Study of frosting diagnosis with Back Propagation network and Elman network

Frosting diagnosis is important in refrigeration system control. This paper focuses on frosting diagnosis in a single-stage vapor compression refrigeration system. According to the reduction of refrigerating capacity, the experimental data of frosting can be divided into three categories: light, moderate and severe frosting. Then, two frosting prediction models are established with Back Propagation neural network (BPNN) model and Elman neural network (ENN) model. The simulation results illustrate that the overall diagnostic performance of the ENN is better than that of BPNN. The mean squared error (MSE) should be paid more attention for affecting the accuracy of models. The presented method in this paper can provide technical reference for frosting prediction.