Efficiency: Parallel vs Sequential Radiator Coolant Flow

[DaveW]

I had a fairly detailed discussion via PM's with another ApexSpeed member on this subject. Since I thought it was generally useful, I reworded it a bit, and here it is.

Efficiency: Parallel vs Sequential Radiator Coolant Flow

The science is basically two-fold:
1. Whether there is parallel flow or sequential flow, if the amount of heat transferred to the air is the same, the difference in coolant temperature from engine exit to engine entry will be the same. If that is true, then that difference occurs whether there is sequential or parallel flow. So, given the above, the temperature difference between the beginning and end of the radiator flow path would be the same in either case. So, that does not result in or explain any difference.

2. The major difference between the two setups is the coolant-flow velocity through the radiators. The sequential setup will have twice the flow velocity through the radiators compared to the parallel setup. This increased velocity increases turbulence and decreases boundary layer thickness, both resulting in increased heat transfer efficiency between the coolant and the radiator tubes. It also reduces “dead” (places where coolant is mostly stagnant) volumes in the radiators. Therefore, since the radiator metal will run warmer, and more heat will be transferred to the air, the sequential setup will give greater cooling capacity for the same size radiators and coolant flow.

In addition, if the two flows in the parallel setup are not equal, even more advantage will be shown by the sequential setup.
----------------------------------------------------------------------------------------------------
Q&A

a) Wouldn't the additional distance for the coolant to flow in the sequential setup reduce the mechanical efficiency of the flow?
-------------------------------
Maybe slightly, but that is mostly controlled by the WP efficiency. With sufficient coolant pipe and radiator sizes, that is not usually significant.

b) Will what holds true for high-speed track racing hold true for low-speed autocross?
------------------------------
Yes – in this case, it is all relative, and the same principles hold.

c) Is it true that maximum coolant temperature entering each radiator will result in more heat transfer?
------------------------------
The temperature entering the radiator(s) is only one value at one point. What is important is the total area of radiator at each coolant-temperature level. With either system the radiator area at any given coolant temperature level is the same, not counting differences in total heat transfer that would change the difference between enter and exit temperatures. At the risk of being redundant, that means whether parallel flow or sequential, area times temperature differential between radiator and air is the same, except for any differences resulting from efficiency.

[EACIII]

I hate to jump into these things since I am usually proven wrong, but heck is it Christmas Eve so I can take it.

The heat exchanger area is the same for both parellel or sequential but the temp difference on the second heat exchanger in a sequential system would be much less then the first. And from what I remember on heat exchanges is the greater the temp difference the more efficient they were. So in a sequential system the first heat exchanger would have the same efficientcy as a parrellel unit but the second one would not. Not sure of that difference would make up for the difference in flow velocity.

Not sure the water pump issue would come into play as modern side rad cars total flow is most likely less then a front rad car.

Ed

[RobLav]

My experience backs up DW's assessment. First time out in my new Citation with one little electric booster pump on the Yamaha and parallel flow rads, the engine got hot. Replumbed to sequential flow between sessions and all was fine.

[Dick R.]

Dave or anyone else:

Autocross related questions since cool down between short runs can be a concern, especially in Pro Solo's where we have 4 runs in rapid succession with between 30 seconds and 5 minutes between runs with the engine idling or briefly off.

1) What about with the engine off? No water or air flow. Just "radiation" from the radiators.

2) What about at low rpm with little or no airflow?

Thanks,

Dick
CM 85

[LJennings]

Thermodynamics is such a complex subject matter and the many equations can be configured in a multitude of different ways to build the proper model of the "system."


Heat travels from hot to cold and there are three basic forms of heat transfer

• Conduction - the interaction of heat transfer on a molecular level - A hot cup of coffee in your hands.
• Convection - Heat transfer by a moving fluid. - A fan blowing during the summer time or cool water flowing over a hot object
• Radiation - Heat transfer by electromagnetic rays - sun burn

All three forms exist in every "system."

The most effective method for a automotive radiator is convection.

For a Solo car, I would think the most effective method to control the temperature between runs would be a combination of a continuously running electric fan and electric water pump, with a large battery.

Also you would need to size the electric water pump to produce a turbulent flow pattern. The electric water pump would address your low rpm and air flow.

