纳米气体传感器灵敏度-温度曲线的拟合

Fitting of the Sensitivity-temperature Curves of Nanosized Gas Sensor

通过试验测试了苯纳米气体传感器的灵敏度随温度变化的规律.在所得实验数据的基础上,应用Matlab软件对苯纳米气体传感器的灵敏度-温度特性曲线进行了拟合.对最小二乘法的分段多项式与三次样条曲线两种方法在传感器特性曲线拟合中的特点进行了理论分析及比较,发现两种方法的拟合精度虽然均为10-3,但分段多项式的拟合相对误差为±1.5％,各段拟合曲线连接点处出现断点现象;而三次样条函数的拟合相对误差仅为±0.5％,各段拟合曲线连续性好,整条曲线光滑.

The variation of the sensitivity of nanosized benzene-mixed gas sensor with temperature was measured by experiments. The sensitivity-temperature characteristic curves of the sensor were then simulated on the basis of the experimental data and by using Matlab software. Theoretical analysis and comparison of the characteristics of the two methods, namely, the polynomial fitting method and the spline-polynomial fitting in least square method, was carried out. It is found that the fitting accuracies of the two methods are all 10-3. The relative error of polynomial fitting is ±1.5%, and there are break points in the section connections of the fitting curve, while the relative error of spline-polynomial fitting is only ±0.5% , and the fitting curve is smooth without any break point in the connections.