Arrhenius.jl is a differentiable combustion simulation package built with the design principle of treating auto-differentiation as a first-class citizen. It aims to facilitate the development of hybrid physics and AI models, paving the way for advancements in automonous combustion research. This README offers a detailed overview of the project, including installation instructions, related publications, applications, and more.
To install Arrhenius.jl, enter the following command:
pkg> add https://github.com/DENG-MIT/Arrhenius.jl- Arrhenius.jl: A Differentiable Combustion Simulation Package: overview of Arrhenius.jl and applications in deep mechanism reduction, uncertainty quantification, mechanism tuning and model discovery. Slides in NCM21, Vedio for NCM21.
- Machine Learning Approaches to Learn HyChem Models: demonstrate 1000 times faster than genetic algorithms using commercial software for optimizing complex kinetic models.
- Neural Differential Equations for Inverse Modeling in Model Combustors
- SGD-based Optimization in Modeling Combustion Kinetics: Case Studies in Tuning Mechanistic and Hybrid Kinetic Models
- Sensitivity analysis for auto-ignition | repo | Features: auto-differentiation, multi-threading, sensitivity to all of three Arrhenius params A, b and Ea, active subspace based uncertainty quantification
- Sensitivity analysis for one-dimensional flames | repo | Features: auto-differentiation, multi-threading, sensitivity to all of three Arrhenius params A, b and Ea.
- Automonous learn kinetic mechanism using neural network | repo | Features: Chemical Reaction Neural Network (CRNN), Neural Ordinary Differential Equations.
- Deep Reduction | repo | Features: Two-stages mechanism reduction with deep learning.
Example
Note that some of the examples are in development and you can have early access by contacting Weiqi Ji
- ReactionMechanismSimulator.jl The amazing Reaction Mechanism Simulator for simulating large chemical kinetic mechanisms
- Cantera A comprehensive C++ based combustion simulation package and with great python interface. Arrhenius relies on Cantera when it is applicable.