DIY Intercooler Water Spray Testing

KERMA

Vendor , w/Business number
Joined
Sep 23, 2001
Location
here
TDI
99 beetle and 04 jetta
the biggest problem is cooling a very large volume of air very quickly. Assume surface area of 1ft2

60 mph = 100 ft/sec
x 60 sec = 6000 cfm
this is about 1 lbm/min of air passing the intercooler @ 60 mph at STP (less at altitude)

It all comes down to how quickly you can cool 100ft2/sec @60 mph for example. And then you still have air close to ambient, and it's still air. 0.001225 g/cm3 vs 1 g/cm3 for water

and everyone who has ever taken a thermodynamics class in college knows mc delta t so even if c is the same for water and air (which it's not) then air is already 1/800 as effective as water so even if you managed to cool it a couple degrees is how effective...

If you kept the intercooler wet maybe more effective, but then why not have air/water IC at that point. When I did these experiments it was most effective if you drenched the intercooler but the drag strip did not like that at all.
 

CraziFuzzy

Veteran Member
Joined
Sep 1, 2011
Location
Jurupa Valley
TDI
'09 JSW (GoneBack) - replaced with '15 Azera and '16 Fiat 500e.
A fine mist spray into hot dry air will very rapidly cool the air down, at a diminishing rate as the air approaches saturation. The rate at which the water evaporates is based primarily on the surface area of the interface between the water and the air and the dryness of the air (the amount of water the air can absorb). Water collecting on the surfaces of the intercooler, where it would pool into larger droplets will not evaporate faster than the actual mist would. At best, if you have enough stay time between mist entering the hot dry air, and the humid cool air hitting the intercooler, you are looking at getting the air temp down to about 5-7 °F above ambient wet bulb temperature.

Ideally, instead of spraying water onto the cooler, you probably will get better effect, and certainly better use of each ounce of water, by spraying as fine a mist as possible into the incoming air charge - if you can duct this air to make sure you get all the cold humid air you are making over the intercooler, that's even better. Spraying the mist INTO the wind (away from the intercooler) would also increase the stay time, as well as increase relative velocity between water and air.

In a ducted mist or steam humidifier, the mist or steam is most often sprayed upstream in the duct for this very reason.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
...Edit: Oh, specs for the intercooler...
188 mm wide, 210 mm tall, 88 mm deep
12 rows of fins 205 mm long with 115 fins each, aka ~14.25 fpi...
the biggest problem is cooling a very large volume of air very quickly. Assume surface area of 1ft2

60 mph = 100 ft/sec
x 60 sec = 6000 cfm
this is about 1 lbm/min of air passing the intercooler @ 60 mph at STP (less at altitude)

It all comes down to how quickly you can cool 100ft2/sec @60 mph for example. And then you still have air close to ambient, and it's still air. 0.001225 g/cm3 vs 1 g/cm3 for water
I did post the frontal details of the intercooler though. With 25 equal segments across the intercooler face, 12 of which will let air flow through, we have ~29.4 in^2 to work with, which gets us ~1167 cfm @ 65 mph with the engine turning over at 2500 RPM +/- 21 RPM. On the stock camshaft, the VE is in the neighborhood of 86% from the data I have.

However, there is the limitation to the ducting as the opening is only about 19.4 in^2 so I'm not sure how to determine the amount of air passing over the intercooler.
and everyone who has ever taken a thermodynamics class in college knows mc delta t so even if c is the same for water and air (which it's not) then air is already 1/800 as effective as water so even if you managed to cool it a couple degrees is how effective...

If you kept the intercooler wet maybe more effective, but then why not have air/water IC at that point. When I did these experiments it was most effective if you drenched the intercooler but the drag strip did not like that at all.
It's a good thing that even at 60 ml/min, the frontal area of the intercooler is saturated then, no?
...I know how much coverage I have due to a wet intercooler ;). After nozzle placement is adjusted and tested, I'm going to bump down to the smallest size to see what the effect is. Currently I'm seeing about 2-3 °C drop in IATs at 75 mph max but a better cool down time after boost is evident too...
On that point, I went to town on cleaning the intercooler and given how slowly water "drains" from the fins due to adhesive forces, It seems that now quite flooding the IC, but keeping an even coat of water is fairly doable. More on this in a sec.

Psychrometrics, a topic I've been acquainting myself with since I started this.
A fine mist spray into hot dry air will very rapidly cool the air down, at a diminishing rate as the air approaches saturation. The rate at which the water evaporates is based primarily on the surface area of the interface between the water and the air and the dryness of the air (the amount of water the air can absorb). Water collecting on the surfaces of the intercooler, where it would pool into larger droplets will not evaporate faster than the actual mist would. At best, if you have enough stay time between mist entering the hot dry air, and the humid cool air hitting the intercooler, you are looking at getting the air temp down to about 5-7 °F above ambient wet bulb temperature.

