ARTIFICIAL NEURAL NETWORK MODELLING APPROACH FOR ASSESSMENT OF STRATIFIED THERMAL ENERGY STORAGE TANK

Abstract

bases and university campus. Hence, the TES performance is important to be monitored. Various methods to measure the performance of the TES covering both numerical and analytical have been published. In this paper artificial neural network (ANN) is used to measure the performance of the TES tank in terms of thermocline thickness and halfcycle figure of merit. The ANN with 14 temperature sensor data as input and the thermocline thickness and half cycle figure of merit as the outputs is proposed. The model is based on 14-90-2 configuration using backpropagation LavenbergMarquadt. The data of one year has been used in modelling. Based on the trial and error the number of neurons were used and the optimum numbers of the neurons found were 90. The overall 𝑅 2 for the model was 0.99 and predictions compared with the actual data gave a 0.94𝑅2