Availability modelling of stochastic power sources

Energy sources which are stochastic in their behaviour include solar radiation, wind energy and run-of-the-river hydroplant. In addition to conventional considerations of plant availability, reliability models for these sources must also represent the stochastic character of the prime energy supply. The paper initially develops probabilistic models for representing power availability and then proceeds to obtain stochastic transition rates for predicting changes of availability. Power gereration is written as a function of intrinsic variables, which describe the state of the relevant elements of the physical environment. Computations of conditional probability are employed to evaluate the probability density profile of power availability. Transition-rate evaluation depends on relationships of time variation for intrinsic variables, in the presence of uncertainty which is modelled by noise functions. The results of these evaluations can be used to integrate stochastic power sources with the reliability index computations of the entire power system to which they may be connected, or used individually, as in the case of isolated sources.