Utilization of LSSVM strategy to predict water content of sweet natural gas
Autor: | Alireza Baghban, Mohammad Navid Kardani |
---|---|
Rok vydání: | 2017 |
Předmět: |
business.industry
Chemistry 020209 energy General Chemical Engineering Energy Engineering and Power Technology Predictive capability 02 engineering and technology General Chemistry Function (mathematics) Geotechnical Engineering and Engineering Geology Support vector machine Fuel Technology Temperature and pressure 020401 chemical engineering Natural gas Approximation error 0202 electrical engineering electronic engineering information engineering Range (statistics) 0204 chemical engineering business Process engineering Water content |
Zdroj: | Petroleum Science and Technology. 35:761-767 |
ISSN: | 1532-2459 1091-6466 |
Popis: | The presence of water in the natural gas causes numerous operating problems. Hence, in order to have satisfactory operational conditions in the gas production units, we should determine the accurate amount of water content of natural gas. This study aimed to develop a low parameter predictive tool based on the least square support vector machine for estimating water content of natural gas as a function of temperature and pressure under a wide range of conditions. Results obtained from the suggested model indicated its lower deviation than other existence correlations and confirmed its acceptable predictive capability. In addition, the obtained values of R-squared and mean relative error were 1.00 and 0.24%, respectively. This tool is simple to use and can be of help to gas engineers to determine an accurate approximation of water content of natural gas. |
Databáze: | OpenAIRE |
Externí odkaz: |