An Ultra-Short Term Load Forecasting Method Based on Improved Human Comfort Index

In order to solve the problem of calculating speed and accuracy of the large-scale ultra-short term load forecasting in micro-grid engineering practice, the characteristics of various kinds of load types are analyzed. The optimal local shape similarity model is established. The similarity of the two local similarity curves is expressed by similarity coefficient. This paper introduces the human comfort index which is used to characterize the meteorological factors, and improves the human comfort index formula according to the influence of the air quality index on the human behavior. According to the optimal local shape similarity series and real-time data, an optimal local shape similarity ultra-short term load forecasting method based on weighted average is proposed to obtain the initial value of the ultra-short term load forecasting. The factors that affect the accuracy of ultra-short term load forecasting are analyzed, and the calculation methods of various influencing factors are developed. According to the improved human comfort index and various influencing factors, the initial value of the ultrashort term load forecasting is firstly corrected. Based on the deviation between the real-time data and the forecast data, the ultra-short term load forecasting value is secondly corrected using the super-stable adaptive control theory. The second correction value is the final ultra-short term load forecasting value. The feasibility of the proposed method is analyzed by an example. The method has good adaptability to the calculation speed and calculation accuracy in the large-scale ultra-short term load forecasting, and can meet the actual demand of the field engineering.