Chapter1Introduction方案.ppt

(4) Repeatability (重复性） Repeatability is the closeness of agreement between successive results obtained with the same method under the same conditions and in a short time interval. Quantitatively, the repeatability is the minimum value that exceeds, with a specified probability, the absolute value of the difference between two successive reading obtained under the specified conditions. (5) Zero drift （零漂） Temperature drift （温漂） Zero drift describes output variations when the input is zero. Temperature drift describes output variations when temperature changes. Resolution is the minimal change of the input necessary to produce a detectable change at the output. Normally, resolution is less than the maximal absolute error of the sensor. (6) Resolution (Discrimination 分辨力） 6. Methods for Data Process Data process is the process from acquiring data to drawing conclusion. Arithmetic average method (算术平均值法）, Discard the abnormal data （剔除异常数据）。 (1) Arithmetic average method Method: set the arithmetic average of all measured data as the result of measurement. x1, x2, … xn --- measured data Residual error (剩余误差，残差） Random error （随机误差） Arithmetic average method Standard error (标准误差) 1. Defined with random error (根据随机误差计算）: 2. Defined with residual error（根据残差计算）: (2) Discard abnormal data Theory：When measured value exceed the normal scope, discard the abnormal data. Rule：Discard those measured value whose discrepancy between the measured value and the arithmetic average value exceeds 3 times of standard error, say, Limitation: 1. The random errors follows a normal distribution. 2. No. of measured value are more than 10. Grubbs rules When the number of measurement is less than 20, the abnormal error is defined based on Grubbs rules; Grubbs rules use Grubbs coefficient to discard those measured value whose discrepancy between the measured value and the arithmetic average value exceeds standard error multiples