# Air-to-Water Heat Exchanger - Car radiator

organiclatte
TL;DR Summary
I have a steady 16KW~ 17.5KW energy source summered in a liquid with a thermal conductivity of .1396 and a specific heat of 2.054 @ 40C°. I want to use an air-to-water heat exchanger to keep the liquid below 50C, with a worse-case scenario ambient temperature of 32 C. How do I select the proper radiator?
I'm looking for a way to keep a liquid under 50°C. This liquid has a thermal conductivity of .1396 and a specific heat of 2.054 both at 40°C. and a constant energy source submerged in it ranging from 16KW to 17.5KW.

I'm barely starting to read about this topic, according some forums, a car generally transfer 1/3 of the engine power through heat to the coolant that gets circulated, this would mean that even for a small car engine ( say a 90's corolla ) the radiator would have to dissipate at most 24 KW, assuming that the engine produces about 100 hp.

My questions are, is the previous scenario correct? How do you calculate the amount of heat any particular radiator can dissipate? Would an liquid-to-air heat exchanger be the right solution for my original problem? If not what would suggest?

Sorry for the vague question, but hopefully someone can help out and point me in the right direction. Thanks!

Welcome to PF.
Can you think of a good use for 15 kW of waste heat?
Are you in tropical or temperate environment?
That may drive the answer to the question of technique and size.
Cross-flow heat exchangers might work better with water as the coolant.

organiclatte
organiclatte
Welcome to PF.
Can you think of a good use for 15 kW of waste heat?
Are you in tropical or temperate environment?
That may drive the answer to the question of technique and size.
Cross-flow heat exchangers might work better with water as the coolant.

We are in a very tropical environment, high humidity, high temperature, however, I've been monitoring the temperature under shade ( where the exchanger would be located ) and the highest temperature I've recorded is 33°C.

I thought briefly about cross-flow, since we have an above ground 5000L plastic water tank in the property. if we were to use that as our coolant I would have to figure out how much heat the tank itself dissipates over certain amount of time, accounting for potential usage of the water by people living the property, right? If we used the energy to heat this reservoir it might be better to get a larger one and maybe install it underground, correct?

I would like to go with the cheapest up-front solution which is why was looking into liquid-to-air, however, this could also be an option.

Last edited:
A car radiator is designed to operate with a high airflow. Maximum power is probably at maximum road speed, so a higher airflow is available. Also the car coolant will be above 80°C, so there is a big temperature difference, (80°C–35°C = 45°C). You have a lower temperature difference, (50°C–32°C = 18°C), so you will need to compensate by having three times greater exchange area with a forced airflow. There will also need to be a fluid circulation pump.

If water was free you might consider a heat exchanger and an evaporative cooling tower.

This is the sort of problem faced by cryptocurrency miners.

Gold Member
It's extremely complex to calculate the heat transfer in an object like a car radiator.

Here you can find a very simple article that lays the basis of the different heat transfer processes. The formula for heat transfer by convection is usually the most important one with a car radiator. Seems pretty simple right? You just need to know the heat transfer coefficient and you are in business, right?

On this Wikipedia page, you have some equations to determine those coefficients for very simple cases. Not so simple anymore. Applying those to a car radiator is even worst.

Sadly, radiator manufacturers don't usually give those coefficients easily, but you can try and ask them. They most probably determined them with experimentations rather than calculations anyway. Which is something you might want to consider for yourself.

Also, you will need to know more than "I want to keep a liquid under 50°C". You will need to know what is the initial temperature of the liquid, to begin with.

organiclatte and Lnewqban
It's extremely complex to calculate the heat transfer in an object like a car radiator.
I agree.
I came across an interesting book last week that deals with liquid / air heat exchange.
It gives a listing of HETE, written in standard C, for simulation of compact heat exchangers.
More than half the book is graphs and data tables.
Title; Thermal Performance Modeling of Cross-Flow Heat Exchangers
Authors; Luben Cabezas-Gómez · Hélio Aparecido Navarro · José Maria Saíz-Jabardo
2015. Springer. ISBN 978-3-319-09670-4

organiclatte and jack action
organiclatte
A car radiator is designed to operate with a high airflow. Maximum power is probably at maximum road speed, so a higher airflow is available. Also the car coolant will be above 80°C, so there is a big temperature difference, (80°C–35°C = 45°C). You have a lower temperature difference, (50°C–32°C = 18°C), so you will need to compensate by having three times greater exchange area with a forced airflow. There will also need to be a fluid circulation pump.

If water was free you might consider a heat exchanger and an evaporative cooling tower.

This is the sort of problem faced by cryptocurrency miners.
Correct, I've encounter many systems that are tailored for this purpose (crypto mining), the most affordable ones are, I think, oversized radiators, if that's the case, this systems are a little over-priced for what they are.

