LLK die 10te: Größen und Leistungsstufen | Posts 16+

Hi Ulf!

Quote:

In the case of 2 identical parallel liquid cooling circuits, the internal flow velocity theoretically decreases to half. That means the intake air remains in the intercooler for a longer time, and therefore has more time to release heat. And yet, the goal is to release less heat per unit of air volume. Which (apparently unknown to me) processes are responsible for that?

Well, the connections aren't entirely clear at first glance.
The heat transfer between a fluid (in this case, a gas) and a stationary wall (the heat exchanger, also known as the cold liquid cooler) is influenced not only by the temperature gradient, but also by the formation of a boundary layer and the type of flow that develops.

Here's a small practical example: In the past (with 386 processors), the CPU of a computer could be cooled using a heatsink and natural convection. This represents the worst heat transfer scenario, as there is virtually no air flow (relative velocity between the air and the heatsink). The flow of natural convection is laminar and has a relatively thick boundary layer, which acts as an insulator.
More powerful CPUs (starting with the 486 series) had to be equipped with a fan to improve heat dissipation. The airflow over the fins becomes highly turbulent, which significantly improves heat transfer. The boundary layer is also much thinner.

One should imagine the interior of the liquid coolant (LLK) to be similar. If only a gentle, laminar flow develops there (with a parabolic velocity profile across the cross-section), then the flow itself acts as an insulator, hindering heat exchange with the walls.
At high flow velocities, the flow becomes highly turbulent. This results in a plug-like velocity profile, meaning that the flow velocity is significantly higher even in the boundary layer. This promotes heat transfer with the wall and improves the efficiency of the liquid cooling system.
Never forget that air is a very good thermal insulator!!!

I hope I was able to illustrate the situation a bit.

Regards,
Alex.

donalexo wrote:

The flow of free convection is laminar and has a rather thick boundary layer that acts as an insulator... The flow around the fins becomes highly turbulent, which significantly improves heat transfer. The boundary layer is also much thinner. Bitte gib mir den deutschen Text, den du übersetzt haben möchtest. I hope I was able to illustrate the matter somewhat.

Hi Alex,

I have understood very well what you mean, thank you .

But:
Even at idle, we already have a relatively fast airflow directed towards the engine.
With an estimated inner pipe diameter of 6 cm (resulting in a cross-sectional area of 28 cm²) and a volumetric flow rate of 450 * 1.9 liters per minute, I'm getting a velocity in the range of 5 m/s.

Given that, as far as I know, the LLK (likely referring to a specific component) is the worst obstruction to flow on the suction side, the flow velocities in the area of the internal cooling fins must be at least as high, and likely even higher.

In another thread, a link was posted.
http://www.sigem-elektronik.de/elektro/glossar.htm#W
(Scroll down a bit to see the diagram.)
"The thermal resistance between a heatsink and the air decreases with increasing air flow velocity, but only up to approximately 6 m/s; beyond that, there is no further improvement."

Okay, the heat is transferring from the air to the aluminum, but I suspect that the statement also holds true in the opposite direction.

Therefore, I speculate that parallel 2 LLKs will reach the individual thermal resistance at or around ~2000 rpm, similar to 1 LLK running idle, and that the thermal resistance will not decrease further beyond that point.

This would mean that two parallel turbochargers, starting at 2000 rpm, are actually twice as effective (with half the pressure loss) as a single turbocharger.

What do you think about that?

Okay, I just got back from dealing with BMW...

Vehicle: Skoda Octavia Combi 1.9 TDI ALH Automatic.
Charging pressure: 0.9 bar (1950 mbar absolute).
Outside temperature: 25°
driving profile: accelerator pedal stuck.
Intake air temperature, standard condition: 63°
Supply air temperature (LLK): 93°

With the stuck LLK (low-level kit), the exit from the parking lot was noticeably heavier, and it became much more difficult in the area of maximum speed. I immediately drove to the next parking lot to free the LLK again. It's better that way.

CU Gremlin.

Gremlin wrote:

vehicle: Skoda Octavia Combi 1.9 TDI ALH Automatic. Charging pressure: 0.9 bar (1950 mbar absolute). Outside temperature: 25° driving profile: accelerator pedal stuck. Intake air temperature, standard condition: 63° intake air temperature LLK air supply covered: 93°

Aaah, thank you - some fresh input .

So, the LLT (likely load temperature) is approximately 38K above the ambient temperature.

Quote:

With the LLK (License Plate Light) taped over, the exit from the parking lot was noticeably heavier, and in the Vmax (maximum speed) range, it became significantly tougher.

Did you happen to log the limits (air mass / rotational speed)?

Even with the LLK (Low-Level Knock) taped off, my soot limit was still about ~5mg above the engine's optimal speed, and I couldn't notice any smoother acceleration.

Did you happen to log the limits (air mass / rotational speed)?


No, because the rotating part is already fixed, and the air mass reading with the diode isn't really meaningful.
btw: it seems that the maximum power (Pmax) is reached at 4250 rpm. Up to that point, it revs up, but then it almost stops increasing... however, it also doesn't decrease in power when going slightly uphill.

