic PI controller for a class of multi input multi output systems having a polynomial input non linearity. A good performance for tracking is obtained using fuzzy logic PI controller. The validity and robustness is tested on a simulation. The approach presents a solution to the problem of robust control of MIMO non linear systems. For each MIMO subsystem a local fuzzy PI control is given. The effect of nonlinear dynamics, disturbances and cross-coupling is compensated by projecting it on Chebyshev polynomials (Ougli and Boumhidi 2009). The Ziegler Nichols and the Cohen Coon tuning techniques are used for the PID tuning. Metamodeling techniques are utilized to tune the PID controller parameters quickly for linear and non linear model. The results for linear model take without disturbance and non linear model with considerable disturbance input. The radial basis function neural network metamodel is to minimize the time in tuning process and able to give a good approximation to the optimum controller parameters in both linear non linear models (Ab Malek and Mohameed Ali 2009). Proportional-derivative type fuzzy logic control was developed for cart position control of a gantry crane that incorporates input shaper control schemes for anti-sway control of the system. A significant reduction in the system sways had been achieved with the hybrid

10 controllers regardless of the polarities of the shapers (Ahmad and Mohamed 2009). 17 The PID controller parameters keep on changing constantly. The online GA based PID controller for liquid level tank system using the reproduction, crossover and mutation creates the new population for other parameters. The controller maintains the process in steady state. (Mohammed Obaid Ali et. al. 2009) An efficient and powerful design method for calculating optimal PID controllers for automatic voltage regulator system by using reinforcement learning prove efficient results in terms of optimality, computation burden and less sensitive to the ranges considered for the design variables. (Mohammadi et. al. 2009) Software incorporates Laboratory virtual instrumentation engineering workbench graphical programming language and MATLAB/Simulink fuzzy logic toolbox to design a light controller. Lighting control system is includes hardware and software. The control circuit applies reed relay in digital control way to adjust the variable resistor value of the traditional dimmer. The rule base of the fuzzy logic controller, either for the single input single output system or the double inputs single output system is developed and compared, based on the operation of the bulb and the light sensor (Mou Lin Jin and Ming Chun Ho 2009). The temperature control of plastic extruder with fuzzy genetic algorithm is proposed by Ismail Yusuf et. al. (2010). The genetic algorithm determines the membership function of FLC with fast

11 18 processing in reducing the overshoot problems of FLC. The experimental results show the feasibility and effectiveness of the proposed method. A fast on line learning method for neural network structures, by using genetic algorithm tuning process, which adjusts interconnection weights of the back propagation algorithm is developed (Ho et. al. 2010). This learning algorithm has faster convergence ability and better performance on reducing mapping error in the on-line learning neural network structures. This leads to the improvement of the transient response of neuro adaptive systems. Chamsai et. al. (2010) proposed a combination of a classical sliding mode control and a PID tuning technique with low-pass filter is developed for a position tracking control of a direct current servo motor. Uncertainty and nonlinearity of the servo motor system can be surmounted by the sliding mode control while the system response can be fine adjusted through the PID gain tuning. The performance strongly depends on the specified control gain in PID portion and sliding function. A multivariable adaptive predictive model based control strategy was simulated on a validated mechanistic transesterification model. The recursive least squares algorithm was used for process model adaptation in the GPC framework and the results revealed the superiority of the proposed centralized adaptive predictive control scheme as compared to the decentralized conventional PID controllers in terms of set point tracking, process interactions handling (Ho et. al. 2010). The consolidated literature review is shown in Table 2.1.

12 19 Table 2.1 Consolidated Literature Review S.No Title Author Details Consolidated Results 1 Control of Plastic Extruders with Multiple Temperature Zones Using a Microprocessor Based Programmable Controller System 2 Temperature Control of a Plastic Extrusion Barrel Using PID Fuzzy Controllers 3 Fuzzy Supervisory Predictive PID Control of a Plastic Extruder Barrel 4 Adaptive Decoupling Predictive Temperature Control for an Extrusion Barrel in a Plastic Injection Molding Process WilliamWare.E (1984), IEEE Transactions on Industry Applications, Vol.IA-20, No.4, pp Taur.J.S, Tao.C.W and Tsai.C.C (1995), Proceedings of the International IEEE/IAS conference on Industrial Automation and Control Emerging Technologies, Taipei, Taiwan, May 22-27, pp Ching Chih Tsai and Chi-Huang Lu (1998), Journal of the Chinese Institute of Engineers, Vol.21, No.5, pp Chi Huang Lu and Ching Chih Tsai (2001), IEEE Transactions on Industrial Electronics, Vol.48, No.5, pp The problem is eliminated by an integration of the variety of individual controllers into a single system accomplishes a high degree of co-ordination and extruder performance. Temperature control of a plastic extrusion barrel using proportional integral derivative fuzzy controllers implies a traditional fuzzy controller. The problem of tuning of PID controllers eliminated and proposes a method that makes the weighting term of the PID control and its gains are adjusted by fuzzy controller. A recursive least square estimation method is implemented by TMS320C31 processor. This improves the capabilities of set-point tracking, disturbance rejection, and robustness by appropriate adjustments to the tuning parameters in the criterion function.

