Wind power forecasting in China based on a novel variable order grey Bernoulli model with weighted time-driven term

Wind power has the advantages of being clean, environmentally friendly, renewable, and widely available. Accurate wind power prediction is crucial for grid operation and energy planning. In order to deal with regional imbalance and nonlinear characteristics of wind power data series, we propose a Variable Order Grey Bernoulli Model with Weighted Time-driven Terms, namely VOGBMWTT. The introduction of the weighted time-driven term provides the model with a strong adaptive structure, significantly improving its applicability to time-varying and nonlinear data sequences. Additionally, the ɛ−constrained Marine Predator Algorithm is employed to optimize the model parameters, which improves the estimation accuracy and optimization speed of parameters. Generally, satisfactory results can typically be achieved within 300 iterations. In particular, the model’s dynamic performance, convergence, and robustness are verified through dual validation methods of online and offline performance. In comparison with the Artificial Bee Colony (ABC) algorithm and Particle Swarm Optimization (PSO) algorithm, the ɛ−constrained Marine Predator Algorithm optimization algorithm demonstrates significant advantages in terms of faster convergence and higher accuracy. The model’s predictive capability is validated using wind power generation data from four representative administrative regions in China, including Xinjiang Uygur Autonomous Region, Inner Mongolia Autonomous Region, Jiangsu Province, and Guangdong Province. It is found that the model has good adaptability to data sequences with various characteristics, including slow fluctuations, drastic fluctuations, smooth to abrupt transitions, and saturated features. By forecasting for the foreseeable short term (2024–2026), it can be observed that China’s wind power development will maintain a steady growth momentum. The research results provide valuable insights for achieving accurate wind power prediction and informing energy policy formulation based on it.