Whether the data is robust or not is the prerequisite for data analysis. This section proposes a statistically based method for the robust judgment of rolling bearing performance data. These criteria are based on the experimental data obey a certain distribution, and the judgment method is cumbersome to calculate, and the gross error is only part of the discrete data. Robust data can yield reasonable and correct estimation results, while unsound data can lead to incorrect or even catastrophic estimation results.
Its judgment criteria include Wright criterion, Chauvigne criterion, Romanovsky criterion, Nair criterion, Grubbs criterion and Dixon criterion. Therefore, how to judge whether the experimental data of rolling bearing performance is robust is an important content of rolling bearing performance analysis. Wholesale wheel bearing kits Gross error is a part of discrete data. However, during the experiment, due to environmental, instrument, temperature and other factors, it will fluctuate, deviate from the true value, cause data pollution or generate discrete data, and cause the distribution of data. Variety.
The steps are as follows: (1) A large number of rolling bearing performance data is obtained through experiments. The robustness of rolling bearing performance data is the prerequisite for rolling bearing performance research. Due to the large rolling bearing performance data, it can be assumed that the data obeys a normal distribution; (2) According to According to the principle of statistics, the data is grouped, the data frequency histogram is made, and compared with the normal distribution diagram, the difference between the experimental data and the normal distribution can be roughly judged;
(3) The probability distribution of the rolling bearing performance data is established, and Compared with the normal distribution, analyze the agreement between the probability of the experimental data and the probability of the normal distribution; (4) If the cumulative frequency of the rolling bearing performance data and the normal distribution probability are on a straight line, the rolling bearing performance data is robust, otherwise the rolling bearing performance data is not Robust; (5) According to the deviation of the data from the normal distribution, analyze the range of the discrete data distribution of the rolling bearing performance
asdc