Experimental design and regression analysis for performance of a chiller system

Abstract

In this paper, we perform experimental design and regression analysis for performance of a chiller system. First, we complete design of experiment (DOE) on response so that determine the main and the interaction effects of the pre-defined factors. Second, we omit the insignificant factors from the analysis and, develop regression functions by considering the significant factors (inputs). Four factors are considered in DOE, these are: the water temperature, water flow rate, electronic expansion valve (EEV) opening percentage and compressor speed. Eight responses (outputs) are considered in DOE. They are the coefficient of performance (COP), capacity of evaporator, capacity of condenser, power consumption of the compressor, temperature of condensing, temperature of evaporation, superheating, and sub-cooling. The DOE results are analyzed, and regression functions are developed by MINITAB - a statistical software - at a 95% confidence level. ©2008 TIBTD Printed in Turkey.