Sintered Alumina Raw Slurry Ingredients Expert Systems Research And Development

With the economy developing, the Alumina demand is greater than supply, and how to improve the quality of alumina is paid more attention to in alumina manufacturers. Blending is the basis of the alumina production by sintering method (APSM), and it forms the pulp whose quality directly relates to the sintered mixture that influences significantly the technical economy results of the whole technology.Howerver, because of the process of pulp in APSM, in which time-delay and many environment effects exists, is too difficult to describe with precise and analytic mathematical models, and it can not be controlled properly by traditional methods. Based on the analysis of pulp in APSM and long-period knowledge of expert experience, a blending expert system that adapts productive features of pulp in APSM has been developed in this paper.Based on the blending theory, the knowledge of experts' experience derived from long-period blending production, the knowledge abstract method is described, the Backus-Naur Form (BNF) is adopted as the knowledge representation, the classified principle of rule collection and knowledge refinement are studied, and a blending expert knowledge base of pulp is designed after the organization principle is defined. Then, according to the knowledge characteristic and the knowledge BNF, the inference engine is realized through the hybrid of Heuristic Search among the rule collection, Forward Chaining in the interior rule collection and Hash algorithm between the precondition and the conclusion of the rule, which implements the match computation.Then, the function module and development strategy is discussed in the paper. The blending expert system has been put into service in an alumina smeltery, whose actual running results show that the system implements the computation of blending ratio fast and efficiently, improves the quality of pulp and stabilizes industrial production.At last, the conclusion about the research work and the future work that need be deeply researched is given.