Repair Method of Chiller Power Consumption Monitoring Data Based on Multiple Linear Regression Model

With the gradual advancement of energy-saving work in public buildings, the number of energy monitoring systems continues to increase. However, due to the low level of information management, quality problems are common in monitoring data; there are cases where data is missing or abnormal, which hinders the well progress of energy conservation. Therefore, the method for quickly and accurately repairing problem monitoring data needs to be studied. The core part of the power consumption of HVAC is the freezing station, in which the power consumption of the chiller occupies a major part, needs to be concerned. This paper proposes to repair the data by establishing energy consumption model. The research object is the chiller of a campus education office building. Firstly, the factors that affect the power consumption of the chiller are mastered. Then, the operating parameters are simulated by e-Quest software, and a multivariate regression model of power consumption with meteorological parameters is established. Finally, the model is used to repair the problem data. The results show that the correlation coefficient between the leaving supply temperature of water and power consumption is about 0.9, which has the greatest impact and is of great significance to the retrofit of the systems. And, the repair accuracy for missing data reaches 2–3%, indicating that the method is fast and effective and can be applied online.