Descending in an elevator after a Mother’s Day rooftop brunch, I had a thought I couldn’t shake: “How is this thing getting past the gigantic column of air below?” I stood there, completely still in one reference frame and plummeting in another, my eyes open wide staring fiercely, wondering how I had never thought of this before. I tried remembering if I’d ever felt a gust of wind shoot out from gaps in the lobby door as an elevator descended to pick me up. “I don’t think so?” I wondered if there is a big vent at the top and bottom of the shaft to equalize pressure. But where does all that air shoot out? Wouldn’t it be loud? I struggled to answer these crucial elevator-piston-related questions that you too, dear reader, likely have grappled with on a daily basis, which is why I’ve dug into it a bit so that you and I can finally relax.
I’ll begin by making it clear that this is a car website, and because the metal cubes you stand in while vertically ascending and descending are technically called “elevator cars,” this article is totally justifiable. Plus, in 2016 Jason Torchinsky asserted that an elevator is a vehicle, and when has he ever been wrong?
Anyway, allow me to restate the dilemma: As an elevator moves up or down in a shaft, it has to displace (i.e. move) air. Because it’s acting just like a piston (or plunger), it wants to try to compress air as it moves towards either end of an elevator shaft, creating an area of high pressure in the direction of travel and low pressure in the other direction.
This, as you could imagine, is a problem, because even though air can be compressed, and even though I’m sure the air on the backside of an elevator is happy to pull a vacuum, this doesn’t happen easily. There’s quite a bit of energy needed to try to compress fluids. And in fact, this energy expense is pretty much how shock absorbers work. Basically, as a piston in fluid moves, it either compresses the fluid or it has no choice but to move the fluid. Compressing the fluid would essentially slow the piston to a stop, and moving the fluid takes away energy/slows things down considerably (which is the point):
So in the case of an elevator, obviously the fluid (air) isn’t going to be perfectly sealed into the shaft (we’ve all seen the gaps in elevator doors), but still: The fluid is being forced to move by the elevator, so isn’t there a bunch of energy being wasted like in a shock absorber?
My search for answers revealed that this “piston-effect” issue that was on my mind is actually way more serious than even I initially thought.
The Piston-Effect And The Smoke Problem
John Klote — a world-renowned smoke-control expert — wrote An Analysis of the Influence of Piston Effect on Elevator Smoke Control, a 1988 U.S. Department of Commerce paper that discusses the “feasibility of using elevators for the evacuation of the handicapped during a fire.”
As it turns out, the piston-effect of elevators has been considered a humongous problem in a fire, as it tends to pull smoke into the elevator shaft, where it can then be sent to other floors, as Klote notes:
The transient pressures produced when an elevator car moves in a shaft are a potential problem for elevator smoke control. Such piston effect can pull smoke into a normally pressurized elevator lobby.
Using A Fan To Mitigate The Piston Effect
The paper goes on to mention those giant elevator door caps we all know, and it talks about a potential fan system (shown above) that could help with this smoke issue caused by elevator piston-effect:
Most elevator doors have large gaps around them [5]. Such large leakage areas around the doors result in lobby and shaft pressures that are nearly equal under most conditions. Thus if pressurization air is supplied to the elevator shaft, the lobbies will be pressurized indirectly to almost the same pressure as the shaft. A concern with such systems is that a few open doors might result in significant loss of pressurization. The first paper [2] of this project demonstrates that this problem can be overcome by use of a system with feedback control. The flow rate of air into the shaft is controlled by a differential pressure sensor to maintain a constant pressure difference across the elevator lobby door on the fire floor. One method of varying the flow rate is a fan bypass system.
The plot below is cool because it shows the pressure difference between the elevator shaft and the lobby as a function of elevator car speed, and in this case the results are for a pressurized system that is trying to minimize the piston effect via a fan that force air at 25 PSI below the elevator car so as the car goes up it’s not pulling a vacuum from the lobby.
