A novel concept which couples ID and 2D CFD models in a simulation of unsteady lC
engine flows was investigated, and such a coupled model was developed.
Two unified solution procedures which are capable of predicting mixed compressible and
incompressible flow fields found in an engine were developed and comparatively studied.
One is the pressure correction algorithm, the other is the block implicit algorithm. They
provided platforms for the implementation of coupled models. Second order spatial and
Euler backward time differencing schemes were adopted. The comprehensive comparative
studies were performed on a variety of benchmark flows ranging from steady to unsteady,
incompressible to compressible. The data documented have shown that the prediction
qualities of the two algorithms were comparable in all calculations. The block implicit
procedure required more storage memory generally but it converged faster in all cases
except the incompressible flow calculations.
General strategies to couple the ID CFD model with the 2D CFD model in one calculation
were proposed. They were successfully incorporated in both of the unified solution
procedures. The predictions from these coupled models for a series of unsteady
benchmark flows were competitive in quality with those from single 2D CFD models,
however, the computing costs involved were comparatively much lower. In these
calculations, the coupled models integrated in the block implicit procedure produced
faster convergence than those in the pressure correction procedure, but required more
computing resource. In addition, the implicit coupling stragety was more efficient
compared to the explicit counterpart.
A ID and 2D coupled model integrated in the pressure correction procedure was applied
to simulate a realistic cylinder-valve-pipe flow. The overal prediction quality is
satisfactory compared with experimental measurements. Some discrepancies which
occurred were largely attributed to numerical representations of valve mechanism and the
lack of turbulence models. For this engine application, the coupled model has shown
advantages in computing cost or straightforwardness over a conventional uniform 2D
model or boundary condition model.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University