Table 1 Overview of key literature on ML applications in gas adsorption.
From: Toward accurate prediction of N2 uptake capacity in metal-organic frameworks
References | Material type | Gas | ML technique | R2 | Input parameters | Main contribution |
|---|---|---|---|---|---|---|
Graphene oxide (GO) | CO2 | Extra trees regressor Gradient boosting regressor RF regressor, XGBR regressor, SVR | 0.989 0.987 0.981 0.979 0.965 | Pressure (bar) BET surface area (m2/g), Temperature (K), pore volume (cm3/g) | Optimized porous GO structures for efficient CO2 sequestration | |
Metal organic framework (MOF) | CO2 | XGBoost LightGBM Cat Boost RF | 0.995 0.983 0.965 0.946 | Pressure (bar), temperature (K), surface area (m2/g), pore volume (cm3/g) | Intelligent modeling of non-linear CO2 adsorption in MOFs | |
Activated carbon (AC) | CO2 | Gradient Boosting ANN-MLP Gaussian K-Near Neighbor RF DT SVR | 0.996 0.976 0.946 0.946 0.981 0.982 0.841 | BET surface area (m2/gr), mean pore diameter (nm), pore volume (cc/gr), temperature (K), pressure(bar) | Development of high-performance adsorbents for sustainable CO2 capture | |
Shale | CH4 | XGBoost RF SVR | 0.978 0.908 0.841 | Pressure, temperature, moisture, TOC | Modeling methane uptake in shales | |
Metal organic framework (MOF) | CH4 | RF | 0.98 | Interpenetration capacity, void fraction, dominant pore diameter, surface area (m2/g), Density (g/cm3), maximum pore diameter,Number of interpenetrations framework | Rapid and precise estimation of methane uptake in MOFs | |
Metal organic framework (MOF) | H2 | DT (Fine Tree) Random Forest Tri-layered NN Boosted Trees Quadratic SVM Linear SVM Exponential GPR Matern 5/2 GPR Bi-layered NN Narrow NN Medium NN Wide NN Rational Quadratic Kernel GPR | 0.98 0.98 1 0.98 0.98 0.84 0.99 0.99 0.99 0.99 0.99 0.99 0.99 | Void fraction, pore limiting Crystal density (g/cm3), Pore volume (cm3/g), diameter(Å), largest cavity diameter(Å) | Prediction of gravimetric and volumetric H₂ adsorption in real MOFs with validation against simulated and experimental data | |
Porous carbon material (PCM) | H2 | GRNN ELM LSSVM ANFIS | 0.970 0.939 0.991 0.968 | Temperature(T), Pressure (MPa), surface area (m2/g) Measured weight percentages (wt %) of oxygen (O), hydrogen (H), carbon (C), and nitrogen (N), Total pore volume (cm3/g); | Forecasting hydrogen retention ability in PCMs via ML techniques | |
Metal organic framework (MOF) | H2 | CNN DNN GPR | 0.998 0.997 0.999 | Pressure (bar) BET surface area (m2/g), total pore volume (cm3/g temperature (K), | The combined impact of input variables on predicting hydrogen storage capacity in MOFs | |
Metal organic framework (MOF) | H2 | Ada Boost Bagging with DT Bagging with RF Boosted DT DT ET Gradient Boosting K-Nearest Neighbors Linear Regression Nu-SVM with (RBF) Kernel RF Ridge Regression SVM- (RBF) Kernel SVM with Linear Kernel | 0.911 0.963 0.964 0.965 0.936 0.967 0.955 0.926 0.913 0.913 0.963 0.913 0.913 0.907 | Largest cavity diameter (Å), void fraction, density(g/cm3), Gravimetric surface area (m2 /g), pore volume (cm3/g), pore limiting diameter (Å), Volumetric surface area (m2 /cm3), | Applying ML models to expedite the search for MOFs with optimized volumetric characteristics | |
Metal organic framework (MOF) | CH4 | SVM RF GBRT | 0.968 0.980 0.983 | Gravimetric surface area (m2 /g), volumetric surface area (m2 /cm3), Largest cavity diameter(Å), Henry coefficient (mol kg− 1 MPa− 1), density (g /cm3), void fraction | ML-enabled acceleration of high-efficiency MOF discovery for CH4 storage applications. | |
Metal organic framework (MOF) | H2 | SVM RF XGBoost LightGBM CatBoost ANN- 5 CMIS-Before Leverage CMIS-After Leverage | 0.912 0.931 0.939 0.936 0.958 0.935 0.935 0.982 | BET surface area (m2/g), pressure (bar), temperature (K), total pore volume (cm3/g) | Modeling framework for MOF selection and process optimization in H2 storage systems | |
Metal organic framework (MOF) | CO2 | LSSVM, MLP, LSSVM-Growth optimization), MLP-Growth optimization, LSSVM- PSO, MLP-PSO | 0.923 0.900 0.992 0.961 0.955 0.939 | Temperature (K), Total pore volume (cm3/g), pressure (bar) Langmuir-specific surface area(m2/g) | Cost-effective ML prediction framework for adsorption capacity in MOF design | |
Metal organic framework (MOF) | CO2 | PSO-ANFIS, DE-ANFIS, RBF ANN, LSSVM | 0.957 0.934 0.999 0.998 | Temperature (K), total pore volume (cm3/g), pressure (bar), surface area(m2/g), | Finding a suitable method for estimating CO2 uptake in MOFs |
