RePlaChem: A dimensionality reduction library for plasma chemical mechanisms

In this work, we present \REDLIB, a software library for reducing detailed large-scale plasma chemical mechanisms to smaller skeletal ones. The library parses a plasma chemical mechanism in the well-established format compatible with the software ZDPlasKin, runs the reduction algorithm, and generates automatically a skeletal chemical mechanism. This feature allows the seamless implementation of the skeletal chemistry in well-known solvers including ZDPlasKin. In turn, the reduction of the chemistry accelerates otherwise computationally expensive numerical simulations. Furthermore, \REDLIB\ can be used as an analysis tool to shed light into the underlying reaction physics by identifying the dominant species and associated reactions. In order to validate and demonstrate its capabilities, \REDLIB\ is used to substantially reduce a large-scale methane-hydrogen plasma chemical mechanism with 77 species and 4404 reactions at various orders with a relatively small loss of accuracy. Program Title: \REDLIB CPC Library link to program files: (to be added by Technical Editor) Developer's repository link: \url{https://github.com/znikolaou/RePlaChem.git } Licensing provisions: GPLv3 License Programming language: Fortran 90 Nature of problem: Detailed plasma chemical mechanisms include numerous species and reactions which describe all possible interactions between electrons, molecules, radicals, and other species. The time-scales associated with these species typically range from ns to ms, and numerical simulations are often plagued by limiting time-steps required to resolve the fastest (smallest lifetime) species. For hundreds of species and thousands of reactions, it becomes almost impossible to identify the dominant species and corresponding reactions which really govern the chemical dynamics of the system. Reduction helps to identify the important species, and at the same time reduces the computational cost of plasma numerical simulations. In turn, this allows the wider adoption of numerical simulation as a tool for designing and developing novel plasma-based technologies in a wide range of industries. Solution method: Reduction is performed using Directed Relation Graphs with Error Propagation, a powerful, well-established, and computationally efficient method which has long been used in combustion research [1] . The user provides: (a) a detailed chemical mechanism in ZDPlasKin format, (b) the individual reaction rates obtained by solving a canonical problem for specific user conditions of interest, and (c) a control file with the reduction parameters of interest. Any solver or canonical problem can be used to obtain the reaction rates. With this information alone, \REDLIB\ performs the reduction, and generates a skeletal reaction set.\bkmquery{CE: Please check keywords are not available and we have captured from client pdf.}