The inclusion of dynamic factors in statistical power system cost models. II. Part loading and reserve costs

For pt.I see ibid., vol.4, no.2, p.419-25 (1989). Techniques are developed for analyzing the costs of maintaining adequate operating reserver on power systems, within probabilistic production costing models and other based on load distribution functions. The aim is to show that such models can be developed to approximate the full range of dynamic penalties associated with practical power system operation. Algorithms for costing predetermined part loading are presented, taking into account the need for bringing plant online out of strict merit order. The author describes how a prediction error function can be used to estimate the probability that the specified spinning reserve is insufficient and hence approximate for the cost of gas turbine or storage utilization for dynamic control. This allows estimation of optimal reserve levels. The author discusses the analysis of longer term reserve from banked thermal units. Some results of the model are compared to those reported from hourly simulation studies