No abstract

No abstract

No abstract

In this work, a data-driven methodology for modeling combustion kinetics, Learned Intelligent Tabulation (LIT), is presented. LIT aims to accelerate the tabulation of combustion mechanisms via machine learning algorithms such as Deep Neural Networks (DNNs). The high-dimensional composition space is sampled from high-fidelity simulations covering a wide range of initial conditions to train these DNNs. The input data are clustered into subspaces, while each subspace is trained with a DNN regression model targeted to a particular part of the high-dimensional composition space. This localized approach has proven to be more tractable than having a global ANN regression model, which fails to generalize across various composition spaces. The clustering is performed using an unsupervised method, Self-Organizing Map (SOM), which automatically subdivides the space. A dense network comprised of fully connected layers is considered for the regression model, while the network hyper parameters are optimized using Bayesian optimization. A nonlinear transformation of the parameters is used to improve sensitivity to minor species and enhance the prediction of ignition delay. The LIT method is employed to model the chemistry kinetics of zero-dimensional H2–O2 and CH4-air combustion. The data-driven method achieves good agreement with the benchmark method while being cheaper in terms of computational cost. LIT is naturally extensible to different combustion models such as flamelet and PDF transport models.

Concerning the geometrical effect of inner cylindrical hot pins, the natural convective heat transfer of nanofluid in a homogenous porous medium in a squared enclosure is numerically studied, using lattice Boltzmann method (LBM). In order to investigate the arrangement of inner cylinders for better heat transfer performance, five different configurations (including one, three, and four pins) were compared, while the total heat transfer area of inner pins were held fixed. Squared cavity walls and inner cylinder's surfaces were constantly held at cold and warm temperatures, respectively. In our simulation, Brinkman and Forchheimer-extended Darcy models were utilized for isothermal incompressible flow in porous media. The flow and temperature fields were simulated using coupled flow and temperature distribution functions. The effect of porous media was added as a source term in flow distribution functions. The results were validated using previous creditable data, showing relatively good agreements. After brief study of copper nano-particles volume fraction effects, five cases of interest were compared for different values of porosity and Rayleigh number by means of averaged Nusselt number of hot and cold walls; and also local Nusselt number of enclosure walls. Comparison of different cases shows the geometrical dependence of overall heat transfer performance via the average Nusselt number of hot pins strongly depending on their position. The four pin case with diamond arrangement shows the best performance in the light of enclosure walls' average Nusselt number (heat transfer to cold walls). However, the case with three pins and downward triangular arrangement surprisingly gives promising heat transfer performance. In addition, the results show that natural convective heat transfer and flow field is intensified with increasing Rayleigh number, Darcy number, and porosity.