Development of Artificial Neural Network Models for Sorption Chillers
Abstract
Solar cooling is still a young and small but growing market with a large potential. Up to now there exists no standardised performance test method for solar cooling or combined solar cooling and heating systems. Also for these innovative systems it is important that performance determination is carried out in a standardised way in order to compare their performance with the one of a well defined reference system (conventional system). In this way energetic and environmental benefits in terms of primary energy savings and CO2 emission reductions can be determined. For this reason and due to the fact that one established procedure to determine the performance of solar thermal systems is the CTSS-method (Component Testing – System Simulation), already standardised in European Standard series CEN/TS 12977, an extension of this method applicable for solar cooling systems and SolarCombiPlus systems (systems which provide domestic hot water, space heating and space cooling) was found to be the most promising way. With this method the annual performance of the whole system can be calculated for defined boundary and reference conditions (meteorology, load profiles) by means of a dynamic simulation of the whole system. For the suggested extension of the CTSS-method towards solar cooling systems (Frey et al., 2010) dynamic simulation models for thermally driven chillers (sorption chillers) are necessary. The main target of the work presented in this paper is to develop appropriate sorption chiller models which can be used for the extended CTSS-method. One promising way is the experimental system identification based on artificial neural networks (ANN). In this approach experimentally measured data are used to derive an ANN model which is able to predict the outlet temperatures of a sorption chiller. In the work presented, measured data of an adsorption chiller were used to develop such a model which is suitable to predict the outlet temperatures of the three hydraulic loops of the adsorption chiller. The model was validated with measured data under real working conditions. The simulated output temperatures show good agreement with the measured temperatures.