CU Gremlin.

@ Ulf:

Regarding our problem with heat transfer from the charge air to the intercooler.

Inside the intercooler, a flow is generated. The heat transfer from the charge air to the aluminum body of the intercooler is more intense the higher the flow velocity and the greater the temperature difference between the charge air and the surface of the intercooler.
Considering only the internal heat transfer between the charge air and the coolant, a higher flow velocity leads to a higher turbulence level in the flow. The laminar boundary layer, which acts as an insulator, becomes thinner with increasing Reynolds number, meaning that heat transfer improves.
Imagine it this way: At low flow rates, many air molecules do not come into contact with the walls because they can move along a path that remains relatively far from the walls. Despite the longer residence time of the molecules in the low-temperature coolant (LTC), they have only a small chance of transferring heat to the LTC.
When the turbulence level is higher, this laminar flow is disrupted by so-called "eddies" (turbulent vortices). These small "swirls" occur chaotically and cause a significant exchange of energy between the core zone of the flow and the edge regions, both in terms of kinetic energy and heat transfer.

In order for this "advantage" in internal heat transfer to be realized, the external heat transfer between the coolant and the airflow must be able to dissipate the heat flow.

Regards,
Alex.

donalexo wrote:

Imagine it this way: At low flow rates, many air molecules do not come into contact with the walls because they can move along a path that remains relatively far from the walls. Despite the longer residence time of the molecules in the low-temperature coolant (LTC), they have only a small chance of transferring heat to the LTC. When the turbulence level is higher, this laminar flow is disrupted by so-called "eddies" (turbulent vortices). These small "swirls" occur chaotically and cause a significant exchange of energy between the core zone of the flow and the edge regions, both in terms of kinetic energy and heat transfer.

Hi Alex,

These are some really good explanations from "niche areas" that very few people know about, but which can sometimes help us move forward -> thank you.

One question, though:
The faster the flow, the more turbulence there is in the liquid coolant (LLK), and the more air molecules come into "very close" contact with the heat-dissipating metal. I now understand that.

But at the same time, the residence time of individual molecules in the liquid coolant and, consequently, the time for heat release becomes shorter. And a shorter cooling time, in my opinion , fundamentally reduces the energy exchange, which works against the "success" of the turbulence.

Let's imagine an intercooler in a laboratory setting: the side exposed to the airflow is constantly supplied with air at 150 km/h and 20°C, while the air entering the intercooler on the charge air side always measures 100°C.
The airflow velocity of the charge air is now being changed.

Does the intercooler actually cool the intake air down to a temperature closer to the 20°C of the airflow side, the faster it flows through the intercooler?
And that the speed of the intake air can be increased "unlimitedly" without its initial temperature rising above a certain point (assuming a constant intake air temperature)?

It's hard for me to believe, but if you clearly say "yes" right now, I will adjust my intuition accordingly...


EDIT / Addendum:

My logs, however, don't seem to confirm your descriptions.

During a measurement, the LLT (likely referring to a temperature or pressure reading) increases steadily from 2000 to 4300 rpm, from 24 to 74°C (with a regulated boost pressure of 1.3 bar).

Assuming a constant boost pressure, the intercooler inlet temperature should also remain largely constant (approximately 75K multiplied by the boost pressure in bar, plus the intake air temperature; for example, at 1.3 bar and 13°C, this would be around 110°C). Therefore, during normal driving, only the flow velocity increases within the intercooler – on both the charge air and the coolant sides.
The driving speed ranged from approximately 75 to 170 km/h.

It's quite obvious to me that the transferred power from the intercooler to the airflow increases in this process.
But the increase in cooling performance is, as a result, apparently smaller than the increasing charge air mass flow , so that each charge air molecule can release less and less thermal energy.

How do you explain the results of my measurements?

@ Ulf:

Unfortunately, I don't have much time right now, but I'll try to give a brief explanation on the spot:

In your analysis, you are failing to consider the entire system. The turbocharger's compressor operates in a completely different operating range at higher RPMs compared to lower RPMs. Furthermore, the compression of the air in the compressor wheel occurs much faster, which alters the polytropic exponent (polytropic compression). Consequently, the air outlet temperature from the compressor is higher at high RPMs than at low RPMs, even with the same boost pressure!
You cannot include this measurement in your line of reasoning in this way.

Regards,
Alex.

How do you explain the result of my measurements?

Caution:

The temperature read from the diagnostic system is SIGNIFICANTLY DELAYED!

Please take a picture.

CU Gremlin.

donalexo wrote:

Consequently, the air outlet temperature from the compressor is higher at high speeds than at low speeds, even with the same charging pressure!

Aha
Das erklärt wieder einiges . . . "Thank you!"
Is there perhaps a rule of thumb for how much to add to the 75 kPa (bar) pressure, depending on the pump speed, to get an approximate value for the low-pressure cooler inlet temperature?

gremlin wrote:

The temperature read by the diagnostic system is SIGNIFICANTLY DELAYED!

