The evolution of computational methods in chemical engineering: From finite elements to artificial intelligence

This article presents a systematic review of the evolution of computational methods in chemical engineering, analyzing the transition from classical numerical solutions to the integration of artificial intelligence (AI). Under the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodological framework, it examines how the discipline was initially founded on first-principles methods, such as finite differences and finite elements, essential for solving systems of differential equations in transport phenomena and reactors. With the increase in computing capacity between 1980 and 2010, Process Systems Engineering (PSE) was consolidated, formalizing flowsheet simulation and large-scale mathematical optimization. The study delves into the current paradigm shift, where AI and machine learning expand the capabilities of physical models to address uncertainty and the high dimensionality of nonlinear and multiscale industrial systems. To conclude, challenges regarding interpretability and the need for hybrid models that combine phenomenological rigor with algorithmic efficiency are discussed. This work offers a historical and technical perspective that identifies the trends guiding chemical engineering practice in the digital era, optimizing prediction and decision-making in complex environments.