Future of NVH Technology: Adapting to New Industry Demands
Key points
Electric cars lack engine noise, exposing quiet cabin distractions like tire tread and rattles.
Acoustic testing chambers use foam or fiberglass to measure high-frequency vehicle sounds accurately.
Automation and robotics help less experienced technicians perform precise and repeatable tests.
When the engine sound disappears in an EV, what is an NVH engineer to do? Randy Rozema, director of acoustics and vibration at ACS, tells SAE Expert Insight that even when the elimination of ICE noise exposes new, high-frequency, and structural sounds, engineers have plenty of strategies to provide expected customer comfort. The discussion also explores NHV brain drain as the industry loses senior engineers, how real-world sensor data can ‘calibrate’ a digital twin so engineers can actually trust its virtual acoustic predictions and how to update testing chambers for an exhaust-free world.
Transcript
00:00:01 Hello everyone, welcome to another SAE Expert Insight. My name is Sebastian Blanco, I’m the editor of Automotive Engineering Magazine, and I’m talking today with Randy Rozenboom, director of acoustics and vibration at ACS, about the future of NVH technology. Randy, how are you doing?
>> Great, how are you?
>> Good. This is an interesting topic
00:00:21 because NVH is changing so rapidly, and we want to talk today about how to adapt to the industry demands. Um, what can you tell me about how the elimination or the reduction or the changes that are happening with the internal combustion engine noise, and how that exposes new sounds that are happening coming from today’s new vehicles?
>> Yeah, sure. So, you know, historically
00:00:43 combustion engine noise has been a masking agent, if you will, in the automotive industry, and what we’re seeing is with the advent of electric vehicles or EVs, we’re seeing new noises pop up in the in the driver cabin and in the passenger areas. So, you know, some of the things that we’re seeing today are
00:01:05 an increase in notice of tire noise. And there’s some parameters there that are influencing the tire noise. It’s the the the tread design, the the material compound of the tire, and factors like that, and those are things that can be tested in the lab and and can be evaluated, and changes can be made. Other things though
00:01:28 to account for are interior noises such as as the industry terms it, buzz, squeak, and rattle, where we’re seeing plastic on plastic surfaces that are now generating noises, and a real good example, related to a project recently done, was um airbags have a capsule in them that has an accelerant that makes the explosion,
00:01:54 if you will, when the airbag deploys. Well, with the quieter interiors, those little pellets that make that reaction are being heard in the dashes of the vehicles while they’re going down the road. So, we’re we’re seeing these quieter environments in the automotive sector or in the the vehicles, I guess you should say.
00:02:16 And as a result, we now are looking at needing to align test facilities that can deal with this.
>> So, it sounds like what you’re saying is when designing or when developing EVs, if you if you’re working on NVH, you know, aspects of a vehicle, you’re really this is the expert mode. You’re not playing at the beginner level anymore where you have some background
00:02:37 to kind of cover the noises. What are some of the things that you need to do in a lab? Like what from the design perspective, from the machine perspective, that anticipate vehicle design changes, you know, in the in the 10 15 years down the road. Do we need to totally rebuild our labs for this sort of testing?
00:02:55
>> No, not necessarily. We’re we’re doing a couple of different things. When we’re when we’re looking at utilizing an existing lab space, we need to keep in mind a NVH lab is designed to have a life cycle of 30 to 40 years. I mean, it’s not made up of components that are going to go obsolete or or wear out over time.
00:03:17 Yeah, you’re going to you’re going to do maintenance. You’re going to buy new data acquisition systems, but the chambers that are used in and of themselves largely are the same. But, historically, you know, in the the ’80s, ’90s, early 2000s, the chambers have been getting fabricated or built with perforated metal wedges.
