Friday, July 18, 2008

Duramax Oil Cooling, Part 1

The fish are in the freezer, and with that, it is time for my annual reel cleaning and vehicle servicing regimen; an oil change is always part of it. I am reminded of what that oil went through on all those fishing trips.
It all started with a realization that my thermo-viscous fan should not be such a common occurrence. I seemed to fly effortlessly up steep mountain grades with my big camper load, on a 103ºF Arizona day, in air conditioned comfort, and in complete complacent silence, dreaming of bass boils. Then I awoke, eyes wide and startled, to the sound of a 747 landing on top of me… THAT FAN!
I wanted to silence that fan, and all heat issues its aural signature represented. I remember the first time I heard it, going 70 MPH on slight, rising terrain with NO load. On a 114ºF sunny day, it doesn’t seem to shut off. Adding insult to injury I know I am losing two to three miles per gallon whenever it is in full spin. I have heard the statement, “that fan is normal, be glad you have it,” a hundred times. I am glad. Glad I don’t have to believe that any more. What if I said that you are an oil overheater, and you have never seen it? Crazy am I? Do you have an oil temperature gauge? If you knew what your oil temp was when you towed through my back yard, you would own one. The oil overheats on every summer camping trip. Presently, oil is cooled indirectly at the stock engine mounted oil-water cooler. Oil heat must be conveyed to the coolant, then to the radiator, then to the atmosphere, making it indirectly cooled.

I have heard the statement, “that fan is normal, be glad you have it,” a hundred times. I am glad. Glad I don’t have to believe that any more.

Below, you see the Stacked Plate Duramax Coolant-to-Oil Heat Exchanger. If it looks unfamiliar, you are looking at the side that fastens to the motor. The coolant passes over a series of plates which contain the flowing hot motor oil. On the far side (left), the filter thread is visible. This heat exchanger mounts to the driver’s side of the motor and houses the filter element. The oil inlet and outlet are the parallel ports on the bottom. The coolant comes straight from the water pump and enters on the upper right (not visible), and emerges at the large plate opening at the left.

Duramax LLY Engine Oil Heat Exchanger

Simplified, it looks like this:

Simplified Diagram of the Duramax LLY Engine Oil Heat Exchanger

Heat Transfer for Dummies

We should look at what is expected of the oil. But before we do – since oil interfaces with the coolant in the heat exchanger – it will be helpful to understand the cooling system first.
The cooling system either rejects the heat the motor produces or it doesn’t. If coolant rises 15ºF in the motor, it must be cooled 15ºF in the radiator. When this does not happen, engine coolant temperature (ECT) rises over time in direct correlation with the cooling shortage, until, once again, the heat off-loaded equals the cooling system heat generated (see ECT Increase Chart on page 20). To demonstrate, consider the motor as a single mass of 1,000 pounds, with an average heat capacity of 0.20 BTU per pound-foot. If ECT rises at a rate of 5ºF per minute to the overheat stage, then the shortage can be easily calculated. The simple formula is:
Mdot*Cp*deltaT = Qdot
M is the flow rate of the coolant, usually expressed in pounds per hour.
Cp is the fluid’s capacity for holding heat energy, specific heat.
DeltaT is the temperature change of the coolant or oil.
Q is the heat in BTU per hour
1000 pounds * 0.20 (BTU/Pound-Foot) * 5 (ºF/min) * 60 (min/hour) = 60,000 BTU/hour
If ECT rises doubly 10ºF per minute, instead of five, then the radiator heat rejection deficit is also double: 120,000 BTU/hour.
Because of limited airflow, an aerodynamic reality, the rejection ability for the OEM radiator on a hot day is around 400,000 BTU/hour (Qdot); add 30,000 if the AC is off. At times of extended high loads, however, we need more like 500,000. Just take 80 GPM of coolant that weighs 8.3 pounds/gallon, and a heat capacity of 0.8 BTU/pound-foot, and some quick math, solve for deltaT in the formula above and we get 16ºF. All the radiator has to do is cool the coolant 16 degrees. Can that be so challenging? Considering that a drop of coolant spends exactly 0.75 seconds in the radiator; sure, that can be a challenge. Even worse, that highway-cruise 60 MPH air parcel is slowed down 82% with all the obstacles to air flow. A 65 MPH vehicle speed produces, at best, 11 to 12 MPH kinetic airflow through the radiator. This reduced quantity of cooling air is heated significantly, before ever reaching the radiator and in cascade fashion, by:
  • Transmission Cooler (4ºF)
  • AC Condenser (12ºF)
  • Charge Air Cooler (48ºF with stock boost, 95ºF with augmented boost).
Added up, they total 64ºF. On a 100-degree day, the little air that even makes it to the radiator is at a nicely pre-heated 164ºF! Understanding this should make the challenge more clear.
Yet the only difference between the vehicle that overheats and the one that does not is often only two degrees coolant temp exiting the radiator. Two degrees is 60,000 BTU/hour and that’s all we have to come up with to make a noticeable difference.

