An essential element in the optimization of vehicle interior noise is the correct understanding of the various noise sources which are present, and how they propagate to the critical receiver. This article, synthesized from a paper by Akinori Iwama, and Tatsuo Osawa, of the Nissan Motor Company, and Stef Goossens, of LMS Engineering Services, presents an experimental source-transfer-receiver method that allows the quantification of the structure borne and airborne source strengths of the intake system components and their contribution to the interior noise. The approach was used to identify the noise contribution of the air intake system to the interior noise of an 8-cylinder luxury vehicle.
Two different mechanisms exist in which the intake system contributes to the interior noise: structure borne transfer and airborne transfer. Both can be integrated in the same source-transfer-receiver model.
Structure-borne Transfer Path Analysis (TPA), describes a receiver microphone signal for a given operating condition in the car as a superposition of contributions coming from different structural paths. By mechanically cutting the car body from the vibrating mounts of the intake system one can equivalently describe the effect of the original sources by operational reacting forces that are active at the interface between intake system and car body. By combining estimates of these operational reacting forces and vibro-acoustical transfer functions on the car body, the total structure borne interior noise can be described as a superposition of contributions of the individual paths.
Since it is difficult to make direct measurements of the operating forces, the latter need to be estimated by the use of matrix inversion techniques, or, as in this study, from a knowledge of the mounting elements. This requires measurements of the relative operational displacement over the mount, and the mount stiffness FRF, from which the operational forces are simply calculated.
The vibro-acoustic transfer function is measured after disassembling the source (intake system) from the car body. The car body is excited at the structural connection points by means of a hammer and pressures are measured at the receiver locations in the cavity. In practice, the FRFs are measured making use of the vibro-acoustic reciprocity principle describing the coupling between vibratory response at the structural boundary of an enclosure and acoustical excitation in the cavity, and between pressure response in the cavity due to structural excitation at the receiver position, at the mechanical enclosure boundary. This relationship allows the derivation of vibro-acoustical FRFs based on acoustical excitation using calibrated volume velocity sources.
Airborne Source Quantification (ASQ), quantifies the acoustic sources and the airborne transmission paths from source to receiver. The method is very similar to the structure borne tpa, but the data required are now acoustic source strengths, characterized by their volume velocities and acoustic-acoustic transfer functions between source and receiver.
A critical issue is the determination of the operational acoustical source which, as in the case of structure borne forces, cannot be measured directly. Two different source quantification methods are used: Panel Normal Acceleration (PNA) for radiating panels, and Pressure Indicator Inversion (PII) for the nozzle noise. PNA is used to estimate the point source strengths of the different parts on the shells of inlet, air cleaner and hoses by multiplying the normal acceleration with the surface of the patches. PII uses near-field acoustical transfer functions between monopole sources at the nozzle and indicators in the vicinity of the nozzles.
The operational source strength from the nozzles is estimated by inversion of the acoustical transfer matrix. Pressures are measured atindicator positions in the vicinity of the source. Acoustic transfer function measurements are then measured between indicator position and source position. The far field as well as near field acoustic-acoustic transfer functions are measured using calibrated volume velocity sources. Acoustical reciprocity is used to make faster measurements.
Case Study
The experimental TPA and ASQ technologies were applied to quantify the intake system noise of an eight cylinder luxury vehicle.
Operating measurements were made on a chassis dyno in road load simulation when accelerating at full throttle in 2nd gear from 1000 to 6000rpm: great care being taken to make measurements with a very good day-to-day repeatability. Transfer function measurements were made in a semi-anechoic room with the vehicle in a warm condition, but without the engine running.
There are six structural connections between the intake system and the car body, making 18 structural transfer paths in all. Accelerations at the source and body sides of the mounts were measured for the transfer paths when making a fast runup: double integration gave the relative displacements over the mounts. The complex dynamic stiffness of the mounts between intake and car body were measured on a standard test bench. Multiplication of mount displacement and mount stiffness gave the operating forces.
The vibro-acoustic transfer functions were measured using the vibro-acoustic reciprocity principle. A volume velocity source was placed at each receiver location in the car and artificially excited the car body. Accelerations are measured at the interface points between intake and car body with entire intake removed from the car to meet reciprocity boundary conditions. Both linearity and reciprocity were checked and found to be very good.
The airborne source quantification required the characterization of 214 shell sources and 3 nozzle sources. The shells of the inlet, air cleaner and hoses have an average surface of 30cm2 and the operating normal acceleration was measured using lightweight accelerometers mounted each. The nozzle sources were quantified by measuring operating pressures at 5 positions in front of them. A 3x5 near field acoustic-acoustic transfer function matrix was measured by reciprocally exciting with a volume velocity source at the 5 indicator positions and measuring the sound pressure at the 3 nozzle positions. Inversion of this matrix and multiplication with the operating pressures gave the source strength of the individual nozzles. The far field acoustic-acoustic transfer functions were measured by artificially exciting the car cavity with a volume velocity source at the receiver positions and measuring the sound pressure at the shell and nozzle sources.
Analysis and Results
In this case the structure borne contribution to the interior noise was negligible, while the airborne contribution was fully responsible for the interior noise at full throttle acceleration. The nozzle noise contribution was dominant at low engine regimes, while the shell radiation dominated at higher regimes. A pumping mode of the upper part of the air cleaner was responsible for this shell noise.
It is found that the high interior noise levels are caused by a combination of source strength (shell and nozzle) and acoustic sensitivity of the body.
Validation
Validation of the TPA-ASQ model with physical decoupling tests has been made for three particular cases:
- 6 mounts between intake and body are removed, i.e. no structure borne contribution.
- Entire intake system is removed and replaced by a rigid hose, i.e. no structural and no airborne contribution
- The flexible hose is replaced by a rigid hose, i.e. no airborne contribution from the flexible hose.
For all three cases a very good correspondence is found between the estimated contribution using the TPA-ASQ model and the measured contribution with the validation experiment.
The figure shows the synthesized interior noise and the measured interior noise. The correspondence is very good except in the 5300 rpm area where the repeatability of the operating measurements were rather limited.
Interpretation
Knowing the mechanisms that produce the high interior noise levels for each of the problem frequencies allows effective countermeasures to be applied.
Acoustic modes in the duct causing high nozzle noise and forced vibration of the intake system panels can be minimized changing the geometry of the piping and chambers or adding Helmoltz resonators.
Numerical acoustic radiation prediction software (e.g. SYSNOISE) allows to make these optimizations.
Structural experimental and numerical modal analysis of the intake system components allow for structural optimization avoiding high shell vibration levels and shell noise generation.
