【作者】 杜庆林；

【导师】 贾亚洲；

【作者基本信息】 吉林大学 ， 机械电子工程， 2004， 硕士

【摘要】 一．目的本文针对国家高技术研究发展计划课题<数控装备与系统可靠性关键技术开发及应用>岢龅摹Ｖ饕攵允叵低辰锌煽啃匝芯浚萁荒甓允叵低呈匝榭己说幕∩希范ǚ鲜叵低匙陨硖氐愕囊惶卓煽啃匝芯糠椒ê屠砺厶逑担⑹叵低车目煽啃云兰厶逑岛凸收戏治龇椒ǎ⒅贫ㄊ叵低车墓ぷ鞴娣叮故叵低车目煽啃匝芯壳饔诒曜蓟?二．数控系统可靠性抽样方法可靠性分析的一般程序，以产品总量为依据或取一定数量的样本，样本在规定的条件下以一定的程序参加可靠性试验，根据试验的失效数据，用统计的方法进行评估和分析。在一般的试验情况下，一般采用截尾试验。在以往的数控机床的可靠性试验中，主要采用定时截尾试验，这种试验方法的周期长，投入的人力、物力大，给实际生产带来不便；数控系统属于电子产品，其可靠性程度较高，在现行的可靠性试验中也采用定时截尾试验，但由于试验条件的限制，试验的时间不能太长。因此，在试验周期内可能出现以下几种情况：(1)对于受试设备每台发生两次以上故障；(2)对于受试设备有一台或几台发生一次或不发生故障；(3)对于受试设备都没发生故障。如果根据数控机床的可靠性评价方案，对于发生一次或不发生故障的机床，在可靠性评价时，应该去除，这必然使可靠性的评价水平过于保守。对于数控系统，出现失效个数较少甚至无失效样品情况时常发生，所以采用经典的方法估计其可靠性特征量会引入较大的误差。故本文探讨改进的定时截尾试验下可靠性特征量的计算，根据确定的失效产品的数量确定样本的定时截尾时间。从而保证在合理的试验时间内使对数控系统的可靠性进行准确的统计分析。 <WP=69>三．数控系统可靠性指标评价方法根据抽样的原理，即失效产品数量较少的情况下，常规的数理统计方法计算的可靠性指标存在较大的偏差。因此，本文采用故障总时间法确定数控系统的失效率函数，从失效率函数推导出数控系统的其他可靠性指标函数。并利用熵权综合评价方法进行可靠性特征量综合指标的评价，熵的引入使得故障分析更加有效，综合指标评价将指标数值和故障的危害程度通过规范化处理统一在一起，使评价更科学、合理。通过确定综合指标，可以对数控系统的可靠性情况进行前后对比分析，并且还可以对不同厂家、不同型号的数控系统进行横向的对比分析，分析影响数控系统可靠性的关键零部件，及其故障模式、原因等，找出自身不足，确定可靠性的研究方向。由上述方法确定的数学模型是否合理，主要利用d检验法进行检验，经过反复试验验证确定合理的理论方法。通过此种评价分析方法，使得数控系统的可靠性试验在较短的时间内，就能有效的评价出数控系统的可靠性水平，对指导可靠性研究方向，缩短研究周期，减少试验的投入有着重要的意义。四．数控系统规范的建立经过与几家数控系统生产厂家的合作，借鉴其可靠性管理经验和企业的管理方法，在对数控系统的可靠性评价和故障分析的基础上，总结故障试验的方法、指标体系、设计和制造等方面的可靠性保证措施，建立了关于数控系统的可靠性工作规范，使得以后进行数控系统的可靠性研究工作时，有一个统一的参照，初步制定了规范的标准体系。对于此工作规范本文重点讨论了软件可靠性的内容，在数控系统可靠性研究方面，软件可靠性的研究还是首次。因此，本文初步从软件设计的需求分析、概要设计、详细设计、实现阶段和测试阶段几个方面进行了规范化，给出了统一的可靠性管理方法。对于数控系统软件可靠性是比较重要的部分，软件的可靠性与硬件的可靠性有很大的不同，软件是无形的，研究其可靠性比较困难，有一些软件故障大多表现为硬件的故障，因此，给故障的诊断带来很大的困难，是今后应重点研究的一个内容。五．结论通过本文的研究，充实了数控装备可靠性工程的内容。对于受试产品数<WP=70>量较少，故障点少的情况进行了理论的研究，确定了关于此种情况的可靠性评价及分析方法，并引用了熵的概念进行了综合指标的评价，使进行数控系统的综合评价成为可能。最后建立了数控系统的可靠性工作规范，补充加入了软件的可靠性保证措施，从数控系统的试验、设计、评审及可靠性分析等方面建立了统一的工作标准，完善了数控装备的可靠性评价系统。更多还原

