Design of an adaptive calibration technique based on LSSVM for liquid level measurement

Design of an adaptive calibration technique using Least Square Support Vector Method (LS-SVM) for liquid level measurement using Capacitance Level Sensor (CLS) is reported in this paper. The objectives of the present work are (i) to extend the linearity range of measurement to 100% of the full scale, (ii) to make the measurement technique adaptive for variations in permittivity of liquid, liquid temperature, and diameter of the storage tank. The output of CLS is capacitance. A data conversion unit is used to convert it to voltage. A suitable LS-SVM model is designed to replace the conventional calibration circuit, in cascade to the existing data conversion unit. The proposed technique provides linear relationship of the overall system over the full input range and makes it adaptive of variation in liquid permittivity, and/or temperature, and/or tank diameter. LS-SVM is trained with simulated data considering variations in liquid level, liquid permittivity, liquid temperatures, and tank diameters. Results show that the proposed scheme has fulfilled the objectives.