Build your own water injection system

I would like to discuss a topic that is often mentioned, but seemingly not implemented in practice by anyone (at least not here in this forum): water injection.

I want to try this out on my fun car, a 1.9 TDI Ibiza (ASV). And I want to build the whole thing myself.
Please refrain from any discussions here regarding the feasibility of this project, especially in relation to the current limits of the engine's performance potential. Here, we will only discuss the feasibility of water injection and the associated physical effects.

Regarding the theory: By evaporating, the injected water removes energy from the hot, compressed intake air. This lowers the temperature, and consequently, the pressure. Since the MSG regulates the pressure constantly, the cylinder filling with oxygen improves.

I've calculated the exact amount using basic thermodynamics. More details about the calculation are provided in the attached PDF (waes.pdf), and the chart shows an overview of the expected temperature reduction. As a rough estimate for those who don't want to look at and read the PDF: With a pressure of 1 bar above atmospheric, 2000 rpm, and a water injection flow rate of 100 g/min, the intake air temperature of 373 K is reduced by 60 K. Please note that this calculation is based on ideal conditions and assumes, for example, perfect cylinder filling (without flow losses).

With this calculation, you get a rough idea of the costs involved in building a water injection system yourself. Commercial kits are far too expensive for me in the Eurozone. After extensive research online, it turned out that the biggest challenge is finding a waterproof pump with sufficient pressure ( > 10 bar) at a 12 V input voltage. I plan to use a vibration pump from an old espresso machine (picture below), which delivers approximately 15 bar with a flow rate of 100 g/min and a power consumption of 48 W. Unfortunately, the pump needs to be operated with 230 V, which I will solve using a 300 W voltage converter in the car.

The entire system will be activated via a simple threshold circuit, depending on the charging pressure. For example, it will activate above 0.8 bar of overpressure. Due to the simple design, without regulation of the water mass flow, more water will end up in the cylinder per piston stroke at low engine speeds than at high engine speeds. This is a problem limited by my pump.

I wanted to take the distilled water directly from the windshield washer reservoir. A check for container leakage should be eliminated for the time being.

It seems to me that the ideal injection point is directly after the air mass sensor, before the charge air temperature sensor. This way, you can immediately see the effect of the temperature reduction via the OBD system.

My biggest question now is:
Where can I find a suitable nozzle that is compatible with the pump's pressure (15 bar) and allows approximately 100 grams of water per minute to pass through while producing a fine mist? How fine does it need to be?


Okay, so please feel free to share your suggestions, information, and ideas. The big advantage for you is that, in this case, only my engine will be damaged .


Best regards,
Georg.

What is the actual temperature of the ASV charge air cooler (the intake air)?

My 1Z carburetor on my G3 (I don't remember if it was the original 0.9 bar or the later 1.1 bar version) quickly produced around 30,000 RPM at full throttle.

My front lower control arm is a bit more effective at 1.5 bar.

[img]/download.php?id=4052[/img]

Background information: Even when boosted to 165 horsepower, the charge air temperature remains within the optimal range for this engine.

Forget that nonsense. The sensor doesn't like moisture. Other measures are more effective.
While the relatively small LLK (likely referring to a liquid cooling system component) located on the driver's side, left, isn't ideal, I would only recommend cleaning it and performing a visual check for leaks.

The Ibiza has a rather small side marker light.

It was a long time ago, but I believe I reached temperatures above 80 degrees Celsius during the peak of summer.

Guys, you're such party poopers.

I want to work on a project over the winter, but you're all telling me it's pointless. I even believe you, and who would I be to doubt people with real practical experience ?

"Nevertheless, I find it interesting and simply want to try it." I also promise not to misuse the forum to cause trouble

Quote:

High summer temperatures exceeding 80 degrees Celsius were recorded.

Then something is wrong; the coolant normally used for this engine is usually fine.

"Common mistakes in car repair include forgetting to adjust the ventilation opening in the bumper or simply not cleaning it properly."