You're probably referring to the "digital low-pass filter" located after the sensor input of the EDC. I'm already familiar with it from recording the characteristic curves of the diesel and water sensor inputs.
Is the LLT input equally slow?

According to the log, the maximum temperature actually occurs approximately 2 seconds after releasing the throttle at 4300 rpm.

However, in the 4th gear, the speed changes and, consequently, the intake air temperature change very slowly, so that the speed and intake air temperature curves are no longer noticeably shifted.

You are probably referring to the 'digital low-pass filter' located after the sensor input of the EDC. I am already familiar with it from recording the characteristic curves of the diesel and water sensor inputs.
Is the LLT input equally slow?


Not really: The measurements and evaluations are performed within the range of 2-10 milliseconds. However, the displayed value is integrated and updated slowly.

CU Gremlin.

donalexandro is clearly a process engineer or a mechanical engineer.

*dust blower*

donalexo wrote:

At low flow rates, a large number of air molecules do not come into contact with the walls, as they can move along a path that is always relatively far from the walls. Despite the longer residence time of the molecules in the liquid coolant (LLK), they have only a small chance of transferring heat to the LLK. When the turbulence level is higher, this laminar flow is disrupted by so-called "eddies" (turbulent vortices). These small "swirls" occur chaotically and cause a significant exchange of energy between the core zone of the flow and the edge regions, both in terms of kinetic energy and heat transfer. In order for this "advantage" in internal heat transfer to be utilized, the external heat transfer between the coolant and the airflow must be able to dissipate the heat flow.

Hi Alex,

I've taken a look at my series of liquid cooling components. In the liquid cooling channels, there are also structures similar to lamellae, resembling the side for cooling air.
They are apparently intended for that purpose.
1. to induce turbulence in the intake air even at low flow velocities.
2. to increase the contact area between metal and air within the liquid coolant system.
This is likely to significantly improve heat transfer in the liquid coolant (IMO), especially at low liquid coolant flow rates, but at the expense of higher pressure drop losses at high mass flow rates.

Currently, I'm considering replacing the water cooler with a narrower one and installing another liquid cooling unit (similar in design to the stock one) in the space that becomes available. This new unit would flow in parallel with the stock liquid cooling unit.
The implementation will likely take some time, but when it's finally up and running, I will post the log data for comparison with the current production trim.

If you like, you can provide an estimate of how the LLT (likely load transfer) will change. The measurement point will remain at the output of the standard LLK (likely load transfer kit).
It only cools the turbocharger down to approximately 50K above ambient temperature when running at 4000 rpm, 1.4 bar of boost, and 160 km/h, despite a large intake of cooling air.
The "smooth" surface area of its LL channels (i.e., without internal and external fins) is likely to be approximately 3,100 cm².

Hello Ulf!

As you may know, EDC reduces the boost pressure at very high LLTs to reduce the thermal load on the engine. This then results in the point at which knocking occurs being reached more quickly at high ambient temperatures .

The thermal efficiency of the process certainly also benefits from a lower LLT (Lowest Limit of Temperature), and NOx emissions are also significantly reduced. This is also the reason why modern diesel engines have a cooled EGR (Exhaust Gas Recirculation) system.

I wish you much success with your renovation project.

Regards,
Alex.

donalexo wrote:

As you may know, EDC reduces boost pressure at very high LLTs to reduce the thermal load on the engine. This then results in the point at which the mixture becomes too rich occurring more quickly at high ambient temperatures

.
High outside temperatures seem to have a relatively... significant impact. During my initial test drives with one of the factory-installed fog lights mounted in the lower cooling air intake grille, the engine temperature reached a concerning 80°C at 4100 rpm, 165 km/h, and an outside temperature of 13°C.
Immediately after, without the NSW (presumably a specific component or setting), the LLT (likely referring to a temperature or pressure reading) only reached 71°C at 4100 rpm, and interestingly, the desired LD (likely another pressure or performance parameter) was almost 100 mbar higher.
In the original 9N (1st version) with ASZ and NSW, it doesn't even require very high outside temperatures, according to Central European standards, to activate the "LLT thermal fuse".

Quote:

I wish you much success with your renovation project.

Thank you... Currently, I'm struggling with the decision of how to implement the T-pieces in the LL lines .

Hi Ulf,

Is this your favorite topic again?

I'm sure you know that the A4 models with the 2.5 TDI engine, including the cabriolet, used a parallel intercooling system. It was a very clever design. You can probably identify the double-flow intercooler in the documentation, which largely eliminated the need for T-pieces. I believe this is the first and only time it was used within the group... and it will likely remain that way.
There are also many publications available on this topic (MTZ, especially in the special issue dedicated to 20J TDI engines).

As your attempts have shown, the most important thing remains a reasonable intake and exhaust of cooling air – even with as many heatsinks as possible or your previous optimizations of the airflow – and the airflow in your front section looks really bad.

Best of luck.