Prediction of frosting process on cold wall surface based on artificial neural network with back propagation algorithm

The artificial neural network with back propagation algorithm is a multi-layer feed-forward network, which suitable to study unsteady frost formation multiple factors. used layer growth on cold flat surface, where four feature variables including temperature of the velocity, relative humidity and air are adopted. experiment generates database, good for training due its fast speed high precision based Levenberg-Marquardt learning rule. establishment model in this paper can quickly accurately predict height surface different control variables, helpful implementation defrosting.