Experimental methods in chemical engineering: Monte Carlo
Ergys Pahija,Soonho Hwangbo,Thomas Saulnier‐Bellemare,Gregory S. Patience
DOI: https://doi.org/10.1002/cjce.25374
2024-06-27
The Canadian Journal of Chemical Engineering
Abstract:Monte Carlo (MC) methods employ a statistical approach to evaluate complex mathematical models that lack analytical solutions and assess their uncertainties. To this end, techniques such as Markov chain Monte Carlo (MCMC), bootstrap, and sequential MC methods repeat the same operations over a specified range of conditions. Consequently, both the frequentist and Bayesian statistical approaches are computationally intensive, depending on the problem formulation. Improving sampling techniques and identifying sources of error reduce the computational demand but do not guarantee that the solution reaches the global optimum. Moreover, efficient algorithms and advances in hardware continue to decrease computation time. MC methods are applicable to a plethora of problems ranging from medicine to computational chemistry, economics, and industrial safety, making them integral to the ongoing industrial digitalization by evaluating the quality of applied models. In chemical engineering, MC simulations are used in four clusters of research: design, systems, and optimization; molecular simulation, including CO2 and carbon capture; adsorption and molecular dynamics; and thermodynamics. There is limited cross‐referencing between the design cluster and the other three, which presents an interesting area for future research. This mini‐review presents two applications within chemical engineering: emissions and energy forecasting.
engineering, chemical