Automatically Generating Association Rules Under Diverse Operating Conditions Of A Large Scale Power Plant

Abstract This paper presents an algorithm for generating association rules from a given large database consisting of inputs and outputs of a 330 MW coalfired thermal power plant. The input parameters include fuel properties, operational parameters of the boiler, and the conditions on the stack. The outputs include emissions and heat losses along with the thermal efficiency of the power plant. The data has been collected at different power ratings of the plant ranging from 145 MW to 330 MW. A counting based algorithm is developed and presented in this paper that automatically determines the underlying correlations between the input and the output parameters of the plant’s operation. Furthermore, the algorithm generates all significant quantitative association rules that can be established between various inputs and outputs. Such association rules can in turn be used for designing an effective and robust control system for the plant.