Session: 10-03-02: CFD Applications - II
Paper Number: 95465
95465 - Pareto-Based Optimization of a Gas Turbine Combustor Design
The presence of combustor turbulence in a gas turbine engine has a significant impact on the stage efficiency of the downstream turbine. Here, the turbulence field at the combustor-turbine interface is optimized through modifications to the combustor geometry. Care is taken to preserve the combustor function through the evaluation of a mixing parameter. We hypothesize that the size, orientation, and positioning of the combustor jets can be modified to promote small-scale turbulence, remove the presence of Reynolds stress, and reduce anisotropy, all of which decrease boundary layer losses in the turbine as a result of combustor turbulence. Improving turbine performance by optimizing the combustor turbulence is a novel concept in the field and has the potential to shift the paradigm of combustor design. We perform a baseline characterization of combustor turbulence against measurements conducted with a combustor simulator. Mixing is quantified through the introduction of a passive scalar to the model. The Improved Delayed Detached Eddy Simulation approach demonstrates a favorable correlation with the experimental results, with the maximum combustor exit plane normalized velocity within 3.6% of that obtained experimentally and flow angles within 1°. To enable rapid optimization of the geometry, the model is simplified by coarsening the mesh and increasing the timestep. An initial sensitivity analysis is conducted, varying the aspect ratio, size, spacing, and protrusion depth of the combustor jets in the primary and secondary zones of the combustor. We determine that, of these factors, only the jet spacing and protrusion depth have a statistically significant impact on the mixing uniformity and turbulence intensity at the combustor-turbine interface. A pareto-based approach using a multi-objective genetic algorithm is used to optimize the jet spacing, size, protrusion depth, and combustor length to produce minimum turbulence intensity at the combustor-turbine interface and maximize mixing uniformity. The factors that promote decreased boundary layer losses in the turbine are likely to be in direct conflict with those that promote engine performance. As an example, higher pressure drop across the combustor walls will increase loss and therefore fuel consumption; however, if the associated higher velocities and smaller jets lead to such an improvement in turbulent boundary layer losses that the increase in combustor loss is outweighed, then such a change may be desirable. We conclude with some recommendations for combustor geometry modifications that can lead to improved turbine performance. Finally, we provide an outlook for future gas turbine combustor optimization efforts, including in the context of gradient-based approaches.
Presenting Author: Jennifer Miklaszewski University of Colorado Boulder
Presenting Author Biography: Jennifer Miklaszewski is a second-year Ph.D. student in Mechanical Engineering at the University of Colorado Boulder. She received both bachelor’s and master’s degrees in Mechanical Engineering from Purdue University. Her research interests include turbulent flows and optimization with application to gas turbine combustor design.
Authors:
Jennifer Miklaszewski University of Colorado BoulderMasha Folk Rolls-Royce Corporation
Peter Hamlington University of Colorado Boulder
Pareto-Based Optimization of a Gas Turbine Combustor Design
Paper Type
Technical Presentation