关键词: 镀膜设备, 状态预测, 灰色系统, 马尔科夫模型

Abstract: A two-stage and multi-level grey state Markov prediction model was proposed to address the problem of multiple states during device operation that make it difficult to accurately predict. Firstly,three predictive indicators were proposed to characterize the status of equipment,namely operating time,fault time,and downtime. Secondly,a new two-stage multi-level equipment state division system was established,and the levels of each equipment state were given. At the same time,the equipment state interval was defined based on the running time of the equipment state. The grey system theory was used to construct a two-stage and multi-level grey state Markov prediction model,which achieved the prediction of multiple equipment states and transformed regular preventive maintenance into precise maintenance based on different equipment states. Finally,taking the actual production status of equipment in a semiconductor panel manufacturing industry as an example,it was found that when the equipment is in the first stage of sub-health,a planned maintenance strategy can be adopted for the equipment. When the equipment is in the second stage of failure and downtime,a preventive repair strategy can be adopted. When the equipment is in a waiting state,maintenance strategies can be implemented for important mechanisms of the equipment. At the same time,the effectiveness and feasibility of the model were verified,and the prediction accuracy of the model was tested and analyzed.

Key words: coating equipment, state prediction, grey system, Markov model