A Machine Learning-Enhanced Benders Decomposition Approach to Solve the Transmission Expansion Planning Problem under Uncertainty

Stefan Borozan,Spyros Giannelos,Paola Falugi,Alexandre Moreira,Goran Strbac

2023-04-15

Abstract:The necessary decarbonization efforts in energy sectors entail the integration of flexibility assets, as well as increased levels of uncertainty for the planning and operation of power systems. To cope with this in a cost-effective manner, transmission expansion planning (TEP) models need to incorporate progressively more details to represent potential long-term system developments and the operation of power grids with intermittent renewable generation. However, the increased modeling complexities of TEP exercises can easily lead to computationally intractable optimization problems. Currently, most techniques that address computational intractability alter the original problem, thus neglecting critical modeling aspects or affecting the structure of the optimal solution. In this paper, we propose an alternative approach to significantly alleviate the computational burden of large-scale TEP problems. Our approach integrates machine learning (ML) with the well-established Benders decomposition to manage the problem size while preserving solution quality. The proposed ML-enhanced Multicut Benders Decomposition algorithm improves computational efficiency by identifying effective and ineffective optimality cuts via supervised learning techniques. We illustrate the benefits of the proposed methodology by solving a number of multi-stage TEP problems of different sizes, based on the IEEE24 and IEEE118 test systems, while also considering energy storage investment options.

Systems and Control

What problem does this paper attempt to address?

This paper attempts to address the computational challenges encountered in power system planning, particularly when long-term uncertainties need to be considered. The transmission expansion planning (TEP) problem becomes extremely complex and difficult to solve. Specifically, the paper focuses on how to significantly reduce the computational burden of large-scale multi-stage stochastic TEP problems without compromising the quality of the solution. Traditional methods usually address computational infeasibility by simplifying the original problem, which may lead to the neglect of key modeling aspects or affect the structure of the optimal solution. To this end, the paper proposes a method that combines machine learning (ML) with the established Benders decomposition to manage the problem size while maintaining solution quality. This method improves computational efficiency by identifying effective and ineffective optimality cuts through supervised learning techniques. The paper demonstrates the advantages of the proposed method through multiple multi-stage TEP problems of different scales based on the IEEE24 and IEEE118 test systems, and considers energy storage investment options.