Ideally, instead of spraying water onto the cooler, you probably will get better effect, and certainly better use of each ounce of water, by spraying as fine a mist as possible into the incoming air charge - if you can duct this air to make sure you get all the cold humid air you are making over the intercooler, that's even better. Spraying the mist INTO the wind (away from the intercooler) would also increase the stay time, as well as increase relative velocity between water and air.

In a ducted mist or steam humidifier, the mist or steam is most often sprayed upstream in the duct for this very reason.
Your statements come out to ~2.8-3.9 °C. This matches up, generally with saturating the air to 70-80% relative humidity, which happens to be where I was expecting diminishing returns. Thanks for confirming.

I extended the nozzle out about 3.5-4", ball parking because I didn't measure the pipe extension. After getting home from steady-state testing, et al, with no more than ~70% relative humidity, I'm quite happy with the results. The back of the intercooler was visibly wet on all but the upper-inner corner, the hardest place for air to travel to from the ducting setup. I'll aim the nozzle about 10-15° higher for the rest of the resting, though it seems that I have no less than 90% coverage at the current state. At least for sub-60 mph driving, which is about the final 2 minutes from the freeway to my driveway, 60 ml/min of flow @ 19-20 °C w/ 70% RH is enough to saturate the surface of the intercooler. My tire was completely dry though. I'm seriously considering switching to the smaller nozzle to see if I can get evaporative cooling.

I did a steady-state test tonight. I'll have the log/graph posted up by 0130 PDT. There is a definitive difference between the sprayer on and off even for a pressure ratio as low as 1.2-1.3. I couldn't get to ambient temps under these conditions, though that I suspect is likely a result of having the water settle on the intercooler and not readily evaporating. Truth be told though, I have a mixture of water, isopropanol, and a light chemical cleaner for the aluminum so I wasn't dealing with a known ratios of fluids with specific heat capacities. The server has about 6-ish min before the email routines render it inaccessible, next post from me will be a graph of the log.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
For anyone interested, this is the final thermistor calibration curve with respect to the MSA15.5 ECU:


I know I didn't run these identically the same, RPM shows where, but the effect is greater than statistical error. Because of that, when I was doing the same stretch of road with the sprayer on, I opted to continue up the hill a bit longer. I stopped at an altitude of ~2000' and the stretch rose from about 910'.

Click for higher resolution
 

nicklockard

Torque Dorque
Joined
Aug 15, 2004
Location
Arizona
TDI
SOLD 2010 Touareg Tdi w/factory Tow PCKG
Nice work and very intriguing.
Best droplet size pattern of homogeneously small droplets with low surface tension probably is a mix of:

  • Distilled or deionized water
  • 2 - 10 drops Dawn ultra dish detergent per gallon
  • 5% - 15% isopropanol (volume %)
  • (and of course) small nozzle at highest practical pressure
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
I went looking for literature about surfactants and surface tensions but couldn't find any. 20% by mass of isopropanol in water seems like a good trade off from that one study in '95. I forget the units off hand but it drops water from 72 to 30 whereas isopropanol is ~22.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Weather was mild today. I'm looking for 80 °F, 27 °C, or hotter with less than 30% RH for testing. In other news, some electronic hardware showed up today so I can start working on a control system. If I can't get less flow from the nozzle then I'll need to have some on-off functionality and duty cycle it.

Part of the hardware is a Raspberry Pi Zero and I plane to pick up an IR camera to help with the sensing and figuring out how to optimize everything when it's hooked up.
 

nicklockard

Torque Dorque
Joined
Aug 15, 2004
Location
Arizona
TDI
SOLD 2010 Touareg Tdi w/factory Tow PCKG
I went looking for literature about surfactants and surface tensions but couldn't find any. 20% by mass of isopropanol in water seems like a good trade off from that one study in '95. I forget the units off hand but it drops water from 72 to 30 whereas isopropanol is ~22.
Scholar.google.com search: https://www.researchgate.net/profil...m_sprayers/links/53f5b5d20cf2888a7491cfcf.pdf

and references articles at end of p.2 into p.3.

Read that section. It gives some basis for understanding the effects of droplet sizes due to surface tension changes. Short version is: lower surface tension means a sheet or stream thins out longer prior to breakup into droplets. Once droplet's form, they're smaller and tend toward greater uniformity of size.

It has an added benefit of keeping road grime, oil films, and soil from sticking to intercooler too.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Thanks Nick!

For everyone who doesn't live in SoCal, we've been expecting a heat wave that starts about now. These temps are normal for August, but not June. We're looking at 90 °F as the lowest high the next few days. It is currently 91 °F @ 16% RH right now. wonderful conditions for the mister :). The hottest day will be monday with an estimated high between 105 °F and 108 °F.

