Abstract:Frost growth was a strong nonlinear and time-varying process,and the experimental data was(usually) interfered by noise;so it was hardly to predict its characteristics precisely.In order to solve this problem,a learning method support vector machine(SVM) which is based on the structure risk minimization principle was introduced;then a set of thermal parameters,time and space coordinates were selected;finally the predication models of frost growth characteristics were presented based on those characteristic vectors.The predicated results were verified by the experimental data,which shows a good agreement.Comparing SVM predication model with LS predication model shows that the SVM model has better ability to deal with nonlinear and noise interference problems.

A Frost Growth Characteristics Predication Model Based on Support Vector Machine