Comparison of methods for building a capacity model in generation capacity adequacy studies

Abstract

The authors provide a comparison of different methods of preparing a capacity outage probability distribution in generating capacity adequacy studies. The recursive method, also referred to as the unit addition algorithm, with different rounding steps and rounding methods is analyzed. Capacity rounding techniques are typically utilized when dealing with units of noninteger capacities to reduce execution time. A new capacity rounding technique is proposed, in which the outage capacities of a unit are rounded to the nearest capacity steps before the unit is added to the system. It is found that this method yields very good results while requiring less computing time. The cumulant method based on the well-known Gram-Charlier or Edgeworth expansion is studied and compared with the recursive method. The convergence behavior of the cumulant method as a function of unit forced outage rate (FOR) values is examined. The accuracy and computing requirements of each method are discussed.<>