论文部分内容阅读
不同于以往温度、应变与光纤Bragg光栅(FBG)峰值波长偏移量之间的数学模型,本文提出利用支持向量机(SVM,support vector machine)补偿温度的影响,并以FBG压力传感器为例,利用标定的压力传感器数据对SVM模型的惩罚系数C和径向基核函数(RBF)核参数γ进行优化,得到SVM温度补偿模型,选择核参数γ为100、惩罚系数C为16。经过补偿后,压力传感器的零位温度系数和灵敏度温度系数由补偿前的34.5%/℃和34.2%/℃减小到1.7×10-5%/℃和7.7×10-5/%/℃。充分说明,利用SVM补偿温度对FBG压力传感器的影响是有效的。
Different from the previous mathematical model of the temperature and strain and the peak wavelength shift of fiber Bragg grating (FBG), this paper proposes to compensate the influence of temperature by support vector machine (SVM). Taking FBG pressure sensor as an example, The calibrated pressure sensor data is used to optimize the SVM model’s penalty coefficient C and radial basis function (RBF) kernel parameter γ, and the SVM temperature compensation model is obtained. The kernel parameter γ is selected as 100 and the penalty coefficient C is 16. After compensation, the zero temperature coefficient and sensitivity temperature coefficient of the pressure sensor decreased from 34.5% / ℃ and 34.2% / ℃ before compensation to 1.7 × 10-5% / ℃ and 7.7 × 10-5 /% / ℃. Fully demonstrated, the use of SVM compensation temperature FBG pressure sensor is effective.