A Motorized Damper ‘Hoistway Vent’
It seems to me that such a pressurized system is not required by law, though “motorized dampers” used to control an exhaust valve at the top of an elevator shaft aren’t uncommon (the image below also shows a fan for odor control):
To learn more, I read the state of Wisconsin’s document from the Department of Safety and Professional Services. Titled Ventilation Requirements for Elevator Hoistways and Machine Rooms, the paper actually mentions the piston-effect in a positive light, specifically because the airflow around the elevator as it moves can help keep things cool:
The piston effect of the elevator moving in the hoistway and the opening and closing of the hoistway doors is often adequate for conditioned building air to maintain temperature and humidity to meet A17.1, 2.7.9.2.
It also mentioned both pressurized shafts and damper valves (hoistway venting) to control airflow:
The commercial building code no longer requires hoistway venting for control of smoke. Other means of smoke control may be required by the building code such as elevator lobbies or smoke doors or curtains. The DSPS Website has an article similar to this with information about smoke doors and curtains.
Hoistway Pressurization for Smoke Control The commercial building code IBC 909.21 allows hoistway pressurization as a method to control smoke. A pressurization system must be tested according to IBC 909.21.1. The hoistway or an adjacent space must be large enough for the ductwork necessary for pressurization and pressurization may not negatively affect elevator equipment in a hoistway that may be sensitive to air movement per A17.1, 2.1.4.
Consulting-Specifying Engineer (a publication by and for engineers), actually says damper vents were indeed required as of 2005:
Incorporated into elevator design in the 19th century, and well established in the codes and standards for many decades, vents have been placed at the top of hoistways to prevent excessive pressure during car ascent, control odors and vent smoke during a building fire.
The International Building Code (IBC) currently requires hoistway vents. These must be located at the top of the hoistway and either open directly to the outside or be routed to the outside in noncombustible ducts that are fire-rated the same as the hoistway.
Originally, these vents were passive openings. But with the rise of energy conservation, vents now are often equipped with mechanical dampers that open upon smoke-detector activation. IBC Section 3004 requires hoistways of more than three stories to have vents of 3 sq. ft., or 3.5% of the area of the hoistway.
But as of the 2005 publication of the linked article, Change is in the air, there was pushback on the need for such vents, and part of that had to do with the piston effect not actually being that big of a deal, per CSE:
One major step toward remedying the smoke-migration problem is realizing that the reasons given for placing vents at the top of hoistways have lost their validity.
Take, for example, the argument that an ascending car produces pressures that require venting for proper car operation. In fact, the pressures developed by elevator car movement in a shaft—called piston effect —are small. A downward-moving elevator car will force air below the car into the shaft above the car. Additionally, air leakage around elevator doors on each floor is significant.
Moreover, piston effect varies with the number of cars in the hoistway and the hoistway area. For example, for a single elevator car traveling at a velocity of 400 ft. per minute (fpm), there is a pressure differential of 0.08 in. H 2 O. For a double-car shaft, it is only 0.02 in. H2O for a car traveling at 400 fpm. In a double-car shaft with a car traveling at 700 fpm, the pressure differential is only 0.05 in. H2O.
Check that out; it turns out the piston effect isn’t as big of a deal as perhaps it once was, and renowned Swiss elevator company Schindler has spent considerable effort minimizing it.
Schindler Talks About The Piston Effect In-Detail
Noise Is A Major Problem
“Depending on the speed, usually faster than [about 8 mph], the piston effect [of an elevator car in a single shaft] becomes an issue. Airflow and turbulence around the car can also generate unpleasant noise and slight vibrations.” Here’s a nice screenshot from the Schindler video showing the situation:
Fairings To Smooth Airflow, Especially Above 8 MPH
Schindler says it has developed special fairings to “smooth the airflow” to mitigate the piston effect due to airflow struggling to get around the cars. Between that and sealed doors, Schindler says its elevator cars are quite quiet:
With that said, the landing doors — i.e. the doors at each level — can also leak air as the elevator car compresses it, and this can make for some loud landings:
Schindler says it’s sealed those doors, too, to keep the landings quiet.
One Key To Mitigating The Piston Effect: Vents Between Multiple Elevator Shafts
The video goes on to mention something quite interesting, and something I hadn’t really considered. On buildings with multiple elevators (like most really tall buildings with high elevator speeds — which per my research usually peaks at around 25 mph), the elevator hoistways are connected via vents!