[DaveW]

...
The heat exchanger area is the same for both parellel or sequential but the temp difference on the second heat exchanger in a sequential system would be much less then the first. And from what I remember on heat exchanges is the greater the temp difference the more efficient they were. So in a sequential system the first heat exchanger would have the same efficientcy as a parrellel unit but the second one would not. Not sure of that difference would make up for the difference in flow velocity.

What I said before in the 1st post...
The temperature entering the radiator(s) is only one value at one point. What is important is the total area of radiator at each coolant-temperature level. With either system the radiator area at any given coolant temperature level is the same, not counting differences in total heat transfer that would change the difference between enter and exit temperatures. At the risk of being redundant, that means whether parallel flow or sequential, area times temperature differential between radiator and air is the same, except for any differences resulting from efficiency.

Not sure the water pump issue would come into play as modern side rad cars total flow is most likely less then a front rad car.

Ed

Agree.

[DaveW]

Dave or anyone else:

Autocross related questions since cool down between short runs can be a concern, especially in Pro Solo's where we have 4 runs in rapid succession with between 30 seconds and 5 minutes between runs with the engine idling or briefly off.

1) What about with the engine off? No water or air flow. Just "radiation" from the radiators.

2) What about at low rpm with little or no airflow?...

What I said before in the 1st post...
The temperature entering the radiator(s) is only one value at one point. What is important is the total area of radiator at each coolant-temperature level. With either system the radiator area at any given coolant temperature level is the same, not counting differences in total heat transfer that would change the difference between enter and exit temperatures. At the risk of being redundant, that means whether parallel flow or sequential, area times temperature differential between radiator and air is the same, except for any differences resulting from efficiency.

That means to me that under almost any condition, the heat transfer to the air would be the same or better with sequential flow, due to the improved radiator efficiency which would result in the radiator metal being warmer even after shut-off.

[DaveW]

Thermodynamics is such a complex subject matter and the many equations can be configured in a multitude of different ways to build the proper model of the "system."


Heat travels from hot to cold and there are three basic forms of heat transfer

• Conduction - the interaction of heat transfer on a molecular level - A hot cup of coffee in your hands.
• Convection - Heat transfer by a moving gas. - A fan blowing during the summer time or cool water flowing over a hot object
• Radiation - Heat transfer by electromagnetic rays - sun burn

All three forms exist in every "system."

The most effective method for a automotive radiator is convection.

For a Solo car, I would think the most effective method to control the temperature between runs would be a combination of a continuously running electric fan and electric water pump, with a large battery.

Also you would need to size the electric water pump to produce a turbulent flow pattern. The electric water pump would address your low rpm and air flow.

No dummy there! IMO, very good points.

[Dick R.]

Thanks guys!

Anyone have any idea of the current draw of "suitable" electric water pumps?

I'm realistically limited to an on board Ballistic lithium battery due to weight, space, and "logistic" considerations.

My car has been doing fine for a long time but temps do get up to 105 - 110 C (220 - 230 F) at times.

Thanks,

Dick

[Numbskull XIV]

I had the same conclusion as Dave when I first got my f500 autox car. Most of the fmod guys run a parallel system and told me the flow was the key and that's why they ran parallel n After doing a lot of reading on the topic, I came to the same conclusion Dave did. What I did was replumb to sequential and take my largest diameter opening (the radiator openings) and use the same size for tubing for the longest span before needing to adapt to a smaller size opening at the w/p and engine in/out ports. I used very thin walled aluminum tubing ran very cleanly in gradual 90° bends as to not reduce flow in the bends. I wired my w/p and fans to run with a aux battery connection and off of the onboard motorcycle battery. Snowmobile motors are very susceptible to heat and don't like to be any higher than around 120°. With 2 drivers running 6 heats each, in ~85° ambient temperature we consistently stayed at 110° even shutting off the w/p and fans in pits in between runs. Hope this helps
Jon

[Rick Kirchner]

My 94 VD FC ran pretty hot, and lugging my lard ass around didn't help. There were a number of things we did to fix the problems, and converting from parallel to series flow was one of them.

Now, at temps below 80 degrees I actually have to tape one of them off a little.

[problemchild]

If anybody is running 90-95 Van Diemens in AutoX, they should look for the original VW rads with the South African VW cores. These had many many fins and cooled incredibly when stationary. When warming FFs up in the paddock, we found it took twice as long to build heat in cars with those rads. This extra cooling performance was not so evident at speed, and arguably had more drag.

