JAUE2018-022 Analysis and Forecasting of Cold and Heat Load in a School Building Based on Machine Learning

Abstract

With the development of intelligent energy system, the accuracy of user load forecasting is also rising. In the common user load, the research of power load forecasting develops faster because of its large amount of data accumulation. Due to the fact that the user's cold and heat load is difficult to collect on a large scale, the prediction of the user's cold and heat load is relatively slow. In the design of the user's cold and hot system, cold and heat load index for designers to use has a great redundancy. This often causes some of the cold and hot system to be useless all the year round. Therefore, accurate prediction of users' cooling and heating load is also a major focus in the field of load forecasting. The cold and heat load of users is mainly affected by temperature, humidity and user behavior. In this paper, neural network fitting algorithm in machine learning is used to predict the cooling and heating load of a school teaching building.