Analysis Of Disc Turning Quality And Optimization Of Key Parameter

As a medium for storing data in mechanical hard disks,disks,together with mechanical hard disks,are developing in the direction of "small size,large capacity".As a result,the requirements for the processing technology of the discs are getting higher and higher,and the processing methods that were originally suitable for large-size discs are no longer applicable.Turning,as a process for processing the size of the disc,has the greatest impact on the size of the disc.Originally,production lines often used DOE(Design of Experiment)tests and other methods to improve quality,but the limitation of this method is that it cannot be applied to assembly line production and can only solve some more obvious problems.Therefore,how to further improve the turning quality of the disc has become an urgent problem for the production line.Based on the above question,this paper proposes a disc turning quality analysis method based on machine learning algorithms.Integrated all aspects of the production line data to find a combination of risk factors that affect the quality of disc turning,that is,the risk path.And give the sequence of key parameters.According to the characteristics of key parameters,different parameter optimization methods are adopted to achieve the purpose of improving the quality of turning.The main research contents are as follows:(1)The overall framework of disc turning quality analysis and key parameter optimization is constructed.First,sort out the disc machining process of the total process and turning.Secondly,through business analysis of disc turning,the data source for subsequent analysis was determined.Then introduce the method of evaluating the turning quality of the production line-CPK(Process Capability Index)in detail,which is the actual process capability index.Finally,a method for disc turning quality analysis and key parameter optimization for the purpose of improving CPK is proposed.(2)Strongly correlated data subset mining and risk path analysis of disc turning.Aiming at the "high pass rate and low value density" characteristics of disc turning data,a double Cart decision tree disc turning quality analysis method is proposed.First,through the first Cart decision tree analysis of the total data,a subset of data that is strongly related to the quality of disc turning is mined.Then the second Cart decision tree analysis is performed on the data subset.By analyzing the split node information of the two Cart decision trees,a combination of risk factors that have a bad influence on the quality of disc turning is obtained,which is the risk path.And get the ranking of key influencing factors.According to historical data and actual production verification of the production line,it is known that the quality of discs produced under this risk path is lower,which verifies the accuracy of the risk path.(3)Optimization of key parameters for disc turning.The key factors affecting disc turning are divided into non-process parameters and process parameters.For non-process parameters,take management actions to improve.According to the feedback of the production line,the quality of disc turning has been improved after the improvement.For the process parameters,the gray correlation method is used for analysis.Different combinations of optimal process parameters have been obtained for different machines.And through actual data verification,it is known that the quality of the disc produced under the optimal process parameter combination is higher.