Sensitivity Study of a Turbo-Charged SI-Engine at Rated Power Jens Neumann Andrei Stanciu Bodo Banischewski BMW BMW Bertrandt GT-SUITE User s Conference Frankfurt, October 08, 2007
Sensitivity Study of a Turbo-Charged SI-Engine at Rated Power Introduction. Modeling Aspects. Knock. Results. Summary. see also in: Motorprozesssimulation und Aufladung II, H. Pucher, J. Kahrstedt (Hrsg.), Berlin, 2007
Introduction. Robustness Reserve. speed e.g. scavenging robustness reserve surge compressor map choke robustness reserve as generalized margin of a turbo-charged SI engine at rated power. special importance for down-sized highly turbo-charged SI-engines. questions about impact of different impaired conditions. how long can rated power be hold and at which fuel consumption. gradient of power decrease of unsustainable rated power.
Introduction. Simulation Strategy. variation of different quantities at rated power (engine speed = const.) in order to find sensitivities. environmental conditions (ambient pressure => altitude, ambient temperature). geometrical conditions along air/exhaust (pressure losses, valve cross-sections, compression ratio ε, intercooler-performance). ξ_p_llk ε intercooler ξ_a_intv ξ_a_exhv pt p4 air exhaust
Modeling Aspects. Stationary Simulation Model..6l-4Cyl-DISI engine with cam phasing (VANOS) and TwinScroll-turbo-charger. GT-Power WOT-model adjusted with Vibe and SI Turbulent Flame combustion model. boost pressure control with safety limits (maximum turbo-speed, maximum compressor outlet temperature T2). vehicle based intercooler modeling. mp_air [kg/s] mp_cool [kg/s] dq W /dt [W/K] Q W =dq W /dt*(t2-t_amb) v_veh [km/h] P_eng [kw]
Modeling Aspects. Knock. empirical knock model of Spicher, Worret (FVV, 2002). calibration at measured point. controller for spark timing of SITurbFlame combustion model (for anchor angle of Vibe combustion model, respectively). combustion (two zone approach) Knock Index I plus additional conversions (e.g. Spicher, Worret) dt = KBG K t0 τ Knock Index [-].5.0 0.5 0.0 α K α K2 knock CA ign α E no knock -5 TDC +5 +30 =f(α Comb50%, λ) critical autoignition level same knock sensitivity as calibration spark adjustment (e.g. PID-ctrl.) valid for DISI-Turbo? (determination of engine specific model constants and additional conversions).
Modeling Aspects. Knock. implemented as generalized knock assembly in GT-Power. 2 arbitrary levels of rated power at 5500RPM. P.MOT and P.MOT2 = P.MOT + 0%. Comb50% [ CA].09P.MOT 2 CA 2P.MOT 4P.MOT 7P.MOT 9P.MOT 20 30 40 50 60 70 80 T2' [ C] Measurement Simulation Calibration Measurement P.MOT Simulation P.MOT Knock + Vibe Simulation P.MOT Knock + Entrainment Simulation P.MOT2 Knock + Entrainment Simulation Performance Potential Knock + Entrainment
Modeling Aspects. Interaction Knock/Combustion Modeling. from this and other correlations: correct knock prediction with an empirical knock model requires () a combustion model which is able to predict the burn rate shape as function of spark timing => un-corrected Vibe is not sufficient. entrainment model, e.g. SI Turbulent Flame. corrected Vibe with implemented relations, e.g. Comb0-90%=f(Comb50%), Elmquist et al. (SAE, 2003). (2) a knock model which includes link to combustion progress.
Results. Classification of Data in Phase A and B. example: decrease of ambient pressure (altitude). ~ P.MOT [kw] P.MOT/P.MOT [-] 5 Phase A 5 T2_max T2, T2 [ C] Phase B Phase A Phase B 0 000 2000 3000 0 000 2000 3000 alt [m] = f(p_amb) alt [m] = f(p_amb) T2 T2 0 000 2000 3000 alt [m] = f(p_amb) n_tc [RPM] P.MOT P.MOT2 n_tc_max Phase A: no decrease of rated power Phase B: power decrease, boost pressure at safety limit (T2_max, n_tc_max)
Results. Phase B. P.MOT/P.MOT [-] P.MOT/P.MOT [-] P.MOT/P.MOT [-] 5 5 5 5 5 5 0 000 2000 3000 alt [m] = f(p_amb) P.MOT/P.MOT [-] P.MOT/P.MOT [-] P.MOT/P.MOT [-] 0 00 200 300 5 5 5 5 5 5 0 20 40 60 T_amb [ C] 8 9 0 2 3 0.4 0.5 0.6 0.7 0 500 000 500 5 5 0.5 0.6 0.7 0.4 0.5 0.6 0.7 pt [mbar] ξ_p_llk [-] ξ_a_intv [-] 5 5 P.MOT P.MOT2 ε [-] ξ_a_exhv [-] p4 [mbar] P.MOT/P.MOT [-] P.MOT/P.MOT [-] alt = f(p_amb) ambient pressure T_amb ambient temperature pt total pressure loss before compressor ξ_p_llk intercooler cooling power ξ _A_IntV cross-section of intake valves ε compression ratio ξ_a_exhv cross-section of exhaust valves p4 pressure loss after turbine
Results. Phase A (P.MOT). T2 0 C Comb50% 5 KW φ φ _ base φ _ norm = φ _ margin φ _ base alt = f(p_amb) ambient pressure T_amb ambient temperature -0.5 0 0.5.5 mp_wg/mp 0.05 φ_norm -0.5 0 0.5.5 φ_norm -0.5 0 0.5.5 φ_norm BSFC 40g/kWh -0.5 0 0.5.5 φ_norm pt total pressure loss before compressor ξ_p_llk intercooler cooling power ξ_a_intv cross-section of intake valves ε compression ratio ξ_a_exhv cross-section of exhaust valves p4 pressure loss after turbine
Results. Sensitivities. environmental conditions geometrical conditions Phase A % BSFC-increase* ( BSFC) caused by: +75m altitude alt (~20mbar p_amb) +.5 C ambient temperature T_amb +20mbar total pressure loss before compressor pt -4% less intercooler cooling power (ξ_p_llk) +0.2 more compression ratio ε +32mbar pressure loss after turbine p4 Phase B % power-decrease ( P_eff) caused by: +0m altitude alt (~5mbar p_amb) +2.4 C ambient temperature T_amb +3mbar total pressure loss before compressor pt -4% less intercooler cooling power (ξ_p_llk) +0.2 more compression ratio ε +77mbar pressure loss after turbine p4 * non-linear!, analyzed within φ_norm = 0.0 0.5
Results. Comparison between Turbo and N.A.-engine. example: decrease of ambient pressure (altitude). P.MOT/P.MOT [-] 5 5 0 500 000 500 2000 2500 3000 alt [m] = f(p_amb) turbo-charged engine: P.MOT P.MOT2 naturally aspirated (N.A.) engine: P.MOT P.MOT2
Summary. classification of results in phase A (without decrease of rated power) and phase B (power decrease). phase A: progressively increasing BSFC (knock tendency, boost pressure, TC-efficiency, etc.). phase B: power decrease mostly linear. altitude behavior in phase B similar to naturally aspirated engine. outlook: improved modeling of knock and combustion. impact/sensitivity of fuel quality (RON).