Ramp Rate Constrained Unit Commitment by Improved Adaptive Lagrangian Relaxation

Abstract

This paper proposes an improved adaptive Lagrangian relaxation (ILR) for ramp rateconstrained unit commitment problem. The proposed ILR minimizes the total supply cost subject tothe power balance, 15 minute spinning reserve response time constraint, generation ramp limitconstraints, minimum up and new down constraints. ILR is improved by new minimum down time toaccount for startup and shut down ramp constraints, new initialization to obtain a high qualityinitial solution, dynamic economic dispatch to include the operating ramping limits, and adaptiveLagrangian multipliers to speedup the convergence. If the 24 hour schedule is feasible, dynamiceconomic dispatch by quadratic programming is used to minimize the production cost subject to thepower balance and new generation ramp operating frame limit. For hours with insufficient 15minute response time spinning reserves, repairing strategy by quadratic programming is used tominimize the generator fuel cost subject to power balance, generator operating ramp rate limit, and15 minute spinning reserve response time constraints. The proposed ILR algorithm is tested on the26 unit IEEE reliability test system. It is shown that ILR could obtain a higher quality solution thandynamic economic dispatch based on artificial neural network.