Proceedings of the National Seminar & Exhibition on Non-Destructive Evaluation NDE 2011, December 8-10, 2011 MODELLING AND EXPERIMENTS FOR THE DEVELOPMENT OF A GUIDED WAVE LIQUID LEVEL SENSOR Subhash N.N and Krishnan Balasubramaniam Centre for Non Destructive Evaluation Indian Institute of Technology Madras, Chennai 600036 ABSTRACT The paper describes the method of liquid level sensing using Ultrasonic Lamb waves. Lamb waves propagating through thin Aluminium sheets can be effectively used for liquid level sensing. Previous research shows that the various modes of Lamb waves are influenced by the presence of liquids. Thereby, the change in the characteristic of the waves can be used as a function of liquid level. Here, in this study an attempt has been made to know the characteristics of lamb wave propagation through an Aluminium plate. Finite Element Analysis is carried out using ABAQUS 6.10 to analyse the behaviour of lamb waves in isotropic Aluminium plate, and the interaction with defects. The study concludes the significance of contact type level sensors over non contact type. Keywords: Ultrasonic, Guided Waves, Lamb Waves, Liquid Level Sensor, Dispersion Curves, Finite Element Analysis. 1. INTRODUCTION Level sensors as an integral part of process control gain attention after the Three Mile Accident in 1979. The accident took place due to the malfunctioning of coolant system, caused by improper coolant level sensing. So, level sensing accounts more than just sensing presence or absence of liquids. Level sensors find their application in petrochemical, chemical, water and waste water, food and beverage, pharmaceutical and power generation industries. Sensors are also important in monitoring the level of water, slurry, chemicals, petrochemicals and solids. Fuel level sensing techniques are of great importance in modern engineering scenario. Their application ranges various from two wheelers to aircrafts and ordinary fuel storage tanks to aviation fuel tank systems. The techniques that are recently used are not expected to yield results of high accuracy. Selection of the sensors depends on various constraints like pressure, temperature, density, dielectric constant, composition, vibration and dynamics of the operating environment. The response rate, non-invasiveness, accuracy, monitoring and calibration capabilities of the level sensor drastically influence its outcome. Process control and prevention of industrial hazards need much research in this field. The paper is a review of various techniques used for level monitoring. Lamb wave propagation in Aluminium plate and their characteristics was used for the present study. The selection of operating mode is carried out using Disperse. Finite element analysis of wave propagation in aluminum plate is carried out using Abaqus 6.10. The numerical model helps in predicting the influence of various parameters like operating frequencies, angles of transducers and their relative positions on wave generation, interaction with defects and detection of lamb modes. Both Aluminum plate with defects and without defects are analysed. Finally the paper presents the advantages in using contact type level sensor over the non contact type. 2. LEVEL SENSORS IN PROCESS CONTROL The level sensors available in market fall under the categories of visual, pressure, electrical, radiation, optical thermal and vibrational type sensors. A brief comparison is shown in the Table 1. 2.1 Lamb waves in level sensing Lamb waves are guided waves that can be generated in materials having thickness of the order of a few wavelengths. Infinite number of modes are associated with a given plate wave problem. Lamb waves can be used to develop a guided wave liquid level sensor as shown in Figure 1. Lamb waves propagate through the thickness of the plate, and have complex vibrational patterns. The propagation of Lamb waves depends on: operating frequency, material thickness, density and elastic properties of the material. Different modes occur when the frequency and wave entry angle is varied. The basic modes are Symmetric(S) and Antisymmetric(A) modes. In symmetric mode, the wave appears to streching and compressing the plate in wave motion direction (Figure 2). This mode can be efficiently produced when the excitng force
NDE 2011, December 8-10, 2011 241 Table 1 : Comparison of different level sensing techniques. No Classification Areas of Application Disadvantages 1 Visual level sensors Oil level sensing using Very poor response rate, dipsticks in automobiles inaccurate 2 Float type Fuel tanks, water tanks. Poor response rate, difficult in interpolation 3 Air bubbler Used for corrosive and Inaccurate measurement slurry type liquids 4 Optical level sensor Used in environment of Point level measurements hydrocarbon vapors 7 Capacitance level Liquids having less variable Varying dielectric constants sensors dielectric constant 8 Conductive Conductive liquids Maintenance problems Fig. 1 : Level sensing of liquids using Lamb waves propagating through Aluminium plate Fig. 2 : Symmetric mode S0 of Lamb wave in 2mm thick Al plate. is parallel to the plate. Antisymmetric mode of lamb waves is also known as flexural mode. Here, the most of the particle vibration takesplace prependicular to the plate (Figure 3). As the two surfaces move in same direction, the antisymmetric mode appears to bend the plate. Fig. 3 : Anti-symmetric mode A0 of Lamb wave in 2mm thick Al plate. 2.2 Selection of optimum operating mode Dispersion curves generated by Disperse, helped to identify and evaluate the possible modes of guided waves and its propagation characteristics in the sample. The analysis was done for an Aluminium plate of thickness 2 mm, density is
NDE 2011, December 8-10, 2011 242 Fig. 4 : Group velocities of various modes as a function of frequency for Al plate of 2 mm thickness Fig. 5 : Phase velocities of various modes as a function of frequency for Al plate of 2 mm thickness 2700 Kg/m 3 and Youngs modulus is 70 GPa. The plots of group velocity vs frequency and phase velocity vs frequencythickness were generated. The resulting dispersion curves are shown in Figure 4 and Figure 5. Selection of the optimum operating mode, required a compromise between lesser number of modes, less dispersion and shorter wave length. It is done to improve the sensibility and simplicity on the signals received. In the present case, S0 mode at a frequency of 100 KHz was selected. The results of this study were used for further analysis using Abaqus 6.10. 3. FINITE ELEMENT STUDIES OF REFLECTIONS IN AN ALUMINIUM PLATE The solution to a guided wave problem must satisfy the governing equations as well as some physical boundary conditions. It is tedious to solve guided wave problems analyticaly due to the introduction of boundary conditions. Also, a infinite number of guided wave modes are supported by a finite body.
