中文摘要：

组建了沼气检测的实验系统,采用国家标准混合气获取大量的浓度标定数据,分析了目前广泛应用的甲烷浓度预测算法及影响预测结果的因素,讨论了支持向量机在CH4浓度预测中的应用,在此基础上研究了将多通道探测器的电压输出及环境温度共同作为支持向量机的输入,实现CH4浓度的预测。将该方法与线性插值法、多项式回归法、神经网络法等多种方法进行比较,预测结果的平均绝对误差减小了0.44%～1.99%。初步试验结果表明该方法在CH4浓度检测中具有一定的应用前景。

英文摘要：

An experimental system was built for biogas detection to acquire abundant concentration calibration data by using national standard mixture-gas.Some widely used algorithms for predicting CH4 concentration are analyzed.At the same time,some elements affecting these algorithms are discussed.On the basis of the above analysis,a prediction model using support vector machine（SVM） is proposed to predict CH4 concentration.The input module of the proposed algorithm includes temperature（T） and outputs of multichannel detector.Compared with linear interpolation method,polynomial regression,neural network,etc,the mean absolute deviation is reduced by 0.44%~1.99%.Preliminary results indicate that this method has definite application prospect in CH4 detection from methane.