A reliability treatment applied to an analysis of optimal generation plant size

This paper describes the methodology and initial application of techniques which directly assess optimal generating plant size within the context of power system planning. The contemporary issue motivating this work is the question of optimal coal plant size, in which the economy of scale of large units is potentially offset by the higher per-kilowatt reliability contribution of smaller plants. The techniques used to analyze this question represent a significant extension of a linear programming generation planning model, into which inherently nonlinear reliability constraints are introduced in a linearized form. The method uses an external reliability calculation and iteratively converges to a reliability-feasible solution, though very little iteration has been needed. Preliminary results were obtained using load and economic forecasts supplied by the New York State Energy Office and with a unit data base provided by the New York Power Pool. These findings indicate that, given the large initial reserve in the New York System, reliability improvements from small plants have only minor effects on the optimal generation expansion. Further, the initial studies demonstrate the economy with which system reliability may be treated by the new planning model, with complete individual 20-year planning studies performed at the computational cost of around $55.