Combination of Lagrangian-relaxation and linear-programming approaches for fuel-constrained unit-commitment problems

The paper presents a new method for determining the unit-commitment schedule of a power system which operates under fuel utilisation constraints. The proposed method employs the Lagrangian relaxation approach to determine a feasible suboptimal schedule. Then, linear programming is applied to improve the feasible solution, as well as seeking the optimal economic dispatch of the committed generators. The algorithm is implemented on a VAX 11/780 machine to solve the unit-commitment problem of a system which consists of 26 thermal units and 6 fuel-constrained units. Numerical results presented in this paper describe the usefulness and practicality of the proposed method.