Application And Study Of Regional Building Cooling And Heating Load Forecasting Model Based On Bayesian Theory

In recent years, the number of researches and applications of regional energy supply technologies are increasing, among which, the cold and heat load prediction of regional architecture research and analysis methods is an important foundation and technical support for regional energy technology development and promotion. In the research group based on analysis of the existing load forecasting methods of building, proposing a prediction model of regional building cooling load based on Bayesian theory, which is a regional building cooling load prediction method designed for the lack of detailed information on the construction during the regional energy planning stages. The method is lack of practical engineering verification and analysis, so the accuracy and the applicability of this method will be tested and analyzed with the analysis of actual cases.This paper first gives a brief introduction of the regional building cooling load prediction model based on Bayesian theory, and uses Design Builder simulation software as the simulation software and Matlab mathematical programming software as the numerical calculation software. Then, selected respectively three cases as the verification cases: Case 1 – a residential area in a certain city of cold area; Case 2-- a school in a certain city of hot summer and cold winter zone; Case 3 –an integrated living area in a certain city of cold area. Through the investigation and statistics of the basic information of the three cases, respectively establish the respective regional standard single buildings, and dynamic the load simulation to obtain hourly /daily dynamic load.The obtained load of each case are treated as prior information, the hourly / daily load value surveyed and statistics of the other regions corresponding to the case are treated as the sample information. In each case substitute the prior information and the sample information into the Bayesian regression model, then calculate the posteriori information, as the load forecasting factor. The correction results and the corresponding simple superposition area predicted result are compared to the measured load. And the accuracy for each case is analyzed and predicted by the evaluation index of the building load prediction method(hourly / daily relative error, root mean square error and the maximum relative error ratio)Through the analysis and comparison of the error of the results can be verified, it proves that the accuracy of the method is better than the simple superposition of the area. By comparing three cases of relative accuracy, applicability of the method: 1) prediction on the regional building cooling load in summer and winter heat load are available; 2) it is suitable for China's various types of building thermal region prediction;3) the area building load forecasting in different types and scale are applicable;4) the better the quality of the sample information is and the more the number of cases is, the better the prediction is.