The procedure is part of the Vehicle Benchmarking Acoustic Program and is well suited for comparison of competitor cars and for sensitivity analyses.
A reciprocal transfer function synthesis procedure is applied to characterise and summarise the structure-borne performance of the body and the Noise Vibration Harshness (NVH) treatment of a group of vehicles independent of the powertrain type. Structure-borne transfer functions or Noise Transfer Functions (NTFs) are useful to quantify vehicle acoustic sensitivity to structure-born noise. In modern cars, direct NTF measurements are difficult with full vehicle configuration due to congested engine bays. This often requires engine dismantling to account for individual structure-borne routes; traditional impact testing is also time consuming and is limited to low frequencies.
In the reciprocal approach, the complete vehicle is acoustically excited in the interior passenger compartment with low frequency and middle-high frequency calibrated point sources to cover the whole structure-borne frequency range of investigation, simultaneously activating and measuring all transfer paths. This is significantly quicker and easier than the direct testing procedure. The mechanical responses at the powertrain body mounting points from the two sources are summarised with a magnitude phase-less technique and synthesised in a single curve in the 40-1000Hz range; a wider frequency range compared to direct measurements capabilities.
The same acoustic sources and a similar vibro-acoustic reciprocal procedure are also used to evaluate the airborne transfer function performance of the vehicle. This measurement technique is part of the Vehicle Benchmarking Acoustic Performance Program (VBAPP) and complements a reciprocal airborne transfer function procedure and standard road measurements (enabling a quick ranking of different vehicles with respect to both air and structure-borne noise filtering performance), and is used in the vehicle benchmarking process at Rieter.
Direct measurements performed under operating conditions (on road and dynamometer) with natural excitations such as third gear Wide Open Throttle (WOT) run up for powertrain noise assessment or constant speed for rolling noise, are convenient to assess the final NVH performance of a vehicle from a customer perspective. However, natural excitations mix structure-borne noise sources / paths with airborne noise sources / paths, and source strength with noise transfer paths.
With structure-borne sensitivity to powertrain excitation, the state-of-the-art measurements use hammer impact or shaker excitation at powertrain body attachment points. Hammer impact measurements provide reliable results only up to 300-400Hz due to decay of the source power spectral density with increasing frequency, whereby the structure-borne related issues are relevant up to 1000Hz. Moreover, with direct external force excitations, access to powertrain mounts is often limited due to confined engine compartments. Alternatively, the engine can be dismantled and all powertrain mounts excited in the three directions, but this requires about four mechanic man-days and specific test vehicles. The procedure is also cumbersome and time-consuming, since each transfer function is measured independently as an average of several impacts performed by a trained operator due to the risk of double hits and overloads.
Here Rieter presents a reciprocal transfer function measurement method and synthesis procedure with respect to powertrain structure-borne noise. The vibro-acoustic reciprocity principle evaluates the NTFs by acoustically exciting the vehicle in the interior passenger compartment at the predetermined receptor positions and by measuring the mechanical response via accelerometers placed at the actual powertrain body attachment points. The convenient positioning of accelerometers (compared to external force excitation methods) makes access to powertrain mounts easy and enables measurements on full vehicles in operating conditions without dismantling the powertrain. The combination of two differently calibrated volume velocity sources extends the investigation frequency range up to 1000Hz. The use of reciprocity also allows all transfer paths to be excited and measured simultaneously, which significantly reduces the time needed and simplifies testing (compared to direct methods).
Vibro-acoustic reciprocity and its advantages in structure-borne NVH testing have already been discussed. Rieter implemented a multiple acoustic source technique to widen the investigated frequency range. A synthesis procedure based on a magnitude phase-less technique enables the vehicle structure-borne powertrain noise attenuation performance to be summarised in a single curve. This procedure is robust and saves time.
