ANN-BASED TEMPERATURE MEASUREMENT USING DIFFERENT SENSORS

Abstract

ANN have capability to solve problems with high speed. It can handle nonlinear functions also. Temperature measurement using different sensors- -are formulated in the frame work of ANN model. The multilayer perceptron model using back propagation training algorithm makes possible to train the proposed ANN with training pattern for different sensors. In the present work. the standard pattern chrome] Alumel for thermocouple and platinum for RTD and thermistor have been chosen. 5 layers for thermocouple and 4 layers for both thermistor and RTD's have been taken for ANN. The training pattern act as a input to the ANN and these are varified with desired temperature giving the percentage error in each case.