Solar absorption chiller performance prediction based on the selection of principal component analysis

Abstract

In this paper, a method to predict the performance of an absorption chiller using solar thermal collectors as the energy input is analyzed rigorously. Artificial Neural Network (ANN) is developed based on experimental data to predict the performance of the solar absorption chiller system at Universitas Indonesia. In order to perform ANN accurately, some parameters such as chilled water inlet and outlet temperatures, cooling water inlet and outlet temperatures, solar hot water inlet and outlet temperatures, hot water inlet and outlet temperatures, ambient temperature and fuel consumption flow rate are chosen as the input variables. In addition, a Principle Component Analysis (PCA) is used to reduce the number of input variables for performance prediction. Without sacrificing the ANN's prediction accuracy, PCA identified the sensitive variables from all input variables. The developed ANN model combined with PCA (ANN + PCA) shows good performance which has a comparable error with ANN model, specifically the configuration 9–6-2 (9 neurons, 6 inputs, 2 outputs) of the ANN + PCA model leads to a COP root-mean-square error of 0.0145.