Identification of Natural Gas Components Using the Support Vector Machine Model

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.

Automated summary

The study proposes a method based on principal component analysis to accurately classify different types of natural gas and obtain reliable data for thermal metering methods. By reducing the dimensionality and pre-classifying the physical parameters of the target natural gas, and using new samples as input for the support vector machine, the method improves the efficiency of recognition and ensures accurate classification.