Towards using general purpose graphics processing unit (GPGPU) units for accelerating the batched perfectly stirred reactor (PSR) calculations

TL;DR

This paper presents a GPU-based parallel implementation for solving the complex, nonlinear equations of perfectly stirred reactors (PSRs), significantly accelerating combustion modeling and pollutant prediction tasks.

Contribution

It introduces a highly parallelized GPU algorithm for batched PSR calculations using a robust nonlinear solver, enhancing computational efficiency in combustion simulations.

Findings

GPU implementation significantly reduces computation time

Parallel processing enables handling complex chemical kinetics

Improves accuracy and robustness of PSR simulations

Abstract

Detailed analysis of efficiency and pollutant emission characteristics of practical turbulent combustion devices using complex combustion kinetics often depend on the interactions between the combustion chemistry involving both gasses species and soot, and turbulent flow characteristics. Modeling of such combustion system often requires the use of chemical kinetic mechanisms with hundreds of species and thousands of reactions. Perfectly stirred reactors (PSR) are idealized reactor environments, where the reacting species have high rate of stirring, and the combustion products are uniformly distributed inside the reactor. PSRs have been found very useful in the study of flame stabilization, prediction of pollutants such as NOx formation, development and testing chemical reaction mechanisms, and investigation of soot formation and growth. The fundamental equations describing a PSR constitute systems of highly nonlinear algebraic equations, due to the complex relationship between the net production rate of the species and the species concentration, which ultimately makes the equations stiff, and the solution of such equations become highly compute-intensive leading to the need for a efficient and robust solution algorithms. Graphics processing units (GPUs) have widely been used in the past as a cost-effective alternate to central processing units (CPUs), and highly parallel threads of GPUs can be used in a efficient manner to improve the performance of such algorithms for speeding up the calculations. A highly parallelized GPU implementation is presented for a batched calculation of PSR model, using a robust and efficient non-linear solver for gas phase chemical reactions and is further coupled to one of the widely used moment