Diesel Engine Tribology

Recent years have seen an increase in the wear rate of engine bearings, subsequently followed by bearing failure, for the large two-stroke diesel engines used for ship propulsion. Here, the engine bearings include main, big end and crosshead bearings, with the bearing type used being the journal bearing, belonging to the class of ‘hydrodynamic bearings’.

This implies that the load carrying capacity is generated by a relative movement of the involved components, i.e. avelocity-driven operation. For the engine application, the velocity stems from the engine RPM. However, to comply with the latest emission requirements as well as attempting to minimise fuel expenses, the engine speed has been lowered together with an increase in the engine mean pressure which in terms lead to larger bearing loads.

With worsened operating conditions from two sides, the encountered problems are understandable as the design criteria for the bearings are no longer valid, albeit still not desirable. To come up with a solution, the operating conditions of the bearings have to be understood. The main challenge is to supply sufficient with lubricant to avoid metal-metal contact under time-varying combustion load.

This project has therefore revolved around the investigation of the tribological performance of the dynamically loaded journal bearing, both theoretically and experimentally.The theoretical work covers two approaches to the modelling of the bearing;a traditional finite element based solver for Reynolds equation, and a moregeneral finite volume discretisation of the Navier-Stokes equations.

In this way the influence from assumptions usually made in regards to supply grooves canbe verified. A test rig has been constructed for replicating engine-like conditions. Anuni-directional load can be applied in both static and dynamic modes, while another key feature being that of a transparent polymer bearing enabling the study of film rupture and re-forming.Paper [P1] describes the development of a suitable finite volume mesh for dynamic loading, while Paper [P2] contains the perturbation implementation used for the dynamic loading.

Resorting to Gümbel boundary conditions, very similar orbits are predicted for a given bearing using the two methodsas demonstrated in Paper [P3]. Good agreement is also obtained between thenumerical and experimental results.Finally, some suggestions to improvements of the modelling as well as the experimental set-up, are made.