Properties of water injection in the diesel cycle

[nicklockard]

Hmmm, I just intuitively thought it should show up in the pressure relationships somehow. Oh well.

 

[TDIMeister]

P6 is usually taken to be atmospheric pressure plus exhaust back pressure. If this value is significantly higher than P1, the exhaust system simply needs to be freer-flowing; I think injecting water downstream of the turbine would be of limited value compared to what it can do injected elsewhere in the engine cycle.

Cp,exh*T5 would almost certainly be lower with water injected in the intake -- T5 reduces to a greater degree than Cp,exh increases. I think this is easy to see.

So, the specific work may come down, but unclear is what happens to the isentropic efficiencies of the turbocharger as a result of the water injection.

The mass flow going through the turbine will increase by the amount of water injected (plus any further, albeit relatively minor, increase in air mass flow because of increased density from evaporative cooling). Whether the water is injected pre- or post-turbo also affects the freewheel balance.

 

[TDIMeister]

Hmmm, I just intuitively thought it should show up in the pressure relationships somehow. Oh well.

If the pressure relationship is constant, yes, it would, but as has been shown, it's not.

However, it's logical that an increase in turbine inlet pressure will also follow with an increase in the compressor outlet pressure. Maybe taking derivatives of F_U_B's EMP and IMP pressure (after dealing with the buzziness of the data) and measuring the time deltas between a positive turbine pressure gradient and the corresponding positive compressor pressure gradient.......

 

[TDIMeister]

Maybe taking derivatives of F_U_B's EMP and IMP pressure (after dealing with the buzziness of the data) and measuring the time deltas between a positive turbine pressure gradient and the corresponding positive compressor pressure gradient.......

Scratch that thought.

 

[nicklockard]

So should we coat our headers with IR reflective oxide coatings to maintain highest possible T5?

The idea of injecting water post exducer was to take advantage of the natural 1.5 bar and 350C pressure and temperature drops...rapid expansion and cooling should make droplets condense, creating negative pressure on that side (equivalent to 1- the partial vapor pressure of water at 'X' temperature,) increasing P5/P6. P6 is not really 1.0 bar if you have a stock exhaust (restricted at downpipe.)

Wait, is P6 after the exducer or before the turbine?

Note: I'm thinking of pairing post turbine water (only) injection with high pressure (fine fog) precompressor intake water/meth/acetone/surfactant injection triggered by intake air temps and pressures.

 

[TDIMeister]

P5, turbine inlet; P6, outlet.

Radiative heat losses at the exhaust manifold are small compared to the total heat content in the exhaust gases. Some people, notably purveyors of coatings and exhaust thermal wraps, say they make a big difference. IMO, I just shrug and say it's not a bad thing to do...

In an automotive application with limited space and weight to carry water around, injecting the water in the inlet as a charge cooler would have far more performance gains than doing it post-turbine. Besides, the post-turbine pressure cannot drop below ambient, and if more boost is not requested, the extra available turbine power from the larger pressure drop across it would be wasted. Far better to have a less-restrictive exhaust all-around.

 

[nicklockard]

[gratuitous gory crushed hardware post having hardly anything to do with topic]

You know that science demo where a 55 gallon drum is filled with hot steam and capped, then someone douses the outside with cold water and the drum is crushed like a beer can?

Jeez, I love that demo. Here's a pic of it on a larger scale:

[/]

 

[CentralOregon]

Water/Meth

I am putting in a Water/Meth kit and was very pleased to see all these articles,info and links.

Not free, but looks to be along similar lines of what you are looking into:

http://www.sae.org/technical/papers/940326

And here is the kit I am getting: I will be more than happy to run tests for you after I install it if you wish?


http://www.snowperformance.net/product.php?pk=8

 

[nicklockard]

http://forums.tdiclub.com/showthread.php?t=209796

 

[nicklockard]

http://www.aquamist.co.uk/phpBB2/viewtopic.php?t=1364[/url]]



WI on a duramax
I have spent much of the last year working through a problem with coolant overheating on GM diesel work vehicles. Since some of my findings are applicable to this effort, I will elaborate.

What I found was that the smallish compressor was working in the northeast part of the map, and in some thermal conditions, "off the page". On further investigation, the temperature of the compressure discharge emerged at near 600 F. I calculated efficiency around the 50% mark. Needless to say, this is in a word, ridiculous! Further investigation led me to question what the effect of this heatup (and expansion) is on the downstream plumbing. Something told me that I was seeing extaordinary head losses with this huge heatup. In other words, the velocity in the IC/CAC plumbing would be increasing dramatically, leading to higher frictional losses. It turned out, the difference between the compressor discharge pressure (work) and the intake plenum pressure (downstream) was nearly 6 psi! This at a plenum requirement of 32 psi. On the diesel this is around 60-65 lb/min of airflow.

