Wide-Field Fluorescence Measurement of Oil Film Distribution on a Two-Stroke Marine Engine Cylinder Liner: Methodology and Injection Parameter Effects

Authors:
Preprint
Publisher website
Google Scholar
Abstract
Low-speed two-stroke marine engines inject lubricating oil directly onto the cylinder liner wall, forming the initial film that piston rings redistribute. Full-field measurements on large curved liners remain scarce, so lubrication simulations prescribe rather than measure this boundary condition. This study establishes a wide-field fluorescence method on a full-scale transparent cylinder liner, with an automated pipeline performing background subtraction, quasi-steady interval selection, cylindrical unwrapping and fixed spatial sensitivity correction. A beta-independent upper-envelope procedure selects calibration candidates solely by spatial containment and temporal stability, criteria entirely independent of the calibrated quantity, avoiding circularity and quantifying between-condition dispersion inaccessible to single-anchor approaches. Four complementary morphology metrics are extracted, each motivated by a distinct cylinder-liner lubrication failure mode: effective area, effective mean film thickness, circumferential coverage and uniformity index. Across 45 parameter combinations with three repeats each, elevation angle dominates all four responses, with partial η² = 0.57–0.64, p < 0.001; larger elevation angles reduce wetted area and circumferential spread. Spray cone angle strongly affects circumferential coverage, with partial η² = 0.42, and significant diameter-by-cone and cone-by-elevation interactions show parameter effects are not reducible to independent one-factor trends. Quadratic response surface models achieve leave-one-condition-out Q² from 0.42 for uniformity to 0.71 for circumferential coverage, confirming sufficient regularity for domain-limited trend interpolation. The reconstructed fields provide experimentally informed, spatially non-uniform post-injection oil film distributions as initial boundary-condition data in lubrication simulations, with a partial expanded uncertainty of approximately 21%, k ≈ 2, for a transparent liner, a single injector and room-temperature operation.

This publication has 20 references indexed in Scilit: