论文部分内容阅读
为了使用蓝宝石晶体高温传感器对工业烟气温度进行长期在线监测,设计了蓝宝石晶体高温传感器的解调系统,并对其进行了标定。对蓝宝石晶体高温传感器的白光偏振干涉测温原理进行了理论分析,并利用离散腔长变换(DGT)解调算法对光程差信息进行解调。在此基础上建立了一套以热电偶为参照的标定系统,使用S型高温热电偶采集温度数据,得到了光程差-温度样本。分别利用二次多项式拟合法与BP神经网络法对传感器的输出曲线进行了拟合与泛化,并进行了对比。实验结果表明:在800~1 300℃温度范围内,与二次多项式拟合方法相比,BP神经网络的拟合精度较高,拟合残差均值达到0.33℃;泛化能力强,多次泛化结果误差均值为0.56℃,均方误差为0.55℃。最终使用BP神经网络方法对传感器进行标定,使得传感解调系统满足了工业测高温的精度要求。
In order to use sapphire crystal temperature sensor for long-term on-line monitoring of industrial flue gas temperature, a sapphire crystal high temperature sensor demodulation system was designed and calibrated. The principle of white-light polarization interference temperature measurement of sapphire crystal high-temperature sensor is theoretically analyzed, and the optical path difference information is demodulated by discrete cavity length transform (DGT) demodulation algorithm. On the basis of this, a set of calibration system based on thermocouple was established. The temperature data was acquired by using the S type high temperature thermocouple, and the optical path difference temperature sample was obtained. The quadratic polynomial fitting method and BP neural network method were respectively used to fit and generalize the output curve of the sensor, and the comparison was made. The experimental results show that BP neural network has better fitting precision than quadratic polynomial fitting method in temperature range of 800 ~ 1 300 ℃, and the mean value of fitting residuals reaches 0.33 ℃; the generalization ability is strong, multiple times The mean error of generalization results is 0.56 ℃ and the mean square error is 0.55 ℃. Finally, the BP neural network method is used to calibrate the sensor so that the sensing demodulation system meets the precision requirements of the industrial temperature measurement.