Deep kernel learning approach to engine emissions modeling

Abstract

We apply deep kernel learning (DKL), which can be viewed as a combination of a Gaussian process (GP) and a deep neural network (DNN), to compression ignition engine emissions and compare its performance to a selection of other surrogate models on the same dataset. Surrogate models are a class of computationally cheaper alternatives to physics-based models. High-dimensional model representation (HDMR) is also briefly discussed and acts as a benchmark model for comparison. We apply the considered methods to a dataset, which was obtained from a compression ignition engine and includes as outputs soot and NO x emissions as functions of 14 engine operating condition variables. We combine a quasi-random global search with a conventional grid-optimization method in order to identify suitable values for several DKL hyperparameters, which include network architecture, kernel, and learning parameters. The performance of DKL, HDMR, plain GPs, and plain DNNs is compared in terms of the root mean squared error (RMSE) of the predictions as well as computational expense of training and evaluation. It is shown that DKL performs best in terms of RMSE in the predictions whilst maintaining the computational cost at a reasonable level, and DKL predictions are in good agreement with the experimental emissions data.

Perspectives for regulating 10 nm particle number emissions based on novel measurement methodologies

“…Finally, the behaviour is captured via the MGA workflow consisting of kinetics & SRM Engine Suite [Etheridge et al 2011] for physicochemical modelling and MoDS [Yu et al 2020] for parameter estimation and surrogate model generation. The workflow is validated against measurements for particle size distributions at multiple load-speed engine operating conditions.…”

Automated Calibration of a Poly(oxymethylene) Dimethyl Ether Oxidation Mechanism Using the Knowledge Graph Technology

“…It is structured as an instantiation of the agent template proposed by Mosbach et al A unified modeling language (UML) activity diagram of the agent is provided in Figure , with the agent template surrounding the model development suite (MoDS) software package. MoDS is an integration of multiple tools developed for various generic model development tasks, such as parameter estimation, surrogate model creation, , and experimental design …”

Universal Digital Twin - A Dynamic Knowledge Graph

“…The calculations performed by agents can take any form, including physics-based models with a theoretical structure, gray box models that combine some theoretical structure with data-driven components, and pure data-driven models. See Yu et al (2020) for examples of these approaches in the context of emissions modeling. The availability of semantically structured machine-queryable data makes the World Avatar well-suited to be coupled to machine learning applications, where it is wellknown that the time spent on data preparation and curation typically adds a significant cost to machine learning.…”