00:03:40 And over the years, there’s been a lot of debate over are the high-frequency noises that we would be seeing from EVs, as an example, being measured correctly. And and the reason for that question is the the the feeling with the experts is you’re getting reflection of sound off of that metal surface, and it’s getting reflected right back to where the sound
00:04:06 is generating from, and it’s resulting in sound levels that are potentially skewed or greater than what they should be. So, in today’s environment, we’re now seeing a lot of applications where if a new chamber is being built, they’re going to use a fabric-wrapped fiberglass, or they’re going to use some
00:04:29 type of a foam compound to undertake that absorption. And therefore, they’re getting rid of that reflection, and we’re getting more accurate data. So, really, it’s just a matter of, you know, what does the internal of the room looking like? And the other piece to keep in mind is
00:04:50 back when we had IC engines, we had to deal with engine exhaust. That’s not a thing today. Now we have batteries. So, we have chambers that have features that really aren’t necessary. They They aren’t needed. And that’s okay. They can still be in the room. The rooms have been qualified to perform with those things
00:05:12 in there, and we can still use the room as it is. Or if if people really want to get into the the the nitty-gritty, we can go in, make a small retrofit, and remove some of those potentially reflective surfaces that comprise engine exhaust or things of that nature.
>> One of the things that that brings up is the idea of sort of, um, knowledge
00:05:35 within the people who are doing the testing and the engineers who are who are working on all this. Um, if you know the way the things worked in the ’80s, ’90s, and 2000s, you can adapt them to today. But the industry is losing, um, a lot of NVH engineers, you know, just the normal cycle of people retiring and everything. And we have a lot of new tools from
00:05:55 robotics and machine learning and automated auto automation. How are these new tools helping some of the younger, less tech experienced technicians successfully run these tests that you’re talking about?
>> Yeah, there there’s a couple of benefits here. Like you say, we have these new tools,
00:06:11 we have technologies, we have robots. So, we can automate tests now, meaning a technician can be brought in who understands the product that’s being tested more so than the test that’s being run because a a knowledgeable engineer can come in, program the measurement surface that needs to be scanned, or program whatever the test cycle is, and then the
00:06:37 technician literally can come in, put the product in, get it operational, set the sensors where they need to be, and push a button and let the the automated test run, let the data be collected. And on the back end, then the the knowledgeable engineer can evaluate, analyze, and post-process that data. And the other piece of it is, you know, like you say, we have robots and things
00:07:00 like that. That that really helps in terms of we have, say, a sound intensity measurement where you have to move a sensor probe over discrete locations or across a dis uh continuous path, you have a an automated repeatable means to do that now, whereas in the past, it’s been all
00:07:23 human interaction, and you know, technician A may may line up on the center of a square, whereas technician B might line up on the intersection between two squares. So, all of the randomness and variable kind of goes away, and your data becomes a higher quality output as well.
>> Yeah, I never trusted uh technician B. I I thought they were looking a little
00:07:47 shady, you know, they didn’t know quite what they were doing.
>> That was the intern.
>> Um Yeah, exactly. Um so you we have these new tools. Um what about something in the realm where we’re dealing with virtual um you know, projects and things and how that it interacts with the real world. Um an AI
00:08:04 system an AI system, for example, could you know, look at the data and design an acoustically perfect part in a computer, but that doesn’t then lead to something you can actually manufacture. How do you deal with some of these, you know, fanciful ideas or things that could be done in the virtual space then then converting them over to to real world technologies?
00:08:23
>> I think before we get into that question, we want to actually look at how do you know or how do you trust the output that’s coming from that digital and you know, they they refer to it as a digital twin because at some point along the line you’re going to build the prototype. You’re going to measure that prototype and really the crux of it are taking that real world data, that real
00:08:46 sensor data and using that to calibrate your digital image. And now when you run those those simulations on that digital image, you’re going to say, “Hey, I have an output I can trust.” But to your point, AI and or or the the the virtual model isn’t necessarily something that’s manufacturable. And so there does need to be a human review aspect to this to
00:09:13 say, “Is what is being optimized or what is being designed something that we can physically produce, whether it be through 3D printing or some other uh existing process such as stamping or or you know, machining, things of that nature. But at the end of the day, you do need to be able to break the part that is a virtual part into components
00:09:38 that is manufacturable and assemblable.
>> Mhm. I think if we have any takeaway point from this discussion today is that, you know, the institutional knowledge is very important as well as understanding how the new technologies can blend and lead us towards better NVH systems in the future. So, Randy, thank you so much for speaking with us today and we’ll see
00:09:59 everyone next time on Expert Insights.
>> Great. Thank you. Appreciate the invite to join you today.
