Research On Equipment Key Component Condition Assessment Based On Multi-source Information Fusion

The advent of the era of intelligent manufacturing provides new opportunities for the development of manufacturing industry,on the other hand,also put forward higher requirements for the performance and reliability of mechanical equipment and its key component.In the practical production process,from the perspectives of economy and safety,to ensure the operation reliability of equipment and continuous production of system,accurately assessing condition of the equipment and its key component is the key point of current research.In addition,the development of multi-sensor systems also provides effective data support for key component condition assessment.In order to effectively utilize multi-source information to assess condition of the equipment and its component,this paper constructs a three-layer condition assessment framework based on information fusion.In the data analysis layer,multi-source information is collected and analyzed.In the initial diagnosis layer,the model based on intelligent algorithm and the model based on evaluation system are selected as the methods of condition assessment.In the decision fusion layer,the improved evidence theory based on the optimal support factor coefficient is used to carry out the fusion of initial assessment results.Based on the framework,a condition assessment model based on improved evidence theory is established.It aims to effectively improve the validity and accuracy of condition assessment and has important academic and practical significance.The main work of the paper is divided into three aspects:Firstly,this paper summarizes the status quo of the current condition assessment,describes the concepts and methods of multi-source information fusion,explores the significance of condition assessment and multi-source information,introduces the application of multi-source information fusion in condition assessment domain,and present the research issues.Secondly,according to the explanations the information and methods in condition assessment domain,a multi-method condition assessment framework based on multisource information fusion is summarized and established.On the basis of the framework,a condition assessment model based on improved evidence theory is established.Through the establishment of a corresponding evaluation system and neural network structure,fuzzy synthetic evaluation method and BP neural network method are utilized to analyze and process the multi-source information to get the initial condition assessment results,and finally use improved evidence theory to fuse the initial results to get the key component condition.Thirdly,classical evidence theory is difficult to deal with conflicting evidence,so this paper starts from the research hotspots of evidence theory,analyzes the causes of evidence theory paradox,and classifies the metrics of evidence.Then,according to the fact that the improvement of evidence theory in the previous research is based on a single metric and fails to fully express the nature of the evidence,this paper not only considers the degree of conflict and degree of difference between the evidences,but also pays attention to the dispersion of the evidence,and proposes an improved evidence theory method based on the optimal support factor coefficient.In the end,through case studies and theoretical studies,the superiority and effectiveness of the improved method are verified.Due to the preservation of the good features of the fusion rule,and the evidence improvement factor has been described in many aspects,so the fusion results are more accurate.