The Design And Implementation Of The RBI Expert System Based On C/S

RBI technology consists of a series of systematic mechanisms of analyzing the risk of equipments or components, identifying the failure causation, and scheduling detection plan. It can extend the device operation cycle, shorten the maintenance period, and optimize equipment inspection plan, thus provides scientific decisions for the safe production. Over the past decade, RBI technology is widely used by international petrochemical enterprises to significantly reduce the operation risk and maintenance costs of equipments. Currently, RBI methods are widely used across many industrial areas, such as aviation, aerospace, petrochemical, pressure equipments or pipelines, oil or gas pipelines, and so on. Applying RBI methods for equipment inspection and maintenance management is becoming a trend. There are two issues to be addressed in RBI risk detection technology. On one hand, due to the lack of scalability in RBI software, it often needs to change the source code and recompile because of the modification of RBI standard and risk prediction model. On the other hand, risk detection of RBI technology makes high requirements for computer hardware for large quantity of data processing and complicated computation, and in practice it faces the problems of multi-object detection, scattered equipment locations identification, as well as multi-users collaboration. The low computation efficiency isn't sufficient for practical requirements. As a result, the study on RBI technology implemented by expert system is very meaningful.In this thesis, we build a knowledge database according to API581 standards which is stored in the form of XML, thus can address the modification of program source code and re-compilation issues when RBI standards updates and risk prediction model changes. In addition, the system uses hierarchical knowledge database structure, application data caching, and task manager based on thread pool. It greatly reduces the knowledge matching and data acquisition time and the server thread creation and destruction time. It also achieves application tasks parallelism so as to satisfy multi-users on network.In this thesis, since the design and implementation of RBI expert system is based on C/S model, it resolves the issue of RBI expansion to some extent. The system enables users to update RBI standards and adjusts the model according to actual conditions.