The Research

As a simplified model, this is easy to understand. Since it is not likely that we can get the radiator to do much better, we are left with adding capacity to the shortage, right? Well, yes, but how shall we do this? Which coolant shall we build capacity upon?
Yes he has been drinking anti-freeze again, there is only one ‘coolant’, albeit green, yellow, orange-red… not amber!
Actually, in today’s high performance motors there are two coolants, both equally important in the transport of combustion heat away from the motor: water and oil. The water half is fairly well understood. Oil, on the other hand, is underrepresented, relegated to simple lubrication duties; yet it is now much more. (I am using the term “water” here, since a 50/50 coolant mixture takes on most of water’s thermal properties). The term “coolant” will now apply to BOTH water AND oil in this article.

Oil, The Other Coolant

From SAE Study 2001-01-1073 Engine Lubrication System Model for Sump Oil Temperature Prediction, Steve Zoz and Steve Strepak et al:
The typical approach to engine design has been that very little thought is given to oil temperature and oil cooling until engine dynamometer or vehicle testing has uncovered a problem.
Energy into the oil – the two primary energy transfers to the oil are from combustion and from friction, though energy also enters the oil through contact with blow-by gases… the combustion contribution arrives in the oil through two paths: 1) convection to the piston, conduction through the piston, and then convection to the oil at the rings, under crown and skirt, and 2) convection to the bore wall, conduction down the bore wall, and convection to the oil at the lower portion of the bore.

Engine Coolant Temperature Increase Over Time

The addition of oil squirters was modeled by increasing both the oil flow rate to the under crown and the under crown convective heat transfer coefficients.
Under piston flooding has indeed changed the role of oil as a coolant, doubling and tripling piston heat removal through oil transport on many vehicles. You can see one (label A) in the GM cutaway photo below. What you can’t see is the internal piston design. Oil shoots up and disperses internally, saturating a large area in its conduction and convection role.
For engines with high specific output, piston cooling is now indispensable. Lubricating oil is diverted from the main flow and injected through injection nozzles against the underside of the piston or into piston cooling channels for the piston cooling. Pressure controlled valves prevent heat being unnecessarily drawn from the piston when the engine, and hence oil, is cold… (Internal Combustion Engine Handbook, 283)
Unfortunately, this pressure gated strategy of regulating under piston flow is flawed. As stated in The Handbook, a threshold minimum pressure enables under piston cooling. With a given viscosity, this pressure will be strictly RPM dependent, as pressure increases with flow rate (low pressure at low RPM, and vice versa). In design phase, the assumption was made that low RPM is equivalent to low load and low thermal stress. Good enough. But as we know, viscosity (and oil pressure) drops quickly as oil rises above this design temperature. This drop in pressure has the detrimental effect of reducing or closing off cooling oil flow to the squirters when it is needed most, an odd design paradox. See our problem now? A bit of a self destruct mechanism. Oil temperature control resolves this issue, bringing pressure back to its design point by restoring the intended viscosity to the oil. This is extremely important to piston life.
Thermograph images of a typical TD piston (left), show that the piston underside is approximately 250-400ºF lower than top surface temperatures. With a 1200ºF exhaust gas temperature (EGT) run, this places oil in direct exposure to 500ºF-plus surfaces in high load conditions. Now you can tell anybody who disputes it: oil is indeed a coolant that MUST, in turn, be cooled.


