A novel information enhanced Grey Lotka–Volterra model driven by system mechanism and data for energy forecasting of WEET project in China

• Published in: Energy (Oxford) Vol. 304; p. 132176
• Main Authors: Duan, Tianyao, Guo, Huan, Qi, Xiao, Sun, Ming, Forrest, Jeffrey
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 30.09.2024
• Subjects: algorithms; China; clean energy; electricity; energy; Grey Lotka–Volterra model; Information enhanced; Network; Power forecasting; power generation; West-to-East Electricity Transmission; China; Grey Lotka–Volterra model; West-to-East Electricity Transmission; Information enhanced; Power forecasting; Network
• ISSN: 0360-5442
• DOI: 10.1016/j.energy.2024.132176

Diversified energy power generation is a critical component of China’s West-to-East Electricity Transmission (WEET) project and a key driver of China’s clean energy strategy. Aiming at the complex non-linear relationship of inter-regional energy system and fully exploring the information of system data, our paper proposes a novel information-enhanced Grey Lotka–Volterra model (IE-GLVM). The novel model consists of Grey Lotka–Volterra equations and universal network terms, which achieves a clever fusion of energy system competitive relationships and data information-driven modeling in the modeling methodology. In addition, the Joint Gradient Descent method is used to optimally search for all the parameters of the novel model, and we theoretically prove the stability of the algorithm. Based on this, the IE-GLVM model is used to analyze the competitive and cooperative relationships among the three provinces of Sichuan, Hubei, and Jiangsu in the middle line of the WEET project in China under multiple energy sources for power generation and to forecast the future power generation. Eventually, IE-GLVM was compared with three benchmark models, and it demonstrated superior performance in most cases. An analysis and summary of the power generation relationships of each regional energy source were conducted based on the quantitative results of the IE-GLVM model. •A novel information enhanced grey Lotka–Volterra model is proposed for diversified energy power generation forecasting.•The new model combines the advantages of mechanism modeling with data modeling and improves prediction accuracy.•The model can analyze the competition and cooperation relationship of the electrical structure and predict future.•The joint gradient descent method is proven to smoothly optimize all parameters of the model.•The model is used to analyze the electrical structure in China’s West-to-East Electricity Transmission project.