基于RBF神经网络气压补偿的非色散红外SF6气体传感器

摘要

非色散红外SF6气体传感器具有测量范围广、灵敏度高、抗干扰能力强等优点,在电力系统中具有广泛的应用.在实际检测过程中,环境气压的变化对气体传感器的检测精度有较大的影响,提出利用RBF神经网络建立气体传感器气压补偿模型,运用其泛化和非线性映射能力对环境气压波动引起的测量误差进行补偿.实验结果表明:采用气压补偿模型后的气体传感器在气体浓度3 260 mg/m3～9 781 mg/m3,气压100 kPa～120 kPa范围内,最大测量误差由±646 mg/m3降为±52 mg/m3,测量精度为±0.53％FS.该方法相比于拟合法和硬件电路补偿法具有更高的测量精度和稳定性,降低了传感器的体积和成本.%Non-dispersive infrared SF6 gas sensor has many advantages,such as wide measurement range,high sensitivity,strong anti-interference ability and so on.It has been widely used in power system.However,in the actual detection process,the change of the ambient pressure has a great impact on the detection accuracy of the gas sensor.Aiming at this,an air pressure compensation model of gas sensor was established by using radial basis function (RBF) neural network,and its measurement error is compensated by the good generalization and non-linear mapping ability.The experimental results show that when the designed air pressure compensation model of gas sensor is in the gas concentration of 3 260 mg/m3 ～9 781 mg/m3,and the air pressure is in the range of 100 kPa～120 kPa,the maximum measurement error is reduced from ±646 mg/m3 to ±52 mg/m3,the measurement accuracy is ±0.53％FS.Compared with the empirical formula method and the constant air pressure compensation,this method has higher measurement accuracy and stability,and reduces the volume and cost of the sensor.