基于故障相关性分析的主轴系统关键备件需求预测研究

发布时间：2018-07-12 17:12

本文选题：主轴系统 + 可靠性　；参考：《吉林大学》2015年博士论文

【摘要】：数控装备种类日渐增多，机床维修备件的需求预测也越来越复杂，如何实现备件科学确定与合理优化，以达到保障精确化的目的，已成为数控装备现代化建设道路上日显突出和亟待解决的现实问题。数控机床使用中故障零部件信息是备件需求预测的依据。本文针对传统研究中忽视故障相关性和环境要素对可靠性影响导致备件需求预测精度较差问题，提出综合考虑故障相关性的关键备件评价理论和基于环境协变量可靠性分析的备件需求预测方法，并以数控机床主轴系统为对象展开研究。本文首先进行了主轴系统关键备件的选择，选择依据为故障模式的危害度、中心度和相关度的综合分析值。在进行故障模式危害度分析过程中，为避免小样本数据计算失真现象，在子系统故障模式发生频率计算中引用贝叶斯理论，利用Slice抽样方法，引入辅助变量实现边缘模拟抽样，结合Gibbs抽样的思想来估计综合信息，实现零部件故障危害度计算；然后采用DEMATEL方法，对相关故障子系统中表征其重要度的“影响度”、“被影响度”、“原因度”和“中心度”进行分析计算；根据相关子系统的数量、作用方向和是否具有间接相关关系对相关故障链进行分类，利用故障子系统的综合故障率、独立故障率和相关故障率之间关系，建立相关系数模型，完成主轴系统相关故障链中，不同子系统间的相关系数的求解，由此建立相关度的数学模型，并完成主轴子系统相关度值的分析计算；最后根据故障零件的故障模式的危害度和在相关故障链中的中心度以及故障相关度进行关键备件的选择。其次，针对选择的关键备件，本文采用基于可靠性的方法进行备件的需求量的预测。根据可修系统故障损伤累积程度导致故障系统维修后状态的差别进行分类，并对不同状态下子系统的故障过程模型进行分析，据此进行故障截尾、时间截尾的广义更新过程模型的推导；通过对主轴系统工作环境影响要素分析，在可靠性回归模型分析的基础上，采用比例风险模型建立关键备件的可靠性模型，利用SPSS软件对环境协变量进行筛选，并求得协变量的系数；根据不同类型的维修备件的故障随机过程特征，推导出考虑环境协变量的备件需求预测模型；对主轴系统的关键备件圆柱滚子轴承和主轴，分别进行了基于环境协变量的需求预测，并对广义更新过程预测模型的参数采用最大似然函数进行估计，并利用遗传算法实现，而故障发生随机点的期望值采用蒙特卡罗仿真求得。此部分证明了环境协变量对备件需求量的影响，同时证明了维修度的存在对故障变化规律的影响。本论文运用理论分析、生产数据验证、模拟仿真相结合的方法，从分析故障机理的微观角度出发，揭示了在内外因素的影响下主轴系统零部件可靠性变化规律；提出了基于贝叶斯理论的危害度分析方法，完善了FMECA理论；通过创建关键备件的选择机制，建立了故障相关性分析的理论体系，为数控机床的可靠性精确评估，拓展了新的视角；在考虑环境协变量的情况下，结合故障子系统维修后的恢复状态实现备件的精确预测。研究结果为数控机床的可靠性研究理论、企业资金预算、备件库存理论研究提供了新的理论依据。主要创新： 1.利用故障模式的危害度、子系统的中心度及子系统间的相关度的分析和综合，建立了基于故障相关性分析的系统关键备件选择理论和概念模型。 2.将分析错综复杂的社会相关关系的DEMATEL方法引入数控机床故障相关关系的分析，通过子系统之间的相关作用频率和作用方向，计算了表征子系统在相关故障链中重要性的“四个度”，并以中心度的值作为主轴系统关键备件的选择依据之一。 3.进行主轴子系统的可靠性评估时，考虑了子系统内部故障相关性，针对I.F故障模式进行了深入的研究，归纳了相关故障链种类，创建了基于故障链种类的相关系数求解的理论体系。 4.在关键备件进行需求预测的过程中，考虑了子系统外部操作环境要素的影响，对其可靠性模型设置了环境协变量，并采用可靠性回归模型中的风险比例模型对关键备件进行可靠性建模，并结合备件的维修后状态选择不同的随机过程模型实施备件预测。
[Abstract]:The number of numerical control equipment is increasing, and the demand prediction of machine tool maintenance spare parts is becoming more and more complex. How to realize the scientific determination and rational optimization of spare parts so as to achieve the purpose of ensuring the precision has become a realistic question which is prominent and urgent to be solved on the road of modern construction of numerical control equipment.
The information of the fault parts in the use of CNC machine tools is the basis for the prediction of spare parts demand. In this paper, the problem of poor prediction accuracy caused by the failure correlation and the influence of the environmental factors on the reliability of the spare parts is neglected in the traditional research, and the key spare parts evaluation theory and the reliability analysis based on the environmental covariate are put forward. The method of spare parts demand forecasting is studied based on the spindle system of CNC machine tools.
This paper first chooses the key spare parts of the spindle system, chooses the comprehensive analysis value based on the hazard degree, the center degree and the correlation degree of the fault mode. In the process of analyzing the hazard degree of the fault mode, to avoid the distortion of the small sample data, the Bayesian theory is used in the calculation of the frequency of the subsystem failure mode, and the Sl is used. Ice sampling method, introducing auxiliary variables to realize edge simulation sampling, combined with the idea of Gibbs sampling to estimate the comprehensive information and realize the calculation of the damage degree of parts and components, and then use the DEMATEL method to divide the "influence", "shadow loudness", "cause" and "Centrality" in the related fault subsystems to represent their importance. Analysis and calculation, according to the number of related subsystems, the direction of action and whether there is an indirect correlation to classify the related fault chains, using the comprehensive failure rate of the fault subsystem, the relationship between the independent fault rate and the related failure rate, and establishing the correlation coefficient model, complete the correlation fault chain of the main axis system, and the correlation system between the different subsystems. The mathematical model of the correlation degree is established, and the correlation degree of the spindle subsystem is analyzed and calculated. Finally, the key spare parts are selected according to the damage degree of the fault mode of the fault parts and the center degree in the related fault chain and the fault correlation degree.
