Optimizing Concrete Floors for Building Acoustics

The manufacturing and production of concrete construction building materials creates roughly 10% of global carbon emissions. As a result, architectural engineers are trying to find new ways to reduce the amount of concrete used in their buildings. Less concrete, though, can lead to some major ramifications in terms of the acoustics of a built space. In this episode, we talk to Jonathan Michael Broyles (Penn State) about techniques for optimizing the design of concrete slabs used in floors to reduce materials while improving acoustic performance.

Associated paper: Jonathan Michael Broyles, Micah R. Shepherd, and  Nathan C. Brown. "Investigation of optimization techniques on structural-acoustical shaped concrete slabs in buildings." Proc. Mtgs. Acoust 42, 022001 (2020). https://doi.org/10.1121/2.0001354

Read more from Proceedings of Meetings on Acoustics (POMA).

Learn more about Acoustical Society of America Publications.

 
Music Credit: Min 2019 by minwbu from Pixabay. https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=1022

Transcript

Kat Setzer  00:06

Welcome to Across Acoustics, the official podcast of the Acoustical Society of America's publications office. On this podcast, we will highlight research from our four publications. I'm your host, Kat Setzer, editorial associate for the ASA.

Kat Setzer  00:25

Today I'm talking to Jonathan Michael Broyles about his article, "Investigation of optimization techniques on structural-acoustical shaped concrete slabs and buildings," which appeared in the 42nd volume of POMA and is based on a presentation he gave at Acoustics Virtually Everywhere. Thanks for taking the time to chat with me, Jonathan, how are you?

Jonathan Michael Broyles  00:44

I'm good, Kat. Thanks for having me.

Kat Setzer  00:46

So first, tell us a bit about your research background.

Jonathan Michael Broyles  00:49

Sure. My educational research background is in civil structural engineering, in addition to music performance and applied acoustics, in addition to architectural engineering, and my research background really reflects my interdisciplinary educational background. Simply put, our research lies at the intersection of structural engineering, architectural acoustics, sustainability, and computational design and optimization. So the proceedings paper that we're actually going to talk about is a pretty good example of a lot of those topics.

Kat Setzer  01:17

That sounds so interesting, and so relevant to a lot of our current needs in society. So why are you interested in looking at concrete slabs in floors?

Jonathan Michael Broyles  01:26

Yeah, great question. So this was somewhat subtle in the proceedings paper, but one of the major reasons that we're looking at concrete slabs specifically is because of a large contribution to the carbon footprint of a building. The building and construction industry contributes nearly 40% of global carbon emissions, which is very significant, and roughly 10% of global carbon emissions comes from the manufacturing and production of construction building materials. So building designers can directly reduce carbon emissions by reducing the amounts of building material without sacrificing the functionality of what the builder should be doing. So for our context, concrete floors could still be structurally sound, but may not need as much material. However, the acoustical ramifications because of the loss of material hasn't been thoroughly studied. This paper started to look at how different design space exploration techniques can be actually employed to evaluate material changes or surface density savings, which corresponds to carbon savings in addition to acoustic insulation performance.

Kat Setzer  01:27

Okay, that makes a lot of sense. You want to reduce the amount of concrete because of the carbon savings like you said.

Jonathan Michael Broyles  02:39

Right.

Kat Setzer  02:39

But then more concrete probably helps with sound damping a lot, right? Okay, the research you talk about in this paper is related to optimizing the design of rooms for acoustic performance. Can you give us a bit of background on acoustic performance and design? Why is it important to consider acoustic performance early in the architectural design of the space?

Jonathan Michael Broyles  02:59

Sure, yeah. Optimization, in a broad sense, has really existed throughout many engineering disciplines for a number of years, including architectural acoustics. However, the optimization in architectural acoustics may have been to mitigate an acoustics-related issue through the most economical or feasible solution possible. So optimization in this paper refers to more of this geometrical manipulation or this shaping of a concrete floor or a different room or building components. So this is known as shape optimization and we applied this procedure to the case study of a ribbed concrete floor. 