The structure-borne path for road and powertrain induced noise dominates in low frequency ranges (up to 1000Hz). The vibration energy from the powertrain and tire-road interaction propagates through wheels, bearings, spindles, chassis, body structure; this generates noise in the vehicle interior compartment and excites the acoustic cavity modes. NTFs are the body acoustic responses that correspond to unit forces at all major body attachment points.
Structure-borne powertrain and road noise is controlled mainly by four factors: 1) the source’s vibration level and power spectral density (powertrain/driveline and suspension mounts displacements), 2) the mount’s dynamic stiffness, 3) body dynamic stiffness at powertrain and chassis connections (usually determined by local stiffness or point mobility measurements) and 4) the body structure-borne NTFs, with respect to the powertrain and chassis attachments points. NTFs also include the influence of body stiffness at the powertrain and chassis connection points.
Taking into account all powertrain and chassis connections to the trimmed body, the total structure-borne interior noise at an occupant’s ear location is given by the vector sum of all individual structure-borne routes; each contribution is given by the operating dynamic forces at the vehicle body connections points multiplied by their appropriate NTF:
where
Pj=total structure-born sound pressure at occupant ear location j
Pj,j,k= partial sound pressure at the occupant ear location j due to mount k in direction i
Fi,k = operational force at mount k, direction i
NTFi,j,k = vibro-acoustic transfer function Pj / Fi, k for mount k, direction i to interior location j
Increasing interest in the effect of higher engine speeds and greater awareness of vehicle sound quality with respect to gear meshing and accessory components excitation makes a structure-borne investigation over a wider frequency range (than standard direct measurements) useful. Traditional investigation frequency ranges for the energy transfer from powertrain to vehicle trimmed body are 0-40Hz for ride and handling and ride harshness issues, and 40-300Hz for engine noise and vibration isolation. Isolation capability in the high frequency structure-borne domain (300-1000Hz) plays a secondary role, which needs to be considered if total vehicle sound quality is addressed. The high frequency contribution of the powertrain structure-borne paths is becoming more critical for vehicle NVH performance due to 1) lighter weight unibody designs, 2) high nominal power and rotational speed of diesel powertrains, 3) super charged gas and diesel engines, 4) rigidly mounted subframe architectures that contribute to body dynamic stiffness and crashworthiness, thus dynamically coupled with the body and 5) complex engine mount structures that are more likely to couple with powertrain harmonic excitations.
The mechanical-acoustic reciprocal measuring procedure is based on the principle of reciprocity of energy propagation; the response of a linear, elastic, time invariable, passive system to a disturbance applied by an external agent is invariable with respect to exchange of the points of input and observed response. The principle is very general, and can also be extended to dissipative vibrating systems. The vibro-acoustic reciprocity can be formulated as:
where
P2 = sound pressure level at point 2 due to force F1 acting in point 1.
X1= acceleration at point 1 due to acoustic excitation of volume acceleration acting in point 2, where quantities are expressed as spectra in the frequency domain.
Reciprocal NTFs are defined as:
where
X= Acceleration spectra along X,Y,Z axes at powertrain mounting points, body side [m/s2].
Q= Volume acceleration of point sources at the investigated passenger compartment receptor positions [m3/s2].
For example, with four engine mounts, i=(x,y,z), j=(driver’s outer and inner ear position), k=(left, right, centre, rear engine mount), so that 24 NTFi,j,k are reciprocally measured for each point source; one NTF for each of the three directions x,y,z on each engine mount, for each of the four engine mounts, and for each of the two interior locations in the passenger cabin. Since two volume velocity sources are used, 48 NTFi,j,k are measured for with four engine mounts. With n powertrain or exhaust line connection points to the vehicle body, 12*n NTFi,j,k are measured. This analysis focuses on structure-borne transfer functions from the powertrain to the driver’s ears and can be easily extended to other passengers in front and/or rear seat positions.
Evaluated frequency ranges are 40-400 Hz for the LFVVS, and 250-1000Hz for the MHFVVS.
