Marco Manta
Numerical Simulation of the Knocking Phenomena in a High-Performance Spark Ignition Engine.
Rel. Federico Millo, Luciano Rolando, Andrea Piano. Politecnico di Torino, Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo), 2022
Abstract: |
Nowadays the combined use of downsizing and turbocharging represents one of the most valuable solutions to reduce the CO2 emission of spark ignition (SI) engines and to improve their specific power. Nevertheless, the need for high boosting levels can lead to a dramatic increase in the knock likelihood. In such a framework numerical simulations can be used to properly select the most promising technologies for the minimization of the knock risk. Such an approach requires the availability of reliable knock models capable of identifying the knock-limited spark advance (KLSA) depending on the engine operating conditions and configurations. Therefore the main target of this thesis project is the development of a 3D-CFD model in the Converge environment capable of properly reproducing the occurrence, location and onset of the knocking phenomena in a high-performance turbocharged direct injection spark ignition (DISI) engine. To achieve this final target the thesis project has been divided into three main working activities. First, a comprehensive statistical analysis was carried out considering experimental data recorded from 6 test rig procedures at constant relative air-to-fuel ratio (RAFR) λ, each of them providing 250 in-cylinder pressure cycles. A proper crank-angle window and a band-pass filter were applied to each experimental pressure trace to isolate knock frequencies from background noise. After the choice of MAPO and ID as indexes for the knock assessment, the analysis of knock intensity average values, the identification of knocking cycles and the analysis of knock intensity frequency distribution, from both average MAPO and ID versus spark advance (SA) plots the KLSA has been experimentally identified as the SA corresponding to the knee of the curve. The second part of the work contains the development in terms of setup before for the intermediate cold or motored cycle simulation and then for the firing cycle simulation in Converge 3.0 of a 3D-CFD model able to reproduce the available experimental data. The main steps include the model calibration for the motored cycle using initial and boundary conditions provided by an already validated GT-Power 1D-CFD model and the validation of the simulated firing cycle against two sets of data: the first one contains the thermodynamic and mass flow rate data obtained from the same GT-Power model whereas the second one includes the in-cylinder pressure stress analysis performed in the test rig. In particular, the validation considered as sensitivity analyses different working conditions featuring a spark advance sweep to assess the robustness of the model. Lastly, an already developed methodology for the assessment of knock onset risk based on formaldehyde CH2O mass fraction as a tracer for the end-gases pre-flame chemical reactions progress within the combustion chamber has been followed. According to this methodology, only a local knock risk index represented by the unburned mass fraction reaching a critical CH2O concentration allows to assess the knock onset risk and even with a RANS turbulence model (without performing a CCV analysis). This means a key advantage in terms of saving both computational power and simulation time compared to a LES turbulence model. The 3D-CFD model calibrated in the knock-free working point at the most retarded ST at constant λ was able to reproduce the knock occurrence and onset resulting from the statistical analysis at KLSA and at most advanced ST providing also its spatial location inside the combustion. |
---|---|
Relators: | Federico Millo, Luciano Rolando, Andrea Piano |
Academic year: | 2021/22 |
Publication type: | Electronic |
Number of Pages: | 83 |
Additional Information: | Tesi secretata. Fulltext non presente |
Subjects: | |
Corso di laurea: | Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo) |
Classe di laurea: | New organization > Master science > LM-33 - MECHANICAL ENGINEERING |
Aziende collaboratrici: | Ferrari Spa |
URI: | http://webthesis.biblio.polito.it/id/eprint/22028 |
Modify record (reserved for operators) |