Neural network model and software for an information system to intelligently analyze gas quality

The problem of analyzing the quality of natural gas is solved by traditional methods of gas chromatography. The article proposes an alternative approach using neural networks. An automated information system to determine energy parameters of natural gas and its operation were studied. The system testing was conducted on experimental data obtained from real gas mixtures in laboratory conditions. The gas quality indicators were calculated and the conclusion about system applicability was drawn. The developed mathematics and software allow to provide high performance for the information system in important cases where gas properties can change quickly and constant monitoring is required. Based on experi-mental results, an algorithmic solution was proposed for natural gas quality analysis, that allows to obtain necessary data with lower time and financial costs.