Study And Realization On Chemical Hazop Expert System Based On Petri Net Modeling

Petri net and expert system have become two main active fields in automation and artificial intelligence (AI). Combination of Expert system and Petri net has been studied for the application of chemical HAZOP, furthermore, a chemical HAZOP expert system based on Petri net is developed for the application of chemical HAZOP in a DMF recovery project for a company.The first half of the dissertation presents the principle of Petri-HAZOP expert system, introduces the method of knowledge representation and inference rules based on Petri-HAZOP analysis. The main elements of Petri net are defined using Visual Prolog. The dissertation proposes the realization of knowledge representation and inference rules based on Petri-HAZOP analysis, which is programmed by Visual Prolog. With an actual example, the dissertation also establishes a Petri-HAZOP expert system. A methodology of knowledge representation based on Petri net model is proposed for HAZOP expert system. The first order predicate logic (Horn clauses) is associated with fuzzy Petri net. According to the characteristic of HAZOP, knowledge representation based on fuzzy Petri net and its modeling algorithms with Horn clauses are presented. In addition, logic reasoning with Horn clauses is effectively associated with knowledge representation based on fuzzy Petri net. The methodology is also applied to an illustrative example with the intent of configuring HAZOP expert system for chemical processes. Fuzzy Petri net graphics is automatically analyzed and exported by importing the contents of HAZOP, which makes the expert system more heuristic and transparent.The second half of the dissertation presents the general structure of Petri-ChemHAZOP; every function module is explicated in detail, including system definition, real-time database, Prolog/Horn knowledge representation and inference engine, human machine interface of HAZOP expert system and the realization of Petri-ChemHAZOP by Visual Prolog. The dissertation pays much more attention to the algorithm of inference engine based on Petri-ChemHAZOP expert system, the algorithm of FPN modeling based on Prolog/Horn and the algorithm of knowledge reasoning by Petri net. At last, the principles of the chemical HAZOP expert system based on Petri net modeling is used in a DMF recovery project. The principles and operation rules of the evaporating system, rectifying system and centrifugal pump system were analyzed for the knowledge acquisition. Then the knowledge base module of the evaporating analysis section, rectifying analysis section and centrifugal pump analysis section were developed by Visual Prolog.