The effect of utilising energy derived from fossil sources on the environment has aroused research interest in alternative and sustainable energy sources. Synthesis gas, a mixture of carbon monoxide (CO) and hydrogen can be used as starting materials in hydrogenation reactions to produce chemical intermediates that can be used in various processes. This study investigates the robustness of applying a fine Gaussian support vector machine algorithm for predicting light olefins from catalytic CO hydrogenation using magnesia nanoparticles-based catalysts. The datasets obtained from the CO hydrogenation reaction consist of input parameters such as magnesia nanoparticles contents, reaction temperature, and reactor pressure, and the output parameters which include CO conversions and the selectivity of light olefins (CH₄, C₂H₆, C₂H₄ C₃H₈, C₄H₈, and C₃H₆). The dataset was trained and employed for the prediction of the light olefins using a support vector machine with an inbuilt Fine Gaussian Kernel function. The performance of the support vector machine was evaluated using the coefficient of determination (R²), root mean squared error (RMSE), mean square error (MSE), and the mean absolute error (MAE). The support vector machine showed significant potential in the prediction of CO conversion, CH₄ selectivity, C₂H₆ selectivity, and C₂H₄ selectivity as indicated by R² of 0.770, 0.800, 0.730, and 0.930, respectively. While less predictive performance was obtained for the prediction of C₃H₈ selectivity, C₄H₈ selectivity and C₃H₆ selectivity as indicated by R² of 0.630, 0.610, and 0.320, respectively.

Search results

Modeling the Effect of Magnesia Nanoparticles on CO Hydrogenation to Light Olefins in a Continuous Flow Reactor Using Fine Gaussian Support Vector Machine