The Software Realization And Application Of Causality Diagram

With the increasing improvement of the automatic manufacture, industry manufacture is becoming more and more complicated. These bring out great complex, investment and maintain expenditure. People are forced to enhance integrated reliability and maintainability for great industry complicated system. Fault diagnosis gives the way to solve the question. The rapid development of computer technology and gradual research in AI make people borrow the rapid and exact computational ability of computer to copy with these complicated control system,fault diagnosis and decision-making system. The core to solve those is based of practice of AI. One of AI kernel question is how to express the existent knowledge and how to apply them for analysis and inference to obtain new knowledge. Because environments usually have uncertainty information, the research of uncertainty knowledge presentation and inference is very provided with practice. The developments of modern software engineering supply insurance to build the fault diagnosis based of AI include credibility,stability,easy access and independence between modules.This paper is based on the Chongqing science&technology surmounting project, whose name is 《The Research of the Industry Application Oriented Intelligent Developing Platform&System》.The paper is mainly introduce knowledge presentation and reason based on probability--causality network and technology in the system. The theory, which is found on and brought forward by Pro. Qin Zhang is the causality trees/diagram's knowledge representation methodology and calculating theory based on probability. The research bases itself upon building own knowledge right for our country in AI software and changing the situation which we just follow the outside research in AI.Author's main work is software realization and application for the theory. Now the work is to refactor the system and build a platform which adapt of industry fault diagnosis, and possess standard data format, high stability. To satisfy the above requirements, we introduce Rational Unified Process to control developing process. The paper has details for associated software engineering and represents the infrastructure of the system. At last, the paper introduces an application of vehicle engine lubricate system, which explains some elements of the system based of Causality network and inference. The result proved that it is accorded with real situation, the system positionfault quickly and efficiently.The developing work of《The Research of the Industry Application Oriented Intelligent Developing Platform&System》has been finished and is being system refactoring and commercializing packaged.