Vorträge und Posterpräsentationen (mit Tagungsband-Eintrag):
P. Demircioglu, I. Bogrekci, Y. Cerci, C. Deniz, G. Bas, M.N. Durakbasa:
"Assessment of cylinder liners using computer vision andnumerical thermal analyses";
Vortrag: ISMQC - International Symposium on Measurement and Quality Control,
Krakow-Kielce, Polen;
11.09.2013
- 13.09.2013; in: "Proceedings of he 11th International Symposium on Measurement and Quality Control",
Osaka University,
(2013).
Kurzfassung englisch:
Honing is required when the surface structure parameters
of cylinder liners are varied critically. Wear and tear signs
occur when cylinder liners of both 4-stroke, and 2-stroke
engines are used extensively. Wear and tear is not evenly
distributed. Most wears occur between the reversal points of
the upper and lower piston rings. Scuffing marks are
generally found where the upper piston ring changes its
direction. When honing time approaches for most cylinder
liners, oval shape of cylinder liners appears. In this study,
images of cylinder liners with different surface topography
were taken using 3D digital microscope with high resolution
and different magnification factors. The sampled images
were pre-processed using high pass filters in order to
enhance images to catch high frequency components. Edges
and patterns of enhanced images then were analyzed using
X-Y Sobel algorithm. Then the results obtained from these
image analyses studies were compared to have a better
understanding about the relationship between temperature
distribution contours and patterns on cylinder liners. The
solid models of cylinder liners were drawn using Autodesk
Inventor and imported to ANSYS Design modeler. The
mesh model of the cylinder liner was achieved using
CFX-meshing. With approximated boundary conditions,
heat fluxes and temperature distribution were computed
using numerical thermal analysis in ANSYS. The thermal
analyses results indicated that the temperature of the inner
surface of the cylinder liners increased about two times
when the heat flux rates ranged from 100 to 500 kW/m².
Edge and pattern determination results showed that scuffing
marks and scratches and their morphological properties from
the images of the cylinder liners could be identified using
image processing and analyses techniques developed in this
study.
Schlagworte:
Computer vision, Manufacturing techniques, Roughness measurement, Wavelet transform
Erstellt aus der Publikationsdatenbank der Technischen Universität Wien.