A Multistage Semianticipativity Model for UC via Distributionally Robust Optimization Combined With Principal Component Analysis

In this article, we propose a new multistage
semianticipativity
model based on the distributionally robust optimization framework for dc power flow unit commitment problems with uncertain wind power. The proposed model constructs a flexible ambiguity set considering predicted wind power, which can flexibly adjust the robustness of the model according to the reliability of the predicted information. Moreover, a new approximation method based on the principal component analysis (PCA) is presented. After selecting the appropriate projection dimension, this method can improve the computation efficiency of the proposed model. At the same time, we propose a new linear decision rule based on PCA and provide the methods to address the problem that the time information disappears after using PCA. The characteristics of the proposed semianticipativity model are shown in a one-unit three-hour instance. Meanwhile, the performance of the proposed model and PCA-based approximate method is verified in six instances.