Study on Data-driven Modeling of Energy Bus System Based on Machine Learning Algorithms
The district energy bus system(EBS)integrates a variety of renewable energy and waste heat resources,and has become a core infrastructure for realizing building heating/cooling in the context of the current low-carbon energy transition.To explore the data-driven modeling methods applicable to EBS,this paper took typical hotel buildings in hot summer and cold winter areas as the research objects and used machine learning methods such as multiple linear regression,support vector regression,random forest regression,artificial neural network,and Gaussian process regression to build models based on simulation data in Dymola under typical system working conditions.Then,the applicability of different algorithmic models was compared and analyzed to provide a basis for modeling and optimization of EBS.The results of comparative analysis of data models show that artificial neural network,random forest regression,Gaussian process regression based on Matern kernel,Gaussian process regression based on radial basis function kernel,and support vector regression based on radial basis function are more suitable for EBS.The R2 of the above model can be kept above 0.90 in most cases,so as to predict the performance variation trend of the EBS and make timely strategic responses.
Energy Bus system(EBS)hot summer and cold winter regionsdata-drivenhotel buildings
NETL
NSTL
万方数据
