Skip to Main content Skip to Navigation

DNS for RANS combustion modelling in multiple injection configurations

Abstract : Compression-ignition engines are widelyused, mainly due to their high thermal efficiency andconsequent low CO2 emissions compared to sparkignitionengines. However, this technology has somedisadvantages related to the limited control over autoignitionof the air-fuel mixtures and heat release rate.Hence, in compression-ignition engines at their mostbasic form, the level of combustion noise and emissionsof nitrogen oxides and particulate matter canbecome critical. An effective strategy to tackle theseproblems is to decompose fuel injection into multipleinjection pulses permitting an optimal control of theair-fuel mixture formation and, thus, of the autoignitiondelay and the heat release rate. Multiple injectionstrategies become more and more popular dueto their advantages over conventional single injectioncycles. The physical phenomena involved in suchconfigurations, however, are complex and their modellingremains challenging, especially in the context ofindustrial 3D simulation using the Reynolds-averagedNavier-Stokes (RANS) formalism. The progress ofcompression-ignition applications depends to a greatextent on the capacity of the physical models to predictheat release rate and pollutant emissions. Thelack ofa experimental results at the scale of interestorientated this study towards the use of Direct NumericalSimulation (DNS) providing a model-free insightinto the interaction between turbulent mixing and combustionchemistry. In the framework of this Ph.D. thesis,a DNS database was generated and analysed,covering a range of single and split Diesel injectionrelevantconditions. Then, different turbulent combustionmodels based on tabulated chemistry were evaluatedby comparison with the DNS results. Followingthis analysis, a new modelling approach adapted tomultiple injection configurations was elaborated. Finally,a strategy for the application of the new modellingapproach in 3D RANS was proposed for prospectivemulti-injection compression-ignition engine simulationswith an improved accuracy.
Document type :
Complete list of metadatas

Cited literature [150 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Monday, February 25, 2019 - 12:01:07 PM
Last modification on : Friday, July 3, 2020 - 9:28:04 AM
Long-term archiving on: : Sunday, May 26, 2019 - 2:05:51 PM


Version validated by the jury (STAR)


  • HAL Id : tel-02047897, version 1


Eleftherios Gorgoraptis. DNS for RANS combustion modelling in multiple injection configurations. Other. Université Paris-Saclay, 2018. English. ⟨NNT : 2018SACLC111⟩. ⟨tel-02047897⟩



Record views


Files downloads