New approach to dynamic modelling of vapour-compression liquid chillers: artificial neural networks

This paper presents a new approach to modelling dynamic processes of vapour-compression liquid refrigeration systems. Using a dynamic neural network model for the performance prediction has been proposed. The model uses a generalised radial basis function neural network as inputs require only those parameters that are easily measurable. It then predicts relevant performance parameters such as the coefficient of performance or compressor work input. It was applied to two different dynamic processes of two different chillers and was found to be able to identify all process characteristics.