论文部分内容阅读
对6 mm厚和8 mm厚两种规格汽车大梁钢WL510以及WL590、WL700产品共4组数据进行含二维单侧容限系数的升降法分析。结果表明,参考以往经验数据的二维升降法结果的相对偏差比常规的一维升降法有大幅下降。部分情况下,二维升降法结果相对偏差仅为一维方法的1/5。另外,当试验样本量过小时,即使使用二维升降法,试验结果偏差仍较大。因此,即便使用二维升降法也得保证基本的试验样本量。从实测结果来看,该方法不但给出了很好的方差估计量,解决了传统升降法对方差估计误差较大的问题,与传统的只能利用当前试验数据的升降法相比,其可利用的信息量有了大幅度增加,所以在精度相同的情况下,可以节省大量试件;而在试件数一定的条件下,又能提高预测精度。
A total of four sets of data of WL510, WL510 and WL590, WL700 products of 6 mm thick and 8 mm thick automotive girder steel were analyzed by lifting and lowering method with two-dimensional unilateral tolerance coefficient. The results show that the relative deviation of the two-dimensional lifting and lowering method based on the past empirical data is significantly lower than that of the conventional one-dimensional lifting and lowering method. In some cases, the relative deviation of two-dimensional elevation method is only 1/5 of the one-dimensional method. In addition, when the test sample is too small, even if the two-dimensional lift-off method is used, the deviation of the test result is still large. Therefore, even with the two-dimensional lifting method to ensure that the basic test sample size. From the measured results, this method not only gives a good estimate of the variance, but also solves the problem of the traditional method of estimating the variance of the variance. Compared with the traditional method which can only use the current test data, it can be used Has greatly increased the amount of information, so the same precision in the case, you can save a large number of test pieces; and in the case of a certain number of samples, but also improve the prediction accuracy.