Methodology And Case Study For Determining The Direct Design Radiation Of Solar Heating System Based On Big Data

Solar energy is a typical clean energy source and an important choice for dealing with energy crises and air pollution.The solar heating system based on parabolic trough collector(PTC)can solve the problem of air pollution caused by coal-fired heating in winter.The value of design radiation should be carefully determined during the designing a solar system,as it would commonly affect the solar collector area,and then the economics and performance of the system.However,in the existing publications,most studies ignore the influence of the daily and seasonal variation of radiation and commonly choose the fixed value of design radiation in some typical days,leading to a non-optimal performance of the system with unreasonable setting parameters.In addition,due to Air Quality Index(AQI),economic conditions and equipment maintenance in some cites,solar radiation is distorted or missing.Support Vector Regression(SVR),as a kind of machine learning algorithm,can correct and predict its value.Therefore,focusing on solar heating systems,the main work and conclusions are as follows:1.A method is proposed to determine the design radiation of a parabolic trough collector(PTC)heating system without an auxiliary boiler.This method introduces the concept of non-guaranteed days to determine the system rated capacity.What’s more,the cost of unit heating supply is employed in this method as an optimized object of the heating system.Based on this method,a case study of four typical solar radiation areas in China is carried out.By comparing with the traditional selection method,the rationality of the method is proved.The optimal design radiation of each region is determined.Then,the relationship between non-guaranteed days and the initial investment of the system is obtained.Finally,an economic analysis of the system in Mentougou is performed.2.Considering AQI,based on data from 12 cities in 3 years,the SVR algorithm is used to predict the daily horizontal direct solar radiation in China.It is found that SVR shows good performance for both separate models and countryside general model.In comparison of the two kernel functions,SVR based on radial basis kernel function is more suitable for direct radiation prediction in China;adding AQI could increase the prediction accuracy by 5%-85%.Finally,the Pearson correlation coefficient method is used to understand the relationship between input parameters and output parameters in the separate models and countryside general mode.The results show that the sunshine duration is the most important parameter.3.Life cycle assessment is applied to the parabolic trough collector and the flat plate collector(FPC)to compare the environmental impact of each square meter collector during the whole life cycle.The results show that the global warming potential of PTC is 22.32% higher than that of FPC,and the primary energy consumption is 6.60% higher than that of FPC.The remaining indicators like acidification potential,eutrophication potential,ozone layer consumption and human body are all relatively low.