| In recent years,new technology application projects such as digital workshop and intelligent factory,which are guided by intelligent manufacturing,are facing a construction boom in China.At present,intelligent manufacturing adopts the top-down construction mode.In the process of construction,enterprises lack the understanding of intelligent manufacturing.There are some blindness,and its risk has not been widely recognized.Because the intelligent manufacturing is directly related to efficiency,quality and safety,it is necessary to take risk analysis and risk management into consideration.In this paper,the industrial intelligent maintenance system in petrochemical industry is selected for research.Through some enterprises investigation,the existing problems are thoroughly analyzed.The risk factors at the R&D and operation stages are identified and analyzed.The Bayesian network is selected to model according to the actual situation of the research object.The preliminary structure of Bayesian network is established by using expert knowledge,and the parameters of nodes in the network are learned by using expert data,thus a complete Bayesian network model is obtained.The structure learning algorithm is studied by using the generated simulation data,and the main factors are found by sensitivity analysis and maximum causal chain analysis in GeNIe software.Through the description and analysis of the actual cases of enterprises,the problems in the construction of industrial intelligent maintenance system are deduced by using evidential data,and the results are consistent with the actual situation.The validity of Bayesian network risk analysis is verified.Through the combination of expert knowledge and risk model,put forward some suggestions for risk control.The research of industrial intelligent system is suggested according to the method of Engineering science. |
