Modeling the Relationship Between Air Conditioning Load and Temperature Based on Machine Learning

Air conditioning energy consumption accounts for more than half of the energy consumption of office and living buildings. The analysis of the correlation between air conditioning load and ambient temperature is the main basis for the study of air conditioning load scheduling strategies. This paper proposes a relationship model between air conditioning load and environmental temperature based on support vector machine (SVM). Firstly, the ambient temperature and air conditioning load data were standardized, and the normalized data were regressed by support vector machine to obtain the optimal relationship curve. The results of relational regression were de-standardized. At the same time, the temperature was divided into a range of 0.1 degrees Celsius, and the average load of air conditioning in each temperature range was calculated. Comparing the optimal relationship curve with the ambient temperatures-average air conditioning load data, the coefficient of determination is 0.94. The results show that the model presented in this paper can obtain a higher precision air conditioning load and ambient temperature relationship model.