Computing CHEMKIN Sensitivities Using Complex Variables

Abstract

This paper discusses an accurate numerical approach based on complex variables for the computation of the Jacobian matrix of complex chemical reaction mechanisms. The Jacobian matrix is required in the calculation of low dimensional manifolds during kinetic chemical mechanism reduction. The approach is suitable for numerical computations of large-scale problems and is more accurate than the finite difference approach of computing Jacobians. The method is demonstrated via a nonlinear reaction mechanism for the synthesis of Bromide acid and a H2/Air mechanism using a modified CHEMKIN package. The Bromide mechanism consisted of five species participating in six elementary chemical reactions and the H2/Air mechanism consisted of 11 species and 23 reactions. In both cases it is shown that the method is superior to the finite difference approach of computing derivatives with an arbitrary computational step size h.

Complex variable methods for shape sensitivity of finite element models

“…It reemerged as a tool for engineering analysis with a paper by Squire and Trapp in 1998 [17]. Since then it has been used in a wide variety of engineering fields including computational fluid dynamics, dynamic system optimization and many others [18][19][20][21][22][23][24][25]. In all of these applications, CTSE has offered a great improvement in accuracy over finite differencing.…”

Mathematically Reduced Chemical Reaction Mechanism Using Neural Networks

“…In this paper we introduce a new meshfree method that uses a simple kernel method enhanced to a higher order of accuracy, Wand and Jones (1995). The simple kernel function approximator is combined with the complex-step method (CSM), of computing derivatives of complex functions, Butuk and Pemba, (2003). The main distinguishing feature of the method is the computation of the second-order derivatives in the TDSE, via a novel approach that is independent of grid spacing, i.e., it is truly meshfree.…”

Calculation of machine precision second order derivatives using dual-complex numbers

“…Both methods yield identical machine precision results, and both methods have been used to compute accurate 1 st order derivatives in a number of fields. Examples for the successful application of CTSE in a variety of disciplines include boundary element methods [10]; finite element methods in solid mechanics; [11][12][13], fluid dynamics [14][15][16], chemical reactions [17], structural dynamics [18], and fracture mechanics [19,20], among others. Examples of the application of multicomplex numbers for second order derivatives include: nonlinear finite elements [21], and structural dynamics [22,23].…”