研究目的
To enhance the accuracy of probabilistic PV power forecasting by proposing a novel combined probabilistic forecasting method based on an improved Markov chain.
研究成果
The proposed IMC method demonstrates improved performance in terms of both expected value and probability distribution forecasting compared to conventional methods. It provides more comprehensive and accurate information for power system operations and dispatching, helping to control the risk in power system operations with large-scale PV integration.
研究不足
The study acknowledges that when the sample size becomes massive, the diagonal elements of the one-step transition matrix tend towards 1, which may cause the transition between adjacent time points to converge to a certain mapping. This suggests a limitation in handling very large datasets.
1：Experimental Design and Method Selection:
The study proposes an improved Markov chain (IMC) forecasting structure that considers more influence factors beyond the statistical information of historical data. Rough set theory is used to refine the major factors, and a k-nearest neighbours algorithm is used to select similar samples for building an accurate forecasting model.
2：Sample Selection and Data Sources:
Two datasets are used for simulations: one from DESERT KNOWLEDGE AUSTRALIA Solar Centre and another from GEFCom
3：List of Experimental Equipment and Materials:
20 Not explicitly mentioned in the paper.
4：Experimental Procedures and Operational Workflow:
The proposed method involves factor reduction using rough set theory, selection of similar samples using KNN, and probabilistic PV forecasting using IMC.
5：Data Analysis Methods:
The performance of the proposed method is evaluated using mean absolute error (MAE), root mean square deviation (RMSE), quantile scoring (QS), and forecasting interval width (FIW).
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
