Research Of Large-scale Reverse Osmosis Desalination Fault Diagnosis Expert System

With the development of technology and society, there is an acute shortage of water. Thus,people have been drawing great attention to this issue gradually. Among numerous ways ofseawater desalination, the method of reverse osmosis is advancing rapidly and becomes one of themost widely used techniques for its small investment, low energy consumption and shortconstruction period. Nowadays, there are lots of massive or super scale reverse osmosis seawaterdesalination projects all around the world that alleviate the stress of water shortage in some areaseffectively. And it becomes a hot research topic that how to ensure the safety and stability of theextensive and complex reverse osmosis seawater desalination project. Therefore, in order to serveour daily production better, it is necessary to explore an intelligent fault diagnosis system orientedto the reverse osmosis project. Thus, the main researches are as follow:1) By way of reading many relative literatures, the development trend of the reverse osmosisproject is summarized and research situation of fault diagnosis and expert system techniques basedon reverse osmosis project is analyzed.2) The subject of this study is a100,000tons per day seawater desalination program on theLiuheng Island of Zhoushan. Based on its operating principle, system architecture and process flow,the reason why the abnormal running of reverse osmosis system would worsen water quality isbeing studied intensively, in order to supply enough knowledge for fault diagnosis.3) The logic relation between the most valuable faults with other faults is being presented andqualitatively analyzed by fault tree analysis. And the minimum cut set, minimal ratio set andstructure importance of the fault tree is obtained. The result of the analysis makes foundation for thedesign work of expert system inference engine.4) Based on fault tree analysis, the minimum cut set of fault tree and production rule will becombined to design a knowledge base of expert system. A fault fact table, minimum cut set facttable, fault rule table, fault condition table and fault maintenance table are established in theknowledge base to store knowledge of reverse osmosis seawater desalination. According to thedesigning characteristic of knowledge base, the expert system inference engine, in which aknowledge-based reasoning method, pros and cons mixed reasoning method and search strategy ofdepth-limit-first are used, is designed to improve reasoning efficiency. Then the design of expertsystem is basically finished.5) Instantiating the fault diagnostics expert system of reverse osmosis seawater desalination in the workbench of Visual Studio2008and SQL Server2005. With the real-time data of reverseosmosis project from the software, users can diagnose the fault, explain the fault and offer advice bydiagnosing the fault automatically or inputting the fault symptoms manually. A simulative test offault diagnostics expert system is conducted by simulating the real-time data and inputting the faultsymptoms manually. The result shows that the fault diagnostics expert system, which based on faulttree analysis, is of great relevance to applications...