[DaveW]

If anybody is running 90-95 Van Diemens in AutoX, they should look for the original VW rads with the South African VW cores. These had many many fins and cooled incredibly when stationary. When warming FFs up in the paddock, we found it took twice as long to build heat in cars with those rads. This extra cooling performance was not so evident at speed, and arguably had more drag.

That sound like it was due to a lot more mass (metal) in the radiator construction, meaning that the radiators were a rather large heat sink. That would make it take longer to warm up. So once warm, it would also probably take longer to cool down.

[SOseth]

Dave.

When I asked C&R to recommend radiator sizing for my f1600 fit, the results based on data I supplied to them confirmed that I could get better cooling results with radiators plumbed sequentially.

SteveO

[chrisw52]

Thanks guys!

Anyone have any idea of the current draw of "suitable" electric water pumps?

I'm realistically limited to an on board Ballistic lithium battery due to weight, space, and "logistic" considerations.

My car has been doing fine for a long time but temps do get up to 105 - 110 C (220 - 230 F) at times.

Thanks,

Dick

I have been searching around a bit and summit has electric water pumps and important specs like amperage draw in some cases. The pumps they offer seem to take anywhere from 6-10amps depending on the flow rate. The minimum flow rate was 50gpm for what I think we need.

So how much does a Kent water pump pump? How many gpm with your average ford 1600 old school water pump?

[rickb99]

...Anyone have any idea of the current draw of "suitable" electric water pumps?..

The 1st race car I ever owned (back in 1970) was a sports racer powered by a 4 cylinder Fiat engine. Water pump was a fuel transfer pump out of a B-29 bomber. Worked great and it was a total loss electrical system too.

[Modo]

autocross, no crew, assuming u have enough time to pass by your support vehicle, rig a fishing pole dealy with a 12 volt plug hanging from it at a reachable level, plug and go, of course assumes u have and need juice for e-pump and e-fans on board, cast plug off for the next run ..... can power engine off for a quick, quiet, cerebral review

agree with dave, it's the rads and air flow thru them, ...... the old flow debate ...... can get hot water quickly out of the engine (won't be as hot) and thru the rads quicker (won't be as cool) or can leave the hot stuff in the engine longer (hotter water) and cool longer in the rads (cooler water), the rads will exchange the same heat calories for air flow, cepting what dave said (Note: flow rate discussion is different than parallel-series discussion, was a "to thermostat or not discussion", flow rate thru engine part in series and parallel is the same, water slows in rads in parallel, cools more but goes thru the length of only one (X2) before hitting engine again, beat that topic to death Modo!!

incidently does anyone make a ready to use formula car radiator shroud fan insert, might beat 89 different versions of the same thing, then again, it would eliminate us walking up and saying, "What the hell is that??" less creativity - lol

[Dick R.]

The brass radiators in my car were different from each other and somewhat beat up when I got it. Took them to a local radiator shop and worked with the knowledgeable owner. I told him the needs and he recommended a particular style core which would maximize cooling. However, I suspect that they may have more aero drag than than the best possible core designs but that isn't really relevant in autocrossing afaik.

Mine actually do a good job but changing to sequential flow sounds like a good improvement. Unfortunately the pump and/or fan idea won't work for me.

Thanks,

Dick

[DaveW]

Dave.

When I asked C&R to recommend radiator sizing for my f1600 fit, the results based on data I supplied to them confirmed that I could get better cooling results with radiators plumbed sequentially.

Steve, thanks for the confirmation from a trusted source.

[Roux]

Thermodynamics is such a complex subject matter and the many equations can be configured in a multitude of different ways to build the proper model of the "system."


Heat travels from hot to cold and there are three basic forms of heat transfer

• Conduction - the interaction of heat transfer on a molecular level - A hot cup of coffee in your hands.
• Convection - Heat transfer by a moving gas. - A fan blowing during the summer time or cool water flowing over a hot object
• Radiation - Heat transfer by electromagnetic rays - sun burn

All three forms exist in every "system."

The most effective method for a automotive radiator is convection.

For a Solo car, I would think the most effective method to control the temperature between runs would be a combination of a continuously running electric fan and electric water pump, with a large battery.

Also you would need to size the electric water pump to produce a turbulent flow pattern. The electric water pump would address your low rpm and air flow.