Magnitude Averaging: A magnitude averaging procedure is applied over i,j,k, i.e. with respect to the different directions of each mount (i), the different mounts (k), and each interior location (j) to summarise the NTF performance of the tested vehicle.
Synthesis Technique: The NTFLF and NTFMHF frequency domains overlap in the 250-400Hz ranges. The two curves are merged in this overlapping frequency range to synthesize the structure-borne sensitivity of the vehicle under investigation.
The measurements are presented as dB fine band spectra curves [a/Q’] (ms-2/m3s-2), reciprocally equivalent to [p/F] (Pa/N). Reference values of 50-55dB/N for the average of the summarised NTFs are common values. For reference, two lines at 50dB/N and at 55dB/N are shown on each following NTF plot. Values higher than 55dB/N (red line) are considered potentially critical, whereas values lower than 50dB/N (green line) usually fulfill OEM requirements. Considering the general behaviour of the summarized and averaged NTFs, the resulting frequency curves typically show high levels up to approximately 300Hz, and decrease significantly above 500Hz; the vehicle trimmed body and chassis present a low pass response to structure-borne powertrain excitation. In addition to the summarised average NTF value, it is important to evaluate peaks, especially in the first body cavity modes range.
The performance of four different vehicles is shown in figures 3-6.
Figure 3 shows engine NTF performance for an A-segment vehicle with basically no BIW damping treatment and basic acoustic treatment at critical areas (dash inner, dash outer, floor carpet). The summarised NTF curve has a high mean value, close to 55dB/N, at low, mid and high structure-borne frequencies. Peaks are in the 100-350 Hz range, probably related to poorly damped cavity modes and body structural modes, and also the presence of peaks is important and coupled with a high mean value at very high frequencies. The structure-borne behaviour of the car is below average, confirmed by subjective drive appraisals.
Figures 4-6 show the powertrain NTF performance of three different B-segment vehicles (identified as vehicles 1, 2, 3). Vehicle 1 has a gas engine and vehicles 2 and 3 have diesel engines. The vehicle 1 NTF summarised curve, although featuring a lower mean value and a lower number of peaks with lower amplitude than the A-segment vehicle in figure 3, has pronounced peaks over the whole investigated frequency range, with an overall high amplitude level. Its performance is below average, also confirmed by subjective appraisals and direct measurements.
Vehicle 2 in figure 5 has less number of peaks in the low frequency range and has overall higher structure-borne noise attenuation at higher frequencies, being systematically below the 50dB/N line from 300Hz up. The peaks near 80Hz and 250Hz are similarly present on both vehicles 1 and 2 and are probably cavity mode-related peaks. Vehicles 1 & 2 have comparable body architectures with similar dimensions.
The NTF in figure 6 shows very good attenuation with few peaks present at very low frequencies (40-100Hz), and low frequencies (100-200Hz). The peak amplitudes are limited and well below the potentially critical 55dB/N sensitivity level. Well-attenuated peaks are present up to ~350Hz. Above this frequency, the attenuation performance of the full vehicle is very high, and comparable to a higher segment vehicle. The three B-segment vehicles (1, 2, 3) differ significantly in the NVH treatment at crucial zones (dash inner and passenger floor). It is also notable that vehicle 3 features a decoupled subframe architecture with four engine mounts, whereas vehicles 1 and 2 have coupled subframes, and a three engine-mount layout.
The NTF metric characterises the structure-born noise attenuation performance of a vehicle body and its NVH treatment at any point in the passenger compartment. As part of the Vehicle Benchmarking Acoustic Performance Program it also integrates the airborne attenuation ATF metric. A reciprocal measurement procedure and a synthesis method were applied with respect to powertrain structural excitation. The measurements are quick and extend the frequency range of investigation (compared to traditional direct measurements). The measured NTFs are averaged and synthesised in a single curve that summarises the structure-borne noise attenuation performance of the vehicle to powertrain excitation.
