论文部分内容阅读
为了解决在线油色谱受外界环境和设备误差影响导致数据失真的问题,笔者提出了一种基于萤火虫支持向量机的油色谱在线数据校正的方法。首先将支持向量机中的一组错误惩罚因子,不敏感参数和核参数作为萤火虫个体,通过萤火虫算法对影响支持向量机性能的重要参数进行优化。然后计算油色谱离线数据间的分段函数,当在线数据超出分段函数误差允许的范围时,认为在线数据异常。利用少数准确的油色谱离线数据对支持向量机回归模型进行训练,当在线数据出现异常时,通过支持向量机回归模型对异常的在线数据进行校正。最后通过某台变压器油色谱的在线和离线数据对文中提出的方法进行验证,结果证明了该方法的可行性和有效性。
In order to solve the problem of data distortion caused by the influence of external environment and equipment error on online oil chromatography, a method of on-line data correction of oil chromatography based on firefly support vector machine was proposed. First, a set of error penalty factors, insensitive parameters and kernel parameters in support vector machines are used as fireflies individuals. The firefly algorithm is used to optimize the important parameters that affect the performance of SVM. And then calculate the segmented function between the off-line data of the oil chromatography. When the on-line data exceeds the allowable range of the segmented function error, it is considered that the on-line data is abnormal. A few accurate off-line data of oil chromatography were used to train SVM regression model. When the online data was abnormal, SVM regression model was used to correct the abnormal online data. Finally, the method presented in this paper is validated by online and offline data of one transformer oil chromatogram. The result proves the feasibility and effectiveness of this method.