OptiSMOKE++: A toolbox for optimization of chemical kinetic mechanisms

Abstract

As detailed chemical mechanisms are becoming viable for large scale simulations, knowledge and control of the uncertainty correlated to the kinetic parameters are becoming crucial to ensure accurate numerical predictions. A flexible toolbox for the optimization of chemical kinetics has therefore been developed in this work. The toolbox is able to use different optimization methodologies, as well as it can handle a large amount of uncertain parameters simultaneously. It can also handle experimental targets from different sources: Batch reactors, Plug Flow Reactors, Perfectly Stirred Reactors, Rapid Compression Machines and Laminar Flame Speeds. This work presents the different features of this toolbox together with five different test cases which exemplifies these features.

Combustion, Chemistry, and Carbon Neutrality

“…Acknowledging that both experiment and model may carry large errors in such situations, it may be more valuable to compare trends rather than to infer a perceived “best” model from inspection of individual species profiles. − Mathematical formalisms have been developed to overcome limitations of manual analysis and facilitate comparison of several data sets with a number of available models, regarding position, shape, and maximum values. , Specific frameworks and procedures concern systematic repositories and handling and sharing data from different sources over decades of research. , With respect to assessing model uncertainties, a main problem is that individual uncertainty ranges of the respective parameters might not reflect the overall uncertainty because these are not independent of one another . Systematic uncertainty analysis and constraining of models using mathematical procedures has become an active field of research. ,− Frameworks and toolboxes for mechanism optimization are being developed that can be applied for widely used types of reactors and other combustion-relevant configurations (see Section ). …”

“…Systematic uncertainty analysis and constraining of models using mathematical procedures has become an active field of research. ,− Frameworks and toolboxes for mechanism optimization are being developed that can be applied for widely used types of reactors and other combustion-relevant configurations (see Section ). …”

“…110,291−299 Frameworks and toolboxes for mechanism optimization are being developed that can be applied for widely used types of reactors and other combustion-relevant configurations (see Section 2.1). 299 Uncertainty analyses are reported for fuels and compounds with current and future energy system relevance, including H 2 , 291,293,300 H 2 -enriched natural gas, 301 NH 3 /H 2 blends, 302 biogas, 303 and syngas. 291,304 Useful insight may be gained into overall model uncertainties, recognizing interdependent parameters in realistic systems with numerous species.…”

Laminar flame speed evaluation for CH4/O2 mixtures at high pressure and temperature for rocket engine applications

“…However, complementary sensitivity analyses conducted at higher pressures for 𝐶𝐻 4 /𝑂 2 tend to show that its 𝑆 increases with 𝑃, which makes the additional role of high pressure diluted experimental measurements where this reaction is also sensitive more important for a future optimization process. This work is currently in progress, using the OptiSMOKE++ [19] code to tune the Arrhenius parameters of several sensitive reactions identified here.…”

Top‐down kinetic modeling from power law to elementary steps using KASTER: A methane steam reforming case study

“…The area of software development for kinetic model construction is continuously evolving, aiming at improving capabilities with respect to the simulation of reaction kinetics, reaction network generation and the determination of parameters via regression and/or first principles calculations. [1][2][3][4][5] While microkinetic models are conceptually designed to be valid over a wide range of operating conditions, they also require a higher number of kinetic parameters that need to be determined. This goes hand in hand with increased model complexity and computational cost.…”