Accurate long-term power load forecasting in the grid is crucial for supply–demand balance analysis in new power systems. It helps to identify potential power market risks and uncertainties in advance, thereby enhancing the stability and efficiency of power systems. Given the temporal and nonlinear features of power load, this paper proposes a hybrid load-forecasting model using attention mechanisms, CNN, and BiLSTM. Historical load data are processed via CEEMDAN, K-means clustering, and VMD for significant regularity and uncertainty feature extraction. The CNN layer extracts features from climate and date inputs, while BiLSTM captures short- and long-term dependencies from both forward and backward directions. Attention mechanisms enhance key information. This approach is applied for seasonal load forecasting. Several comparative experiments show the proposed model’s high accuracy, with MAPE values of 1.41%, 1.25%, 1.08% and 1.67% for the four seasons. It outperforms other methods, with improvements of 0.25–2.53 GWh² in MSE, 0.15–0.1 GWh in RMSE, 0.1–0.74 GWh in MAE and 0.22–1.40% in MAPE. Furthermore, the effectiveness of the data processing method and the impact of training data volume on forecasting accuracy are analyzed. The results indicate that decomposing and clustering historical load data, along with large-scale data training, can both boost forecasting accuracy.

This study proposes four kinds of hybrid source–grid–storage systems consisting of photovoltaic and wind energy, and a power grid including different batteries and hydrogen storage systems for Sanjiao town. HOMER-PRO was applied for the optimal design and techno-economic analysis of each case, aiming to explore reproducible energy supply solutions for China’s industrial clusters. The results show that the proposed system is a fully feasible and reliable solution for industry-based towns, like Sanjiao, in their pursuit of carbon neutrality. In addition, the source-side price sensitivity analysis found that the hydrogen storage solution was cost-competitive only when the capital costs on the storage and source sides were reduced by about 70%. However, the hydrogen storage system had the lowest carbon emissions, about 14% lower than the battery ones. It was also found that power generation cost reduction had a more prominent effect on the whole system’s NPC and LCOE reduction. This suggests that policy support needs to continue to push for generation-side innovation and scaling up, while research on different energy storage types should be encouraged to serve the needs of different source–grid–load–storage systems.

Offshore wind energy has been widely exploited as a promising renewable energy source in China in recent years. This renewable source's high curtailment rate due to strict wind energy penetration limits, however, remains a challenge of concern. Also, the difference between peak and valley loads of electric power systems continues to increase with economic and societal development. To address these issues, this study proposes a hybrid system composed of a high-capacity offshore wind farm coupled with a pumped hydrostorage (PHS) power system to absorb rejected wind power. The hydraulic energy stored by PHS system is expected to provide guaranteed energy during peak load demands on the main power grid. This study presents a multiobjective technoeconomic assessment model based on the principle of dynamic economy for optimizing the size of the PHS system. The conclusions can be summarized as follows: (1) the seasonal characteristics affect the utilization of wind energy; (2) PHS operating strategies have an influence on the PHS configuration, energy utilization, and economic benefit; (3) reducing the wind curtailment is the key to improve the system performance and economic feasibility; (4) the drop in water level has a negative effect on economic benefit and (5) provides clear guidance to set external financial parameters for a hybrid power station. The results show that a size-optimizing study is important in determining the system's economic feasibility and predicted performance.