Research On The Performance Of Heat Pipe PV/T System Based On Generalized Regression Neural Network

Solar energy is a renewable energy source with wide distribution,large reserves,easy access,no transportation,and clean and pollution-free.In the context of vigorously promoting energy structure reform,the solar industry has good development prospects.PV/T technology combines photovoltaic technology and photothermal technology to make full use of the entire solar spectrum,simultaneously produces heat and electricity to improve the comprehensive utilization rate of solar energy.Artificial Neural Network(ANN)technology mainly used to solve high-dimensional,nonlinear information processing problems.The purpose of this article is to research the influencing factors of PV/T system and improve the operating efficiency of it,combining generalized regression neural network technology with mathematical statistical methods to explore the effects of various meteorological conditions and operating parameters on the performance of heat pipe PV/T system and the operating characteristics of the system.The main research contents and related results of this paper are as follows:Firstly,the novel test bench of the PV/T system was established.The synthesis experiment for PV/T system of the Power generation and heat collection in typical summer day was designed,which could control the controllable variables(cooling water flow,cooling water temperature)to get the experimental data for influence of controllable variables and various uncontrollable interference variables on the heat,power generation,photoelectric efficiency,and photothermal efficiency of the PV/T system in the typical summer day.Based on the experimental results,5000 sets of typical data are selected as the data set,of which 4900 sets of data are used for training to obtain the heat output,thermal efficiency,power output,and electrical efficiency neural network model,the remaining100 sets of data are used for prediction result verification and error analysis.The actual operating characteristics of the PV/T system were obtained using GRNN model.Singlefactor impact analysis was conducted mainly for cooling water temperature and cooling water flow.Based on the prediction results of the neural network model,a multi-factor analysis of variance is used to perform significance analysis and use range analysis to obtain the primary and secondary order of factors affecting heat output and power output and the optimal level combination of each factor were obtained.The optimal controllable variable value was determined,which could provide guidance to the excellent operation of PV/T system under the influence of uncontrollable interference factors.