Your description of Convection should read, "Heat transfer by a moving fluid..." instead of moving gas. Since convection is also at play on the wet side of the thing we incorrectly call a radiator. Calling these heat exchangers "radiators" is a bad use of the word radiator. there is almost no radiation heat transfer going on in these systems, but somewhere a hundred or more years ago a mistake was made in choosing a name, and we still have it today

Steve.

[DaveW]

Your description of Convection should read, "Heat transfer by a moving fluid..." Since convection is also at play on the wet side of the convector. Calling these heat exchangers "radiators" is such a bad use of the word radiator. there is almost no radiation heat transfer going on in these systems, but somewhere a hundred or more years ago a mistake was made in choosing a name, and we still have it today

Steve.

Very true. And very pertinent to "radiator" design. Many have tried to enhance the air side heat transfer (radiation) by painting radiators black, and all they succeeded in doing was partially insulating them with the paint.

In the past, some radiators were at least partially that, however - steam-heated cast-iron radiators fit that description.

Maybe that's where the confusion over how our "engine cooling heat transfer units" actually work got its start.

[Roux]

My Wyvern design looks a lot like a Swift DB1-3-6 as far as placement of the radiator goes. I struggled with overheating and made a few goofy attempts to troubleshoot the system before getting the pump flow numbers from Honda and deciding that the water flow velocity is high enough even in single pass. Since the aluminum radiator I use is only a hundred bucks, I made a single pass, a double pass and a triple pass version and never got improvements. The thought behind multiple pass radiators is that you make the water flow faster in the tubes thereby ensuring turbulence and better convection of the heat to the aluminum on the wet side.

So my problem clearly exists on the air side. This winter a tapered fuel cell and better airflow design is in the works to feed more air in from the front. Another area of concern is getting low enough air pressure behind the rad to ensure flow. I suspect you guys will see more tufts pasted to the car in 2014 till I get this all sorted.

Cheers

Steve

[David Ferguson]

So my problem clearly exists on the air side. This winter a tapered fuel cell and better airflow design is in the works to feed more air in from the front. Another area of concern is getting low enough air pressure behind the rad to ensure flow. I suspect you guys will see more tufts pasted to the car in 2014 till I get this all sorted.

You're on the right track. Airflow is all about pressure differential -- often it's the low-pressure side that needs the most work. Try some small wickers ahead of any body-side exits to promote a low-pressure region at the exit.

I heard that one FA design actually had the radiator air flowing from rear-to-front due to pressure differentials.

[DaveW]

... So my problem clearly exists on the air side. This winter a tapered fuel cell and better airflow design is in the works to feed more air in from the front. Another area of concern is getting low enough air pressure behind the rad to ensure flow. I suspect you guys will see more tufts pasted to the car in 2014 till I get this all sorted.

Cheers

Steve

Back a few years ago I got involved in a series of posts on a 3rd-generation RX-7's (the '93-'95 255-HP twin-turbo version) forum on how to get the best engine water cooling. The major limiting factor there is the poor air flow through a very cramped engine compartment. That part was so important that the actual fin area in contact with the air-flow and the fin shape were the only really significant factors in the "radiator" design. Radiator material (copper vs aluminum) and radiator size really didn't matter much. If the fins weren't right, a bigger radiator could actually be less effective, especially at lower speeds where the air-flow was marginal.

[DaveW]

You're on the right track. Airflow is all about pressure differential -- often it's the low-pressure side that needs the most work. Try some small wickers ahead of any body-side exits to promote a low-pressure region at the exit.

I heard that one FA design actually had the radiator air flowing from rear-to-front due to pressure differentials.

On my F2000, I made a "tunnel" air exit to enhance cooling for that very reason. Water and oil temperatures decreased >10F with that one change. Photos below.

[dereklola]

I intuitively knew that sequential, rather than parallel, water flow through two radiators was better but never realized it was because of faster water flow rate and more turbulence giving better heat transfer from water-to-metal.

So how about air speed and metal-to-air heat flow? Again intuitively I've always thought that slowing the air flow was better - with air inlets maybe 1/4 the rad area. Is this correct? And why is this the opposite of the water-to-metal situation?

Thx - Derek

[LJennings]

As David pointed out it is all about the pressure drop across the core. I just finished a project and it required a 3 psi pressure drop. 3 psi does not sound like a lot but in practice can prove difficult to achieve

[DaveW]

I intuitively knew that sequential, rather than parallel, water flow through two radiators was better but never realized it was because of faster water flow rate and more turbulence giving better heat transfer from water-to-metal.

