Multi-objective optimisation with hybrid machine learning strategy for complex catalytic processes

• A hybrid machine learning strategy is developed for complex chemical processes. Continuum lumping kinetic embedded into the neural network framework. NSGA-II used multi-objective optimisation of hydrocracking process. Catalytic processes such as hydrocracking, gasification and pyrolysis play a vital role in renewable energy net zero transition. Due to non-linear behaviours during operation, catalytic require powerful modelling tool prediction smart speedy green process routes discovery rapid design. However, challenges remain due lack an effective toolbox, which requires not only precise analysis but also fast optimisation. Here, we propose by embedding physics-based continuum model data-driven artificial This adopted surrogate demonstrated benchmarking The results show that novel exhibits mean square error less than 0.01 comparing with simulation results. well-trained was then integrated non-dominated-sort genetic algorithm (NSGA-II) evaluate optimise yield selectivity Pareto front from able identify trade-off curve between objective functions essential decision-making Our work indicates adopting promising approach various enable accurate computation well