To address the issue of declining prediction accuracy caused by the lack of data in newly constructed wind and solar power stations, this paper introduces a transfer learning-based forecasting approach for wind and photovoltaic power. The method incorporates sensitive meteorological feature selection and utilizes a Temporal Convolutional Network–Attention–Long Short-Term Memory (TCN-ATT-LSTM) model. Spearman’s rank correlation, mutual information entropy, and Pearson correlation are employed to investigate the relationship between meteorological features and power output. Through evidence theory, meteorological features with a cumulative contribution exceeding 85% are selected as inputs for the wind and solar power forecasting model. The TCN-ATT-LSTM network is pre-trained to extract common knowledge, and transfer learning is applied to fine-tune (FT) the model through network parameter adjustments. This enables the adaptive model to be quickly constructed for target wind and solar power stations with limited data, improving the prediction accuracy. Finally, the effectiveness of the proposed method is validated through its application to data from a projected wind and solar power station planned for a region in northwestern China. The proposed method not only enhances forecasting accuracy for emerging wind and solar power stations with limited data but also has significant implications for the renewable energy industry.
The proposed approach involves a method of joint optimization configuration for wind–solar–thermal-storage (WSTS) power energy bases utilizing a dynamic inertia weight chaotic particle swarm optimization (DIWCPSO) algorithm. The power generated from the combination of wind and solar energy is analyzed quantitatively by using the average complementarity index (ACI) to determine the optimal ratio of wind and solar installations. We constructed a multi-objective optimization configuration model for the WSTS power generation systems, considering the equivalent annual income and the optimal energy consumption level as objective functions of the system. We solved the model using the chaotic particle swarm optimization algorithm with linearly decreasing dynamic inertia weight. To validate the effectiveness of the proposed approach, we conducted a simulation using the 2030 power energy base planning data of a particular region in Inner Mongolia. The results demonstrate that the proposed method significantly improves the annual income, enhances the consumption of wind–solar energy, and boosts the power transmission capacity of the system.
This study addresses the integral role of typical wind power generation curves in the analysis of power system flexibility planning. A novel method is introduced for extracting these curves, integrating an enhanced K-means clustering algorithm with advanced optimization techniques. The process commences with thorough data cleaning, filtering, and smoothing. Subsequently, the refined K-means algorithm, augmented by the Pearson correlation coefficient and a greedy algorithm, clusters the wind power curves. The optimal number of clusters is ascertained through the silhouette coefficient. The final stage employs particle swarm and whale optimization algorithms for the extraction of quintessential wind power output curves, essential for flexibility planning in power systems. This methodology is validated through a case study involving wind power output data from a new energy-rich provincial power grid in North China, spanning from 1 January 2019, to 31 December 2022. The resultant curves proficiently mirror wind power fluctuations, thereby laying a foundational framework for power system flexibility planning analysis.
With the rapid development of renewable energy sources such as wind and solar, the net load characteristics of power systems have undergone fundamental changes. This paper defines quantitative analysis indicators for net load characteristics and examines how these characteristics evolve as the proportion of wind and solar energy increases. By identifying inflection points in the system’s adjustment capabilities, we categorize power systems into low, medium, and high renewable energy penetration. We then establish adjustment models that incorporate traditional coal power, hydropower, natural gas generation, adjustable loads, system interconnections, pumped-storage hydroelectricity, and new energy storage technologies. A genetic algorithm is employed to optimize and balance the net load curves under varying renewable energy proportions, analyzing the mechanism behind net load balance. A case study, based on real operational data from 2023 for a provincial power grid in western China, which is rich in renewable resources, conducts a quantitative analysis of the system’s adjustment capability inflection point and net load balancing strategies. The results demonstrate that the proposed method effectively captures the evolution of the system’s net load and reveals the mechanisms of net load balancing under different renewable energy penetration levels.