13 20 Table 2.1 (Continued) S.No Title Author Details Consolidated Results 5 Hybrid Fuzzy Logic Control with Input Shaping for Input Tracking and Sway Suppression of a Gantry Crane System 6 A Genetic Based Neuro-Fuzzy Controller for Thermal Processes 7 Mold Temperature Control of a Rubber Injection Molding Machine by TSK Type Recurrent Neural Fuzzy Network 8 Implementation of MATLAB- SIMULINK Based Real Time Temperature Control for Set Point Changes Ahmad.M.A and Mohamed.Z (2009), American Journal of Engineering and Applied Sciences, Vol.2, No.4, pp Ashok Kumar Goel, Suresh Chandra Saxena and Surekha Bhanot 2005, Journal of computer science and Technology, Vol.5, No.1, April Chia Feng Juang, Shui Tien Huang and Fun Bin Duh (2006), Journal of Neurocomputing, Vol.70, No.3, pp Emine Dogru Bolat (2007), International Journal of Circuits, Systems and Signal Processing, Vol.1, No.1, 2007, pp The hybrid Fuzzy PI for the temperature control improves the performance of industrial process by supplementing conventional controls The method has effectively built accurate linguistic neuro fuzzy models and competes well with other existing approaches. The recurrent structure eliminates the need of prior knowledge of molding machine order and with the use of simple gradient, descent algorithm, practical experiments results in the elimination of the complicated time delay property and make the sampling interval even. The temperature set point with Ziegler Nichols step response method, relay tuning method, integral square time error, and disturbance criterion method. The relay based PID controller provides the best response among other methods.

14 21 The classical PID controllers are sensitive to variations in the system parameters; fuzzy controllers do not need precise information about the system variables in order to be effective. To enhance the controller performance, hybridization of these two controller structures comes to one mind immediately to exploit the beneficial sides of both categories. Hence, it can be concluded that the hybrid fuzzy PID controller is suited for practical applications (Pratumsuwan 2010). The Figure 2.1 shows the observation of research work carried out on temperature control of plastic extrusion system with different controllers. The literature survey reveals that there is a need for new algorithms for temperature control optimization techniques. Neuro fuzzy temperature controller is not implemented for plastic extrusion system. The previous research papers consider a single stage control only and not focused on more number of heating stages. Embedded based intelligent controllers are used to control the temperature of a multistage proto type model of a plastic extrusion system. The drawbacks of the existing methods can be summarized as follows Figure 2.1 Temperature Control Techniques for Plastic Extrusion System

15 22 i) Temperature control of a plastic extrusion proposed by the researchers are not considering the effect of coupling with a large peak time, delay time with peak overshoot and not considering more number of heaters. The time taken for the settling is more. The fine tuning of the controller with the controlled variable to achieve the set point temperature is not obtained and they do not provide contented results for non linear and dead time process. ii) Many papers have reported on the performance of temperature control in extruder of plastic industry with simulation model and not with real time implementation and also not with different control strategies (Taur et. al. 1995, Ching-Chih Tsai et. al. 1998, Chi- Huang Lu et. al. 2001, Yang Yanjuan et. al. 2009). None of these papers has jointly focused on intelligent controller with the embedded controller and considering the effect of coupling effects, set point tracking for more number of heaters. Hence in this research, an attempt has been made to overcome the drawbacks of the earlier research on the multiple set point tracking, disturbance rejection, transient analysis, optimum stability with real time hardware implementation using embedded based neuro fuzzy. The neuro-fuzzy network does not require a priori knowledge about the system and eliminates the need for complicated design steps like manual tuning of input output membership functions, and selection of fuzzy rule base.