|
|
|
|
|
01/03/2010
 Email to a friend
 Comment on this article
|
Engineers are using ever more advanced tools to create the unique sounds of today's vehicles. SAE International's Bruce Morey reports on the technology and processes behind the words
 Driving a car is an experience of the senses. So it is no longer good enough for carmakers to deliver a quiet ride. Today, OEMs are establishing brand identities around an expectation of how cars should sound – just as they do with styling and other attributes.
"To give our customers what they need in sound quality, we first help them design a quiet vehicle and engine – to clean it of objectionable noise. Then we tune in the specific sound quality they want," explains Kiran Govindswamy, chief engineer for NVH, CAE and Driveline systems at FEV.
He makes the point that this is often an iterative process, combining detective work and engineering. "For a given powertrain input, light vehicles present more NVH challenges than heavy vehicles." But that's not all: advanced engine technologies can also cause problems, such as increased combustion noise or direct injector ticking. "Getting the right sound quality is part art and part science," he admits.
Tools that FEV uses to help include a number of NVH test cells and CAE simulation tools, combined with proprietary methodologies. Its CAE tools include: the FEA code (eg NASTRAN); multibody systems software for vehicle-level NVH simulations; and FEV's Virtual Engine suite.
Specifically to aid in the detective work of isolating noises, FEV built its own vehicle interior noise simulation (VINS) software, which creates vehicle sound quality models. Outputs from VINS models are sound files representing the interior noise from all structure-borne and airborne noise paths into the vehicle. NVH and acoustics engineers then perform much of their analysis by representing levels of sound in colour maps plotted against frequency, rather than time.
"These are useful [tools] to clean the sound, but you cannot listen to a colour map," explains Govindswamy. So VINS integrates and converts all data to sounds that analysts can listen to – referred to as time-domain files.
Translation & interpretation
Years of NVH testing, coupled with dedicated databases, give FEV another important tool – scatter band plots. Examples include both time-domain or frequency-domain metrics documenting NVH behaviour of both components (think engines or transmissions) and whole vehicles. These are useful in both target setting and in measuring performance against an industry normalised measure, he states. "We also routinely use a combination of test-based data and CAE-based analyses to conduct 'hybrid' analyses, using our Virtual-VINS process." Why? Because this hybrid process combines modelled forcing functions, such as engine vibration at the engine mount, with measured noise transfer functions – which he describes as "not quite ready for CAE simulation".
Another level of simulation is Bruel & Kjaer's Type 3644 NVH Simulator, which offers a different window into the sound engineer's world. Recognising that sound quality is, above all else, subjective, the NVH simulator stores sound files to create an interactive driving experience. The desktop version presents the sound of the vehicle through headphones to a 'driver', who controls the 'car' through a computer-generated environment, using a steering wheel, pedals and gearshift. Frequently, multiple models are stored for instant comparison – eg, when examining two brands of luxury cars or, at a more detailed level, two intake manifolds.
To make it scalable, Bruel & Kjaer created a three-level data structure: vehicle level, source level and contribution or path level. "It's important for engineers to know that, though our model can be detailed – even isolating the sound of each engine mount or tyre – it's not always required," explains David Bogema, senior application engineer at Bruel & Kjaer North America. "Someone looking at axle noise doesn't care what's coming from the engine mounts or intakes."
Ceating a simple powertrain model for noise would be sufficient for an axle study, for example, so the NVH simulator brings together all the component sounds, while also accounting for phase.
Modelling sound
Modelling sound and NVH using CAE has its advantages, especially for vehicles in the early stages of design or when many modifications are under consideration. However, there are limitations, especially in the mid frequency ranges from 200Hz to 1kHz, points out Terence Connelly, lead engineer Vibro-Acoustics for ESI Group.
"Statistical energy analysis [SEA] is useful above 1kHz, because the modal density is great enough to treat the problem stochastically," he explains. "Below about 200Hz, the sample size of modes is small enough where a full finite element model is needed." Above 200Hz, model sizes become so large as to make a solution impractical. Establishing which parts of the model should use FE and which SEA requires knowing some rules, he says, but the software helps.
Recognising that creating an FE model for NVH and acoustic studies alone can be expensive, however, ESI offers an integrated option for creating such models from existing crash data. "Our PAM-CRASH and VA-One software are widely used in the automotive world," asserts Joe Strelow, director of VE/VA solutions. "Crash models are critical and are created early in a vehicle development programme. By leveraging these, we believe we can create an NVH model where maybe 90% of the work is already done." What then remains is adding NVH-specific details for computing the acoustic and passenger comfort metrics that identify the brand. This is a concept Strelow calls an adaptive computing model, embodied in ESI's Virtual Performance Solution.
Spherical beamforming
So far, so good. But while engineering sound quality uses modelling coupled with measured data, engineers can generally only measure error states, such as buzz, squeak and rattle (BSR). So mapping sound in interiors is particularly challenging. A class of devices to solve that involves a sphere fitted out with microphones and cameras. This creates maps of the energy from directive sound pressure in three dimensions, using a technique called spherical beamforming.
Most often, this uses the acoustic wave equation (sometimes with spherical harmonics) to interpret sound hitting the microphones at the same time – so determining the time or phase differences and thus direction. Spherical beamforming devices accurately map line-of-sight transients, as well as stationary sources. They also map reflections just as easily as sources and are better at higher frequencies, typically above 200Hz.
Nittobo Acoustic Engineering offers a spherical beamforming system called Noise Vision. It is a rigid sphere with 31 microphones and 12 cameras, which maps directive sound pressures on to images to show where the sound is measured. Different diameters of the rigid sphere measure different frequency ranges more accurately, according to Kazuhiro Takashima of Nittobo.
Suite advantages
Bruel & Kjaer is another provider of spherical beamformers and also offering other equipment, such as conformal mapping devices using nearfield acoustic holography (NAH) and planar beam- forming. Using a suite of devices has its advantages, according to Tony Frazer, the company's array acoustics solutions manager. "Spherical beamforming covers a little higher frequency [200Hz to 8kHz range], while conformal goes a little lower, [20Hz to 6.4kHz]," he explains. "Spherical provides a coarse location tool and then conformal provides a fine-tuned, high spatial resolution tool."
Meanwhile, SenSound offers a slightly different take on spherical beamforming, with its acoustic camera, marketed for GFal Tech of Germany.
This is an open system, advertised as acoustically transparent, compared to rigid systems. Instead of mapping sound intensities to pictures that have taken by the camera, the acoustic camera fixes the position of the device, in relation to a CAD model, and then shows the intensities on CAD renderings.
|
|
| |
Author
Bruce Morey
|
| |
| |
Copyright AD Media Europe.
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies contact the sales team.
|
| |
|
|
| |
To access the whitepaper section you need to complete our 60 second registration process. Once completed this then allows you to download any and all white papers, register for e-zines and access our detailed supplier directory for FREE.
If you are all ready a registered user then enter your e-mail address and login.
You will need to have logged in prior to entering your comments in the boxes provided.
|
|
Engineering Recruitment Partner:
|