An Experimental Study of In-Cylinder Heat Transfer from a Pressurized Motored Engine with Varying Peak Bulk Gas Temperatures

Abstract

The variation of in-cylinder heat transfer with parameters such as engine speed, air-to-fuel ratio, coolant temperature and compression ratio were frequently studied in classical research. These experimentally-obtained relationships are important for improving in-cylinder heat transfer models, essential in developing CO2 reducing strategies. In this publication, a 2.0 liter compression ignition engine was tested in the pressurized motored configuration. This developed experimental setup allowed testing of the engine at speeds ranging between 1400 rpm and 3000 rpm, with peak in-cylinder gas pressures from 40 bar to 100 bar. The engine was motored using different gas compositions chosen specifically to have ratios of specific heats of 1.40, 1.50, 1.60 and 1.67 at room temperature. This enabled motored testing with peak in-cylinder bulk gas temperatures ranging from 700 K to 1500 K. This wide variation of peak bulk gas temperature was achievable even at constant peak in-cylinder gas pressure, which gave the possibility of varying the thermal load of the engine independently from the gas pressure load. This experimental setup offered the repeatability and robustness of motored testing, with the benefit of a fired-representative gas pressure and thermal load. Throughout the test matrix, the engine was instrumented with eroding surface thermocouples at two locations in the cylinder head; at the cylinder central axis, and at the periphery in the squish region. The steady-state and transient components of heat flux were investigated separately, along with the average surface temperature and its swing. The transient component of heat flux was computed using the Impulse Response method, coupled with a two-dimensional finite element model of the eroding thermocouples, as presented in SAE 2021-24-0018. This method takes into account the two-dimensional nature of heat flux through the thermocouples and hence presents a more robust analysis than the more common one-dimensional treatment using the Fast-Fourier Transform method.