Assessment of Combustion Mechanical Noise Separation Techniques on a V8 Engine abio Bianciardi*, Karl Janssens, Konstantinos Gryllias, Simone Delvecchio, Claudio Manna Simcenter Symposium 17&18 October 2017 Realize innovation.
. Agenda Powertrain NVH CMB analysis on a errari V8 engine Engine noise localization ge 2
Spark Ignition (SI) engine evolution Downsizing & new technologies: Turbocharger Direct injection & GDI Variable valve timing Lightweight material 2025. 1876 irst SI 4stroke engine 1930 uel improvements 2000 CO 2 regulations Regulation in terms of emissions is driving new technology & innovation to higher efficiency ge 3
SI Powertrain evolutions & NVH Turbo related noise Turbocharger Transmission vibrations Higher torque level ECU control parameters optimization Gear rattle & driveline booming noise Engine knock Direct injection & GDI GDI noise Valve ticking VVT Engine components NVH Lightweight materials Mechanical noise Mount vibrations Engine downsizing ge 4
SI Powertrain evolutions & NVH NVH challenges Increased gear rattle (driveline-gearboxes) Increased radiated noise especially at low speed due to the higher torque delivered New type of noise introduced by the presence of turbocharger (turbo whoosh, turbo whine, BP, ) Large torsional fluctuations in the crankshaft (torsional vibrations shaft torsional modes excitation) 620 Engine radiated noise 10 db Torque (CH4) Nm ge 5 100 1000 Tacho_front_1ppr (T4)
errari V8 California T The errari V8 engine features most of the recent technologies to increase fuel economy and reduce emissions while increasing the performances. irst turbocharger engine with reduced turbo lag for instantaneous throttle response: V8 twin scroll turbocharger low inertia, which optimized the pressure peaks of the exhaust gases into the turbine to reduce lag lat plane crankshaft provides even firing between the 2 cylinders banks Despite delivering an extra 69 hp/l and a massive 49% more torque, fuel consumption has decreased by 15% emissions reduced by 20% ge 6 http://auto.ferrari.com/it_it/modelli-auto-sportive/gamma/ferrari-california-t/
CMB Engine test-rig setup The setup consisted of: - 6 Microphones located around the engine (1m distance from the engine) - In-cylinder pressure sensors (8 in total) - Speed sensor on the crankshaft (incremental encoder 60ppr) - Torque sensor LMS SCADAS Lab NEW VC8-QS board for cylinder pressure measurement 620 Engine radiated noise 650 Hp 600 Hp 550 Hp 500 Hp 10 db bar 0.00 deg Angle Cyl_1 2000.1 Angle Cyl_2 2000.1 Angle Cyl_3 2000.1 Angle Cyl_4 2000.1 Angle Cyl_5 2000.1 Angle Cyl_6 2000.1 Angle Cyl_7 2000.1 Angle Cyl_8 2000.1 720.00 Torque (CH4) Nm 100 1000 Tacho_front_1ppr (T4) 450 Hp 400 Hp 350 Hp 300 Hp 250 Hp 200 Hp 150 Hp 50 Hp 100 Hp ge 7
CMB Classical Wiener filter Classical Wiener filter - Reference in-cylinder pressure measurements required - Transmissibility filters are estimated from the reference signals (in-cylinder pressure) and microphones - The noise is separated in: - Combustion related noise coherent with the incylinder pressure signals - Mechanical noise, everything not correlated with incylinder pressure (mechanical, residual, aerodynamic) H ki ω = S ki ω S ii ω 1 x est k ω = nc i=1 H ki ω r i ω - Can be applied in time domain, by applying a dedicated filter pre-processing ge 8
CMB Classical Wiener filter Time domain filtering approach when the pressure in the combustion chambers of the engine is available is possible to extract the coherent part of the sound by mean of filtering techniques. db / 180.00 Phase R mic1:s/cyl_1 4066.5 H ki ω = S ki ω S ii ω 1 x est k ω = nc i=1 H ki ω r i ω -180.00 0.00 Hz requency 10000.00 H ki (ω) bar Real In-cylinder pressure (Time traces) IR ilter = Combustion related noise (Time traces) Real 10.00 s Time (Throughput) 20.00 10.00 s Time (Throughput) 20.00 ge 9
CMB Classical Wiener filter Classical Wiener filter Results of the CMB analysis: Torque 8000.00 635.76 800.00 - The method has been applied to non-stationary operating conditions in order to better decorrelate the mechanical noise contribution from the combustion related noise (cit. Lee, Bolton et al.) - The solution is also in line with the most common testing maneuver at OEMs - The frequency resolution (4Hz) is maintained over the entire speed run-up ge 10 RPM Amplitude 600.00 6500 6000 5500 4500 4000 3500 3000 2500 2000 B 11.00 s Time (Throughput) 1000 0 1000 2000 3000 4000 Hz requency Amplitude Nm 124:Tacho_front_1ppr only reduced valu 113:Torque 0.00 101.00
