Research On Building Energy Consumption Forecasting Based On GA-ELM Method

With the acceleration of urbanization in China and the increasingly sophisticated construction functions,building energy consumption is increasing.At this stage,building energy consumption accounts for 26.7% of the total energy consumption in the society,and there is huge potential for energy saving.Building energy consumption prediction is an important content of building energy saving.Its fast and accurate prediction results are the basis for building energy efficiency optimization.Therefore,how to construct an efficient and accurate building energy consumption prediction model is one of the research hotspots in the field of building energy efficiency optimization.Taking a smart production office comprehensive building as an example,based on the collection and analysis of building energy consumption data,it focuses on the use of genetic algorithms to improve and establish a building energy consumption prediction model to improve the efficiency and accuracy of energy consumption prediction results.Energy-saving optimization provides a quick and accurate reference basis,laying a foundation for building energy-saving work.The article mainly focuses on three aspects: the construction of energy management system,the establishment of prediction model and the simulation of engineering application.An efficient and accurate building energy consumption prediction system is built.The thesis mainly studies from the following aspects:(1)Building energy management system construction: According to the characteristics of large public buildings,such as large total energy consumption and large proportion of air conditioning system energy consumption,a building energy management platform is established.The energy consumption system can be adjusted according to the building function zone,and it has strong extensibility.And the cooling load of the air conditioner can be predicted in the system,which is beneficial to the management of energy consumption.(2)Building energy consumption prediction model: Based on analysis and comparison of current typical energy consumption prediction algorithms,according to the problems existed in the application of traditional BP neural network in building energy consumption prediction,Extreme Learning Machine algorithm is introduced,and genetic algorithm for optimization is adopted.Then the Genetic Algorithm Extreme Learning Machine algorithm based on genetic optimization algorithm is put forward.The data collected by the building energy management system is pre-processed to establish GA-ELM energy consumption prediction model.(3)Energy consumption prediction simulation and energy saving optimization management: the building energy consumption prediction model is applied to an actual engineering project.A reasonable load type,time period,collection points and evaluation criteria are selected to perform simulation experiments and the results are analyzed.The research results show that,compared with BP,GA-BP and ELM methods,building energy consumption prediction based on GA-ELM method is feasible and can further improve the prediction accuracy.On the basis of this research,a certain energy consumption management strategy is adopted in conjunction with engineering examples to study the rational allocation of energy consumption during the operation phase of the project,which provides a reference for further energy saving and consumption reduction.