Different models for determination of thermal stratification in a solar storage tank

In this work two different models are shown for describing the thermal stratification in the solar storage tank of the solar water heating system. The first model was developed on the basis. of the. measured physical parameters of the system with the TRANSYS software, and the other one is an artificial neural network (ANN) model. The ANN model calculates the temperature of 8 layers in every hour from the average hourly data of the solar radiation, the water consumption, the ambient temperature, the mass flow rate of the collector fluid and the previous temperature of the layers. During the modeling predicted consumption data and measured meteorological and layer temperatures data were used. The describer ANN model uses two layers with 8 tansig and 8 linear neurons. The average deviation in case of TRNSYS model was 1.9 degrees C and in case of ANN model was 0.53 degrees C during the training and 0.76 degrees C during the validation.