基于状态振动特征的空间滚动轴承可靠性评估方法研究

发布时间：2018-01-15 11:12

本文关键词：基于状态振动特征的空间滚动轴承可靠性评估方法研究　出处：《重庆大学》2014年硕士论文　论文类型：学位论文

【摘要】：滚动轴承作为空间活动部件的重要组成部分，它的运行状态直接影响整个空间活动件的运行性能。实践表明，空间活动件的故障大多是出自其轴承的问题，但是在空间场合，由于轴承应用受到环境条件的限制，不可能采用备份来保证轴承的可靠性，所以轴承一旦出现问题，将导致整个空间活动件的性能破坏。空间环境下的滚动轴承要承受低温和交变温度、原子氧侵蚀等极端环境的综合作用，极易造成空间滚动轴承的精度失效，加快空间滚动轴承的损坏。由于空间环境下轴承的失效机理与地面环境下的失效机理存在差异，因此为了保证空间活动件高可靠、长寿命运行，避免一些重大事故的发生，需在模拟空间环境下开展滚动轴承运行可靠度评估。传统的可靠度评估方法，将概率论和数理统计理论作为主要的数学工具，利用大量的具有概率重复性的失效样本，以确定失效分布类型，从而获得宏观意义上一批同类设备共性的平均可靠度。然而，对于空间滚动轴承而言，由于运行工况不同、转速不稳定等因素的影响，各个滚动轴承的损伤、故障程度不同，导致其运行可靠度也必然不同。针对某个具体的空间滚动轴承进行运行可靠性评估是个性问题，而基于大样本条件并依赖概率统计数据得到的平均可靠度难以满足单个空间滚动轴承的运行可靠性评估要求。由于轴承的状态特征量能够提供可靠性评估的重要信息，因此，基于状态特征量的可靠性建模与分析技术是解决单个空间滚动轴承运行状态可靠性评估需求的一个重要途径。目前，反映轴承运行状态的特征量主要有三种，即摩擦力矩、振动和温度，由于摩擦力矩、温度等参数等不能有效反映空间滚动轴承寿命状态的变化，因此论文选用包涵空间滚动轴承寿命特征信息丰富的振动信号作为状态特征量以评估滚动轴承运行过程中的可靠度。比例故障率模型(Proportional hazardmodel, PHM)是其中一种最为常用的基于振动特征的可靠性评估模型，基于比例故障率模型的可靠性评估方法的关键是提取反映空间滚动轴承运行状态的特征指标及确定比例故障率模型的具体数学表达式，由实时获取的振动信号提取振动特征指标，评估空间滚动轴承的运行可靠度。同时，结合性能退化趋势预测理论，在已建立的比例故障率模型的基础上，实现空间滚动轴承的可靠度趋势预测，以确定滚动轴承在未来任意一段时间内的可靠度。具体内容安排如下： ①基于状态振动特征的可靠度评估技术首要解决的是特征指标构建的问题，为此研究了基于多域特征融合的构建方法。提取时域、频域、时频域和威布尔分布特征信息组成高维多域特征集，采用流形学习方法对多域特征进行维数约简，，以解决高维特征集之间存在的冲突、冗余问题，并将约简以后的特征信息作为趋势预测的特征指标及比例故障率模型的响应协变量。 ②针对单个空间滚动轴承运行状态可靠性评估要求，同时克服经典的可靠性分析方法存在的问题，提出了基于振动特征指标的比例故障率模型评估可靠度评估方法。将高维多域特征集维数约简后的特征信息作为比例故障率模型的响应协变量，采用极大似然函数原理估计模型的待定参数，建立空间滚动轴承状态特征指标与可靠度之间的数学模型，实现空间滚动轴承运行状态的可靠性评估。 ③针对单个空间滚动轴承可靠度趋势预测问题，提出了基于性能退化趋势预测的空间滚动轴承可靠度趋势预测方法。将高维多域特征集维数约简后的特征信息作为最小二乘支持向量机的输入，训练并建立趋势预测模型，实现空间滚动轴承性能退化趋势预测。将趋势预测结果代入已建立的比例故障率模型中，即可实现空间滚动轴承的可靠度趋势的预测。 ④在以上理论的基础上，采用C#为开发平台，研发空间滚动轴承性能退化趋势预测、运行可靠性评估等功能模块，通过应用对各模块进行检验，并对本文所提方法进行验证。
[Abstract]:As an important part of the rolling bearing space of the moving parts, its running state directly affects the performance of the entire space activities. The practice shows that the fault space of moving parts mostly from the bearing problem, but in space, because the bearing application is environmental conditions, it is impossible to use backup to ensure bearing so the bearing reliability, if there are problems, will lead to the damage of space activities. The performance of the rolling bearing under the space environment to bear alternating temperature, the comprehensive effect of atomic oxygen erosion and other extreme environment, extremely easy to cause the failure of space rolling bearing accuracy, accelerate the damage of space rolling bearing. Because of the space environment of the bearing the failure mechanism and failure mechanism of ground environment are different, so in order to ensure the high reliability of space activities, long service life, avoid some major accidents It is necessary to carry out the reliability evaluation of rolling bearing operation under the simulated space environment.
The reliability of the traditional evaluation method, the probability theory and mathematical statistics theory as the main mathematical tools, the use of a large number of repetitive failure probability samples to determine the failure distribution, so as to obtain the macro sense of a number of common similar equipment average reliability. However, for the space rolling bearing, because the running condition the different effects of speed, instability and other factors, each bearing damage fault degree is different, the operation reliability is also different. For a specific space bearing operation reliability assessment is based on individual issues, and a large sample and depends on the probability and statistics data from the average reliability assessment is difficult to meet the operation reliability of single space rolling bearing requirements.
