研究目的
To improve the accuracy of photovoltaic (PV) generation forecasts by utilizing spatial-temporal correlation analysis amongst PV generation data of distributed PV systems.
研究成果
The proposed method, integrating spatial similarity and temporal correlation into a Bayesian network-based inference model, consistently outperforms baseline methods in PV generation forecasting accuracy for very short-term horizons.
研究不足
The effective range for the proposed method is within 60 km. The similarity metrics may not reflect the distance between two PV sites on sunny days accurately.
1：Experimental Design and Method Selection:
The study employs a Bayesian network-based inference model for PV generation forecasting, incorporating spatial similarity and temporal correlation analysis.
2：Sample Selection and Data Sources:
PV generation data from distributed PV systems and weather data are used.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned.
4：Experimental Procedures and Operational Workflow:
The methodology includes spatial similarity check, temporal correlation check, and Bayesian network inference.
5：Data Analysis Methods:
Root Mean Squared Error (RMSE) is used for accuracy evaluation.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