CMB Classical Wiener filter Classical Wiener filter Measured noise Results of the CMB analysis: The method proved to correctly separate: - the combustion noise characterized by the engine harmonics generated by the combustion process - from the mechanical noise, which is the residual noise composed of mechanical engine components and turbocharger noise 7000 6500 6000 5500 4500 4000 3500 3000 2500 2000 1500 1000 6500 6000 5500 4500 4000 3500 3000 2500 2000 1000 Combustion related noise 0 500 1000 2000 3000 4000 Hz requency 0 1000 2000 3000 4000 Hz requency 7000 6500 6000 5500 4500 4000 3500 3000 2500 2000 1500 1000 Mechanical noise 0 500 1000 2000 3000 4000 Hz requency run-up test at full throttle ge 11
CMB Classical Wiener filter 620.00 Torque (CH4) Nm 100.00 620.00 Torque (CH4) Nm 100.00 620.00 Torque (CH4) Nm 1000.00 Tacho_front_1ppr (T4) 1000.00 Tacho_front_1ppr (T4) Unrestricted 100.00 Siemens AG 2017 1000.00 Tacho_front_1ppr (T4) ge 12 Measured noise 600 650 Hp 550 Hp 500 Hp 450 Hp 400 Hp 350 Hp 300 Hp 250 Hp 200 Hp 150 Hp 50 Hp 100 Hp Combustion related noise 650 Hp 550 600 Hp 500 Hp 450 Hp 400 Hp 350 Hp 300 Hp 250 Hp 200 Hp 150 Hp 50 Hp 100 Hp Mechanical noise 500 Hp 450 Hp 400 Hp 350 Hp 300 Hp 250 Hp 200 Hp 150 Hp 50 Hp 100 Hp.00.00 600 650 Hp 550 Hp.00 10 db 10 db 10 db 10 db Measured noise original:s 100Nm original:s 300Nm original:s 600Nm 1000 2000 2500 3000 3500 4000 4500 5500 6000 6500 Clear effect of the residual noise affecting the OASPL noise in the speed range 1500-3000 for the high load Combustion related noise 10 db combustion_related:s 100Nm combustion_related:s 300Nm combustion_related:s 600Nm 1000 2000 3000 4000 6000 10 db Mechanical noise 1000 2000 3000 4000 6000 residual:s 100Nm residual:s 300Nm residual:s 600Nm
CMB & Engine SSL Objective: - gain information on the source distribution (strength, location, etc.) based on acoustic measurement of the radiated sound - Applicable for transient and stationary applications with z: distance between camera and engine D: diameter camera : sound wavelength ge 13
CMB & Engine SSL MIC 1 from the array Combustion noise Real 6500 6000 5500 4500 4000 3500 3000 2500 2000 Real s Time (Throughput) 1000 0 1000 2000 3000 4000 Hz requency 6500 6000 5500 s Time (Throughput) Real 4500 4000 Mechanical noise 3500 3000 2500 2000 s Time (Throughput) 1000 0 1000 2000 3000 Hz requency ge 14 x nr. of array microphones
CMB & Engine SSL Residual noise 1. Wiener filter has been used as a pre-processing step to the array microphones recordings. 2. SSL techniques used to identify specific noise phenomenon 7000 6500 6000 5500 4500 4000 3500 3000 2500 2000 1500 1000 0.00 Hz requency.00 low noise Turbo whoosh noise Engine noise localization of a typical noise generated by the Turbocharger for high engine load. 8 db 1200Hz 2500Hz ge 15
Turbo noise identification Turbo whoosh noise 10 db Mechanical noise Turbo whoosh results from turbulence due to flow separation within the compressor when running close to the surge line on the compressor map. Encountered during ull load acceleration Tip-in / tip out 1000 2000 3000 4000 6000 residual:s 100Nm residual:s 300Nm residual:s 600Nm Turbo whoosh is a broadband flow-type noise When present it can: Increase 1m OASPL up to 5 in the range 1500-3000 can result in driver annoyance 8 db ge 16
Conclusions SSL using microphones array technique, clearly shows the capability of identifying the most critical sources. CMB can be considered as preprocessing step before the SSL to enhance the dynamic range of the SSL results A specific turbo noise has been identified (turbo whoosh noise). The technique can be applied to other noise sources. bar Angle Cyl_1 2000.1 Angle Cyl_2 2000.1 Angle Cyl_3 2000.1 Angle Cyl_4 2000.1 Angle Cyl_5 2000.1 Angle Cyl_6 2000.1 Angle Cyl_7 2000.1 Angle Cyl_8 2000.1 The method identifies opportunities for powertrain NVH improvements. 0.00 deg 720.00 The advantage of having time domain separated noise allows to perform further analysis on the decomposed noise (i.e. angle domain processing) 8 db ge 17
Thank you Realize innovation. ge 18