Rolling bearing vibration performance evaluation overview

Rolling bearing performance mainly includes the vibration and noise, friction torque, temperature, rotation precision and so on, these properties have important influence on the performance of the mechanical system. Vibration is an important performance index of the rolling bearing, the comprehensive reflects the manufacture, installation, lubrication of bearings and other factors, influencing the dynamic characteristics of the bearing, the life and reliability. Reasonable evaluation of rolling bearings bearing vibration performance variation has very important application value, can be bearing failure of hidden danger in time, early take measures to avoid a major safety accident. Therefore, based on the principles of modern solid statistical data, this chapter puts forward - Kind of assessment of rolling bearing vibration performance variation method, to test the performance of the rolling bearing during degradation condition. According to the principle of modern solid statistical data, data robustness - is the most basic conditions for data analysis By the evaluation results, the data is stable, the more reliable. As a result, the sound of the data processing method is very important. Huber M estimation and the median estimate is tiny maximization of sound modern statistical data processing under the principle of two optimal estimates. Huber M estimation data can reflect the overall situation, has a critical value, and to zero as the center, data center is about zero symmetric singular function, practical engineering problems, it is difficult to satisfy the conditions lack of practicality. The median estimate is a robust data, can reflect the characteristics of the location data, but data can't reflect the overall situation. This chapter will both optimal estimation for organic integration, complementary advantages, put forward a kind of both can reverse reflection data location characteristics and can reflect the overall data of steady state of the data processing method, with liver evaluation of rolling bearing vibration data variability. In the process of robust data treatment, the median is the most robust estimation, the average is not robust estimation, because the average is easily affected by the outliers. The closer the average median, explain the data more robust. This chapter of the mean and the median proximity as evaluation factors, used to test the performance degradation during the rolling bearing service, belongs to the robust estimation. In this chapter, the rolling bearing performance variation refers to the rolling bearing during experiments and service performance changes with the degree of degradation; Median refers to the rolling bearing vibration data in absolute value, and then by size order, get the absolute value collating sequence, according to statistics to obtain the absolute value of median collating sequence; Average refers to the improvement according to the principle of Huber M estimates obtained the average of the data sequence; Mean and median proximity, with improved average sequence data and absolute value the absolute difference between the median collating sequence. This chapter put forward the evaluation method of rolling bearing vibration performance variation, evaluation factors of variable rate and general eigenvalue interval. In the characterization by mutation rates are rolling bearing performance variation, mutation rate is the number of data and variations according to the ratio of the number, the total variation data is not overall intrinsic interval data; Overall eigen interval is the intrinsic reflect data, processed data distribution interval is steady. At present, the research method of the rolling bearing vibration are mainly bearing vibration signal data of time domain features and neural network, the spectrum of bearing vibration signal analysis method, bearing vibration data model method, based on the Hilbert - Huang bearing vibration characteristic analysis, based on the phase space of the rolling bearing vibration characteristic parameter analysis method, etc. These methods require prior assumptions specific performance degradation model, distribution law, the probability density function and threshold value, and does not involve the robustness problem of rolling bearing vibration data. This chapter puts forward the evaluation of the performance variation of the rolling bearing, do not need to assume that performance degradation model beforehand, distribution law, the probability density function and threshold, the bearing vibration data of actual measurement directly obtained stable after treatment, the overall intrinsic interval, and then the implementation of performance degradation assessment.