A NOVEL-APPROACH TO CIRCUIT SYNTHESIS IN MINERAL PROCESSING

Traditional plant design procedures seldom produce the optimum design for given plant parameters. Non-optimum design is due to a variety of factors including, amongst others, a poorly structured approach, reluc tance to undertake rime consuming iterative design, lack of equipment knowledge and the bias of design due to personal preferences. This pap er presents derails of a new technique for mineral processing plant sy nthesis. Using an approach incorporating various aspects of Artificial Intelligence, including Learning Classifier Systems, the idea is to c reate self-contained plant units that possess knowledge of applicabili ty from within. The process objects then bid for an appropriate positi on in the plant. As a processing plant is a multi-component process th e Intelligent Process Plant Object (IPPO) is used to advise on possibl e plant interactions during the bidding process. Using this approach i t is believed that the proposed technique will use up-to-date knowledg e and resources in a more efficient manner than conventional methods. As part of the development of the synthesis technique the paper descri bes a case study comprising of a three stage crushing and screening ci rcuit used in the quarrying industry. The new synthesis technique is a pplied to the manual plant design in order to determine if the approac h is capable of replicating the original circuit. The result of this a nalysis is presented. The idea of intelligent objects representing the plant and the subsequent synthesis of a circuit via bidding and compe tition has great potential for the minerals' industry. It is hoped tha t this procedure will save the design engineer considerable time, redu cing design expenses, and thus allowing the user to get closer to the goal of optimum plant design for given input and product requirements. (C) 1997 Elsevier Science.