Modeling of a mechanical cooling system with variable cooling capacity by using artificial neural network

Abstract Capacity modification in mechanical cooling systems can be performed by various methods. Changing condenser temperature also changes the capacity of the cooling system. In this study, a series of experiments were performed in order to determine the effects of changing cooling water flow rate (changing condenser temperature) in a mechanical heat pump experimental setup on the cooling capacity of the system. Power consumption, thermal efficiency, coefficient of performance (COP) of the system in various cooling capacities were estimated theoretically by using the data acquired from the experiments performed. Performance values obtained were used for training Artificial neural network (ANN) whose structure was designed for this operation. The Network, which has three layers as input, output, and hidden layer, has one input and four output cells. Six cells were used in hidden layers. Training was continued until the square error became (e ⩽ 0.005) in this ANN, for which back propagation algorithm was used for training. Desired error value was achieved in ANN and, ANN was tested with both data used for training ANN and data not used. Resultant low relative error value of the test indicates the usability of ANNs in this area.