Modeling the Effect on Engine Performance of Heat Transfer and Friction losses in the Turbocharger

Modeling the Effect on Engine Performance of Heat Transfer and Friction losses in the Turbocharger H. Tartoussi, A. Lefebvre, S. Guilain, Renault, A. J. Torregrosa, J. R. Serrano, F. Arnau, CMT

1 INTRODUCTION 2 3 4 TURBOCHARGER MATCHING GT-POWER COUPLED WITH CMT MODEL SIMULATIONS RESULTS 5 CONCLUSIONS 2

1 INTRODUCTION 3

Turbine Compressor Complex interactions between different energy fluxes occur in the turbocharger W T _ Out W Oil _ Out W C _ Out W T WC W Mech _ Loss Q Tur Housing QComp Q T ( Rad, Conv ) Q H / Oil Q H ( Rad, Conv ) Q C ( Rad, Conv ) W T _ In W Oil _ In W C _ In Turbine power Compressor power Mechanical power Oil power Heat power W Q W Efficiency in turbine maps is not enough for a fully C predictive modeling of turbocharged engines: T, map * Ts C

As heat transfer is neglected in the turbine side, direct use of turbine maps efficiency over-predicts turbine outlet temperature. T, map W C Q W * Ts C * * T, map m Ts 1 Q C W C

Analysis of heat transfer phenomena: Heat conduction (testing in specific thermo-hydraulic bench) Heat convection (testing in hot gas stand) External heat flows experimentally obtained Q C W C Q W * T Ts T turbine inlet_min : 350 ºC T turbine inlet_max : 460 ºC

MODEL [K] Lumped heat transfer model combined with 1D gasdynamic model T H1 H2 H3 C 700 650 600 550 500 450 400 350 300 OUTLET TEMPERATURES Turbine Compressor Oil 300 350 400 450 500 550 600 650 700 MEASURED [K]

Characterization of mechanical efficiency: Quasi-adiabatic testing Modeling of friction losses:

Turbine testing in a wide range of operating conditions: Physical equation to extrapolate isentropic turbine efficiency Developed equation is fitted to steady flow data measured in adiabatic conditions 50% VNT open Ts 2 * K1 K2 1 K 2 3 1 1 : blade to speed ratio

2 TURBOCHARGER MATCHING PROCESS 10

TURBOCHARGER MATCHING PROCESS TURBOCHARGER MAPS ENERGETIC HYPOTHESIS LOW END TORQUE TIME TO TORQUE PART LOAD BSFC MAXIMUM POWER. THE CHOICE OF THE TURBOCHARGER SUPPLIER IS DECIDED AT THE BEGINING OF THE ENGINE DEVELOPPEMENT GOOD SIMULATION ACCUARACY IS NEEDED TO REDUCE THE NUMBER OF TESTED TURBOCHARGER PROTOTYPES TURBOCHARGER SUPPLIER CHOICE 11

BENEFIT OF GOOD LOW END TORQUE SIMULATION WHEN TORQUE IS HIGH, WE CAN USE HIGH TRANSMISSION RATIO FUEL CONSUMPTION IMPROVEMENT WHEN TORQUE IS HIGH, WE CAN MOVE HEAVY CAR ENGINE CAN BE USED IN MANY CAR APPLICATIONS KNOWLEDGE OF ENGINE LOW END TORQUE IS IMPORTANT BECAUSE IT DEFINES STANDING START CAPABILITY AND CAR DYNAMICS. 12

3 GT-POWER COUPLED WITH CMT MODEL 13

Model Interface MODEL INTERFACE FILLED BY TURBOCHARGER SUPPLIERS INITIAL PROPERTIES Pipes Turbocharger TURBOCHARGER Geometrical & Materials Properties Data Mechanical Losses Heat Transfer COMPRESSOR Geometrical Data Compressor Map TURBINE Geometrical Data Turbine Map SAVE FILE 14

Model Interface MODEL INTERFACE FILLED BY TURBOCHARGER SUPPLIERS INITIAL PROPERTIES Pipes Turbocharger TURBOCHARGER Geometrical & Materials Properties Data Mechanical Losses Heat Transfer COMPRESSOR Geometrical Data Compressor Map TURBINE Geometrical Data Turbine Map SAVE FILE 15

Model Interface MODEL INTERFACE FILLED BY TURBOCHARGER SUPPLIERS INITIAL PROPERTIES Pipes Turbocharger TURBOCHARGER Geometrical & Materials Properties Data Mechanical Losses Heat Transfer COMPRESSOR Geometrical Data Compressor Map TURBINE Geometrical Data Turbine Map SAVE FILE 16

GT-POWER Model A UserFlowDomain replaces the turbocharger component in GT-POWER An User Subroutine is used for instant information exchange between the model and GT-Power. An User Model Reference is used to include initial data of the model 17

4 Simulation results 18

Model Application Turbocharged three-cylinder 0.9 litre engine Maximum power: 66 kw Maximum torque: 135 N.m from 2000 to 3500 rpm Variable Valve Timing combined with an integrated manifold. Stop & Start 19

Full Load Results 20

Temperature [ C] Full Load Results T H1 H2 H3 C Turbine inlet Turbine case Turbine Node1 Turbine Node2 Turbine Node3 Compressor case 1000 2000 3000 4000 5000 6000 7000 Engine speed [tr/min] 21

Full Load Results 22

Full Load Results Temperature before turbine Pressure before turbine 23

Compressor efficiency ADIABATIC EFFICIENCY GOOD PREDICTION OF THE DIABATIC COMPRESSOR EFFICIENCY AT LOW ENGINE SPEED 24

Torque over-estimation [%] Low End Torque Estimation 30 25 20 15 10 CMT turbo Original supplier data 5 0 1000 1250 1500 Engine speed [rpm] At low engine speed the wastegate is closed on the test bench The maps provided by the suppliers overestimate the turbocharger efficiency at low engine speed By taking into account the heat transfer and mechanical losses in the turbocharger the torque overestimation can be reduced 25

Instantaneous intake pressure The instantaneous static pressure at the engine inlet are well represented in the CMT turbo model. 26

Torque Transient Behaviour 1500rpm Simulation Test Time 27

5 Conclusions 28

Conclusions Much improvement in the quality of results from the raw data of FNR The CMT model provides new features to calculate the heat transfer in the turbo Remains to test the quality of extrapolations for turbo not fully characterized Need to improve suppliers maps coverage over large BSR interval 29