Research on public building energy consumption prediction method based on hybrid analysis of dynamic and static data

Abstract

Energy consumption prediction plays an important role in the energy management of smart buildings. Through accurate prediction of electricity consumption, it can provide smart buildings with precise demand response strategies and promote the realization of smart building operation and maintenance. The measured historical energy consumption data (dynamic data) from the building networking system often has many abnormal values. Therefore, the method of predicting the energy consumption of buildings in the future based on dynamic data is not accurate and has low credibility. Aiming at this problem, this paper proposes a power consumption prediction method based on mixed analysis of dynamic and static data. First, based on historical energy consumption records, establish a time convolutional network to predict and output electricity energy consumption values in the future then use statistical analysis, expert experience analysis and other methods to determine the hourly energy consumption coefficient of office buildings, combined with power-consuming equipment Rated power and other static data are used to estimate the hourly energy consumption value of the office building finally, the hourly energy consumption forecast value of the office building s electricity consumption is weighted and merged with the hourly energy consumption estimate value to obtain the hourly energy consumption revised forecast value. The experimental results show that this method can predict the future electricity energy consumption with high accuracy and certain credibility under the condition of only using the historical energy consumption sequence of the building and simple static data of the building.