基于乏信息的滚动轴承磨削过程可靠性评估

发布时间：2018-03-01 06:14

本文关键词： 可靠性磨削过程泊松过程乏信息融合灰自助最大熵模糊集合非排序灰关系　出处：《河南科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着机械制造技术的发展,为了使轴承零件达到更高的技术要求,滚动轴承磨削过程越来越复杂也越来越难控制,经常遇到概率分布未知和小样本的可靠性分析难题,如果仍采用传统的统计学方法分析滚动轴承磨削过程的可靠性,则结果是不理想的。因此,针对滚动轴承磨削过程的可靠性问题,本课题基于乏信息系统理论,运用灰自助原理、最大熵方法、模糊集合理论、泊松过程和非排序灰关系,对滚动轴承某质量参数进行分析,以完成滚动轴承磨削过程可靠性评估。主要分为4个部分:1融合隶属函数法、最大隶属度法、滚动均值法、算术平均值法和自助法,提出一种乏信息融合技术,以调整机床的加工误差。运用该融合技术融合机床试加工时输出的小样本数据,获取工件的估计真值,参照估计真值对机床的加工误差进行调整;运用模糊集合理论,在置信水平P=95%的条件下,借助机床调整后输出的小样本可靠数据,预测出工件的估计区间,以判断调整后机床的可靠程度。案例研究表明,乏信息融合技术能够实现机床加工误差的调整,且调整后的机床可靠性至少为95%。2针对滚动轴承磨削系统运行状态的可靠性,推荐一种乏信息评估方法,以解决小样本和概率分布未知条件下的可靠性评估问题。基于工件质量检测的小样本数据,用灰自助方法生成大量模拟数据;用最大熵方法处理大量模拟数据,构建磨削系统运行状态的概率密度函数;设定置信水平P,获取磨削系统运行状态的置信区间与扩展不确定度;根据磨削过程中输出工件的质量数据,通过泊松计数过程,计算磨削系统运行状态的变异强度;借助泊松过程的无失效概率,得到磨削系统运行状态的可靠性函数。案例研究表明,置信水平P为95%是磨削系统实现良好运行状态的最佳选择点。该结果为合理调整机床提供了理论依据。3基于磨削系统运行状态可靠性研究,运用乏信息评估方法对磨削系统运行状态演变的可靠性进行实时评估。获取多组检测数据序列,在最佳置信水平P=95%下,绘制磨削系统运行状态演变的可靠性曲线图,以实时预测磨削系统运行状态演变的可靠性。案例研究表明,若检测数据的品质实现可靠度r0.65,检测数据与本征数据的可靠性概率密度函数的面积交集A0.3,检测数据的变异概率PB≤0.7,则说明磨削系统运行状态演变是可靠的,没有发生变异;否则,认为磨削系统运行状态演变是不可靠的,其发生了变异。该结果为判断磨削过程是否可靠提供了理论依据。敏感性分析表明,样本大小不影响其评估结果。4根据非排序灰关系和模糊集合理论,对滚动轴承磨削过程的变化程度进行评估。运用非排序灰关系,设在灰置信水平P=95%下,借助按顺序采集的时间数据序列,求解两个时间数据序列之间的灰属性权重,并构建时间数据序列之间的灰属性权重相似矩阵;运用模糊传递闭包法,求解时间数据序列之间的灰属性权重等价矩阵;通过假设性检验和分段平均等价性系数,描述磨床系统误差的变化过程,以实现磨削过程变化程度的评估。案例研究表明,该评估方法能够准确描述磨床系统误差的变化过程,实现了磨削过程变化程度的评估,且评估结果与实际情况相符。
[Abstract]:With the development of manufacturing technology, in order to make the bearing parts meet the requirements of higher technology, rolling bearing grinding process more complex more and more difficult to control, often meet the reliability probability distribution is unknown and small sample problem analysis, if the reliability analysis of rolling bearing grinding process using the traditional statistical methods, the result is not ideal. Therefore, in view of the reliability problem of rolling bearing grinding process, the lack of information system based on the theory, using the principle of Grey Bootstrap, maximum entropy method, fuzzy set theory, Poisson process and non sort grey relation, on some quality parameters of rolling bearings of bearing grinding to complete the process of reliability evaluation is mainly divided into rolling. The 4 part: the 1 fusion method of membership function and the maximum membership degree method, rolling average method, arithmetic mean method and bootstrap method, presents a lack of information fusion technology to adjust Machining error. The use of small sample data output of the fusion fusion machine test processing, the true value of the estimated acquisition work, according to the estimated true value error of machine tool adjustment; using the fuzzy set theory, the confidence level of P=95% under the condition of small sample, reliable data by means of machine tool adjustment output, forecast the interval estimation of the workpiece, to determine the extent of reliable adjusted machine. Case study shows that the lack of information fusion technology can realize the machining error of machine tool adjustment, reliability and adjusted at least 95%.2 reliability for the running state of rolling bearing grinding system, recommended a poor information evaluation method to solve the problem of reliability evaluation of small sample and unknown probability distribution under the condition of small sample data. The quality of workpieces based on detection of a large number of simulation data generated using the method of Grey Bootstrap; using the maximum entropy method. And a lot of simulation data, the probability density function of the construction operation state of grinding system; setting the confidence level P, confidence interval and extended access operation state of grinding system uncertainty; according to the quality of data output workpiece in grinding process, the Poisson process, variation of strength calculation of running state of grinding system; no failure probability with Poisson process the obtained reliability function of running state of grinding system. A case study shows that the confidence level of 95% P is the best choice of point grinding system to achieve good running state. The results for the reasonable adjustment of machine bed provides a theoretical basis for the research of.3 system based on the operation reliability of grinding using the reliability assessment method, lack of information on the evolution of the running state of the grinding system the real time assessment. Access to multiple sets of test data sequences in the optimal confidence level P=95%, drawing the running state of the grinding system modeling The reliability curve of the reliability evolution of the running state of grinding system real-time prediction. A case study shows that if the test data quality achieving reliability r0.65, the testing data and the intrinsic data reliability is the probability density function of the intersection area of A0.3, detection of mutation probability of PB is less than or equal to 0.7 data, the running state of system evolution is grinding reliable, no change; otherwise, that the running state of the grinding system evolution is not reliable, the mutation. The results for the judgment of the grinding process is reliable and provides a theoretical basis. The sensitivity analysis showed that the sample size does not affect the evaluation result of.4 according to the sorting grey relation and fuzzy set theory to evaluate the changes the degree of bearing grinding process of rolling. Using non sort grey relation, a grey confidence level under P=95%, using the time series data according to the order of acquisition, for two hours Grey attribute weights between the data sequence, and construct the grey attribute weights between the time series data of the similar matrix; fuzzy transitive closure method, grey attribute weight equivalent matrix between solution time series data; through hypothesis testing and piecewise average equivalence coefficient, describe the change process of grinder system error, in order to achieve the evaluation of degree of change the grinding process. The case study shows that the change process of the evaluation method can accurately describe the grinder system error, realizes the evaluation of degree of change of the grinding process, and the evaluation results and the actual situation match.

【学位授予单位】：河南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH133.33;TG580.6

【参考文献】