An operation auxiliary system for power plants based on decision-theoretic planning

Making good operation decisions during abnormal power plant conditions represents in many cases the possibility to avoid a unit trip or having economical losses. This paper introduces a novel way to built deterministic rules for making decisions for (i) maintaining a plant running under safe conditions, or (ii) dealing with process transients when an unexpected event occurs. Given that the power generation process can be seen as a planning problem under uncertainty, we present a variant of the Markov decision processes (MDPs) approach to provide a powerful solution to it. One of the main features of the MDP approach is that it assumes that the effect of a decision on the plant can be probabilistic. It also considers a utility function for a sequence of operation decisions in time, so that the decisions produced are usually optimal or approximately optimal. We tested our solution in a simplified version of the steam generation system. The model has been implemented and tested with a power plant simulator with promising results