TY  - EJOU
AU  - Huang, Nantian 
AU  - Zhang, Jingyuan 
AU  - Ren, Shicheng 
AU  - Zhang, Hao 
AU  - Li, Bingling 
AU  - Wang, Yaoyao 

TI  - Coupled Meteorological-Electricity Behavior Analysis and Multi-Energy Load Forecasting Based on a Combined Model
T2  - Energy Engineering

PY  - 
VL  - 
IS  - 
SN  - 1546-0118

AB  - User electricity consumption behavior analysis and multi-load forecasting in integrated energy systems are crucial for system operation and scheduling. Traditional user electricity consumption behavior analysis fails to adequately incorporate meteorological factors, limiting the accuracy of characterizing user electricity consumption patterns. Traditional multi-load forecasting models do not consider the differentiated coupling relationships with meteorological factors across different seasons, which restricts the improvement of forecasting accuracy. To address the above issues, a method integrating data cleaning and meteorological correlation for electricity consumption behavior and multi-dimensional forecasting analysis is proposed. First, the Akima interpolation method is used to repair anomalous points in the load data. Second, the BK-Means algorithm is employed to determine the optimal number of clusters and initial centers, achieving a coupled analysis of wind-solar power output, load-meteorology clustering, and user electricity consumption behavior. Subsequently, the Kendall rank correlation coefficient method is applied to analyze the correlation between multi-dimensional loads and meteorological factors, constructing differentiated input feature sets for various loads tailored to different seasons. Finally, a combined model is used to generate multi-load forecasting results. The results demonstrate that compared to single forecasting methods, the proposed method achieves higher forecasting accuracy for electricity, cool, and heating loads across different seasons.
KW  - NRBO algorithm; BK-means algorithm; LSTM algorithm; XGBoost algorithm; multivariate load forecasting

DO  - 10.32604/ee.2025.072993