See also:
Okay, I'm ready. Please provide the German text you want me to translate.

Best regards, Rainer.

@Rainer,

Ah yes, your intercooler temperature measurement from the early days of this forum. Yes, they clearly show that water injection at 1 bar is not necessary.

My intercooler temperature was approximately 80 degrees Celsius, with an outside temperature of around 30 degrees Celsius, and this occurred at a boost pressure of approximately 1.2 bar. That roughly corresponds to your approximately 55 degrees Celsius with an outside temperature of 19 degrees and a boost pressure of 1 bar.

Okay, I admit it, this DIY project is completely pointless. It's now Saturday at 2:30 PM, and I'm finally going down to the garage to remove the (ZK-) head. The replacement of the ZKD (presumably a specific component) is currently a more pressing issue.


Best regards, Georg.

Do you know what the 2/1min on the ULKA means? It means 2 minutes of running, 1 minute of cooling. [img] The EP5 might not be the best choice; an Invensys might be better. I also have a 90-degree piece with external threads and a nipple, which might be useful. If needed, you can attach it to the Ulka pump, and you can replace the rear pressure relief valve with this 90-degree piece.

It would be even better to install a membrane pressure regulator from SAECO, as this would make the flow rate much more consistent, because otherwise the ULKA pump delivers the water in pulses rather than a constant flow. However, the connection is then made with a Teflon pipe and a clamping ring.

[/img]http://www.kaffee-optimal.de/WebRoot/Store5/Shops/62456718/4DEF/94FD/5009/50F8/20D7/C0A8/28BD/ADF9/11300_001.JPG{MARKER}

For nozzles, you can find them here: http://www.ps-motorentechnik-shop.de/index.php/cat/c89_Zerstaeuberduesen.html

Haha Deus, somehow, based on your contributions here in the forum, I could imagine that you've already dealt with something like this before.

I didn't understand the part about the 2/1 minute ratio. However, it might actually be relevant in practice, because I'm not really a heavy-duty user (and that's another reason why I don't actually need water injection).

The nozzles you linked are a bit too expensive for me. Not necessarily in an absolute sense, but I find them expensive for a simple nozzle. It's just the premium price you pay for a tuning product. But if it has to be, then it has to be.

Regarding the 90-degree bend, I might get back to you about it. I need to see where I can best position the pump after I assemble my engine. (By the way, is it "z" or "tz" after the "s" in German now?) Sure, here's the translation:

"Please correct the spelling."

It's been around for about 5 years.
has been tested and approved by leading manufacturers.
but unfortunately, it has disappeared into the drawers.

@Georg: No, seriously (I'm not kidding!), I've modified my espresso machine, so I've been doing a lot of research on these pumps, check valves, and pressure regulators.
Since I had previously used the EP5, it was almost like a subtle déjà-vu effect .

Georg_G wrote:

It seems to me that the ideal injection point is the position directly after the LLK (low-temperature coolant), before the charge air temperature sensor. This way, you can immediately see the effect of the temperature reduction via OBD.

Hi,

You've come up with a great winter project, and you'll be pleasantly surprised when it's completed.

Here's a tip regarding the placement of the nozzle:
Mount it as close as possible to the intake manifold inlet. This will allow for better combustion chamber/piston cooling, and it will prevent condensation of the water mist, which inevitably occurs when the water mist has to travel long distances.
A side effect is that coking is significantly reduced.
The nozzles with threaded housings proposed by Deus Violentia are perfect because they can be screwed into both the charge pressure hose and the aluminum housing.
And you only need ONE nozzle.
However, I had installed two nozzles: one for normal operation, and a second one. Nozzle activation during hang-up condition.

