精品文档---下载后可任意编辑张正江 1, 方伟超 1, 张仁 1, 骆雅晴 1,洪雪聪 1, 邵之江 21. 温州大学,物理与电子信息学院, 温州 325035E-mail: zjzhang@2. 浙江大学,工业控制技术国家重点实验室,工业控制讨论所,杭州 310027摘 要:过程系统的控制与优化要求可靠的过程数据。通过测量得到的过程数据含有随机误差和过失误差,当有多组数据时,可以对其进行数据的预处理,减小恒值系统误差的影响。此外,采纳数据校正技术可有效地减小过程测量数据的误差,从而提高过程控制与优化的准确性。针对传统基于最小二乘的数据校正方法和基于准最小二乘的鲁棒数据校正方法,分析了两种方法的优缺点,并提出了一种最小二乘与准最小二乘组合方法,综合前两种方法各自的优点,使得数据校正结果更加准确。将提出最小二乘与准最小二乘组合方法应用于线性与非线性系统的数据校正中,通过校正结果的比较验证了此方法的高效性。关键词:数据预处理,数据校正,过失误差,最小二乘,准最小二乘Data pretreatment and with the least squares and quasi least squares calibration data combination methodZhengjiang Zhang1, Weichao Fang1, Ren Zhang1, Yaqing Luo1, Xuecong Hong1, Zhijiang Shao21. College of Physics and Electronic Information Engineering, WenzhouUniversity, Wenzhou325035, Zhejiang, ChinaE-mail: zjzhang@2. State Key Laboratory of Industrial Control Technology, Institute of Industrial Control, ZhejiangUniversity, Hang Zhou 310027, Zhejiang, ChinaAbstract: Reliable process data are required for process control and optimization. As a result of random and gross errors existing in the measured process data, data rectification is needed to minimize the measurement errors.When there are multiple data,can carry on the data pretreatment and reduce the influence of constant value system error.Therefore, the results of process control and optimization are more accurate. The advantages and disadvantages of methods for data rectification based on weighted least squares and quasi-weighted least squares are analyzed....