A Fast Solution Method for Large-scale Unit Commitment Based on Lagrangian Relaxation and Dynamic Programming

TL;DR

This paper introduces a rapid solution approach for large-scale unit commitment problems in power systems, combining Lagrangian relaxation and dynamic programming to improve efficiency and feasibility.

Contribution

It presents a novel hybrid method that estimates initial solutions with Lagrangian relaxation and refines them using dynamic programming for large-scale systems.

Findings

Effective on IEEE 24-bus system

Validated on Polish 2383-bus system

Demonstrates high efficiency and accuracy

Abstract

The unit commitment problem (UC) is crucial for the operation and market mechanism of power systems. With the development of modern electricity, the scale of power systems is expanding, and solving the UC problem is also becoming more and more difficult. To this end, this paper proposes a new fast solution method based on Lagrangian relaxation and dynamic program-ming. Firstly, the UC solution is estimated to be an initial trial UC solution by a fast method based on Lagrangian relaxation. This initial trial UC solution fully considers the system-wide con-straints. Secondly, a dynamic programming module is introduced to adjust the trial UC solution to make it satisfy the unit-wise constraints. Thirdly, a method for constructing a feasible UC solution is proposed based on the adjusted trial UC solution. Specifically, a feasibility-testing model and an updating strategy for the trial UC solution are established in this part. Numerical tests are implemented on IEEE 24-bus, IEEE 118-bus, Polish 2383-bus, and French 6468-bus systems, which verify the effec-tiveness and efficiency of the proposed method.