Research On Recovery And Remediation Of Failed UID In Product Tracking

The requirement of safety management and quality assurance for the complex product is consistently increasing.Product tracking is the most effective way to achieve the product configuration management and quality tracing.UID(Unique Identification,UID)is the direct and effective method for the product lifecycle tracing.DPM(Direct Part Marking,DPM)working group is established by the international institutes and enterprises,DM(Data Matrix,DM)is selected as a component of unique identification.However,the complexity of environment and production process of the discrete manufacturing enterprise can easily wear,pollute and corrode Data Matrix.Therefore,it is difficult to automatically identify product identification based on the existing algorithms.This is called identification failure,which will lead to track circult.Therefore,it is necessary to study the failure identification recovery.At present,the majority of research has focused on the single angle of process,graphics image,but a set of theoretical system and technological framework about product tracking process of failed identification recovery is rarely seen.There exist problems in the product tracking model of discrete manufactureing enterprise.Such problems include batch sample information homogenization,batch identification partial or total failure,warning and timely remediation,but the existing technical methods could not solve these problems.This paper aims to solve the problems of different failure patterns and enterprise tracking mode.The failure identification recovery of product reliable tracking process is studied.Research works are as follows:(1).Product identification failure form of tracking process is analyzed;recovery or remediation concept is proposed;the problems of failed identification are discussed in product tracing;key issues and solutions are descriped.And then the author has established product failure identification recovery or remediation strategy,has described the strategy and process model,and has carried out from identification of homogeneous product design,failure identification prediction recovery and information fusion remediation.This provides a guidance for the product uninterrupted tracking.(2).Homogenizatied product identification enhanced feature design and identification method is researched.In order to solve the homogeneity problem of the sample information of batche product,feature enhanced direct parts marking is analyzed.The DM code feature design based on marking process and affine transformation were proposed.Manufacturing process texture feature design was proposed based on GLCM.Texture library and Marking feature matrix were established.Finally,using Hamming distance and Euclidean distance method were matched similarity the target characteristics and marking feature matrix and manufacturing process texture library respectively,feature differences between individuals was enhanced from the source.(3).Remediation of failed identification based on product multi-information fusion method is studied.The complexity of environment and process of manufacturing enterprise leads to identification partial or total failure.Part multi-states and time series of product manufacturing process are analyzed,based the product physical and manufacturing process information,the recovery or remediation model of product failed identification is proposed based on multi information fusion.The calculating method of the similarity between the historical database of failure identification parts and product characterization data was proposed.The modified coefficient weighting method is used to make the initial matching.Finally,basic reliability of the distribution features and fusion identification was accomplished by D-S evidence theory.(4).Prediction and remediation of failed product identification was put forward based on manufacturing history data.Firstly,relation between the mounts of history manufacturing process data and identification was analyzed.Product manufacturing history data model based on DPM was proposed.Secondly,the factors of failed product identification were analyzed,and extracted feature was optimized through PCA method.Thirdly,neural network model for prediction failed product identification was established.Using neural network prediction results with identification transfer and inheritance were predicted or remedied failed product identification.(5).In terms of project background,product manufacturing process data acquisition and tracking system using product failed identification recovery or remediation method was built,which illustrated the process and the method validated by the implementation of product data acquisition and tracking system in an aerospace enterprise.