Abstract:Environmental compatibility requires low-emission burners for gas turbine power plants as well as for jet engines. In the Past, significant progress has been made developing low NOx and CO burners. Unfortunately, these burners often have a more pronounced tendency than conventional burner designs to produce combustion driven oscillations. The oscillations may be excited to such an extent that pronounced pulsation may possibly occur; this is associated with a risk of engine failure. The stability of a burner system can be investigated by means of a stability analysis under the assumption of acoustical behavior. The problem with all these algorithms is the transfer function of the flame. A new method is presented here to predict the dynamic flame behavior by means of a full Navier-Stokes simulation of the complex combustion process. The first step is to get a steady-state solution of a flame configuration. After that a transient simulation follows with a sudden change in the mass flow rate at the flame inlet. The time-dependent answer of the flame to this disturbance is then transformed into the frequency space by a Laplace Transformation. This leads, in turn, to the frequency response representing the dynamic behavior of the flame. In principle, this method can be adapted for both diffusion as well as premixed flame systems. However, due to the fact that diffusion flames are more controlled by the mixing process than by the chemical kinetic, the method has first been used for the prediction of the dynamic behavior of turbulent diffusion flames. The combustion has been modelled by a mixed-is-burnt model. The influence of the turbulence has been taken into account by a modified k-ε model and the turbulence influences the combustion rate by presumed probability density functions (pdf). The steady state as well as the transient results have been compared with experimental data for two different diffusion flame configurations. Although the burner configuration is relatively complex, the steady-state results collaborate very well with the experiments for velocity, temperature, and species distribution. The most important result is that the heat release that drives the oscillations can be modeled sufficiently accurately. The effect of using different pdf models has been discussed and the best model has been used for the transient calculations of the dynamic flame behavior. The results for the frequency response of the flame are very encouraging. The principal behavior of the flame—higher order time element with a delay time—can be predicted with sufficient precision. In addition, the qualitative results collaborate fairly well with the experiments.

Linear and nonlinear flame response prediction of turbulent flames using neural network models

Prediction of Efficient Operating Conditions Inside a Heavy-Duty Natural Gas Spark Ignition Engine Using Artificial Neural Networks

A Robust Autoregressive Long-Term Spatiotemporal Forecasting Framework for Surrogate-Based Turbulent Combustion Modeling Via Deep Learning

A Dual-Path neural network model to construct the flame nonlinear thermoacoustic response in the time domain

Physics-aware neural network flame closure for combustion instability modeling in a single-injector engine

Neural Network Flame Closure for a Turbulent Combustor with Unsteady Pressure

Modelling turbulent premixed flames using convolutional neural networks: application to sub-grid scale variance and filtered reaction rate

Predictive Models for Flame Evolution Using Machine Learning: A Priori Assessment in Turbulent Flames Without and with Mean Shear

Assessment of Machine Learning Techniques for Simulating Reacting Flow: From Plasma-Assisted Ignition to Turbulent Flame Propagation

Reconstructing temperature fields from OH distribution and soot volume fraction in turbulent flames using an artificial neural network

Modelling the response of a turbulent jet flame to acoustic forcing in a linearized framework using an active flame approach

Modeling of turbulent flames with the large eddy simulation–probability density function (LES–PDF) approach, stochastic fields, and artificial neural networks

Network- and CFD/CAA-modelling of the high frequency flame response in multi-jet combustors

Real-time parameter inference in reduced-order flame models with heteroscedastic Bayesian neural network ensembles

Development of Machine Learning Models for Studying the Premixed Turbulent Combustion of Gas-To-Liquids (GTL) Fuel Blends

Numerical Predication of the Dynamic Behavior of Turbulent Diffusion Flames

An Improved NO Prediction Model for Large Eddy Simulation of Turbulent Combustion

Neural Differential Equations for Inverse Modeling in Model Combustors

Using an Artificial Neural Network to Predict Flame Spread Across Electrical Wires

Machine Learning of Combustion LES Models from Reacting Direct Numerical Simulation

Data-Driven Fire Modeling: Learning First Arrival Times and Model Parameters with Neural Networks