Because of the important information, characteristics of bearing capacity can provide the reliability evaluation result, reliability modeling and analysis based on the amount of state characteristics is an important way to solve the single space needs assessment of the reliability of rolling bearing. At present, there are three main features reflect the running state of the bearing, the friction torque, vibration and temperature because, friction torque, temperature can not effectively reflect the change of space rolling bearing lifetime state, so the inclusion of space vibration signal of rolling bearing life rich feature information for character in order to assess the reliability of rolling bearing during operation. The proportion of the failure rate model (Proportional hazardmodel PHM) is one of a kind the reliability evaluation model based on the vibration characteristics of commonly used, key reliability evaluation method based on the proportional hazards model is The extraction characteristic index space reflect the running state of rolling bearings and determine the proportional hazards model specific mathematical expressions, the real-time vibration signal by the vibration feature extraction index, evaluation of space rolling bearing reliability. At the same time, combined with the performance degradation prediction theory, based on the proportion of the established fault rate model, reliability prediction the degree of tendency to realize the space rolling bearing, rolling bearing in the future to determine any time reliability. The main contents are as follows:
Based on the state of the vibration characteristics of primary evaluation reliability is characteristic of index construction, this research constructs the multi domain fusion method based on feature extraction. Time domain, frequency domain, time-frequency domain and Weibull distribution feature information of high dimensional multi domain feature set, the manifold learning methods for dimensionality reduction of multi domain features. In order to solve the conflict between the high dimensional collection of redundancy, and the characteristics of information reduction after as the covariate response characteristic index and proportion trend forecast failure rate model.
The single space rolling bearing reliability assessment requirements, and methods to overcome problems of classical reliability analysis, put forward the vibration characteristic index proportional hazards model evaluation method of reliability evaluation based on response covariate Gao Weiduo domain feature set feature dimension reduction as the proportion of the failure rate model, the maximum likelihood the function principle of estimation parameters of the model, establish the mathematical model of state space characteristics of rolling bearing and the reliability of the reliability evaluation, implementation of space running state of rolling bearings.
The single space rolling bearing reliability prediction problem, proposes the prediction method of reliability prediction trend of rolling bearing performance degradation of space. Based on the high dimensional multi domain feature set feature dimension reduction as the inputs of least square support vector machine, training and the establishment of trend prediction model, realize the space rolling bearing performance degradation the trend of trend prediction. Prediction results into the established proportional hazards model, reliability prediction of rolling bearing can achieve spatial trend.
(4) on the basis of the above theories, using C# as the development platform, the functional degradation trend prediction, operational reliability evaluation and other functional modules of the rolling bearing are developed, and the modules are tested by application, and the proposed method is verified.

【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TH133.33;TB114.3

【参考文献】