Jonathan Michael Broyles  03:39

Now to answer your question with why do we need to consider acoustic performance earlier, in architectural design of a space, generally it's to help prevents costly acoustical treatments further down the, I guess, the construction timeline of a building, because many acoustical treatments could be very costly if they aren't dealt with in the construction phase, and potentially could result in some negative health consequences of the building occupants. So considering acoustics during early stages can help prevent those negative consequences. It should be noted, though, that the design practice is evolving, and building practitioners are really pushing the bounds of what conventional building design practice is like. So you see a lot of mass timber buildings out there, non-traditional elements like these shaped slabs, just for some examples.

Kat Setzer  04:28

Okay, got it. Yeah. So kind of just trying to save time and energy and the health of people, like you said. So what are some of the techniques currently used for acoustic optimization?

Jonathan Michael Broyles  04:39

Yeah, generally speaking, acoustical engineers use, I would say, modeling techniques, and like computational ray tracing models is a good example of that, that rely on performance convergence. And this convergence is in a way a result of optimization. Specifically, it's a minimizing of an error until a certain threshold is met, like it has to be below a certain threshold to ensure that convergence has been met. So, acoustics have been able to kind of have optimization in a kind of in a different way. But additionally, acousticians have to balance not just a singular building objective, or an acoustic characteristic, they generally have to balance multiple objectives like the reverberation, the clarity, and the loudness in one single room, which is unique to architectural acoustics, and why optimization strategies could be very applicable. This is generally not done by using a formal optimization technique or an algorithm. But the concept is still kind of there in practice. And same with the different design space exploration techniques that we'll get into as well. Acoustical engineers can evaluate a few different designs before ultimately deciding on a specific final design. But very few people actually use, I would say, these more advanced optimization procedures in design practice, potentially missing opportunities for non-traditional and high-performing designs. And the Paris Philharmonic did use some of these more advanced strategies, and it's a very unique concert hall, a very, very fun concert hall.

Kat Setzer  06:16

Awesome. Cool. So first, before we get into the strategies, let's talk about parametric, structural, and acoustic modeling. What are they, and how did you use each of them in this study?

Jonathan Michael Broyles  06:27

Yeah, so I'm going to be pretty, I guess, I'm going to try to simplify it for each of these. But parametric modeling refers to the modeling process, a specific geometric model, that has the ability to be manipulated by changing the shape of the geometry, as dimensional design variables are modified. So in our floor system, for example, we could have a variable that manipulates how thick the top slab is, in addition to maybe the curvature of the ribs or even the rib depth. 

Jonathan Michael Broyles  06:56

For the structural model, in our case study, it's to... it basically refers to the design and analysis of a structural system, specifically the concrete floor. It should be noted that structural modeling doesn't always have to have a building application but in our paper, there is that building application. So the concrete floor in our paper had to meet and ensure structural integrity, so meet design codes, to actually be considered in our design-space exploration techniques. 

Jonathan Michael Broyles  07:25

Now, acoustically...  acoustic modeling broadly refers to like any analytical or computational model that can simulate the acoustic performance of a building element. And again, acoustics, there's so many broader acoustic modeling definitions out there, but ours is more tailored to a building application. In our study, we use an analytical expression based on the infinite panel theorem to numerically obtain just the transmission coefficients across a broad frequency range. And these transmission loss values were then used to obtain the acoustical objective of sound transmission class, or STC.

Kat Setzer  08:03

Okay, got it. What is sound transmission class, exactly?

Jonathan Michael Broyles  08:07

Yeah, sound transmission class is a single integer value that quantifies the airborne sound transmission performance of a building element like a floor or even a wall.

Kat Setzer  08:17

Okay, got it. Can you describe the computational techniques that you looked at in this study? Why did you choose to look at these techniques in particular?

Jonathan Michael Broyles  08:27

Sure. So the three different computational techniques that we explored in our study are Latin Hypercube Sampling, Multi-objective Optimization, and Constrained Optimization. And these three techniques were chosen because they are commonly used in more advanced computational design frameworks, especially within architectural, I guess, computation and application. Second, they provide different resolutions of design-space exploration. And third, they require varying degrees of computational capacities. And I'll explain that a little bit more in these broader definitions. 