Secondly, in view of the key spare parts selected, this paper uses the reliability based method to predict the demand of spare parts. According to the degree of failure damage accumulation of the repairable system, the difference of the state after the maintenance of the fault system is classified, and the fault process model of the different state system is analyzed, and the time of the failure is truncated. On the basis of the analysis of the reliability regression model, the reliability model of the key spare parts is built on the basis of the reliability regression model analysis. Based on the analysis of the reliability regression model, the SPSS software is used to screen the environmental covariate and obtain the covariate of the covariate. According to the different types of the model, the coefficient of the covariate is obtained. The prediction model of spare parts requirement considering environmental co variable is derived from the random process characteristics of maintenance spare parts. The requirement prediction of the key spare parts cylindrical roller bearing and spindle of the spindle system is carried out based on the environmental covariate, and the maximum likelihood function is used to estimate the parameters of the generalized updating process prediction model. It is realized by genetic algorithm, and the expectation value of the random point of failure is calculated by Monte Carlo simulation. This part proves the influence of the environmental covariate on the demand of spare parts, and proves the effect of the existence of maintenance degree on the law of fault change.
In this paper, using the method of theoretical analysis, production data verification, simulation and simulation, from the analysis of the microcosmic angle of the failure mechanism, the paper reveals the reliability change rule of the spindle system components under the influence of internal and external factors, and puts forward the hazard analysis method based on Bias theory, perfect the FMECA theory, and through the creation of key points. For the selection mechanism of spare parts, the theoretical system of fault correlation analysis is set up, and a new angle of view is developed for the accurate evaluation of the reliability of the CNC machine tools, and the precise prediction of the spare parts is realized with the restoration state after the maintenance of the fault subsystem in the case of the environmental covariates. The research results are the reliability research theory of the CNC machine tools. The theory of capital budgeting and spare parts inventory provides a new theoretical basis.
Main innovation:
1. based on the analysis and synthesis of the damage degree of the fault mode, the centrality of the subsystem and the correlation between the subsystems, the theory and concept model of the key spare parts selection based on the fault correlation analysis are established.
2. the DEMATEL method of analyzing the complicated social correlation is introduced into the analysis of the failure correlation of CNC machine tools. The "four degrees" of the importance of the subsystem in the related fault chain are calculated by the frequency and the direction of the interaction between the subsystems, and the selection of the key spare parts of the spindle system is taken as the value of the center degree. One of the basis.
3. when the reliability evaluation of the spindle subsystem is carried out, the internal fault correlation of the subsystem is considered, and the I.F fault mode is studied deeply. The types of the related fault chain are summed up, and the theory system of the correlation coefficient solution based on the type of fault chain is created.
4. in the process of demand prediction for key spare parts, considering the influence of the external operating environment factors of the subsystem, the environmental co variable is set up for its reliability model, and the reliability model of the key spare parts is modeled by the risk proportion model in the reliability regression model, and the different random processes are selected in the condition of the maintenance of spare parts. The model implements spare parts prediction.
【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TG659

【参考文献】