Now hopefully I have not lost you. The plumbing restrictions when considering the 2.5" IC plumbing and the IC itself, totalled 6 psi. (2.5" is way too small)

After working through the compressible flow equations in a 2.5" conduit, it turns out the increased discharge temp was creating much of this loss, via increased air flow velocity. With fluid flow, there is apoint where the force required to push the fluid (air) through the straw (or IC pipe) becomes exponential, the curve quickly rising vertically at some flow rate. When this happens it is time for a diameter increase from the engineers. (what in fact happened, is that GM reduced the diameter from 3.0" in previous models, and to this day there is no known explanation why)

Adding insult to injury, that high temp product was leading to dramatic ambient temp increases behind the grill. The IC sits in front of the radiator, and measured ambient in front of the radiator, on the hot side of the IC, was 240 degrees! So the CAC was acting like a torch to heat coolant. I calculated something like 290,000 BTU/hr of heat exchange with the IC. Nuts!

But back to hot discharge product. The predictions for this non-adiabatic behavior, show that velocity increases dramatically, and adds a lot of added restriction in the plumbing. Clearly cooling the charge increases density, reduces velocity, and hence pressure losses. This means that, for a given desired intake plenum boost pressure, less work (discharge pressure) is required. This compressor now works less, which means higher efficiency. The improvement is cyclical in nature.

it appears that pre-compressor WI, PCWI, can be a performer or a deterent to performance. If you already operate in the efficient islands, on a humid day, then cooling charge can move you to lower efficiency on the left. It also leads to huge condensation effect in the IC. So PCWI would have limited usefulness on a properly designed forced induction platform with large stock amounts of charge cooling. But the undersized compressor, operating on a dry day, with excessive air box temps, should benefit big.

From my point of view, I do not share the idealized concept that WI provides quasi-isothermal compression. I believe that all the inefficiencies of non-adiabatic compression are in place even with WI. But naturally the beneficial impact of cooler charge can be seen in my explanation. But I don't believe that WI improves the efficiency of the compressor, from a purist sense. This assumes that there is no appreciable evaporation prior to compressor, as is the case in an axial mount nozzle in front of the nut.


_________________
Michael Patton (aka Killerbee)

Insightful analysis on precompressor fumigation, IMO.

I read it as saying: "if your turbo is ever off map, or at the right hand, upper quadrant of the flow map, then PCWI will help (by translating the entire map rightward, putting your operating point back onto a "good" efficiency island--since wet, denser, cooler air moves more efficiently through plumbing (with less pressure drop) than hot dry air.

Now, methanol injection is another matter. Methanol injection serves to effectively advance combustion timing, AFAIK. I'm not sure if it has any effect (or measurable effect) on turbo compressor efficiency/ air mass throughput. I think it's cooling effect is too weak.

Ideally, we could have separate methanol and water tanks and map the injection of each to be ideal to the conditions.

Edit: PCWI should be maximum when your turbo is near or at "choke."

 

[dvldoc]

And very right, But it's always better to look at it on a dyno. But if you have a proper sized turbo it's not a issue.

Example 1gph pre-turbo nozzle, 2gph post turbo nozzle. 2001 starex, Notice the top end does not drop off because the turbo is dropping out of it's efficiency range. The turbos on these are tiny and the engines are basic and don't even have electronics.

Compare the dynos and you will see just what he is talking about and what we have been talking about for a while now.

Now no pre-turbo nozzle on a similar vehicle, this 1997 Mazda MPV does have electronics and a higher stock HP, but the turbo is also just as small for the vehicles weight. It still makes awsome power on the alcohol injection but the turbo looses out at around 3600rpms. Showing that the pre-intake nozzle only effects the compressor when it is in the efficiency range once it drops out of that it still drops off at the same rate and does not make more power even though the EGT's are lower.

 

[nicklockard]

Thanks devil
I also found some very good forum discussion here about triggering PCWI off of the MAF signal instead of MAP because MAF inherently includes load demand (I mean to say that the amount of air your engine ingests is loosely proportional to load requested; even though air/fuel ratios change across load, this shouldn't matter so much.)
http://www.aquamist.co.uk/phpBB2/viewtopic.php?t=267&postdays=0&postorder=asc&start=330
Devil, our stock turbo's have a very restrictive exhaust side. We see EGT's go through the roof often, unless larger downpipes are installed. What EGT changes have you seen with PCWI on a 4 cylinder turbo diesel?

