Author/Authors :
Golpour Kandeh ، Soleiman Department of Industrial Management - Faculty of Management - Islamic Azad University, Central Tehran Branch , Ramazani Khorshid doost ، Reza Department of Industrial Engineering - Faculty of Industrial and Systems Engineering - Amirkabir University , Kabaranzadeh ، Mohammadreza Department of Industrial Management - Faculty of Management - Islamic Azad University, Central Tehran Branch
Abstract :
Expert systems are computer tools that, like an expert, advise on issues related to their area of expertise and support decision-making when required. These systems can be defined as counseling programs to solve complex problems that require experts to be solved. In this research, an expert system was designed to detect faults in the chemical process of polypropylene production. Using this system, all the information and experiences of experts can be accessed and used as a comprehensive resource. First, a diagnostic classification and fault detection is provided, which is prepared from a review of the literature related to the design of expert systems as well as the knowledge available in the polypropylene production process. Also, in this stage, the feasibility of the project was investigated, which was done by holding meetings with experts. In the second stage, the groups and constituent elements of the classification are explained. Likewise, more than 300 system faults were identified and coded to acquire the required knowledge for the system. In the next stage, the main elements of the fault detection system in the polypropylene production process are classified. Information related to Marun Petrochemical Company was used as a case study to further investigate the designed system. Also, the main reasons for defects and faults of the process were investigated and the frequency and percentage of each were calculated and reported. After classifying the reasons for the stoppages, the faults leading to each stoppage were extracted and classified. In the design stage, the prototype was coded for the system using JavaScript programming language and nodeje technology. In order to design the algorithm, each of the faults with one of the causes was considered as a scenario and related to a unique question to act as an intermediary between the expert system and the user in designing the user interface. Factors affecting the evaluation include the cost-consuming nature of the solution, the time-consuming nature of the solution, and the frequency of iteration of the fault. Finally, in the testing stage, the proper performance of the designed expert system was ensured.
Keywords :
fault detection , Expert System , diagnose errors , polypropylene production process
