An intelligent flow measurement technique using ultrasonic flow meter with optimized neural network

Design of an intelligent flow measurement technique using ultrasonic flow meter is reported in this paper. The objective of the work are; (i) to extend the linearity range of measurement to 100% ofthe input range, (ii) to make the measurement system adaptive to variations in pipe diameter, liquid density, and liquid temperature, and (iii) to achieve the objectives (i) and (ii) by an optimal Artificial Neural Network ((ANN). An optimal ANN is considered by comparing various schemes and algorithmsbased on minimization of Mean Square Error (MSE) and Regression close to one. The output of ultrasonic flow meter is frequency. It is converted to voltage by using a frequency to voltage converter. An optimal ANN block is added in cascade to frequency to voltage converter. This arrangement helps to linearise the overall system for 100% of full scale and makes it adaptive to variations in pipe diameter, liquid density, and liquid temperature. Since the proposed intelligent flow measurement technique produces output which is adaptive to variations in pipe diameter, liquid density, and liquid temperature, the present technique avoids the requirement of repeated calibration every time there is change in liquid, and/or pipe diameter, and/or liquid temperature. Simulation results show that proposed measurement technique achieves the objectives quite satisfactorily.