Then you need to install a check valve between the nozzle and the pump, otherwise it will force the water out of the pipes and the pump.
At the water intake point, a filter with the finest possible mesh should be used to filter out impurities; I use a sintered filter.
[img]www.mccoi.de/neushop/images/artikel/24.jpg[/img]

I would reconsider the choice of the current pump if it were to be used for an extended period.
Please note that the pump should be mounted below the waterline/reservoir.
Never underestimate the pump's operating time. For example, when we ride to Dolos with 4 people and 4 bikes on the trailer, the pump provides more than 30 minutes of continuous operation on the journey from Kufstein to Italy. The engine will then have to haul more than 3 tons up the Alps.

The reduction in intake air temperature is an effect of pure water injection, but the main effect/performance gain only occurs in the combustion chamber when the water/methanol mixture "vaporizes" explosively.

Furthermore, it reduces the levels of nitrogen oxides and soot.

Hi,
I don't quite understand the statements.
a)

Quote:

...and there is no condensation of the water mist, which inevitably occurs when the water mist has to travel "long" distances

. Why does condensation inevitably occur? Where should the further cooling of the intake air below the dew point be triggered after evaporation?
b)

Quote:

...the main effect/performance gain occurs only in the combustion chamber when a water/methanol mixture "vaporizes" explosively.

This mixture was likely used in gasoline engines to increase knock resistance, thereby allowing for higher compression ratios. Therefore, and due to the additional fuel, the power increase occurs. The amount of water injected must be adjusted to match the engine's load.
Injecting water into the intake air for evaporation and intercooling, and adding water in liquid form to the combustion process for cooling, are two distinct things.
hg
Herbert.

Hi Herbert,

a) Keyword: Contact condensation. Any type of atomized/nebulized liquid or wet steam will condense when it comes into contact with a surface (e.g., fine particles in the air or the wall of an intake hose). First, droplets form, then larger drops, eventually leading to a "stream" of water. This effect is amplified the longer the distance the water has to travel, and each angle or change in direction further intensifies the effect.
This has less to do with the dew point (saturation) and more to do with fluid dynamics.
In the worst-case scenario, it can even lead to a "water accumulation," for example, when water is injected before a (deeply located) low-pressure well. If larger quantities are drawn in and compressed, the piston will no longer reach top dead center, and the connecting rod will bend or break.

b) Yes, the knock resistance in gasoline engines is significantly improved with a water/menthol mixture, but this application is for a diesel engine, so it is negligible.
Both the pure water injection and the water/menthol mixture result in a power increase in the diesel engine, but the power increase is higher with the water/menthol mixture.

The performance increase achieved with the water injection system described here is due to both the cooling of the intake air (resulting in better volumetric efficiency due to increased air density) and the phase change of the water during the compression/combustion process.
If the water injection reduces the intake air temperature by, for example, 30 degrees, the exhaust gas temperature will be significantly lower, due to the phase change/evaporation process.

Herbert wrote:

.. and the addition of water in liquid form to the combustion process .....

If I understand correctly, this is a different process that has nothing to do with the topic we are discussing. In this process, water is directly injected into the combustion chamber via an injector nozzle. Typically, this is done in a staged process where, for each injection cycle, diesel is injected first, followed by water, and so on, all through a single injector nozzle. Another method using emulsions has not been very successful.

Hi Jürgen,
"Contact condensation" occurs when the surface of the wall is too cold (which then has nothing to do with fluid dynamics). Furthermore, condensation can occur in the wall area if the flow creates a vacuum on the wall surfaces (but then we are back to the dew point). I assume that a water injection system is designed in such a way that neither of these phenomena occurs. Otherwise, the design is unusable. Specifically, you should not inject fuel directly in front of the air intake in a low-temperature combustion (LLC) system, and the amount of fuel injected should be adjusted based on air mass and temperature.
What you're trying to achieve with the methanol additive in the diesel is not understandable. (Methanol, not menthol ). Are you aiming for a performance increase due to the methanol combustion? To achieve this, more diesel fuel is injected, which cools the exhaust gases and increases the torque.
BTW, with a properly designed intercooler, all that effort is unnecessary.
You are using the term "Verpuffung" incorrectly.
hg
Herbert.