 

[dvldoc]

You can easily trigger them off of MAF many kit makers controllers including ours use a 5v input signal so a MAF signal will work just as well. Turbo diesels behave totally different than gasers when it comes to alcohol injetion.

 

[nicklockard]

Thanks Devil.

Can any of the engineering-oriented folks help us understand Fix_Until_Broke's data graph posted in post #57 better?

In particular, why are there 3 distinct regimes of relationship between EMP and IMP?

I think understanding these relationships can help us know when to inject water/meth and how to trigger it best.

 

[TDIMeister]

Nick, post 58 gave a mathematical and theoretical background to turbocharging. EMP does not follow IMP because, as already shown, the relationship is a function of the total turbocharger efficiency, mass flow and turbine inlet temperature T5. All these parameters are constantly changing with engine load and speed. The snap changes in EMP are due to the VNT vanes being snapped open and shut.

 

[nicklockard]

Why are the vanes snapping shut so rapidly? I thought the whole point of VNT was fine control. Is that a consequence of his tune strictly? I mean if someone with a different tune logged data it could look dramatically different?

Questions:

1. How can mass air flow be falling while he is still accelerating? Is the car just overfueling and smoking?
2. Why do the vanes suddenly slam open at 4450 rpm's when he is still accelerating? Is the ECU trying to prevent an overspeed?
3. Between 3300-3800 rpm, EMP>IMP. Is the turbo crossing over efficiency islands/holes?

1. 

Peak acceleration occurs at ~2400 rpm's and not at peak power which is hugely interesting in and of itself.

 

[nicklockard]

Devil, or anyone:

Does pre-compressor water/methanol injection exacerbate turbo surge at low rpm?

 

[Rub87]

Why are the vanes snapping shut so rapidly? I thought the whole point of VNT was fine control. Is that a consequence of his tune strictly? I mean if someone with a different tune logged data it could look dramatically different?

Questions:

1. How can mass air flow be falling while he is still accelerating? Is the car just overfueling and smoking?
2. Why do the vanes suddenly slam open at 4450 rpm's when he is still accelerating? Is the ECU trying to prevent an overspeed?
3. Between 3300-3800 rpm, EMP>IMP. Is the turbo crossing over efficiency islands/holes?

1. 

Peak acceleration occurs at ~2400 rpm's and not at peak power which is hugely interesting in and of itself.

1: Mass airflow's unit is mg per engine volume, so it's clear that at high engine rpm the VE of the engine dropts due pumping losses atc, so the cilinder filling is less complete..
2: And yes indeed at 4500rpm EMP drops, this is because fueling is cut dramaticly I think to prevent engine and turbo overspeed..
3: at the req boost at 3000rpm turbo begins to choke the engine, my experience for example it with my vnt20 at 1.7bar relative boost pressure EMP only became greater than EMP at 3600rpm, so this says alot over the efficiency and size of the little turbo used (vnt15)..

4 this is logical because peak torque happens at 2400rpm..

 

[Rub87]

Devil, or anyone:

Does pre-compressor water/methanol injection exacerbate turbo surge at low rpm?

No, I think it will just do the opposite, because the compressor is more efficient, comp out temp will be lower, so IAT too, so air will be more dense, this means more mass flow at the same PR, this means we move to the right on the compressor map, away from the surge line..

This is also why on a hot day, or when the intercooler is heat soaked surge will occur more frequently..

 

[TDIMeister]

Reducing EMP = opening vanes. I'm sure you picked up on this already. The apparent fast increase in EMP in the first phase over 5 seconds logged doesn't give me any reason to worry.

 

[nicklockard]

No, I think it will just do the opposite, because the compressor is more efficient, comp out temp will be lower, so IAT too, so air will be more dense, this means more mass flow at the same PR, this means we move to the right on the compressor map, away from the surge line..

This is also why on a hot day, or when the intercooler is heat soaked surge will occur more frequently..

Thank you

So, how about this idea:

1. Install an EMP pressure transducer reading 0-5 Volts.
2. Take the voltage signal from an EMP pressure transducer and subtract the 0-5 Volt MAP signal.
3. Whenever EMP-IMP > 0, trigger water/meth injection proportional to the difference signal.
4. Since when EMP>IMP, the turbo is off of an efficiency island, pre-compressor water meth injection will:

a. increase turbo efficiency
b. increase overall volumetric efficiency (turbo breathes better and water droplets going into combustion chamber decreases compression work--recall that water/meth is flashed to steam after turbo but then recondenses inside intercooler. I think we can get it to make fine droplets again through some tricks (try the simple one first (surfactant,) then progress to a harder trick if that doesn't work (Nick's crazy idea #3.)
c. effectively advance combustion timing at a certain water/meth ratio. Many people on the aquamist forums are saying that diesels knock at high methanol ratios, so let's be careful! 12.5% methanol to water volume ratio is suggested widely as a perfectly safe point to start at. 50/50 is said to create knock at high boost
d. presence of water in combustion cylinder prevents excessive cylinder pressures which preserves head gaskets--good for reliability.