This particular Exchanger is rated for 40KW dissipation at 35C and a capacity of approximately 40L ( assumed by its weight difference when dry and full ( 89kg - 50kg ) and it's fitted with a 42 GPM water pump. Just by looking at this I think I could approach this by progressively adding more radiators to the loop until I reach the desired heat rejection. I started with a 1.5KW heat source on a 5gal water reservoir and I'm able to keep the water at around 40C with a 1L Capacity, not very efficient radiator and a 4GPM pump, I'm guessing by moving to a liquid to liquid solution with a higher circulation rate and overall area on the radiator(s) will improve the heat rejection.

I like your suggestion of an evaporative cooler, I was thinking about how I could use the principles of a swamp cooler to cool down the water and this is exactly, that, I didn't know about this types of water coolers, cool suggestion that I will defiantly look into, since they seem to be very good at rejecting heat and very affordable too, I might even be able to DIY it.

organiclatte
It's extremely complex to calculate the heat transfer in an object like a car radiator.

Here you can find a very simple article that lays the basis of the different heat transfer processes. The formula for heat transfer by convection is usually the most important one with a car radiator. Seems pretty simple right? You just need to know the heat transfer coefficient and you are in business, right?

On this Wikipedia page, you have some equations to determine those coefficients for very simple cases. Not so simple anymore. Applying those to a car radiator is even worst.

Sadly, radiator manufacturers don't usually give those coefficients easily, but you can try and ask them. They most probably determined them with experimentations rather than calculations anyway. Which is something you might want to consider for yourself.

Also, you will need to know more than "I want to keep a liquid under 50°C". You will need to know what is the initial temperature of the liquid, to begin with.
Yeah, I'm starting to realize how complex this calculations are, experimentation might be easier at the end of the day.

As far as the initial temperature of the liquid goes, the container configuration allows for approximately 33gal of liquid to be directly above the heat source, at 16KW this amount of water would take 3 minutes to increase its temperature by 5C assuming a starting temperature of 40C, the liquid that goes into the loop will at least be at least 45C.

As I said in my previous answer to Baluncore, I'm now looking to experiment with adding more surface area and increasing the volume capacity of the loop until I'm able to reach the desired temperature. I'm also looking into sourcing a small evaporative cooler or making one myself to test it's efficiency.

jack action
organiclatte
A car radiator is designed to operate with a high airflow. Maximum power is probably at maximum road speed, so a higher airflow is available. Also the car coolant will be above 80°C, so there is a big temperature difference, (80°C–35°C = 45°C). You have a lower temperature difference, (50°C–32°C = 18°C), so you will need to compensate by having three times greater exchange area with a forced airflow. There will also need to be a fluid circulation pump.

If water was free you might consider a heat exchanger and an evaporative cooling tower.

This is the sort of problem faced by cryptocurrency miners.
So... a cooling tower worked, sort of, I didn't measure the output of the flame, but based on the time it took to go from 32 to 60ish I would say its about 2kw, I also added a 1.5kw heating element. The system has been running for a couple of hours now and it seems to be stable. The next step is to get proper fill (I improvised with bubble wrap and gutter foam as fill) and get a more consistent heat source, like a propane torch.

[Video]

#### Attachments

• IMG_20210811_202041.jpg
21.8 KB · Views: 87
jack action
https://en.wikipedia.org/wiki/Cooling_tower#Counterflow
So... a cooling tower worked, sort of, ...
That is what I call a real experiment.
You need something like a shower rose, to produce many small droplets with a large surface area, over the whole of the tower area. Higher towers work better. The tubular sides prevent the water blowing away in the wind.
As the droplets fall, some water evaporates from the droplets, which cools the droplets.
The hot droplets also warm the air, which rises creating a natural draft.
The system works best when there is low humidity in the air.
Air rising from the tower is warm and saturated, so a mist should form above the tower when it is working well.

jack action and organiclatte
Gold Member
Summary:: I have a steady 16KW~ 17.5KW energy source summered in a liquid with a thermal conductivity of .1396 and a specific heat of 2.054 @ 40C°. I want to use an air-to-water heat exchanger to keep the liquid below 50C, with a worse-case scenario ambient temperature of 32 C. How do I select the proper radiator?

according some forums, a car generally transfer 1/3 of the engine power through heat to the coolant that gets circulated,
Late arrival of info here but the circulating oil accounts for significant cooling in high performance cars; they all use oil coolers which accounts for a significant fraction, I believe.
Many motor bikes don't use water as a coolant.

Did you consider passing the liquid through the radiator directly, without the water in between? Perhaps 'the liquid' is too expensive.

I do like the cooling tower idea.

Did you consider using a state change for the initial cooling (boiling and condensing a fluid)? You'd need an appropriate fluid. High power radio transmitter valves often use 'vapour phase' cooling which shifts a lot of heat by boiling water around the Anode (dissipating a few hundred kW!) The water is very pure so it's an insulator as well (tens of kV to deal with too).