A linear AC unit commitment formulation: An application of data-driven linear power flow model

Unit commitment is an important schedule procedure in power system daily operations, and an ideal version is to combine the unit commitment with the optimal power flow, thereby, it achieves an AC-power-flow-constrained unit commitment problem, which is referred to as the AC unit commitment. The AC unit commitment is quite benefit for system operation, however, the AC unit commitment problem is intractable since it is generally a large-scale non-convex and nonlinear program.To overcome this difficulty, this paper proposes a linear AC unit commitment model, which achieves a high -accuracy solution that is close to the solution of the exact AC unit commitment, while it keeps the problem from getting into a huge computational burden. To achieve a high accuracy, the recent data-driven linear power flow model is adopted, and for making the data-driven linear power flow model applicable for the unit commitment problem, a chance-constrained support vector regression approach is proposed to the replace the common least-squares regression in former data-driven linear power flow models, and physical models are used to deduce aid formulations to make the data-driven linear power flow model more robust.The resultant linear unit commitment model is tested on several standard systems, including a 2000-bus system, and the results verify the effectiveness of the proposed method, and show that the accuracy of the proposed data -driven linear power flow model improved about two-folds compared with the classical one.

Automated summary

This paper proposes a linear AC unit commitment model to solve the intractable large-scale non-convex and nonlinear AC unit commitment problem. By adopting data-driven linear power flow model and chance-constrained support vector regression, a high-accuracy solution close to the exact AC unit commitment is obtained, and physical models are used to deduce aid formulations to make the model more robust.