Volume 6, Issue 3 (2022)
IQBQ 2022, 6(3): 41-59 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Eghbal Ahmadi M H, Mosayebi A. Prediction of conversion and products yields in steam reforming of methanol over Cu-Zn/ZrO2 catalyst using Artificial Intelligence methods. IQBQ 2022; 6 (3) :41-59
URL: http://arcpe.modares.ac.ir/article-38-64202-en.html
URL: http://arcpe.modares.ac.ir/article-38-64202-en.html
1- Department of Chemical Engineering, Tafresh University, Tafresh, 39518 79611, Iran , mh.eghbalahmadi@tafreshu.ac.ir
2- Department of Chemical Engineering, Tafresh University, Tafresh, 39518 79611, Iran
2- Department of Chemical Engineering, Tafresh University, Tafresh, 39518 79611, Iran
Abstract: (1213 Views)
Subject
In this study, the steam reforming of the methanol process was analyzed based on three different inputs including temperature, pressure, and H2O/CH3OH ratio with the use of different Artificial Intelligence methods.
Methodology
In the first step, Cu-Zn/ZrO2 catalysts were synthesized via the co-precipitation method, and then experimental tests of steam reforming of methanol were performed at a temperature range of 180 –500 °C, the pressure of 1-11 bar, and the H2O/CH3OH ratio of 0.75-3.75 on the Cu-Zn/ZrO2 catalyst in a fixed bed reactor. Afterward, three different methods of Mamdani fuzzy type-1, Mamdani fuzzy type-2, and Sugeno fuzzy were applied in order to develop the models. Using these methods, the developed models only required the heuristics derived from the expert’s knowledge and some experimental data, without needing the calculation of complex kinetic as well as thermodynamic parameters related to the corresponding process. In addition, the structures of the developed fuzzy models were optimized to improve the model performance according to the analysis of the initial results. The model developments didn’t require a high number of experimental data, and this feature is especially interesting when dealing with the process conditions in which data gathering is expensive or the accuracy of data is low.
The main results
The overall accuracy as well as the properties of the developed models were compared. The type-2 Mamdani fuzzy model proved to be the best model, using which, the methanol conversion, H2 yield, and CO yield were predicted with accuracies of 67%, 91%, and 83%, respectively.
In this study, the steam reforming of the methanol process was analyzed based on three different inputs including temperature, pressure, and H2O/CH3OH ratio with the use of different Artificial Intelligence methods.
Methodology
In the first step, Cu-Zn/ZrO2 catalysts were synthesized via the co-precipitation method, and then experimental tests of steam reforming of methanol were performed at a temperature range of 180 –500 °C, the pressure of 1-11 bar, and the H2O/CH3OH ratio of 0.75-3.75 on the Cu-Zn/ZrO2 catalyst in a fixed bed reactor. Afterward, three different methods of Mamdani fuzzy type-1, Mamdani fuzzy type-2, and Sugeno fuzzy were applied in order to develop the models. Using these methods, the developed models only required the heuristics derived from the expert’s knowledge and some experimental data, without needing the calculation of complex kinetic as well as thermodynamic parameters related to the corresponding process. In addition, the structures of the developed fuzzy models were optimized to improve the model performance according to the analysis of the initial results. The model developments didn’t require a high number of experimental data, and this feature is especially interesting when dealing with the process conditions in which data gathering is expensive or the accuracy of data is low.
The main results
The overall accuracy as well as the properties of the developed models were compared. The type-2 Mamdani fuzzy model proved to be the best model, using which, the methanol conversion, H2 yield, and CO yield were predicted with accuracies of 67%, 91%, and 83%, respectively.
Article Type: Original Research |
Subject:
nano-catalyst
Received: 2022/09/14 | Accepted: 2022/12/31 | Published: 2022/12/31
Received: 2022/09/14 | Accepted: 2022/12/31 | Published: 2022/12/31
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |