基于ABC-PSO的ε-SVM在甲烷测量中的应用
|
摘要
针对红外甲烷传感器在工业现场测量时易受到温度、湿度以及类似气体等非目标变量的影响,提出了一种基于人工蜂群和粒子群混合优化算法(ABC-PSO)的支持向量机模型(ABC-PSO-ε-SVM)对其进行校正。将ABC算法与PSO算法并行组合构成混合优化算法,能够感知非目标变量的变化,快速、准确地搜索到SVM参数。实验中,采用红外甲烷传感器对0%~5.05%浓度的16组标准甲烷气体进行测量,将其中11组数据作为训练集,5组数据作为测试集,建立ε-SVM回归校正模型并进行预测。结果表明:模型的回归拟合效果好,预测精度比单一优化算法的SVM模型高。
A support vector machine( SVM) model based on artificial bee colony( ABC) and particle swarm hybrid optimization algorithm is proposed to have a correction for the measurement of infrared methane sensor aiming at in industrial field which is easy to be affected by temperature,humidity and other gases,and so on. The model combines the ABC algorithm with the PSO algorithm to form a hybrid optimization algorithm. It can detect the change of non target variables,and quickly and accurately search the SVM parameters. In experiments,adopting infrared methane sensor to measure concentration of 16 groups of standard methane gas which is in the range of0 %~ 5. 05 %. Selecting 11 groups of data as training set and the rest of data as test set to establish ε-SVM regression correction model and carry out prediction. The results show that regression fitting effect of the model is good,and the prediction precision is higher than single optimization algorithm of SVM model.
A support vector machine( SVM) model based on artificial bee colony( ABC) and particle swarm hybrid optimization algorithm is proposed to have a correction for the measurement of infrared methane sensor aiming at in industrial field which is easy to be affected by temperature,humidity and other gases,and so on. The model combines the ABC algorithm with the PSO algorithm to form a hybrid optimization algorithm. It can detect the change of non target variables,and quickly and accurately search the SVM parameters. In experiments,adopting infrared methane sensor to measure concentration of 16 groups of standard methane gas which is in the range of0 %~ 5. 05 %. Selecting 11 groups of data as training set and the rest of data as test set to establish ε-SVM regression correction model and carry out prediction. The results show that regression fitting effect of the model is good,and the prediction precision is higher than single optimization algorithm of SVM model.
引文
[1]谭秋林.红外光学气体传感器及检测系统[M].北京:机械工业出版社,2013:1-139.
[2]陈红岩,郭晶晶,刘文贞,等.非分光红外的甲烷传感器设计[J].传感器与微系统,2016,35(12):1-4.
[3]Leis John,Buttsworth David.A temperature compensation technique for near-infrared mathane gas threshold detection[J].IEEE Transactions on Industrial Electronics,2016,63(3):1813-1821.
[4]李卯东,梁永直.基于遗传优化的BP神经网络法在甲烷检测中的应用[J].工矿自动化,2013,39(2):51-53.
[5]柴文光.CPSO支持向量机红外瓦斯传感器动态补偿[J].华侨大学学报:自然科学版,2016,37(3):316-319.
[6]张伟.人工蜂群混合优化算法及应用研究[D].杭州:浙江大学,2014.
[7]秦全德,程适,李丽,等.人工蜂群算法研究综述[J].智能系统学报,2014,9(2):127-135.
[8]唐俊.PSO算法原理及应用[J].计算机技术与发展,2010,20(2):213-216.
[9]刘俊芳.粒子群和人工蜂群的混合优化算法优化SVM参数及应用[D].太原:太原理工大学,2012.
[10]郭天太,洪博,潘增荣,等.改进的SVM在矿井气体定量分析中的应用[J].红外与激光工程,2016,45(6):211-218.
[11]郭瑞,常勇.红外甲烷传感器非线性动态补偿的研究[J].传感器与微系统,2011,30(8):31-33,36.
[12]王宏伟,韩云涛,彭继慎.基于TSPSO支持向量机红外甲烷传感器动态补偿[J].传感技术学报,2013,26(9):1193-1197.
[13]刘文贞,陈红岩,李孝禄,等.基于自适应变异粒子群算法的混合核ε-SVM在混合气体定量分析中的应用[J].传感技术学报,2016,29(9):1464-1470.
[2]陈红岩,郭晶晶,刘文贞,等.非分光红外的甲烷传感器设计[J].传感器与微系统,2016,35(12):1-4.
[3]Leis John,Buttsworth David.A temperature compensation technique for near-infrared mathane gas threshold detection[J].IEEE Transactions on Industrial Electronics,2016,63(3):1813-1821.
[4]李卯东,梁永直.基于遗传优化的BP神经网络法在甲烷检测中的应用[J].工矿自动化,2013,39(2):51-53.
[5]柴文光.CPSO支持向量机红外瓦斯传感器动态补偿[J].华侨大学学报:自然科学版,2016,37(3):316-319.
[6]张伟.人工蜂群混合优化算法及应用研究[D].杭州:浙江大学,2014.
[7]秦全德,程适,李丽,等.人工蜂群算法研究综述[J].智能系统学报,2014,9(2):127-135.
[8]唐俊.PSO算法原理及应用[J].计算机技术与发展,2010,20(2):213-216.
[9]刘俊芳.粒子群和人工蜂群的混合优化算法优化SVM参数及应用[D].太原:太原理工大学,2012.
[10]郭天太,洪博,潘增荣,等.改进的SVM在矿井气体定量分析中的应用[J].红外与激光工程,2016,45(6):211-218.
[11]郭瑞,常勇.红外甲烷传感器非线性动态补偿的研究[J].传感器与微系统,2011,30(8):31-33,36.
[12]王宏伟,韩云涛,彭继慎.基于TSPSO支持向量机红外甲烷传感器动态补偿[J].传感技术学报,2013,26(9):1193-1197.
[13]刘文贞,陈红岩,李孝禄,等.基于自适应变异粒子群算法的混合核ε-SVM在混合气体定量分析中的应用[J].传感技术学报,2016,29(9):1464-1470.