Identification of natural gas components is vital for the natural gas measurement and determination of the gas flow. The accurate evaluation of the natural gas composition is particularly vital for thermal methods of measurement. The thermal measurement principle is the focus of the current research and development trend of natural gas control technology. Therefore, in this study, we propose a method based on principal component analysis. To eliminate the partial correlation error between the samples and retain maximum information, the dimensionality reduction and pre-classification methods are performed on the data obtained by analyzing the physical parameters of the target natural gas. To reduce the input into the network and ensure recognition efficiency, the new samples are used as the input for the support vector machine. The method can be applied for providing an accurate classification of the existing types of natural gas and obtaining reliable data for thermal metering methods.

天然气成分的识别对于天然气测量和确定气体流量至关重要。天然气成分的准确评估对于热测量方法尤为重要。热测量原理是当前天然气控制技术研究和发展趋势的重点。因此，在本研究中，我们提出了一种基于主成分分析的方法。为消除样本之间的偏相关误差，保留最大信息，对目标天然气物理参数分析得到的数据进行降维和预分类方法。为了减少对网络的输入并保证识别效率，新样本被用作支持向量机的输入。