Thermodynamical and artificial intelligence approaches of H2S solubility in N-methylpyrrolidone
| Autor: | Mahdi Koolivand Salooki, Mohammad Shokouhi, Jafar Sadeghzadeh Ahari, Morteza Esfandyari |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Materials science
Artificial neural network business.industry Vapor pressure General Physics and Astronomy Experimental data 02 engineering and technology 021001 nanoscience & nanotechnology 020401 chemical engineering Operating temperature Genetic algorithm Artificial intelligence 0204 chemical engineering Physical and Theoretical Chemistry Solubility 0210 nano-technology business Cubic function Topology (chemistry) |
| Zdroj: | Chemical Physics Letters. 707:22-30 |
| ISSN: | 0009-2614 |
| Popis: | Artificial Intelligence approach is utilized for developing the auto-design model of neural network based genetic algorithm (GA) to manipulate the obtained experimental data of H2S solubility measurements. The solubility of hydrogen sulfide (H2S) in low vapor pressure organic solvents N-methylpyrrolidone (NMP) has been measured experimentally. The experimental input data included operating temperature and pressure, and solubility values regarded as output data. Design of topology and parameters of the neural networks as decision variables have been achieved by GA, which enhances the effectiveness of the forecasting scheme. The precision of the GA is compared with experimental data and also with thermodynamical modeling approach in which Peng-Robinson-Stryjek-Vera (PRSV) cubic equation of state has been chosen for this purpose. The results reveal that the thermodynamical and GA–artificial neural networks (ANN)-based model, both have sound agreement with experimental values. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect