Conditioning Turbocharger Compressor Map Data for Use in Engine Performance Simulation

摘要

Turbocharger compressor maps are used in engine performance modeling and simulation to predict engine air system operating conditions. Errors in compressor map data can result in inaccurate engine performance prediction. A method is described for conditioning compressor map data for use in engine performance simulation, by detecting and replacing suspect data points, and interpolating and extrapolating the map data. The method first characterizes enthalpy rise through the compressor, after removing data points likely influenced by heat transfer from turbine to compressor, using energy transfer coefficient vs. impeller outlet flow coefficient. This is done concurrently with estimating impeller outlet conditions using simplified geometry assumptions and a modified definition for compressor stage reaction. A single value for this modified stage reaction is determined to collapse all of the compressor map data onto one curve for energy transfer coefficient vs. impeller outlet flow coefficient. Next, the method characterizes pressure recovery corresponding to enthalpy rise, using simplified models for flow friction and incidence losses. To verify capability, the method is applied to compressor map data historically resulting in accurate engine performance prediction. The modeled compressor map predicts speed within ±2% and specific enthalpy rise within ±3% of measured data for more than 90% of the evaluated, engine data points. To demonstrate usefulness, compressor maps with suspected issues are conditioned using this method. The conditioned maps reduce speed prediction error to measured engine data by up to 5 percentage points. However predictions using the conditioned map for the extrapolated region to low compressor speed may not be sufficiently accurate for all compressor data. Future work will focus on improving method assumptions for this operating region.