To enhance the efficacy of UC results, this paper proposes an ML-based adaptive robust optimization method to solve the UC problem in the power system connected to wind power generation. The ML techniques are first used to calculate the Calinski-Harabasz index, leading to the separation of forecasting errors of intermittent renewable energy sources. Then, the uncertainties are grouped using the spatial clustering tool and the average density of the K-means distribution. The nonparametric algorithm for clustering points in space and incorporating noise, the discrete uncertainty in renewable energy sources, and the outliers in the comprehensive uncertainty set are taken into account accordingly. Improving and combining a few strategies and frameworks and utilizing Density-Based Spatial Clustering of Applications with Commotion (DBSCAN) presents an unused fashion of vulnerability clustering so that each cluster compares to the RES vulnerability information. A Principal Component Analysis (PCA) set uses these normal boxes as a nonparametric method of estimating Dirichlet density and analyzing different dimensions of large datasets. Finally, by implementing the proposed model in the widely used power system based on renewable energy sources, IEEE 39-bus and 118-bus systems, the solution process is changed, and the proposed technique is analyzed and compared, considering the uncertainty of the clusters in each subset of ML.
The rest part of this paper is arranged as follows: the problem formulation is defined in Section II. Section III presents the proposed method. Test results and discussion are provided in Section IV. Finally, the work conclusions are described in Section V.