So that means, even if the elevator my wife and I were on was trying to squish air below us as we descended, it was almost certainly shooting that air into a neighboring elevator hoistway via a vent:
There are situations where this wouldn’t work out well, like if two elevators in a two-elevator building were parallel and ascending/descending at the same time, but per Schindler, its control system doesn’t allow this.
What’s more, as Schindler notes, car-to-shaft ratio is key. “They provide sufficient space, especially in a single shaft, for the air to flow around the car.”
So What’s The Takeway About The ‘Piston Effect’?
So that was a deep rabbit-hole I just went down, but it was an important one. I needed to know the answer to how an elevator can fight against all that air, and it turns out: There are a few methods. There are exhaust vents, there are vents between parallel elevator shafts, there are fans, there are fairings, there are just nice big gaps between the elevator and the h0istway walls.
Finally I can sleep at night. This rabbit-hole was worth it.
Top Image: Schindler
As an architect I’ve never thought about this, because A. I’ve known since the very start* that there are significant gaps between the car & the shaft (a small car is only about 2/3 the size of the shaft), so it never struck me as that piston-ey, and B. I’ve never worked on an elevator shaft taller than 3 stories, plus they’ve mostly been (drumroll) piston-driven, which is to say: slow.
So now I know that this is an issue to be dealt with in taller buildings, but also that passive measures can handle things (I like the fairings). Good stuff!
*one of the very first things I worked on out of school was an elevator retrofit for a historic library
I’ll admit that I kind of hoped that “when has he ever been wrong” was going to be links to 6 articles about Torch insanity and/or saying things that David definitely doesn’t agree with but passes along out of friendship and/or the demands of the content gods.
Careful, you risk opening the “is it more energy efficient to live in a stand alone house or an apartment block” question.
Apartment blocks say houses waste space, need more heating, are symbols of capitalism etc.
Houses say as soon as you need an elevator the energy efficiency arguments tilt in favour of houses. Add to that the need for controlled air and ventilation in blocks, and the environmental factors lean even more towards houses.
And elevators are transport systems in French law, the most commonly used transport system too.
Fun fact: to reduce pumping losses inter-cylinder vents are sometimes added to engine cylinder blocks.
I added that to the engine in my Merkur once. Introduced via an emancipated connecting rod.
It’s high time we brought back the perilous ‘man engine or the Paternoster lift’: https://en.wikipedia.org/wiki/Man_engine
https://en.wikipedia.org/wiki/Paternoster_lift
You cannot compress fluids. Gases yes, fluids no.
Both are fluids but different states of matter. “A fluid is a liquid, gas or other material that may continuously move and deform under an applied shear stress or external force”
Gases are fluids, mate.
This is like the opening statement in every Fluid Mechanics 001 class.
Yeah, and when I repeated “fluids are incompressible” at my engineering job I got crucified. A little education is a dangerous thing.
More a language issue. I read “fluid” as “Flüssigkeit” (liquid) but that is obviously not what is meant.
To be fair, the common English meaning of the word ‘fluid’ for most people is ‘liquid’. But in physics/engineering it has a specific meaning which includes liquids and gases.
Liquids are compressible, just nothing like as compressible at gasses.
In fact everything this side of Neutronium is compressible, including hydraulic oil, steel and diamonds.
Compression is how sound waves pass through materials, so it’s pretty easy to demonstrate the compressibility of liquids and solids.
Liquids can be considered to be non-compressible for all practical intents and purposes. Yes, sound waves are longitudinal so their propagation relies on compression and expansion of the carrier medium, but liquids are not compressible at scale. Just consider what happens when water enters the cylinders of a combustion engine.
I design engines. One of them had a variable compression ratio device that could run up to 40:1. We had a design that operated that via a hydraulic ram, but the compression of the hydraulic fluid made a significant difference to the position of the doodad, and therefore the actual compression ratio under combustion loads.
Fluids being what they are the degree of movement depended of the total system volume, so having longer hoses to a remote actuator would have made it significantly worse.
