In situ multi-tier auto-ignition detection applied to dual-fuel combustion simulations

Abstract

Here we use an anomaly detection methodology that is centered on analyzing fourth-order joint moments (co-kurtosis), particularly focusing on its application in auto-ignition of combustion problems with large numbers of species. Unsupervised anomaly detection is challenging to generalize across problem types and domains. A recent technique, centered on analyzing information in the fourth-order joint moment co-kurtosis, has shown promise, especially for high-dimensional scientific data. In this work we present developments to the co-kurtosis based anomaly detection method needed to make it effective and scalable for large-scale distributed scientific data, such as those generated by massively parallel simulations. An in situ co-kurtosis algorithm is employed as the anomaly detection method for identifying ignition kernels in simulations of turbulent combustion. We extend an existing methodology which identifies regions of the domain where anomalies are present, and add another tier of anomaly detection where the individual samples contributing to the anomaly are identified. We apply this algorithm on-the-fly to a variety of turbulent reacting flow problems and compare it to the widely used (but significantly more expensive) chemical explosive mode analysis (CEMA). We demonstrate the ability of the method to detect and identify the onset of low and high temperature ignition which can be used for computational steering, as chemical and combustion anomalies occur intermittently at spatio-temporal locations unknown a priori. Finally, we apply our lightweight in situ algorithm to an exascale high-fidelity simulation with a total of 2.4 Trillion degrees of freedom, performed using an adaptive mesh refinement solver. Furthermore, through a scalability analysis, we show that the relative computational cost of this in-situ anomaly detection algorithm compared to an iteration of the reacting flow solver is negligible.

Novelty and Significance Statement

Techniques to identify occurrence of auto-ignition have hitherto relied on the specifics of the chemical kinetics, and criteria have been based on ad-hoc thresholds. The novelty of this work lies in adopting a purely statistical viewpoint of auto-ignition, as one signified by higher-order joint moments, instead of a chemical kinetic viewpoint which does not generalize from one fuel to another or under different conditions. While previous work has demonstrated the accuracy and generalizability of the higher-order joint moments (co-kurtosis) approach in identifying regions of auto-ignition, this work builds on the concept and presents additional tiers of identifying auto-ignition, specifically individual samples within regions. Such multi-tiered detection is demonstrated to be accurate in simulations with spatio-temporally complex ignition behavior. Moreover, the technique is shown to be significantly more computationally efficient and scalable when deployed in situ in exascale class simulations.