Towards design optimization of high-pressure gasoline injectors using Genetic Algorithm coupled with Computational Fluid Dynamics (CFD)
| Autor: | Dominique Thévenin, László Daróczy, Karl Georg Stapf, Paul Jochmann, Robin Hellmann, Erik Schuenemann |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2017 |
| Předmět: |
Spray characteristics
Optimization Genetic Algorithm Computer science business.industry Nozzle Flow (psychology) Injector Shadowgraphy Computational fluid dynamics Metamodeling Multi-objective optimization Automotive engineering law.invention High Pressure Injector law Genetic algorithm business CFD Nozzle flow |
| Zdroj: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia instname |
| Popis: | [EN] The spray pattern of high-pressure multi-hole injectors as well as the atomization process are of uttermost importance regarding efficiency and emissions in gasoline combustion engines. Ensuring optimal homogenization while meeting the engine individual specifications regarding spray targeting and massflow is a crucial development goal. High effort is put on the layout of the nozzle seat to meet the engine requirements. Success is only possible with a deep knowledge of the influencing quantities, considering that many design parameters affect the inner nozzle flow. Based on this knowledge improvement in spray penetration length and atomization can be achieved. In the current investigation a segment model of the injector is considered. A fully automated, highly parallelized workflow enables a systematic examination of the constrained design space with acceptable computational time. The CFD workflow is implemented in the OPtimization Algorithm Library++ (OPAL++) developed at the “Otto von Guericke” University of Magdeburg. First, inner nozzle flow 3D-CFD calculations of two selected nozzle geometries are validated by comparison with shadowgraphy and Long-Distance-Microscope (LDM) measurements. Using these simulations, correlations between nozzle flow parameters and the key spray characteristics, serving as optimization objectives, are analyzed. Second, a Design-of-Experiment (DoE) is created to understand the interdependency between design variables and objectives. Based on the DoE, metamodels are constructed, validated, compared with each other and used for optimization. Afterwards, a direct 3D CFD-optimization is carried out for the nozzle geometry. It relies on a Genetic Algorithm in OPAL++ to identify the Pareto front of the multi-objective problem. Finally, the Pareto front is analyzed and conclusions are drawn for future research. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|