A Probabilistic Production Costing Methodology for Seasonal Operations Planning of a Large Hydro and Thermal Power System

This paper describes the results or several years' work in designing and implementing a probabilistic production costing methodology in the model for seasonal operations planning at the Swedish State Power Board (SSPB). The probabilistic production costing methodology can handle a detailed representation of the thermal system and the interchange of power to neighbouring systems. The nuclear units are modelled as a multistate and multiblock unit in combination with certain regulating rules for the off-loading process. The fossil fired units are in general multiblock units and the successive blocks for a given unit are not sequentially merit order loaded. The load is represented by three duration curves for one week. The probabilistic production costing simulation is based on a two-stage approach using the cumulant method. The first stage works in an iterative mode with the optimizer and the second stage is employed when the optimizer has converged. The model of seasonal planning including the probabilistic production costing model has been implemented in an information system (POP) at the SSPB.