CFD Methodology Assessment for the Investigation of Convective Heat Transfer Properties of Engine Coolants under Boling Conditions

The paper presents a combined experimental and numerical program directed at defining a cost/effective methodology for conjugate heat transfer CFD simulations of engine water cooling jackets. As a first step in the process, deficiencies in current numerical strategies for the analysis of conjugate heat transfer problems under typical engine operating conditions are exposed and commented. Results are shown form a wide validation program based on the comparison between experimental measurements from a test facility at Villanova University and CFD predictions at the University of Modena. On the experimental side, the test apparatus consists of a test section, pump, accumulator tank, rejection heat exchanger and required pumping. The test section is provided with a constant volumetric flow rate, and consists of a cylindrical aluminum body with a drilled horizontal flow channel. The section is heated by ten cartridge heaters located at a constant radial distance from the cylinder axis. The test section is connected to the flow loop by means of two calming sections, respectively at the cylinder inlet and exit. Twenty thermocouples are used to measure the test section local temperature along a radial plane cutting the cylinder. Water / ethylene-glycol binary mixture and pure water are tested and compared during the experimental program, in order to reproduce a set of thermal situations as close as possible to actual engine cooling system operation. On the CFD side, an extensive program reproducing the experiments is carried out in order to assess the predictive capabilities of some of the most commonly used eddy viscosity models available in literature. Both non-evaporating and evaporating conditions are tested, showing severe limitations to the use of simplified boiling models to correctly capture the complex interaction between turbulent boundary layer and vapor bubble dynamics. In order to overcome the above stated deficiencies under boiling conditions, a methodology is then proposed to both improve the accuracy of the CFD forecasts and reduce the computational costs of the simulations. A few preliminary results from the validation process are shown and briefly discussed at the end of the paper.