This paper describes predictive models and validation experiments used to quantify the in-chamber heat transfer of LiquidPiston’s rotary 70cc SI “XMv3” engine.
The XMv3 engine is air cooled, with separate cooling flow paths for the stationary parts and the rotor. The heat transfer rate to the stationary parts was measured by thermal energy balance of that circuit’s cooling air. However, because the rotor’s cooling air mixes internally with the engine’s exhaust gas, a similar procedure was not practical for the rotor circuit. Instead, a CONVERGE CFD model was developed, and used together with GT-POWER to derive boundary conditions to estimate a ratio between rotor and stationary parts heat transfer, thus allowing estimation of rotor and total heat losses.
For both cases studied (5000 and 9000 rpm under full load), the rotor’s heat loss was found to be ∼60% that of the stationary parts, and overall heat losses were less than 35% of supplied fuel energy.
The significance of this work relates to the following facts:
It represents the first time that heat transfer was quantified for the “X” engine architecture;
Preliminary experimental and modelling results show reasonable correlation
The predictive models developed will inform future engine cooling system optimization work, leading to higher power densities and thermal efficiencies. Results for these two metrics are already market competitive in the 3 horsepower engine size.
