基于神经网络的瓦斯传感器的非线性校正

The Nonlinear Correction of Methane Sensor Based on Neural Network

文章提出了一种基于改进型BP神经网络的瓦斯传感器的非线性调校方法,该方法利用神经网络良好的非线性映射能力逼近反非线性函数,从而完成非线性校正。仿真结果表明:与传统的分段线性方法和BP算法相比,改进型BP神经网络收敛速度快、逼近精度高,准确度由原来分段线性校正的±5.02%提高到现在的±0.130%,且易于动态调校。

The nonlinear correction method of methane sensor based on improved BP neural network was introduced, which approached inverse nonlinear function to finish the nonlinear correction by use of nonlinear mapping ability of neural network. The experimental results showed that the improved neural network has quick convergence rate and high approaching precision, the accuracy of the sensor could be greatly increased from 15.02% to ±0. 130% compared with traditional method of piecewise linear and BP algorithm.