Data generation method for power system operation considering geographical correlations and actual operation characteristics

Data generation methods for power system operation are the basis for many studies such as linear power flow models and data-driven models. On the one hand, the generated data can be used to test the performance of the model. On the other hand, the generated data can be used to train the model. In actual power systems, the operating data is not completely random, but conforms to certain operating laws, and has some typical characteristics, including geographic correlations and actual operation characteristics of thermal units. Taking the geographic correlations of renewable outputs as an example, the light density or wind speed in a certain area has similar characteristics. For another example, restricted by the unit cost of generating electricity, thermal unit outputs are at a relatively stable level rather than completely random. However, existing studies have not sufficiently considered these geographic correlations and actual operation characteristics. Therefore, this paper proposes a new data generation method considering geographic correlations and actual operation characteristics of thermal units. Specifically, the proposed method includes four steps: parameter estimation, sampling, generation benchmark settings, and system operation data calculation. To accurately characterize geographic correlations and actual operation characteristics of thermal units, the Eigendecomposition and unit commitment model are introduced into the proposed method. Numerical tests verify that the proposed method can guarantee the geographical correlations and actual operation characteristics of generated data.