Adaptive calibration of turbine flow measurement using ANN

This paper aims to design an adaptive flow measurement technique using the turbine flow meter. The objectives of this work are (i) to make the system linear over the full scale, (ii) to extend the linear range of flow measurement, (iii) to make the proposed flow measurement technique adaptive of variations in (a) Number of turbine blades (b) density of liquid and (c) mean radius of the turbine. Output of turbine flow meter is converted to current by using magnetic pickup and additional data conversion circuit. A suitable Artificial Neural Network (ANN) block is added in cascade to the data conversion unit. This arrangement helps to linearize the overall system and make it adaptive to variations in liquid density, number of turbine blade, and mean radius of turbine. The proposed work is tested with variations in input flow, liquid density, dimensions of the turbine. Results show successful achievement of the set objectives. Measurement by the proposed technique results 0.283% as root mean square of percentage error.