【作者】 陈晓颖； 宋爱国； 李建清； 王晓蕾； 朱益民；

【Author】 CHEN Xiao-ying~(1,2),SONG Ai-guo~2,LI Jian-qing~2,WANG Xiao-lei~1,ZHU Yi-min~1 (1.Institute of Meteorology,PLA University of Science & Technology,Nanjing 211101,China; 2.School of Instrument Science and Engineering,Southeast University,Nanjing 210096,China)

【机构】 解放军理工大学气象学院； 东南大学仪器科学与工程学院；

【摘要】 为了提高所研制的新型数字式探空仪上的硅压阻气压传感器的精度,必须对传感器的输出进行温度补偿。针对所选传感器,分别采用多项式拟合方法与遗传算法建立传感器温度误差的校准模型,并对两种方法进行了比较、分析,提出了一种简单有效的软件温度补偿方法。研究结果表明,结合传统回归方法和遗传算法,对于减少用传统回归方法拟合出现的振荡及欠拟合现象,提高校准模型的准确度,有较好的效果。所建立的温度补偿模型实质为传感器特性曲面的非线性校正,其输入为二维变量,也适用于其他传感器特性曲面的校正。更多还原

【Abstract】 To increase its precision,the output of the silicon piezoresistance pressure sensor used by the radiosonde should be compensated.For the selective sensor,polynomial regression and genetic algorithm were applied to set up calibration model of sensor temperature error.By the comparison and analyze of two methods, a simple but effective software temperature compensation method is proposed.Based on the combination of regression method and genetic algorithm,the numerical model is established.The result indicates that the genetic algorithm can greatly reduce the oscillation and poor fit in traditional regression and improve the prediction accuracy of calibration model.The essential of temperature compensatory model is the nonlinear correction of characteristic curving surface of transducers.There is a two-dimensional characteristic for the input of sensor.And it provides an important reference value for the nonlinear correction of the characteristic curving surface of transducers.更多还原