I hope to have logs, and possibly video from the intercooler exhaust, for your viewing pleasure within the next week. SoCal traffic is my limiting factor for repeatable testing.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
A mix of 32 floz of 91%, by volume, isopropanol mixed with 1 gallon of distilled water producing approximately a 20% isopropanol and 80% water molar solution was used. The solution's temperature was not directly taken before the experiment was conducted, but can be estimated to be a few degrees below ambient @ 37-42 °C.

On 06.18.2016 I drove from 36 °C w/ 10% RH to 35 °C w/ 30% RH without the use of the sprayer. Cruize control was not used and I'm just posting the IAT graph currently:


On 06.20.2016 I drove from 44 °C w/ 9% RH to 39 °C w/ 11% RH with the use of the sprayer and at 100% duty cycle. Cruize control was not used and I'm just posting the IAT graph currently:


I'm currently trying to figure out a route that I can use and drive on two days in a row with near equal conditions that will produce repeatable results. I have picked up a lower flow nozzle for testing to see if reducing the water volume can yield better evaporation and then possibly cooler results, but I won't be surprised if this effect is rendered moot.

Current observations from the driving and testing I've done are this. As only evident by a faster responding thermistor, IAT's are buffered by the intercooler, big or small. A longer intercooler give more time for this buffering to occur though I'd expect greater heat transfer from a long and narrow one than from a short and wide one. That guesstimate aside, what is being controlled through this experiment is the buffering capacity of the aluminum. No immediate results will discernable by suddenly turning on the sprayer. This is caused by the fact that the heat has to travel through the aluminum core and that is inherently a slow process when compared to the mass of air traveling through to the engine. However, temperature variations will be reduced and the mean temperature will definatively be lower. Figuring out how much lower requires tighter controls on driving conditions.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
Note: Due to resting air and self-heating properties of the thermistor, the log was begun with the engine off and at rest. Several seconds of data were captured after the engine was turned over to allow for the thermistor to register ambient temperature. This is the first flat spot on both graphs before any temperature change occurs from driving and is equivalent to current ambient air temps. I started in a cooler garage for the 2nd test ;)
 

nicklockard

Torque Dorque
Joined
Aug 15, 2004
Location
Arizona
TDI
SOLD 2010 Touareg Tdi w/factory Tow PCKG
Hi Digital,

I suggest to pick a nice long stretch of uphill with no traffic so you can really stress test.

Then run an alternating test sequence to minimize heat soak biasing:

control1
test1
control2
test2
control3
test3
control4
test4

capture datalogs as text file and pass to me. I'll find area under curves and compare tests to controls.
 

john.jackson9213

Top Post Dawg
Joined
Mar 14, 2007
Location
Miramar, Ca. (Think Top Gun)
TDI
1996 B4V
Hi Digital,

I suggest to pick a nice long stretch of uphill with no traffic so you can really stress test.

<snip>
Cajon Pass will provide the nice, long stretch of uphill. It is the main route from L.A. to Vegas. Rises to 4100 feet from just under 1000 feet? But no traffic is a problem in Southern California? Especially where D.C. lives - not going to happen. The 215/15 interchange has about 150,000 vehicles per day and includes 21000 trucks per day.
Even with all the traffic, that section of freeway should be excellent for testing.
 

Digital Corpus

Top Post Dawg
Joined
Mar 14, 2008
Location
Ontario, California
TDI
'97 B4 w/ 236K mi body, 46K mi soul
The closest non-traffic road I can work with is GMR, Glendora Mountain Road. The west-to-east road to Mt Baldy can work too and then there is Big Bear, but I'm not familiar with it's traffic and would wager that it is worse. The Cajon Pass is best, but driving it at night is easiest for stable temp, relative humidity, and traffic, though the former two are less desirable around then. I have a PM out to Nick and I'd appreciate any email or PM correspondance to help figure out a testing regiment and location. This Google Maps link will load up my city dead center for anyone who wishes to chime in.
 

CraziFuzzy

Veteran Member
Joined
Sep 1, 2011
Location
Jurupa Valley
TDI
'09 JSW (GoneBack) - replaced with '15 Azera and '16 Fiat 500e.
The closest non-traffic road I can work with is GMR, Glendora Mountain Road. The west-to-east road to Mt Baldy can work too and then there is Big Bear, but I'm not familiar with it's traffic and would wager that it is worse. The Cajon Pass is best, but driving it at night is easiest for stable temp, relative humidity, and traffic, though the former two are less desirable around then. I have a PM out to Nick and I'd appreciate any email or PM correspondance to help figure out a testing regiment and location. This Google Maps link will load up my city dead center for anyone who wishes to chime in.
Most the roads up and around Big Bear and Lake Arrowhead would be far to unpredictable, and frankly, are far too curvy to get repeatable tests. Cruise Control is certainly not an option up there.
 
Top