Process engineering is facing new grand challenges, adapting to the many changes in society induced by the struggle for resources and the need for better ways of using them. For example, process design needs to include evolutions in material feedstocks (e.g., renewable materials, waste) as well as energies (e.g., various sources, decarbonized), in environmental considerations (e.g., carbon, water, biodiversity, chemical pollution in water, air and soil); while performing new services such as intrant and product flexibility or polygeneration.To address these challenges, a broad variety of technical innovations offers more possibilities, in addition to traditional well-established functions, such as new materials (e.g., ionic liquids, composite materials, biomolecules), production techniques for materials (e.g., additive manufacturing), use of non-conventional driving forces (e.g., light or microwave), hybridization of existing functions (e.g., membrane distillation or adsorption bioreactor), and engineering paradigms (e.g., modularization or decentralized production).To achieve sustainable design of processes and match challenges and opportunities, process designers therefore need to make complex decisions among an expanding set of possibilities at all scales (material, unit operation, process system) within a complex environment.
Fortunately, recent advances in algorithmic techniques, such as artificial intelligence (AI), have gained in popularity and applicability, and could be used as decision-aiding tools to assist process designers cope with increasing complexity when faced with new horizons. The growth in computing capacity and discoveries in cognitive and digital sciences are leading to the popularization of new algorithms for automation and assistance with certain tasks (e.g. optimization, machine learning), discovering new horizons (e.g. molecular design) and multi-criteria decision analysis methods.AI covers a wide range of techniques at the intersection of the data science and operational research communities. Thus, a broad variety of "intelligent" algorithms could be used, exploiting experiences from both reward-driven optimization techniques (global optimization, meta-heuristics, nature-inspired algorithms, etc.) and data-driven learning techniques (Machine Learning, Deep Learning, etc.). Today, AI is being used in the field of chemical engineering for addressing new molecular separations, reaction pathways, innovative processes conceptualization, improving sustainability of conventional technologies, or re-purposing them to new challenges.
This Research Topic concerns the use of AI -in the general sense of the automating decision-making on a complex engineering problem by mimicking intelligent behavior-to achieve sustainable process design. It collects contributions illustrating different machine-learning concepts that can be purely data-driven of coupled with physics-based approaches, and applied at various scales -from the product to the process-of chemical engineering.Contributions are grouped into three categories: property prediction, hybrid modelling, and flowsheet generation.The first set of contributions concerns a popular use of Machine Learning in chemical engineering: the predictions of properties of interests (thermochemical, kinetics, etc.) by data-driven techniques. These approaches could either use data based on experimental measurements or generated using physics-based numerical simulations.Wang et al.'s paper, "Predicting xylose yield from prehydrolysis of hardwoods: A machine learning approach", explores the prediction of a reaction yield, with an application to biomass conversion (e.g., biorefining). Database is based solely on experimental samples taken from the literature. Authors used three types of Machine Learning algorithms (ridge regression, support vector regression and artificial neural networks), and compared their predictions with a conventional kinetic model, such phenomenological models being often limited to specific species and operating regimes.Komori et al.'s paper, "Stress Prediction of the Particle Structure of All-Solid-State Batteries by Numerical Simulation and Machine Learning", deals with the stress at the contact surface between microparticles within all-solid-state batteries, with an application to energy storage (e.g., electric and hybrid vehicles). Authors used a discrete element method to generate 2D images of particles structures and corresponding predictions. A convolution neural networks is used to make predictions from the cross-sectional images. ML prediction compares well with physics-based simulations, and links the structure of porous materials to stress properties.The second set of contributions concerns an emerging hybrid modelling paradigm, combining knowledge-based modelling (often called white box), and data-driven modelling (often called black box). The former (white box, often called "first-principle models" in the context of chemical engineering) is based on a fundamental understanding and modelling of physico-chemical phenomena, it exhibits good physical interpretability and extrapolability but offers high model complexity and is computationally intensive. The latter (black box) on the contrary is purely based on data; it exhibits lower model complexity and computational time, but lacks in physical interpretability and extrapolability. The two contributions are both review articles on the use of hybrid modelling, respectively regarding biopharmaceutical processes and predictive maintenance.Narayanan et al.'s paper, "Hybrid modeling for biopharmaceutical processes: advantages, opportunities, and implementation", deals with the use of such modeling paradigm for biopharmaceutical industry. Authors discuss the interest of such a technique for this industry. They present the two common architectures for hybridization (serial data/knowledge, or parallel) and report successful implementations of hybrid models for different types of bioprocesses. A step-by-step sequence is then proposed for developing hybrid models.Cai et al.'s paper, "Review of synergy between machine learning and first principles models for asset integrity management", discusses another aspect of the chemical process industry. Authors present how hybrid models combining first principles and machine learning models can be applied to the asset integrity management. They discuss especially the interest for the improvement of predictive maintenance, with an emphasis on advances on digital twins, and existing software solutions.The last original contribution is related to another emerging field: the generation of process flowsheets. The use of generative AI for molecular design is an extensive line of work in the literature, that could be a source of inspiration for process flowsheet design.Nabil et al.'s paper "Data-driven structural synthesis of supercritical CO2 power cycles", deals with the design of innovative power cycles using supercritical CO2 as working fluid for waste-heat recovery. Authors use Long Short Term Memory networks (common in natural language processing) to generate process flowsheets that exhibit high efficiencies. Data are initially random cycles, then the model is fine-tuned using efficient cycle layouts optimized with a physics-based simulator.