ICS solar water heater study using artificial neural networks

Abstract

In this paper we present a study in which a suitable Artificial Neural Network (ANN) and TRNSYS are combined in order to predict the performance of an Integrated Collector Storage (ICS) prototype. We use the experimental data that have been collected from outdoor tests of an ICS solar water heater with cylindrical water storage tank inside a CPC reflector trough, to train the ANN. The ANN is then used though the Excel interface (Type 62) in TRNSYS to model the annual performance of the system by running the model with the values of a typical meteorological year for Athens, Greece. In this way the specific capabilities of both approaches are combined, i.e., use of the radiation processing and modelling power of TRNSYS together with the “black box” modelling approach of ANNs. We present the details of the calculation steps of both methods that aim to the accurate prediction of the system performance and we show that this new method can be used effectively for such predictions