GA and RBF Based Real-Time FDD for Refrigeration Units

Purpose of this research is to describe the application of a radius basis function (RBF) network to the problem of real-time fault detection and diagnosis (FDD) in a vapor compression refrigeration system. First, we analyze the common refrigeration system faults and their dominant symptoms. Next, an FDD strategy for the refrigeration system is proposed which adopts a RBF network to model the causation of symptoms and faults. Gaussian functions are selected as the basis functions of the hidden layer neurons. The parameters of the Gaussian functions and the weights of the network are obtained by using a novel network training method which combines genetic algorithm (GA) and pseudo-inverse matrix algorithm. Finally, a real-time FDD program in C language based on the proposed strategy is developed. The FDD program is tested with the refrigeration system installed in a laboratory and successfully identifies each of the six faults artificially introduced at the laboratory.