Kinetics-Informed Neural Networks

TL;DR

This paper introduces a neural network-based framework for solving inverse kinetic ODEs in chemical reaction modeling, enabling the estimation of kinetic parameters from transient data and aiding in reaction mechanism elucidation.

Contribution

It presents a novel algebraic framework and a multi-objective optimization approach for training neural networks to estimate kinetic parameters from synthetic transient data.

Findings

Neural networks can accurately estimate kinetic parameters from synthetic data.

The methodology is robust against statistical noise in the data.

It facilitates reaction mechanism elucidation from transient kinetic measurements.

Abstract

Chemical kinetics and reaction engineering consists of the phenomenological framework for the disentanglement of reaction mechanisms, optimization of reaction performance and the rational design of chemical processes. Here, we utilize feed-forward artificial neural networks as basis functions to solve ordinary differential equations (ODEs) constrained by differential algebraic equations (DAEs) that describe microkinetic models (MKMs). We present an algebraic framework for the mathematical description and classification of reaction networks, types of elementary reaction, and chemical species. Under this framework, we demonstrate that the simultaneous training of neural nets and kinetic model parameters in a regularized multi-objective optimization setting leads to the solution of the inverse problem through the estimation of kinetic parameters from synthetic experimental data. We analyze a set of scenarios to establish the extent to which kinetic parameters can be retrieved from transient kinetic data, and assess the robustness of the methodology with respect to statistical noise. This approach to inverse kinetic ODEs can assist in the elucidation of reaction mechanisms based on transient data.