So we ditched hydraulics and had to go fully mechanical instead.
Hence my sensitivity to the word “incompressible”.
My question is how do subways work!? Those trains are much closer to pistons. Sure, you can feel a blast of air as the train approaches, but it’s not continuous, so there must be regular vents in those long tunnels?
Regular vents and lack of piston rings. Think the famous Marilyn Monroe picture for an example of the street level effect. There are also regular crossovers for maintenance and emergency use, as well as the stations themselves provide paths for air to get around the train.
The World Trade Center had an amazing elevator system with multiple cars per shaft. and adjacent shafts pumping air into each other. Also, the express elevators were really fast. It almost always took two or three cars with transfers to get to your floor.
Unfortunately, they became a little too piston-like on 9/11, but that was hardly a design flaw.
Interesting…now I’m curious if there are any differences to the elevators on cruise ships
Two words: Schindler’s Lift
I’ve worked with Schindler, Otis, and Thyssen-Krupp for decades, yet I never thought of this quip.
Shame on me.
Elevator shafts are quite complicated from a fire protection standpoint…its one reason why that once a fire starts…don’t use an elevator.
Why…
Building code has pages and pages on elevator shaft requirements. Fire management issues for elevators is well studied and understood.
Exit stairwells are deigned to specifically pressurize with outside air to permit safe egress during a fire. Also…doors to exit stairwells must remain closed and are air tight…so that the pressurized air does not feed an existing fire.
Rabbit hole for sure.
Noticed the elevator to the underground garage always returned to the lobby and locked its doors open (with a significant breeze from the open doors) when the fire alarm went off. Figured building/fire code was the reason, but makes sense from a smoke management perspective. I assume the elevator shaft was getting positive pressure from an auxiliary/emergency ventilation fan.
Interesting observation.
Separating the parking elevators from the main is typically done for security reasons.
Sometimes building owners rent out parking spots to folks who do not live/work in the building.
Separating the elevators provides access control for this type of building.
So … elevators kind of top out at 25 mph (37-ish in Taiwan).
But, why – oh why?! – are 2- to 3-story motel elevators sooo slooow, you could ride all the way up and down Taipei 101 before reaching the next floor?
Low rise buildings use hydraulic elevators, which are pushed up and down with a big fluid-filled piston.
They are slow as fuck, but.. as you guessed it, also the cheapest.
The Flatiron Building in NYC was the tallest in the world at 22 stories, or 285 feet tall, and originally was equipped with Otis Hydraulic elevators. Additionally, it only had one restroom per floor, all of which were designated for men until the men were assigned the even floors and the women were assigned the odd-numbered floors.
The hydraulic elevators were replaced in 1997. I don’t know when the women got restrooms.
The Sunsphere, a relic of the 1982 World’s Fair in Knoxville, Tennessee (and most famous for its role as the Wigsphere in “Bart on the Road” episode 20 in season 7 of The Simpsons https://en.m.wikipedia.org/wiki/Bart_on_the_Road), has a hydraulic elevator which has or used to have signs cautioning passengers not to jump or otherwise jostle the elevator because then it will get stuck. I’ve been told that it’s because the elevator is hydraulic but I’ve never seen such signs in any hotel/motel elevators.
While it’s indeed fairly small there’s just something about that elevator that somehow seems more claustrophobic than most elevators. Unfortunately the stairs are not generally open to the public, alas. Those of you wanting to visit the Wigsphere, beware!
This is all well and good for normal up/down elevators. But what about those fancy angled elevators in The Luxor?
Or the Wonkavator?
And the turboshafts in Star Trek.
The Gateway Arch in St Louis, with its cable car elevators – Lots of airflow around those. But they might be more like a ferris wheel than an elevator/piston
Elevator shafts have never been precision honed. There’s a lot of space in there for wires, plumbing and even ladders, which all provide more than enough space to equalize the pressure above and below the car.
I guess it never would’ve occurred to me that an elevator would be fit tightly enough in the shaft for this to even be an issue.
Schindler?! I’m more of an Otis man.
