论文部分内容阅读
准确可靠的测量数据是工业过程优化和控制的重要前提,实际过程中由于过失误差的存在导致测量数据不准确,而传统的过失误差侦破方法常会出现过失误差分摊到其他测量数据上的现象。本文在研究广义似然比法(GLR法)过失误差侦破的原理及步骤的基础上,对GLR法的检验统计量进行改进,提出了2个变量同步侦破补偿的两两同步侦破法,即每次同时用2个变量计算检验统计量,并对相对调节量大的测量数据进行补偿。两两同步侦破法与GLR法的实例考核结果表明,在相同显著性水平下,两两同步侦破法的总体检验能力和选择性比GLR法高,且犯第1类错误的概率比GLR法小。因此新方法提高了过失误差侦破率,降低了犯第1类错误的概率,能有效地对过失误差进行侦破和识别,且同步补偿后得到的补偿值更准确,有利于获得较为准确的过程数据。
Accurate and reliable measurement data is an important prerequisite for optimization and control of industrial processes. In the actual process, the measurement data is inaccurate due to the existence of fault errors. However, the traditional method of detecting fault errors often shows the error distribution to other measurement data. Based on the study of the principle and steps of GLR error detection, this paper improves the test statistic of GLR method and proposes a two-two simultaneous detection method of two variables synchronization detection compensation, that is, At the same time, two variables are used to calculate the test statistic, and compensation is made for the measurement data with a relatively large amount of adjustment. The results of two-sample simultaneous detection and GLR test show that under the same significance level, the detection ability and selectivity of two-by-two synchronization detection method are higher than GLR method and the probability of making type 1 error is smaller than that of GLR method . Therefore, the new method improves the detection rate of fault errors, reduces the probability of making a type 1 error, effectively detects and identifies the fault errors, and obtains more accurate compensation values after synchronous compensation, which is in favor of obtaining more accurate process data .