【作者】 陈炳锟；

【导师】 郑午；

【作者基本信息】 吉林大学 ， 工业工程， 2008， 硕士

【摘要】 可靠性是数控机床用户最关心的技术指标之一,被人们视为数控机床的“大脑”与“核心”的数控系统的可靠性更是用户关注的焦点。数控系统可靠性的评价体系和评价软件已经基本趋于成熟,但是多数是从故障间隔时间的角度切入。本文主要以国内某厂生产的数控系统为研究对象,对现阶段数控系统的可用性评价进行了研究。主要方法是在比较成熟的可靠性研究上增加了维修性分析、可用性分析和评价及GO法分析。分别从整体和系统的角度分析了该数控系统可用性情况,为提高其可用性提供了理论的指导。首先建立故障间隔时间模型和故障维修时间模型,确定MTBF和MTTR的观测值和点估计值。再根据可靠性、维修性研究的结果对该数控系统进行可用性分析,求出数控系统的瞬时可用度模型及稳态可用度的值,并对其进行可用性评价。为从系统角度分析数控系统的可靠性,本文将再使用GO法原理对该数控系统进行可靠性分析。GO法是近几年快速发展起来的一个新的方法,已经在天然气管道系统、核电站电源系统和配电网络等领域取得了很好的成绩,然而在数控系统可靠性研究领域尚未得到广泛应用,因此本文将使用GO法对数控系统进行系统的可靠性分析。更多还原

【Abstract】 1. IntroductionCNC system is the Control core of the CNC machine. The reliability of the CNC system is the quality attribute by the most concerned when most users want to purchase products. And if the reliability is well or not is the key of our homemade NC industries survival. The availability research on the CNC system can increase the level of the homemade NC industry and direct decrease the expense of the suppliers’after service\three guarantees costs\downtime losses and machine maintenance. So this research has an important historical and practical significance.This thesis is just put forward under this kind of background. It is a part of the JiLin province Scientific and technological development project-The key technology of the computer-assistant NC machine tool’s availability research for the life cycle（Contract No.: 20050535）.At present, the reliability research technique of the CNC system has been basically mature, but most of them are from the point of view of the time between failures. This paper will add maintainability analysis\availability analysis and GO methodology analysis. And it analyses the availability of some factory CNC system which is produced in recent years. This paper’s research provided theoretical guidance for the improvement of CNC system availability.2. The reliability of CNC systemAccording to the large quantity of data collected to establish time between failures model, and then analyses CNC system’s reliability.CNC system belongs to electronic products. Because of what is called“no memory”, its failure data should be exponential model, and we usually call it exponential production. Basing on this point, the paper must check up the data to find out if the data contain some uncommon data. After this work, I found a uncommon big one, so I will eliminate this data in order not to affect the whole data’s authenticity. By dealing with the data, this paper will establish the mathematic model first, and use the optimal distribution pattern as criterion to scale the fit on the graph later. The index of correlation is 0.99064 by calculating,and the model has passed the test. So the eliminating of the data is rational.The paper gives out the mathematic model of four basic functions, which are the failure probability consistency, f（t）, accumulated failure probability, F（t）, instantaneous failure probability,λ（t）, reliability grade, R（t）.The observation value of MTBF is 1228.7 hour.The spot estimation value of MTBF is 941.2 hour.3. The maintainability of CNC systemIf the reliability is well or not is measured on maintaining time. The maintenance goes better as maintaining time goes shorter. The paper collected a mass of data to analyze the maintenance of the CNC system, and make sure that the distribution function model of maintaining time is lognormal distribution. The paper will establish maintainability function model. Using the relatively accurate d inspection and taking the significance levelαequal to 0.1, the hypothesis test statistic D_n （=0.080043） less than the critical value D_n,α that equals to 0.2799, is obtained. So these maintaining data are matched to the lognormal distribution model.The paper gives out the mathematic model of three basic functions, which are the maintenance probability consistency, （m|^） （t）, maintenance time distribution function, （M|^） （t）, maintainability probability function, （μ|^） （t）.The observation value of MTTR is 5.47 hour.The spot estimation value of MTTR is 5.94 hour.4. The availability of CNC systemThe availability contains reliability、maintainability and maintenance indemnification. Because of the maintenance indemnification data is rarely few in CNC system test data, so the paper will not consider the maintenance indemnification, only analyze on the basis of reliability and maintainability. Using the conclusion of reliability and maintainability analysis, the paper analyzes the CNC system’s availability to establish instantaneous availability model and the inherent availability. And evaluate the CNC system’s availability.The CNC system’s the inherent availability Ai is 0.993728 and instantaneous availability model is A （ t）= 0.993728+0.006272e-^0.169413t. Draw the two curve in one coordinate chart （Chart 1）, and the curve 1 is the instantaneous availability model and the curve 2 is the inherent availability. We can see that the instantaneous availability is falling gradually from the zero times, and after 30⁴0 hours the instantaneous availability and the inherent availability will be nearly superposition. Evaluating the CNC system’s availability, the paper found that the estimate is 0.995189 and the unilateral lower confidence limit of the interval estimate is 0.9925. So, it is obvious that the CNC system’s availability is well.5. GO Methodology analysis on CNC systemFor analyzing CNC system’s reliability from inner and system point of view, the paper will use GO methodology to analyze CNC system’s failure. GO methodology is a new reliability method, which has been developing fast in recent years, but has never been used in CNC system’s reliability area in-depth. So the paper will use it in CNC system’s reliability research in-depth. According to the inner configuration and principle, we divide the CNC system into several parts. Analyze CNC system’s reliability according to the communication transfer among the parts. So the GO methodology can represent its reliability more accurately. In the GO operation, we use the steady availability of every part as normal state probability, and work out the CNC system’s every state value: the average normal working probability is 0.9920779, the average un-normal working probability is 0.0079221, the failure probability is 0.0015319, and the maintainability probability is 0.1918381.6. ConclusionThe paper analyzed the CNC system’s reliability first, and found out the observation value and spot estimation value of MTBF. It also analyzed the CNC system’s maintainability, and educed the observation value and spot estimation value of MTTR. And then use these results to analyze and evaluate the CNC system’s availability. For analyzing CNC system’s reliability from inner and system point of view, the paper used GO methodology to analyze CNC system’s reliability, and educed reliability value according to the inner parts’logic connection.更多还原