So how about air speed and metal-to-air heat flow? Again intuitively I've always thought that slowing the air flow was better - with air inlets maybe 1/4 the rad area. Is this correct? And why is this the opposite of the water-to-metal situation?

Thx - Derek

Again, back to basics:
The faster the air flows through the radiator, the cooler it stays on its way through, enhancing heat transfer, and the thinner any boundary layers will be. Also, again, the turbulence with faster flow will aid heat transfer to the air. So, faster flow through the radiator is always better.

The throttled air inlet is to match the inlet air mass to what will be able to flow through the radiator. Too big an inlet just leads to excess air spilling out into the exterior airstream, ruining the flow both into and around the inlet. Also, the smaller the inlet, the less parasitic drag is created by the flow inside the bodywork. So what you're trying to do is to balance the air flow volumes for optimum efficiency in both cooling and drag.

So, with all that in mind, it is not at all opposite of the water-to-metal situation.

[dereklola]

Lloyd - yes I understand that (about pressure drop) - but I was trying to understand the fundamentals - if faster fluid flow is better on the water side - why is it not also better on the air side? EDIT: DAVE JUST BEAT ME TO IT - SO IGNORE THAT COMMENT - AS USUAL WHAT HE SAYS MAKES EMINENT SENSE.

I'm assuming that having air inlets much smaller than rad area does in fact slow down the air - seems obvious to me - previous car (vintage Atlantic) had forward facing 27 sq.in. and rad 150 sq.in. (5.7:1) - current car (F3000/Interserie) has forward facing 24sq.in. and rad area 180sq.in. (7.5:1). Current Stohr DSRs seem to have about 20sq.in. forward facing inlets and about 150sq.in. rad area (7.5:1). All numbers approx.

Current project (vintage 2L CanAm renovation) has twin sidepod rads with side inlets (not forward facing) so this whole subject is very pertinent right now. Also maybe I need to actually measure press drop when it's running but right now it's all mental work - and very interesting.

Thx - Derek

[Marty Nygard]

Since, as a former boiler engineer, I have been around a lot of air to water heat exchangers, I have read this thread with great amusement. Unless you are prepared to get into the LMTD and the heat transfer coefficients (as they vary with velocity, pressure and fluid type) of each heat transfer coil, it might be a good idea to stop making up new (rule of thumb) laws of physics.
Instead, I suggest re-reading the three Carroll Smith books; especially Chapter 8 of Prepare to Win.
Happy Holidays
Marty

[DaveW]

Since, as a former boiler engineer, I have been around a lot of air to water heat exchangers, I have read this thread with great amusement. Unless you are prepared to get into the LMTD and the heat transfer coefficients (as they vary with velocity, pressure and fluid type) of each heat transfer coil, it might be a good idea to stop making up new (rule of thumb) laws of physics.
Instead, I suggest re-reading the three Carroll Smith books; especially Chapter 8 of Prepare to Win.
Happy Holidays
Marty

So, on the original subject of parallel vs sequential flow cooling efficiency, and other related things, please help us better understand the subject. What, specifically, have we gotten wrong?

I am not claiming to be an expert. I, and the others commenting here, are going on our general engineering knowledge and personal experience. Any better, more accurate, information is very welcome.

[Marty Nygard]

We developed our own heat transfer coefficient curves using an in-house wind tunnel and found them to be sufficiently different from published data that we were able to better design our product. Unless you are buying radiator cores from a vendor that is willing to provide laboratory tested curves for their product, anything you do is just an educated guess. I have never come across better practical information than the Smith books, that is why I recommended re-reading them.
To answer the question, if your radiator configuration works, good for you. There is no reason why sequential or tandem cores would be better than parallel cores if designed correctly. The advantage of parallel cores is you can have symmetrical ducting and plumbing configurations. An optimized tandem flow configuration (where heat transfer is identical on both sides of the car) would ultimately have either different cores in identical (symmetrical) air flow paths or matching cores in two different air flow configurations. The best practical reason for tandem cores is when one side pod has a single large radiator while the other one has a smaller radiator and separate oil cooler.
If you want specific instances of statements I believe to be wrong, "The faster the air flows through the radiator, the cooler it stays on its way through, enhancing heat transfer, and the thinner any boundary layers will be". Almost all air flow through the core will be fully turbulent (see definition of Reynolds number). This does not mean that all air flowing through the core will contact the fins to convect heat from them. Fin type, fin spacing and core depth also enter into the performance. That's why space velocity has such a large effect on the heat transfer coefficient.
Marty

[RobLav]

Marty,
Before I plumbed my FB Yamaha coolant in parallel flow, I had consulted the usual suspects (Smith's books). My at the track change to sequential immediately solved the cooling issue. Personally, I believe the problem was that parallel flow was not anywhere near equal between the two rads.

