A practically validated adaptive calibration technique for temperature measurement using resistance temperature detector

This paper presents a calibration technique for a Resistance Temperature Detector (RTD) used in measurement of temperature. Soft calibration circuit is designed using an optimized Artificial Neural Network (ANN), optimization of ANN is to choose a particular neural network scheme, algorithm, transfer function, and number of hidden layers. The objective of the present work is design an adaptive calibration circuit using optimized ANN model which produces (i) a linear output over 100% of full scale input range, (ii) accurate output even when the RTD is replaced with a different RTD (different refers to variation in parameter like reference resistance and temperature coefficient). Resistance temperature detector is cascaded to the designed neural network model through a suitable data conversion circuit. The designed system is evaluated for its performance in simulation and practical domain. Results obtained show that the set objectives are fulfilled.