 

[Fix_Until_Broke]

Why are the vanes snapping shut so rapidly? I thought the whole point of VNT was fine control. Is that a consequence of his tune strictly? I mean if someone with a different tune logged data it could look dramatically different?

Questions:

1. How can mass air flow be falling while he is still accelerating? Is the car just overfueling and smoking?
2. Why do the vanes suddenly slam open at 4450 rpm's when he is still accelerating? Is the ECU trying to prevent an overspeed?
3. Between 3300-3800 rpm, EMP>IMP. Is the turbo crossing over efficiency islands/holes?

1. 

Peak acceleration occurs at ~2400 rpm's and not at peak power which is hugely interesting in and of itself.

Nick - Yes, the ecu in my brain is trying to prevent overspeed by lifting my right foot . My appologies for not correcting the above graph in the other threads that it's in. The above picture has about a 2 second offset in the analog pressure measurements versus the VagLog data. This thread has all the results from the 4-5 setups that I've tested thus far and they are much more synchronous (same data, just with the time offest removed).

Sorry for the confusion. I have some spare exhaust flanges that just need to be drilled out so I can measure post turbo pressures as well. Current setup will measure any 4 pressures, but can easily be expanded up to 16 with no problem. I have all these things that I want to do, but just not enough time to do them all .

 

[nicklockard]

Nick - Yes, the ecu in my brain is trying to prevent overspeed by lifting my right foot . My appologies for not correcting the above graph in the other threads that it's in. The above picture has about a 2 second offset in the analog pressure measurements versus the VagLog data. This thread has all the results from the 4-5 setups that I've tested thus far and they are much more synchronous (same data, just with the time offest removed).

Sorry for the confusion. I have some spare exhaust flanges that just need to be drilled out so I can measure post turbo pressures as well. Current setup will measure any 4 pressures, but can easily be expanded up to 16 with no problem. I have all these things that I want to do, but just not enough time to do them all .

Oh!

I'm interested in duplicating a setup for data logging if it's not very expensive.

Thanks

 

[Fix_Until_Broke]

PM Sent

 

[nicklockard]

Technical forum discussions:

http://www.eng-tips.com/faqs.cfm?fid=811
http://www.eng-tips.com/viewthread.cfm?qid=88490&page=57
Rate constant arguments:

http://www.mayfco.com/injchem.htm

Did you ever think about what happens in the combustion chamber when you inject water into the charge air? In addition to cooling the air, they are many other things going on. The presentation below, taken from a web site with the authors permission, I might add, describes in understandable terms what happens in the combustion chamber. Wish I was smart enough to have written it!
I wish to thank the author for letting me use his words. Every little bit of knowledge helps us to run faster without blowig any thing up and scattering car parts all over the track.
"[Reformatted slightly for readability, but otherwise as posted.]
From: Robert Harris
To: DIY_EFI@lists.diy-efi.org
Subject: Water and its effect on combustion.
Date: Mon, 10 Jul 2000 10:24:08 -0700
Message-ID: <9ptjms0uu4oe292mpk6a6vhm2hn8bu9h1j@4ax.com>

Let us take a quick look at ignition. Those who have a Heywood can look it up
- mines on loan so going by memory. The first thing that happens is a plasma
cloud is formed by the arc consisting of super heated electron stripped atoms.
When this cloud "explodes" a ball of high energy particles is shot outward.
The highest energy particles are the hydrogen atoms - and they penetrate the
charge about 5 times as far as the rest of the particles. As they lose energy
and return to normal temps - about 5000 k - they begin to react chemically
with any surrounding fuel and oxygen particles. The effectiveness of spark
ignition is directly related to the availability of free hydrogen. Molecules
containing tightly bound hydrogen such as methanol, nitromethane, and methane
are far more difficult to ignite than those with less bonds.
During combustion - water - H2O ( present and formed ) is extremely active in
the oxidation of the hydrocarbon. The predominate reaction is the following:
OH + H ==> H2O
H2O + O ==> H2O2
H2O2 ==> OH + OH
Loop to top and repeat.
The OH radical is the most effective at stripping hydrogen from the HC
molecule in most ranges of combustion temperature.
Another predominate process is the HOO radical. It is more active at lower
temperatures and is competitive with the H2O2 at higher temps.
OO + H ==> HOO
HOO + H ==> H2O2
H2O2 ==> OH + OH
This mechanism is very active at both stripping hydrogen from the HC and for
getting O2 into usable combustion reactions.
Next consider the combustion of CO. Virtually no C ==> CO2. Its a two step
process. C+O ==> CO. CO virtually drops out of early mid combustion as the O
H reactions are significantly faster and effectively compete for the available
oxygen.

