Large-scale vortex structures and their effects on the dispersion of particles in turbulent free shear flows are very important in many industrial applications, such as combustion, pollution control, and materials processing. In order to understand large-scale vortex structures and particle dispersion in depth, as well as their interaction effects, a two-way-coupled three-dimensional mixing layer laden with particles at a Stokes number of 5 initially located in the upper half region is studied numerically. A pseudospectral method was used to directly simulate the flow fluid, and the Lagrangian approach was used to trace particles. The concept of computational particles is introduced to vary the mass loading of particles. The momentum coupling effect introduced by a particle approximates to a point force. The simulation results show that coherent structures are still dominant in the mixing layer, but the flow dynamics and particle dispersion are modulated. The length of large-scale vortex structures is shortened and the pairing is delayed. Higher mass loading results in lower energy of the fluid in the phase of Kelvin-Helmholtz rolling up, while in the pairing process of large-scale vortex structures, the energy of the fluid increases as the mass loading increases. Higher mass loading also leads to larger mixed fluid thickness and Reynolds stresses of the flow. In addition, the particle dispersion along the transverse direction differs from that along the spanwise direction, which indicates that the effects of the addition of a particle on the spanwise large-scale vortex structures are different from those on the streamwise large-scale vortex structures.

In this work, an effective method based on artificial neural network (ANN) and genetic algorithms (GA) is suggested for modeling the carbon burnout behavior in a tangentially fired utility boiler and optimizing the operating conditions to achieve the highest boiler heat efficiency consecutively. When carbon burnout behavior under various operating conditions are experimentally investigated, the comparison between the output of ANN modeling and the experimental data shows satisfactory agreement. A genetic algorithm is employed to perform a search to determine the optimum solution of the neural network model, identifying appropriate setpoints for the current operating conditions.

The present work introduces an approach to predict the nitrogen oxides (NOx) emission characteristics of a large capacity pulverizedcoal fired boiler with artificial neural networks (ANN). The NOx emission andcarbon burnout characteristics were investigatedthroug h parametric fieldexperi ments. The effects of over-fire-air (OFA) flow rates, coal properties, boiler load, air distribution scheme and nozzle tilt were studied. On the basis of the experimental results, an ANN was used to model the NOx emission characteristics andthe carbon burnout characteristics. Comparedwi th the other modeling techniques, such as computational fluid dynamics (CFD) approach, the ANN approach is more convenient and direct, and can achieve good prediction effects under various operating conditions. A modified genetic algorithm (GA) using the micro-GA technique was employedto perform a search to determine the optimum solution of the ANN model, determining the optimal setpoints for the current operating conditions, which can suggest operators' correct actions to decrease NOx emission.

A new approach has been developed for the study of dense gas-solid flows. The approach uses Eulerian and Lagrangian coordinates and takes account of inter-particle interactions. A numerical algorithm is presented and a dense, circular, particle-laden jet has been simulated numerically using the new model. Comparison of the calculations with published experimental data shows satisfactory agreement.

A two-dimensional particle population balance model has been developed for the particle size and density distributions in a circulating fluidized bed (CFB) boiler furnace based on analysis of particle properties and the core-annulus hydrodynamic model. The model, incorporating modules to consider fuel particle fragmentation, char combustion, particle attrition and gas-solid separation, is part of an overall model developed earlier by the authors to simulate the operation of a 12MW CFB boiler. The model predictions for particle population in a CFB furnace agree well with the measurement data.