Bayesian Energy Consumption Prediction Method For Campus Buildings Based On Monitoring Data And Its Application

With the continuous development of higher education in China,the energy consumption of university buildings is increasing continuously.In the urban energy consumption composition,the energy consumption of university buildings with a floor area of only 3%-7%accounts for 30%.Therefore,China vigorously promotes the concept of green campus and the construction of energy consumption supervision platform.The severe energy consumption situation and national policies require us to accurately predict and master the energy consumption of campus buildings.Therefore,as an important way to study campus energy consumption and design and transform campus energy consumption system,the establishment of high precision campus building energy consumption model has attracted extensive attention and research.Based on energy consumption monitoring data,this paper improves the traditional bottomup building energy consumption prediction model establishment method.According to the Bayesian theory,the simulated energy consumption density of the prototype building is combined with the monitored energy consumption density of the sample building,and the revised predicted building energy consumption density is obtained.Finally,the high-precision bottom-up campus building energy consumption model is established.At the same time,Matlab program is used to simulate the influence of the number of monitoring sample buildings on the prediction accuracy under the accurate value of energy consumption density of different research objects and different prototypes.Under the condition that the prediction accuracy is guaranteed and the cost is considered,the optimal sample number in the process of Bayesian correction is determined.Taking Dalian University of Technology as an example,this paper establishes a bottomup campus building energy consumption model using the proposed method.At the same time,using the campus energy consumption monitoring platform data to verify the prediction results.In the verification area,taking office buildings as an example,the maximum monthly energy consumption error of the traditional method is 54.8%,and the average monthly absolute error is 14.54%.The maximum error of the improved method is 14.86%,and the mean value of the absolute error is 7.16%.Compared with the traditional model,the prediction accuracy of monthly energy consumption of the improved model for a single type of building is greatly improved.In the prediction results of the whole region,the maximum monthly error of the traditional model is-14.13%,and the mean monthly absolute error is 8.11%.The maximum monthly error of the improved model is 5.22%,and the mean value of monthly absolute error is 2.19%,indicating that the error is significantly reduced.By comparing the actual energy consumption platform data,it is proved that the improved model has good prediction accuracy at both the single building type and the whole region level.In the application part of the model,a distributed energy system suitable for campus energy consumption is established by using the high-precision bottom-up model annual hourly forecast data and the data related to Dalian’s renewable resources.At the same time,through the economic analysis,the static investment payback period of campus reconstruction is calculated to be only 5.23 years.More than 28.96 million yuan of electricity and heating costs can be saved every year.It not only greatly saves the operating cost of the campus,but also makes full use of the renewable resources,which is an effective way to build a green campus and implement the policy of energy conservation and emission reduction.Finally,based on the establishment of renewable power generation system and electric storage heating system,this paper explores a variety of strategies to improve the matching degree of renewable power-load in the system and the utilization rate of renewable electricity in the system.By calculating the matching degree of renewable power-load,it is proved that adopting the ice storage air conditioning system and adding electric vehicle charging piles can improve the matching degree to different degrees and reduce the power abandonment rate of the renewable power generation system.This paper provides an optimization method for the energy consumption prediction model of campus buildings in cold areas.This method can effectively improve the accuracy of the model prediction,and can provide quantitative reference and implementation path for the optimization design of green campus and campus energy conservation and carbon reduction work.