Jonathan Michael Broyles  09:04

So, Latin Hypercube Sampling of a design space really is basically looking at random points in a, I would say, a design catalog, to provide insight on design performance of the designs that I'm looking at. So imagine I'm reading a catalogue, I'm looking at a few random, but strategically random, designs to kind of see how they perform for the design objectives that I'm curious about. In our study, we're looking at, again, mass density and STC. So Latin Hypercube Sampling can give us a broad definition while exploring parts of the design space. 

Jonathan Michael Broyles  09:43

Multi-Objective Optimization is a pretty common optimization strategy utilized when evaluating designs based on two or more competing objectives. Again, our study is looking at minimizing surface density and maximizing sound transmission class performance. So we wanted to find a series of high performing designs that had great performance for STC and very low surface density. So these shapes were, again, had the least amount of material, but provided good acoustic insulation performance. 

Jonathan Michael Broyles  10:19

And I want to emphasize a good Multi-Objective Optimization found a series of high performing designs, so more than just one, but Constrained Optimization only converges to one. It strategically investigates the design space to find the single best design when optimizing for one objective, while there's constraints applied for the other objectives. So in our study, we used constraint optimization for different STC ratings like an STC of 50 or 55, to find the one design that had the least amount of surface density, while constraining to that STC rating.

Kat Setzer  10:59

Okay, got it. So it's kind of like, you have varying levels of fine tuning for what you're searching for, I guess, right?

Jonathan Michael Broyles  11:08

Yeah. And the like the the context of this catalog, your being... it... there's different ways that you can read the catalog. And I think that's maybe the best analogy to these explorations. Because Latin Hypercube Sampling is kind of you're just flipping a little randomly, you kind of know what designs work and what designs aren't the best. But Multi-Objective Optimization and Constrained Optimization is a little bit more strategic, you're starting to look at finding those best designs.

Kat Setzer  11:34

Okay. Yeah. So to go with your catalog example, your going to a section of the catalog and looking at every option, or, you know, constraining what section of that catalog is even further, like this page, or whatever. Is that kind of it? 

Jonathan Michael Broyles  11:49

Yeah, yeah, you're starting to get like, there's this subsection of this catalog that are like maybe a high-performing design section versus others that are maybe not so great, but still viable options.

Kat Setzer  11:59

Got it. Okay. So how did you end up employing the design-space exploration techniques?

Jonathan Michael Broyles  12:05

Sure, yeah. So we applied the three techniques using an open access Grasshopper plugin called Design Space Exploration. And that's free on, through rhino.com. So anybody who's familiar with like, I'd say computational design and parametric modeling, Rhino and the Grasshopper interface, that's essentially what we use. So our parametric, structural, and acoustic models were all coupled with the framework to these design-space exploration techniques.

Kat Setzer  12:32

Okay, so what did you end up learning about the different design-space exploration techniques? How good were they at helping you optimize your concrete slabs?

Jonathan Michael Broyles  12:41

Yeah, we learned quite a bit. So I'll start with Latin Hypercube Sampling. And we essentially saw from using this technique that it was a really good first pass for kind of understanding those broad objective trade offs in the design space. So Latin Hypercube Sampling is a good representation of possible designs. However, they don't necessarily... the technique didn't necessarily give us the best-performing designs in this sampling. As a result, we can't, you know, quickly, I'd say ascertain, which was the best-performing design for given STC ratings. However, we again did see broadly, that increased mass density generally increased STC rating. So that was very, I guess, identifiable with Latin Hypercube Sampling. 

Jonathan Michael Broyles  13:29

Multi-Objective Optimization was able to actually tell us more about those, or show us more where those best-performing designs were in the design space. So we used an evolutionary optimization algorithm to converge to a series of best-performing designs, but one of the trade offs was in using this technique, it required more computational time. And I'll maybe preface with, it had to also, we were under the, I guess, constraints of having these simulations run pretty much at the same computational time. So if LHS, or if Latin Hypercube Sampling took five hours, we'd have MOO also run for five hours, because you can run these different optimization and design-space exploration techniques for a while and ultimately get to the same results. But that's very, not practical in design. So, so MOO had more, it required more computational time to kind of get to this set of Pareto front, or these best performing designs. And what I mean by Pareto front again, is just that series of non-dominated or these best-performing designs for two or more objectives. 