Then consider that pure CO and pure O2 burns very slowly if at all. Virtually
the only mechanism to complete the oxidization ( Glassman - Combustion Third
Edition ) of CO ==> CO2 is the "water method".
CO + OH ==> CO2 + H
H + OH ==> H20
H2O + O ==> H2O2
H2O2 ==> OH + OH
goto to top and repeat.
This simple reaction accounts for 99% + of the conversion of CO to CO2. It is
important in that fully two thirds of the energy of carbon combustion is
released in the CO ==> CO2 process and that this process occurs slow and late
in the combustion of the fuel. Excess water can and does speed this
conversion - by actively entering into the conversion process thru the above
mechanism.
The peak flame temperature is determined by three factors alone - the energy
present and released, the total atomic mass, and the atomic ratio - commonly
called CHON for Carbon, Hydrogen, Oxygen, and Nitrogen. The chemical
reactions in combustion leading to peak temperature are supremely indifferent
to pressure. The temperatures and rates of normal IC combustion are
sufficient to cause most of the fuel and water present to be dissociated and
enter into the flame.
As can be seen above, water is most definitily not only not inert but is a
very active and important player in the combustion of hydrocarbon fuel.
Ricardo and others have documented that under certain conditions ( normally
supercharged ) water can replace fuel up to about 50% and develop the same
power output, or that the power output can be increased by up to 50% addition
of water. This conditions were investigated by NACA and others for piston
aircraft engines. It is important to note that these improvements came at the
upper end of the power range where sufficient fuel and air was available to
have an excess of energy that could not be converted to usable pressure in a
timely manner.
As a side note - Volvo recently released some SAE papers documenting the use
of cooled EGR to both reduce detonation and return to a stoic mixture under
boost in the 15 psi range - while maintaining approximately the same power
output. Notice - they reduced fuel and still get the same power output.
When you consider that EGR consists primarily of nitrogen, CO2, and water ( to
the tune of about two gallons formed from each gallon of water burned ), you
might draw the conclusion that it also was not "inert". They peaked their
tests at about 18% cooled EGR - which would work out to about 36% water
injection and got about the same results under similar conditions that the
early NACA research got."

Thinking of adding hydrogen peroxide to water also during summer months. (water + H2O2 + surfactant.)

Does anyone have either of these books? http://www.amazon.com/s/ref=nb_ss_g...keywords=+Glassman+-+Combustion+Third&x=0&y=0

 

[nicklockard]

I have some spare exhaust flanges that just need to be drilled out so I can measure post turbo pressures as well. Current setup will measure any 4 pressures, but can easily be expanded up to 16 with no problem. I have all these things that I want to do, but just not enough time to do them all .

LOL, I want to do the same things you are...I'd like to track post turbo temp/pressures also...let's make our cars into rolling dynos FUB, hehe.

 

[nicklockard]

Injecting water into exhaust stream:

http://www.msdpowersports.com/pwc_waterinjection.html

 

[ATBio]

Adding water, surely we want to do that mainly towards the end of the expansion stroke.

Diesels run at part load and lower temperature have less NOx, so a buit of water on intake air is usefel, but the potential gains come from reducing engine block cooling by spraying water onto the piston head and the cylinder sides as well. The diesel should also be emulsified to increase burn of particulates at TDC and also reduce peak temperature? The goal is a municipal diesel genset with a DC output, no after treatment of NOx required, no turbocharging? and a water heat-to-power process in the cylinder. Municipal gensets and hybrids run commercially can obtain a water supply readily and transport with the fuel.

A steadily run hybrid could also run on a standard emulsified fuel mix, and the operator can still use red low duty diesel.

The other thing, if we are not using exhaust pressure for compression, we can run a turbo-generator and make electricity, again something that can operate nicely in a diesel run constantly. Adding water to the exhaust during the end of the expansion phase should deliver more pressure for such a turbine?

 

[mojogoes]

So in laymans terms is water methanol overall a good way to go in our 1.9L 2.0L tdi's so at what mix level in your eyes is safe ..............you seamingly being a bio expert.

 

[ATBio]

I'm afraid Im not a biofuels expert, the bio is from elsewhere