Validation of flow simulation model using particle image velocimetry data and dimensionality reduction techniques

Shen, L

Oxford Thesis Collection

Thesis

Validation of flow simulation model using particle image velocimetry data and dimensionality reduction techniques

Abstract:: In-cylinder air motion has a profound influence on the mixture preparation and subsequent combustion in the internal combustion engines. The particle image velocimetry (PIV) technique provides a quantitative tool to measure the in-cylinder air velocity with high spatial and temporal resolution. Additionally, the in-cylinder flow velocity field can also be simulated by computational fluid dynamics (CFD) tools, of which the Reynolds averaged Navier-Stokes (RANS) method predicts the overall flow structures with acceptable accuracy and affordable computational cost for widespread industrial applications. In order to provide a more accurate and robust prediction, the simulation results need to be validated against the experimental measurements. In this thesis, the crank-angle resolved in-cylinder air flow fields were measured using the PIV technique in multiple planes of a near-production optical engine under different working conditions for a few hundreds of consecutive cycles. The measured flow fields show a high consistency over multiple measurement planes and across different test runs, and hence they can be used to validate the flow simulation model. Vector field comparison metrics were used to compare the measured and the simulated flow fields, and to identify the regions of interest and differences in flow characteristics. Due to the nature of averaging physical parameters in RANS, its validation against experimental results obtained by particle image velocimetry (PIV) requires consideration of how best to average or filter the measured turbulent flows. During the validation process, it was found that the use of the conventional PIV ensemble mean flow field may lead to biased conclusions in some cases. Therefore, two dimensionality reduction techniques, namely the proper orthogonal decomposition (POD) and the kernel principal component analysis (KPCA), were used to preserve the large-scale cycle-to-cycle variations while filtering out the small-scale turbulent fluctuations in experimental measurements, and hence enabled a fair comparison to the RANS simulations. Recommendations are made on how to provide an appropriate validation between measured data and simulation results in highly fluctuating flow fields such as the in-cylinder air motion.

Actions

Email

Email this record

Send the bibliographic details of this record to your email address.

Your Email
Please enter the email address that the record information will be sent to.

-
Your message (optional)
Please add any additional information to be included within the email.
Share
Cite

Cite this record

APA Style

Shen, L. (2021). Validation of flow simulation model using particle image velocimetry data and dimensionality reduction techniques [PhD thesis]. University of Oxford.

MLA Style

Shen, L. Validation of Flow Simulation Model Using Particle Image Velocimetry Data and Dimensionality Reduction Techniques. 2021. University of Oxford, PhD thesis.

Chicago Style

Shen, L. 2021. “Validation of Flow Simulation Model Using Particle Image Velocimetry Data and Dimensionality Reduction Techniques.” PhD thesis, University of Oxford.
Print