ON THE MODELS OF PEAK LOAD FORECAST UNCERTAINTY IN PROBABILISTIC PRODUCTION COSTING ALGORITHMS

This paper presents a model and related methodology for the problem of reliability calculation and production costing simulation with peak l oad forecast uncertainty. The model uses conditional load duration cur ves, each of them represents a realization of the random peak load. Th e conditional load duration curves are arbitrary but identical over th e interval of the base load domain, and linear in the peak load domain . The final load duration curve is a mixture (the integral) of the con ditional load duration curves. The following cases of the peak load fo recast random variable are considered: univariate gamma, beta and tria ngular distributions. The paper develops the calculation method for th e parameter(s) of mixing probability density functions in these cases, and the formulae for the final load duration curves are given. The me thodology can be used with any analytical or segmentation based convol ution methods.