【Abstract】 1.Purpose Based on the subject of Hi-Tech Research and Development Program of China --The research and development on the key reliability-technology of Computer Numerical Control(CNC) equipment and system , put forward the scientific research of this thesis. The thesis aims to study the reliability of CNC system, construct the reasonable study method and standard system info according to the characteristic of CNC system, and establish the fault analysis method.2.The sample selection method of CNC system The general analysis procedure of reliability is that choose some products based on the amount of this set. The sample has been proceeding the relevant experiment before leave factory. The fault data of this paper are all come from working environment. Assess and analyze the faulty data by using statistic theory . In general, the experiment of reliability adopt time-limit or quantity-limit test. To the test on reliability of CNC machine tool mainly adopt time-limit experiment. This method will take long time, use a mount of manpower and material, it is not very fit the practical manufacture. The CNC system is belong to electronic products, it’s reliable level is very high, as the limit of experimental condition, it can not take for a long time. In the period of test, there are several case will be take place, such as below:The test equipment at least occurred two failure.There is one or more test equipment has no failure.The all test equipments have no failure. According the reliability assess of CNC machine tool, to the case of (2) and (3), we must delete the corresponding data, that will make the result be lower than factual reliability level. To the experiment of CNC system, these cases will often take place, adopt the method that using in machine tool will have more error, so in this thesis, we discuss a better technic that integrate the characteristic of time-limit and amount-limit experiment, based on the amount of failure products, select the reasonable deadline of test, so as to make the result exact and reasonable. <WP=72>3. The assess method of CNC system reliability index According to the sample theory, in the cases of less sample and lower failure rate, the general statistic method has some deficiency. So in the thesis apply total failure time method to build up the failure probability function λ(t), then calculate the other reliability function and index. Based upon the comprehensive evaluation index system constructed, this thesis uses entropy method to do a comprehensive valuation of CNC system reliability index. This method overcome the shortcoming of frequently used method which gives weight subjectively. The entropy method integrated the reliability index and failure effect, it makes the evaluation reasonable and objective. Using comprehensive evaluation index , we can analyze the reliability level of the before and after reliability test, or contract reliability level of other manufacturing factory, find out the deficiency of own and improve the design, manufacturing, etc. through this method, in short period , we can assess the reliability level of CNC system, reduce the payout of reliability test.4. Build up the criteria of CNC system After the fault analysis and reliability assess of CNC system, summarize the method of test, the index system, the reliable measure in design and manufacture, build up the reliability criteria of CNC system. In this criteria, add the content software reliability, constitute the reliability measure of software from requirement analysis, general design, detailed design, the carry out and test phase, and consolidate the management of software reliability. Software reliability is different from the reliability of machine tool, it is difficult to study it’s reliability, as an important content of fault analysis and diagnose, it is an crucial direction in further study.5. ConclusionThrough the study of this thesis, enhance the reliability system of CNC equipment. For the case of less sample and lower failure rate, build up the correspondi更多还原