In this cutaway shot of the Duramax engine, we see an oil squirter (A) and the oil heat exchanger (B) that transfers heat from the oil to the radiator fluid.

The RPM Connection

The most notable result from the 2001 SAE Study is the effect of RPM on oil temperatures. The effect of doubling RPM, with no additional workload, raises oil sump temperatures 54 degrees! This effect is explained with a couple of phenomena. There is more friction, obviously. But also, oil flow rate (and thus heat transfer) rises almost linearly with RPM. In contrast, the lossy nature of centrifugal water pumps does not afford the same capability. In fact, most water pumps realize reduced flow rate at redline RPM, due to low pressure suction and partial cavitation anomaly. This reduced flow rate occurs with the Duramax water pump.
So, while oil flow increases sharply with RPM, water flow stays flat or even reduces moderately. This increase in oil flow increases the oil’s heat transfer coefficient (a fancy way to describe less resistance to conduction) significantly, while the water’s heat transfer coefficient may actually decline after peak pump RPM is reached. Plainly stated, at higher engine RPM, the proportionate role of oil as a coolant increases significantly.
WELL, so what, all the oil’s thermal energy is dumped into the water, at the OEM cooler, correct?
The best standard industry-wide efficiency for this heat exchange device is 80%. That means at least 20% of the energy picked up by the oil in the motor does not transfer to the water, leaving the oil warmer than the water on each cycle.
“What does this mean?”
The oil temperature continues to rise with high RPM and workload, until it becomes such a thermal burden at 80% heat exchange, that water cannot contain it at the 210 degree regulated water temperature. Under load, oil temperature heads north of 280ºF before there is any sign of trouble on the ECT gauge.
280 degrees. It gets much worse, up to 380 degrees at coolant purge stage. 380 degree oil, after all, will do a job on the temperature of anything at 210 degrees, and it can’t stay that way for long. This is the downside of indirectly air cooling the oil with water, regardless of radiator capacity.
But back up a bit. This is NOT a liquid cooled vehicle. Boats are liquid cooled. Cars are air-cooled. ALL CARS! No I haven’t been sniffing VW exhaust, that is why they are equipped with a heavy duty fan, to increase air flow for heat removal. True, we use coolant, but the coolant only transports all the heat to the radiator for release into the air. Coolant is merely the transport component. Liquid-to-air heat exchange is the only way this vehicle stays cool, (aside from a small radiation factor). Only 30% of the fuel we burn performs work, the rest is wasted as heat. Of the 70% of waste combustion heat that does not perform wheel work, 30% must find its way to the atmosphere via the radiator. Off-loading oil heat into the coolant does not help this cause, nor does it effectively cool the oil (if expectation is oil at 205ºF, the optimum oil temperature hydrodynamic lubrication and oxidation resistance; 275ºF is the max specification). For conventional oil, this is a point at which lubrication is still significant, though quite degraded. The cooling media (water) itself is at 200ºF. We know the engine mounted cooler can not get close to 100% efficiency, so, with the exception of warm-up, the oil will always leave the cooler at a higher temperature than the coolant. (And, by the way, water will always be heated by the oil also; let’s not forget what we are seeing past our white knuckles on that ECT gauge).
Back to our 60,000 BTU/hour need. We have to make a choice. Cool the oil? Or try to cool the water more?
The water, under the best high-altitude circumstances, boils at 260ºF under pressure. But just as bad, when it gets to 244ºF (248ºF on some vehicles), the system commands the motor to start killing power through a phased in PCM (ECU) defuel schedule. That’s only 20 to 30 degrees past where the thermostats are regulating the water.
On the other hand, the PCM could care less what my oil temperature is. (A little secret: GM doesn’t either, as long as there is oil; if they cared, they would give us an oil temp gauge). The DMax oil is seen at 350 to 400ºF local oil temps in some cases, well above the spec maximum, and certainly in the “worthless lubrication” category.
So, on to answer the question: which option will give me a better temperature differential for heat exchange with 100 degree air? 350 degree oil, or 240 degree water? You may know the answer by now: the oil.
Tip: Get a gauge that goes to at least 280ºF; 320ºF is even better.

No comments:

Post a Comment