Jonathan Michael Broyles  14:39

For Constrained Optimization, we did see that it was the best technique for finding a single best design, but it's very hard again to kind of see broad trade offs in the design space again, because you're ultimately converging to a single design. So to actually get kind of this broad trade off set to find like that Pareto front approximation or a series of best-performing designs, multiple iterations and multiple simulations are needed. 

Jonathan Michael Broyles  15:05

So I guess in summary, we kind of see these different design resolutions, or different ways to explore this catalog with different efficiencies depending on what we need and our design problem.

Kat Setzer  15:17

That's really interesting. What are the next steps in research related to optimizing the design of rooms for acoustic performance?

Jonathan Michael Broyles  15:25

Yeah, research wise, one clear next step is to include not just STC, but also the impact sound insulation objective, or IIC, impact insulation class, in an optimization framework to further explore how this trade off between reducing material consumption and carbon saving affects, you know, acoustic insulation performance for both air- and structural-borne sound. So that's, that's a clear next step. But these design-space exploration techniques could also be used to further explore other architectural acoustics phenomenon, including, like reflections, sound reflections, echoes, the sound absorption of different elements. And this could be applied at different scales, not just like the building-elements scale, but potentially at a full-room level or even a full building. So there's a lot of different ways that these optimization strategies can be used to really find some, I would say, novel acoustical solutions in our field. 

Jonathan Michael Broyles  16:25

And I'll note as well that design practitioners should be... there is a little bit of a ceiling, or I guess, there's a floor to jump on to get, you know, familiar with these computational design techniques. But once you're able to get educated on these techniques, the opportunities to, you know, to further research into, like, I guess, converge towards better performing design solutions is endless. So, I'll also maybe finalize, I'll conclude with, I think the majority of the profession is starting to become more educated on these opportunities, too, which is very exciting.

Kat Setzer  17:00

Yeah, it sounds like we might see some big changes in how we engineer our buildings, hopefully better impact acoustics and better impact the environment and so on. 

Jonathan Michael Broyles  17:09

Right. 

Kat Setzer  17:09

Do you have any closing thoughts?

Jonathan Michael Broyles  17:11

So I'll I'll kind of say this because this, this proceedings, paper, I guess, ties in a lot of seemingly independent, but again, interesting how these all these different disciplines correlate. Like there are many... this is a very multidisciplinary and multifaceted project. And I would say for anybody who is curious about any of these topics, whether that's, you know, building acoustics, computational design, optimization, sustainability, like now is the time to go in to study these topics in graduate school. Like acoustics is becoming a very, very hot topic. There are lots of things you can do with acoustics, and sustainability, or computational methods and strategies and machine learning. There's actually a recent JASA special issue on machine learning, and there's even our current special issue on like climate change and how acoustics, you know, is affected by all these things. So, yeah, I would encourage people who maybe are interested to get some more education on some of these topics.... Go pursue a graduate degree. Now is the time.

Kat Setzer  18:18

Sounds good, amazing. Well, I'm not going to lie. When I first saw this article, I thought it might be a bit esoteric, but it's interesting to learn how researchers are looking for ways to reduce the amount of concrete used in projects since it's so ubiquitous, but not great for the environment, while still keeping in mind that we need other ways to attenuate sound if we're going to have less total concrete. It's a really interesting question that touches people's lives more than they may think, or at least more than I thought, or realized. Thank you again for speaking with me today. And I wish you the best of luck in your future research. 

Jonathan Michael Broyles  18:48

Of course, Kat. Thanks. Thanks again.

Kat Setzer  18:53

Thank you for tuning into Across Acoustics. If you'd like to hear more interviews from our authors about their research, please subscribe and find us on your preferred podcast platform.