A Profit-Oriented Expert System for Coal Washery Optimization

Abstract

When designing a coal preparation plant, a process flowsheet should be determined to maximize profits. For a given constraint such as clean coal ash or yield, and by applying twelve basic flowsheets for coal preparation recently used in China, the optimum ash and yield composition of the coal products--coal product structure--can be obtained with the maximum profitability. In this article we studied the method of optimization in a coal plant designing expert system to acquire a required coal product structure. By comparing three fitted mathematical models of the raw coal washability curve and seven models of the partition curve, the optimum models are obtained. Based on the models, applying the profits of coal product as an objective function and utilizing the method of generalized Lagrange operators to conditionally constrain the yield and ash content of the coal product, the optimum coal preparation flowsheet was achieved using the Zangwill method. It provides an efficient theoretical basis for defining a technical plan in coal preparation plant design. This paper also studies the computer programming development and implementation of coal product structure optimization by applying an object oriented programming method in a coal plant design expert system. The overall scheme of the coal plant designing expert system, an explanation of the reasoning method, and a discussion of the self-learning mechanism are presented in this paper.