Optimal response of a thermal unit to an electricity spot market

This paper addresses the optimal response of a thermal unit to an electricity spot market. The objective is to maximize the unit profit from selling both energy and spinning reserve in the spot market. The paper proposes a 0/1 mixed-integer linear programming approach that allows a rigorous modeling of (i) nonconvex and nondifferentiable operating costs, (ii) exponential start-up costs, (iii) available spinning reserve taking into account ramp rate restrictions, and (iv) minimum up and down time constraints. This approach overcomes the modeling limitations of dynamic programming approaches and is computationally efficient. Results from realistic case studies are reported.