NDE 2011, December 8-10, 2011 243 Fig. 6 : Wave propagation through aluminium plate without notches Fig. 7 : A- Scan for determining mode velocity An Aluminium plate of thickness 2 mm, length 1 m and width 381 mm is used for the finite element analysis using ABAQUS 6.10 software. The material properties were mentioned in the previous section. The total time period for wave propagation was found out using the velocity of shear waves in Aluminium plate and the length of the plate. Inorder to analyse the wave propagation through the entire length, shear waves which travels at a lower velocity was considered. The shear wave velocity is 3160 m/s. The time was found as 315 micro seconds. For analysis, a time period of 400 micro seconds was used for better simulation. The step time used was 0.01 micro seconds. Hexagonal elements with mesh thickness 1mm were employed which resulted in about 887298 hexagonal elements. The results of analysis is shown in Figure 6. A point 100 mm below the point of excitation was selected for monitoring the velocities of various modes. The velocity was found out from the A-scan and the location of monitoring point. Figure 7 shows the A scan at the monitoring point The velocity obtained was 5405 m/s. This velocity was almost close to the velocity of S0 mode. The velocity of S0 mode obtained from the disperse was 5424 m/s. The same analysis is repeated with three rectangular holes of varying dimensions. The results of analysis is shown in Figure 8 and Figure 9. From the analysis it is possible to find out the nature of waves after reflections and their characteristics. While performing analysis, it is possible to create monitoring points at prime locations to determining the velocities. Using Fig. 8 : Wave propagation through aluminium plate with notches Fig. 9 : A Scan for determining mode velocity
NDE 2011, December 8-10, 2011 244 the A scan shown in Figure 7, it is possible to determine the velocity. The result obtained is to be compared with that of dispersion curves to determine the mode and other wave characteristics. So, by making various combinations we can compare the accuracy of results obtained from them. Finaly the system with proper accuracy, response rate and monitoring controls can be optimised. 4. CONTACT TYPE LEVEL SENSOR USING ALUMINIUM PLATE WITH RECTANGULAR HOLES In the case of non contact type ultrasonic level sensors, the transmission of sound between the transducer and the process liquid across an air gap is very inefficient. This is because of the large impedance mismatching between materials. Majority of the sound wave get reflected at the transducer to air boundary. The signal levels obtained from the non contact sensors tends to be very low. Frequent averaging is required to overcome this problem, which is inconvinient to both industrial and process control applications. A contact type level sensor can be developed by using an aluminium plate to guide the lamb waves. Unlike non contact type ultrasonic level sensors, the use of aluminium strip will reduce the accoustic impedeance mismatching. From the finite element analysis conducted, it is found out that the characteristics of waves changes as they encounters rectangular holes. A contact type sensor using an aluminium plate having holes of varying size at different locations can be employed. By using FEM analysis we can determine characteristics of the waves while encountering the rectangular holes. The accoustic impedance of water is higher when compared with that of air. So, research should be focussed in such a way that the return signal can be expressed as a function of liquid level. Research in the areas of impedance mismatching between solid-air and solid-water yield enough results for developing calibrated datas and algorithms to find out varying liquid level. Once the algorithms and calibrated data are developed, same procedure can be applied to various combinations. Instead of rectangular holes, we can try rectangular grooves, notches etc. CONCLUSION The work in development shows the potential of lamb waves in liquid level sensing. The available numerical methods can be effectively used to predict the propagation of lamb waves and their interactions with defects as shown in the paper. The change in characteristic of lamb waves, when it encounters liquid can be developed to a function of liquid level. Much research are yet to be done in determining the optimum conditions for level sensing using the lamb wave propagation in aluminium plate. Future studies in the field of dynamic environments like fuel storage tanks in aircrafts is of great significance. Hence, a study in this scenario will be useful for the process control applications in future. REFERENCES 1. M. Castings and P. Cawley, The generation, propagation, and detection of Lamb waves in plates using air coupled ultrasonic transducers, J. Accoustical society of America, Vol. 100, No. 5, pp. 3070-3077 (1996). 2. V. E. Sakharov, Liquid level sensor using ultrasonic lamb waves, ultrasonics 41, pp. 319-322 (2003). 3. A. B. Gillespie, A new ultrasonic technique for the measurement of liquid level, ultrasonics, pp. 13-17 (January, 1982) 4. C. Willberg, Simulation of piezoelectric induced lamb waves in plates, PAMM Proc. Appl. Math. Mech. 9, 503 504 (2009). 5. J. L. Rose, Ultrasonic waves in solid media, Cambridge University Press (1999). 6. Omega, Flow and level measurement, Transactions in measurement and control, Vol. 4.