Data-Driven Analysis and Modeling of Collective Generation Behaviors in an Electricity Market: A Perspective From Market Participants

To interpret and characterize the collective behaviors of power plants from the perspective of market participants, this article proposes a data-driven method by utilizing the public data of power plants and electricity market to analyze the relationships among the outputs of different power plants. The proposed method includes three major stages. In stage 1, a soft dynamic time warping-based clustering method is proposed to group generators with similar operation behaviors. In stage 2, the clusters' behaviors are analyzed in detail using a data-driven method. This stage utilizes the public data of power plants and the electricity market to analyze the relationships among the outputs of different power plants using the autoregressive model and Cholesky decomposition. In stage 3, the analysis results obtained from stage 2 are used in power system analysis considering collective generation behaviors by adjusting the cost models of generators. Finally, a comprehensive case study is carried out on a modified IEEE 118-bus system and a modified ACTIVSg2000 system to verify the proposed modeling approach. The simulation results show that the proposed method is valid and effective in modeling the observed collective behaviors of generating units and can be easily expanded to a large-scale power system. The impacts of different collective behaviors on locational marginal prices and transmission congestions are also analyzed in the article.