(278b) Multiscale Modeling and Control for Lignin-First Biomass Fractionation

This work presents multiscale modeling approaches and a control framework for optimizing lignin-first biomass fractionation, addressing key challenges in lignin valorization industries [1, 2]. The framework connects molecular-scale reaction mechanisms with reactor-level dynamics. This will serve as a powerful tool to simulate lignin reaction dynamics and enable predictive operation throughout the process.

The first part of the framework integrates molecular simulation techniques to capture the detailed chemistry of lignin depolymerization, condensation, and demethoxylation reactions under various conditions. Ab initio molecular dynamics simulations identify temperature-dependent lignin conformations, while density functional theory provides activation energies for key reactions. These inputs inform kinetic Monte Carlo (kMC) simulations that describe the evolution of molecular weight distributions (MWd) and monolignol compositions [3]. However, the kMC algorithm requires iterative calculations to simulate the system over time.

Furthermore, to address the high computational cost of kMC, a Gaussian mixture model is used to parametrize lignin MWd, followed by an artificial neural network (ANN) training to reproduce reaction rate distributions. While the original kMC model requires iterative rate calculations to execute single event for each time step, the proposed ANN-acceleration circumvents the computational bottleneck, thereby capturing lignin properties rapidly while maintaining mechanistic fidelity [4].

Building on these developments, the framework has been applied to identify optimal operating strategies that target specific lignin structural outcomes. This lays the foundation for the future implementation of model-based control strategies that can regulate molecular properties during fractionation. The integrated methodology provides a scalable platform to advance lignin valorization through predictive modeling.

Literature cited:

[1] Shorey R., Salaghi A., Fatehi P., & Mekonnen T.H., (2023). Valorization of lignin for advanced material applications: A review. RSC Sustainability, 2, 804-831.

[2] Ali S., Rani A., Dar M.A., Qaisrani M.M., Noman M., Yoganathan K., Asad M., Berhanu A., Barwant M. & Zhu D., (2024). Recent advances in characterization and valorization of lignin and its value-added products: Challenges and future perspectives. Biomass, 4, 947-977.

[3] Lee C.H., Kim J., Ryu J., Won W., Yoo C.G., & Kwon J.S.-I., (2024). Lignins structure dynamics: Advanced real-time molecular sensing strategies. Chem. Eng. J,, 487, 150680.

[4] Kim J., Ryu J., Zhang M., Yang Q., Yoo C.G., & Kwon J.S.-I., (2024). Real-time model predictive control of lignin properties using an accelerated kMC framework with artificial neural networks. Ind. Eng. Chem. Res., 63, 20978-20988.