This study was performed in order to better understand air-cooled engine heat transfer and to determine and carry out a methodology for analyzing the thermal flows in an air-cooled engine. In the project, experiment and engine cycle simulation were used together.

The main focus of the thermal analysis methodology was to make experimental measurements of flow rates, component temperatures, and heat flows, and then rely on simulation where appropriate. It is the coupling of simulation and experiment that can lead to reductions in experiments, reductions in analysis time, and an increase in model accuracy if simulations are chosen carefully.

The thermal analysis methodology consists of three phases:

Phase 1 Gas Dynamics
Phase 2 In-Cylinder Analysis
Phase 3 Finite Element Analysis

In Phase 1, the internal combustion engine cycle simulation code WAVE was used to model the gas flows through the engine. A simple flow network was created which could accurately predict the transient pressure behavior in the intake port. This pressure was then a boundary condition for the in-cylinder simulation model. Also in Phase 1, some experimentation was used to aid in developing a more accurate model. Measurements include air and fuel flow rates, exhaust port temperature, and intake pressure, temperature, and composition.

Phase 2 consisted of the detailed modeling of the in-cylinder flows with the code IRIS. IRIS predicts spatially resolved boundary conditions through the use of two-zone thermodynamic model and a four zone in-cylinder flow model. A spatial model was chosen so that emphasis could be placed on the causes of temperature gradients in some engine components, such as the cylinder head and bore. Also in Phase 2, experiments were run which analyzed the in-cylinder thermal loading in detail. The cylinder head of the test engine was instrumented so that simultaneous measurements of local instantaneous heat flux, in-cylinder pressure, and flame arrival at the heat flux sensor could be taken. Data was taken over a wide range of air/fuel ratios and for both the stock intake system and the CMS.

For both the stock system and CMS, strong correlations were found between flame arrival time and location of peak heat flux, and location of peak pressure and location of peak heat flux. The largest values of peak heat flux typically occurred at or near the peak power mixture for both fueling systems. The integrated heat flux over the entire cycle peaks near stoichiometry, but the integrated heat flux over compression and expansion peaks near the peak power mixture. When heat flux was integrated over the gas exchange process, it peaked at a mixture slightly lean of stoichiometric. This is due in part to the higher exhaust gas temperatures caused by differences in burn rate.

The final phase of the thermal modeling consisted of finite element analysis of the critical engine components. It was determined that due to the asymmetry in the system, three dimensional modeling of engine components would be necessary. ProE was used to model the components and for creating and refining the mesh. Specifically, the valves, head, rocker cover, fin pack, and piston were modeled. Due to computation limitations at UW-Madison, only the head and fin pack were used in the final analysis.

Experiments were used in Phase 3 to aid in the determination of external boundary conditions. The engine was instrumented with thermocouples which measured temperatures such as oil, fin-pack, cylinder liner, and cooling flow.

After two iterations on the finite element mesh and boundary conditions, a temperature distribution resulted which had agreement to within 50 °F at 14 locations in the head and fin-pack.