That’s about what I assumed, but I never thought about the elevator drawing smoke and fire as a reason not to use them. I always just figured it was to prevent people from getting trapped as locating and rescuing would be a lot more difficult than a stairway.
Having lived 3/4 the way up in a 19 floor apartment building for 20 years and having worked in high-rise offices for many of those years – I must say….
….I never until today gave this a moment’s thought.
“… in 2016 Jason Torchinsky asserted that an elevator is a vehicle”
An elevator is also a room.
If you’ve ever watched Neal Simon’s “Murder by Death” (1976) you’ll know what I’m talking about.
(And did you ever notice the cars in that film?)
If you want an even deeper rabbit hole, look into high speed elevators for ultra tall buildings. For example the express observation deck elevators in Taipei 101 run at up to 60kph and actively manage the air pressure within the cabin.
Kind of funny that the 25 mph or ~11 m/s of a quick elevator is right around the upper end of mean piston speed of a typical commercial vehicle engine, just without the constant, high acceleration from reciprocation.
Just install the piston rings on the elevator with the gaps lined up to let the air through…
Highest mean piston speed in a production engine is 26.9 m/s (Audi R8 V10). That’s 60.17mph.
I love this stuff. It is always amazing how much information there is to be learned about basically everything we use day to day.
An architect friend has always said that elevators are always huge deals when designing a building because they are basically like adding chimneys for any potential fire, and airflow is a big part of fire safety in building design.
One other interesting thing that impacts all of this is the desire to use every square inch of a commercial building. Increasing the gap between the elevator and the shaft would be easy and solve the piston issue. But on a 30-story building, increasing a shaft from 12’x12′ to 13 ‘ x13’ could mean $25k less in rent every year per elevator shaft.
I once sat in a meeting where the developer was exploring ideas for fire suppression systems (with an eye to saving money, of course). The proposal was to include a rooftop pool so obviously someone suggested that the pool also function as a reservoir for a wet suppression system. I don’t know if it ever came to fruition since there are other interfacing systems to be concerned with, but the fire department at least was on board.
Very cool!
I used to install residential elevators, which tend to be simpler than commercial units, so my experience might not be valid. However, at least on a residential elevator, there is nothing akin to piston rings to prevent fluids or gasses from passing right by the car.
h0istway would be a great industrial band name.
Fun fact – in South Carolina elevators are inspected and certified by the “Bureau of Amusement Rides and Elevators” or something similar. Always gives me pause when I see it on the certicificate when in state.
BARE?
More concerned about the amusement ride aspect of it.
Curious if you’d care to further articulate your concern. Most amusement rides use components similar if not identical to those in elevators, as well as sharing similar concerns for rider safety. I had thought expertise to inspect and certify such equipment would easily translate between those two fields but I am happy to learn more if I’m mistaken!
Related side note: Disney’s “Twilight Zone Tower of Terror” ride (now Guardians of the Galaxy: Mission Breakout) isn’t just similar to an elevator, it IS an elevator. Albeit one intentionally built to ride neither smoothly nor comfortably. Reportedly Otis the elevator company was involved in its engineering design.
Oh, I’m an engineer. Intellectually I know all of this, emotionally I do NOT want my elevator to impersonate an amusement ride.
Any elevator can become an amusement ride in the right circumstances.
Elevators tend to break down quite frequently. It makes me wonder if amusement parks inspect and repair their rides at night while the park is closed. If Otis makes amusement park rides that break down as often as elevators, that would be a reason to not ride them.
Because you’ve heard horror stories of uninspected rides causing serious injuries or worse? Cuz sorry to break it to you but most elevators in the US are way past when they ought to have been inspected too, whether that falls under the same department as amusement park rides or not.
When New York was building their giant sightseeing ferris wheel (in Staten Island!!!) the developer had a hard time figuring out which NYC Department of Buildings division they should seek approval from. Ultimately it went to the elevator division.
It shall come as a no surprise that the project died.
Some elevator inspectors have cult followings!
Not to be ‘first’ w/o reason, I’ve bookmarked this article because I’m sure I’ll want to read it carefully, but have to get offline and do some actual physical things before the entire day gets away from me. 😉