SUMMARYThe use of the biharmonic operator for deforming a mesh in an arbitrary-Lagrangian-Eulerian simulation is investigated. The biharmonic operator has the advantage that two conditions can be speciÿed on each boundary of the mesh. This allows both the position and the normal mesh spacing along a boundary to be controlled, which is important for two-uid interfaces and periodic boundaries. At these boundaries, we can simultaneously ÿx the position of the boundary and ensure that the normal mesh spacing is continuous across the boundary. In addition, results for deforming surfaces show that greater surface deformation can be tolerated when using biharmonic equations compared to approaches using secondorder partial di erential equations. A ÿnal advantage is that with the biharmonic operator, the integrity of a grid in a moving boundary layer can be preserved as the boundary moves. The main disadvantage of the approach is its increased computational expense.

No abstract

Testing was conducted on a prototype automobile exhaust thermoelectric generator (AETEG) installed in a 1999 GMC Sierra pick-up truck. The system consisted of the generator, its power conditioning unit, and the interfaces to the test truck's engine coolant and exhaust systems. The objective of the test was to measure the AETEG's performance and its effect on the truck systems as well as to determine which factors are important for optimizing an AETEG design. Testing was performed in a dynamometer-equipped wind tunnel at Delphi Corporation's Harrison Thermal Systems Division in Lockport, New York. The first tests established the benchmark data set. Then the prototype AETEG was installed and three configurations of the system were tested in succession: the AETEG alone, the AETEG with portions of the exhaust pipes leading to it insulated, and the AETEG with insulated upstream exhaust pipes and with a pre-cooling heat exchanger operating to lower the inlet coolant temperature to the generator. Some of the important outcomes of the tests were: insulating the exhaust and lowering the coolant temperature had a significant positive effect on the power, parasitic losses resulting from the AETEG weight and the coolant pumping power were significant but manageable, and the increased exhaust flow resistance and the additional heat load from the AETEG were not significant effects.

No abstract

No abstract

No abstract

No abstract

Experimental and numerical results are presented on the process of horizontal ribbon growth (HRG) of single-crystal silicon. Experimental data on the leading edge position of the growth front as a function of pull speed is compared to model predictions with and without solidification kinetic effects. Without kinetics, the numerical results predict leading edge positions which are completely different than that observed in the experiment. With kinetics, the leading edge position is predicted typically within 1 mm and the change in position with pull speed also is well predicted. Conclusions from the kinetic model are that the growth occurs through a faceted process where the leading edge is a {111} facet that requires significant supercooling to maintain the growth. An outcome of the model is that the leading edge position versus pull speed response shows a turning point beyond which there are no steady growth solutions. This is consistent with all previously reported experiments on this process, which have reported maximum attainable pull-speeds. These results