[Modo]

if Smith's angle (can't recall who) was the one that said that in the case of series radiators, the second one has cooler water and won't have the same temperature difference so it won't cool as well, I believe there is a misunderstanding, I believe it should have been the temperature difference is cooler air not hotter water, you are not going to get cooler water out of a radiator by putting hotter water in with the same air/temp flow, not on this planet, ...... or up the efficiency with hotter water for the temp difference, hotter in, hotter out, less hot in as in series setup second radiator, less hot out

the parallel so called advantage that even had me jumping was that the two rads would get hot water directly from the engine so the temperature difference would be greater in both .......... ho ho ho as the season goes ..... yea, in parallel, the water slows, but then goes right back to the engine (heater creator), series hookup goes thru two rads and back to engine, end result similar except for quirks being discussed, maybe this way .... engine output-same, radiators same, air temp and air flow thru rads same, results similar, the rads performance fixed by design ... the cooler radiator in the series setup, do de same calories per airtemp-airflow regardless of what temps are in the rad ..... IMO


by the way, if the air is hotter than the rads, you will heat the water in the rads, cold soaks heat energy from hot ..... Dinner time yet, LOL

[Steve Demeter]

The rate of heat transfer from any body to any other body is directly proportional to the temperature difference between them.

In the discussion, the bigger the temp diff between the water and air (assuming same heat exchangers) the faster the heat transfers from the hotter to the cooler body (water is the one and air is the other).

Now of course other things enter into the parallel vs series set up like Reynolds number which is a measure of the turbulence of the fluids in question. Once that gets above a particular value, heat transfer rate increases significantly all other things being equal. Series set up flows the water at higher speeds through the core and therefore higher Reynolds number and consequently higher rates of heat transfer.

If you paralleled them, the speed of the water through the core will be more or less half (half of the flow that the water pump moves thus half the speed) than series set up and thus lower Reynolds number and lower rate of heat transfer .

[Neil_Roberts]

A thought experiment for discussion:

The performance of a series radiator configuration is (or is not) equal to a parallel configuration with the water flow through both cores changed to double pass.

Which is correct, and why?

[Jnovak]

175hp, 1 radiator, single pass, single row. 1 oil cooler. The side pods are 5" tall maximum. The openings are 2.5" tall and 16"+ wide.

Never overheated ever.

[Modo]

Damn, that's a nice Formula Vee in the banner, color and all, Johnny U's number as well .....

tnx Steve, will look the formulas up, sounds good as well ..... but the rate of transfer and total calories needed may be the key, ie. does the increase in transfer rate make the rad now get rid of the extra heat calories to make the out water or system water cooler?? remember, the water is hotter (more calories) not the air cooler for the thermal difference ,,,,,, perhaps total system cooler because one rad 's rate of transfer is more, compared to series second rad, think that's what Smith said. (that's a tough one to swallow since it came from hotter water/difference in rad II but more calories) though hot water boils quicker than cold, now that's a calorie quantity problem

anyway, good in-law visiting thread diversion!! Neat banner FV as well, air-cooled too, can use a built in fan too, LOL

[S Lathrop]

Having built cooling systems of both parallel and sequential for the same car and engine combination, I haven't seen much in this discussion that I can argue with.

One issue I have with the parallel system and a 4 cylinder engine is how do you assure even distribution of hot water between the 2 radiators. It is quite possible for the water to flow through one radiator more than the other.

Based on experience, I am now advocating a series arrangement for the radiators.

A vee type engine is another issue because you can use one radiator per bank of cylinders.

[DaveW]

A thought experiment for discussion:

The performance of a series radiator configuration is (or is not) equal to a parallel configuration with the water flow through both cores changed to double pass.

Which is correct, and why?

Is not. Because you can usually not guarantee that the parallel flow system will deliver coolant equally to both radiators. see Steve L.'s post above.