Skip to content
dcurreli edited this page Sep 26, 2018 · 3 revisions

What is CRANE?

CRANE (Chemical ReAction NEtwork tool for plasma chemistry) is a new software for the general treatment of plasma chemistry problems with an arbitrary number of reacting plasma species and arbitrary number of plasma reaction pathways. The mathematical kernel of the software is a coupled system of ODE rate equations.

CRANE has been specifically developed as a part of the MOOSE framework to address the non-equilibrium, multi-species reaction needs of Zapdos. In general a CRANE input file follows the same syntax as other MOOSE applications; however, the input file has been streamlined to allow the system of coupled chemical species, rate equations, and rate coefficients to be written into the input file as reaction equations in the form of rate equations, allowing reaction networks to be built in an intuitive, human-readable form.

As a MOOSE application, CRANE inherits the same classes and syntax as the MOOSE framework and can therefore be natively interfaced with a Zapdos plasma simulation. Thus CRANE is designed with two possible use cases:

  1. As a standalone software, it is designed to solve a general system of zero-dimensional reaction equations;
  2. Coupled into Zapdos, CRANE solves a system of one-dimensional, time-dependent reaction pathways as part of the fluid plasma solver.

The software is capable of utilizing rate coefficients that are either constant or energy-dependent, with rates derived from electron energy distribution function solvers, allowing the inclusion of experimental cross-sections (e.g. such as those from the LX-Cat database) and experimental rate coefficients.

General Features

CRANE is a general reaction network solver including support for both 0D and spatially-dependent simulations. As a MOOSE application, CRANE casts each term in the governing equations as classes which compute each individual residual and jacobian term to be solved with finite elements. These classes, named ``Kernels" in the framework, contain the governing equations required to solve a system of coupled first-, second-, and third-order reaction equations. Individual species are treated as nonlinear variables in the reaction network by being cast within the Variable class (or ScalarVariable class, if the system is zero-dimensional) and coupled into the relevant Kernels. Other MOOSE classes are utilized to attach rate coefficients to each reaction; spatially varying information is provided by the Materials class, and scalar information is stored within a UserObject class. Utilizing the MOOSE framework's Action system, reaction equations may be written directly into a MOOSE input file in a human-readable format, drastically simplifying the construction of a reaction problem in MOOSE. The Kernels, Materials, and UserObjects required by the MOOSE framework are automatically added to the solver by the Action. Use of the Action is optional; the input file may be built by adding each component of a reaction network manually.

Within the MOOSE framework are systems for both multidimensional problems and scalar problems. The Kernel system requires mesh information (generated by MOOSE directly or externally through a third-party software such as Gmsh) to be included in the governing equations for 1D-3D cases, while problems that are independent of the mesh may utilize the ScalarKernel system. Thus, CRANE is split into two separate use cases: 0D simulations are carried out utilizing the ScalarVariable and ScalarKernel systems (intended as a standalone chemistry solver), and 1D-3D simulations utilizing the Variable and Kernel systems with the intention of being coupled to other MOOSE applications. The Action developed in CRANE is capable of treating both sets of problems.

Acknowledgements

This project was developed as part of NSF Award 1740310, National Science Foundation.

Clone this wiki locally