东南大学误差处理与数据处理复习课程序.ppt

Dr. Jinbo Fu Chapter1:Preface Theory of Error & Data Processing误差理论与数据处理 The purposes of error studying研究误差的意义 Understand the properties of errors and investigate the sources of errors, so as to reduce or remove them; 认识误差的性质，研究误差的来源，以减小和消除误差； Properly handle measuring and data processing, and correctly analyze the results of measurement, in order to obtain outcomes as close to true values as possible; 正确处理测量和实验数据，合理分析所得结果，以便得到更接近于真值的数据； Correctly arrange the experiments, design or select instruments and apply measuring methods, thus construct the optimal system to obtain the best result. 正确组织实验，合理设计或选择仪器和测量方法，以便在合理优化的系统下得到理想结果。 Definitions The sources of error (误差来源) Classifications Correctness, Precision & Accuracy Significant Digit For any approximate number, the left most non-zero digit is called the first significant digit. From the first significant digit to the last digit, all digits, zero or non-zero, are called significant digits. 对于任何近似数，从左边起的第一个非零的数字称为第一位有效数字，从第一位有效数字起到最末一位数字为止的所有数字，不论零或者非零，都是有效数字。 Significant Digit Examples of rounding-off operation数字舍入运算示例 Original data: 3.14159 2.71729 4.51050 3.21550 6.378501 7.691499 5.43460 Rules of data computation数据运算规则 The figure shows normal distribution curve and related parameters. 正态分布密度曲线和相应的各种参数如图所示 σ is the X value of inflexion point A; θ is the X value of point B , the gravity center of the right half of the curve; line parallel to Y-axis at x=ρ evenly splits the area of the right half of the curve. σ是曲线拐点 A 的横坐标; θ是 B 点的横坐标, 也就是右半部曲线下方面积的重心; 在 x=ρ处与Y轴平行的直线等分曲线右半部分下方面积. Properties of Normal Distribution 正态分布的特性: Arithmetic Mean 算数平均值 When measuring the same object with equal precision, the results are different due to random errors, and the arithmetic mean should be used as the final measurement result. 对某一量进行多次等精度测量，由于随机误差的存在，测量结果各不相同，应以全部测量值的算术平均值作为最后测量结果。 单次测量的标准差 Standard Deviation of Arithmetic Mean算数平均值的标准差 Extreme errors Review of last lesson 前课复习 Section5 Assignment of errors 误差的分配 Assign