Water Injection Project
(Was Water Induction Project)

Began: February 5, 2002

Objective: To design and implement the most efficient water injection system for a naturally aspired engine.

Reason: A lot of people have asked me why I am doing this, or told me that it will not work.  My main reason for building this is not for cooling the air, but for the increase in power from the expansion of the water spray into steam (which is much more than the exploding fuel mixture alone). Converting water to steam is a product of the heat generated from the fuel ignition. This lowers the temperature and modifies the characteristics of the expanding gasses in the cylinder during the power stroke.  The ignition flame front will travel slower with the water injection and as such timing can be advance for even more power gains. The slower moving flame front causes the power stroke to be accomplished with a lower peak pressure, because energy is being absorbed from the heat generated by fuel ignition to convert the water in to steam. Though the power stroke duration is longer, there is more energy because of steam expansion, creating considerably more torque with little increase in horsepower. Which will be an advantage when accelerating.

Status: Ongoing, working on modeling right now (update 02/20/2002)

Project: Initially this project was the Water Induction Project, but after some research I determined water injection was the way to get the most efficient water injection system for a naturally aspired engine.  Below was my first design idea/experiment which I made after a lot of research (but no experiments or modeling....bad Matt), you can click on it for greater detail:

I decided against using water induction (water being sucked into the intake of the engine) and opted for water injection (water being forced / pushed into the intake of the engine) because after doing a lot of research I saw that the most efficient way of giving the engine more power from water would be with a precise mixture of water/fuel (the water/fuel ratio) between 10% to 30%.  The most efficient mixture will depend on a lot of factors (intake, exhaust, engine heat range, engine temperature, engine efficiency, rpm), so I will have to run experiments and models to determine what water fuel ratio is optimal.  If I am going to have a constant water/fuel ratio (like the air/fuel ratio of 14.7) for maximum efficiency then I will need to match the water injection to RPM.

The first experiment I did was to determine how much water injection could cool hot air.  I used a 10 inch long, 2 inch wide tube with air being blown into it at a rate of 0.66 pph (lbs per hour, wait...what?  you didn't know air has weight?  yup, 14.7 PSI) and a temperature of 248° F.  The water I used was cooled to 77° F and injected from a high mist nozzle with the use of a 100 PSI electric water pump.  I recorded results with a digital lab grade thermometer.  I was able to set three different water flow rates by varying the pressure of water being pumped out of the electric water pump, here is the data I recorded:

Conclusions I have drawn: 26.4 pph of water was able to entirely cool off a 6.283 square inch inlet over 10 inches.  Ideas are forming in my head for locations to mount the water injector, and 10 inches after from the cylinder is where I am thinking of locating the water injector (right next to the EFI in other words).  The 248 ° F is much hotter than even a stock intake on a NA engine.  I used the vent on the stove as the heat source because it was the only heated air source I could find that I was able to regulate the temperature within a degree or two.  I played around with the variables a bit, I increased air temperatures up to 350° F the temperature coming out of the tube with 6.6 pph and 13.2 pph, but 6.6 pph still cooled the air by 20.83% and 13.2 pph by 41.84%, 26.4 pph still cooled the air to 77° F, so it obviously can absorb a lot of heat.  I also changed the length of the pipe from 6 inches to 24 inches, the longer the pipe, the more the air was cooled, the shorter the pipe the less it was cooled, pretty simple.  The 26.4 pph was able to entirely cool up to 300° F with a 6 inch long pipe, impressive.  What have we learned?  The farther back in the air system the water injector is placed, the more it will cool the incoming air, giving you a denser charge (but at the same time less will enter the cylinders to increase your torque and to retard timing/knock, so there must be some balance between the two).  Water injected at a specific rate will give you a fixed % change in temperature drop based on the water/fuel ratio (though if that is allowed to change by not matching water flow to fuel flow then % change in temperature drop will not be fixed)

Next experiment: I will experiment with injecting fixed rates of water into a NA engine and record data of what happens.

Possible ideas: I need some way of matching the flow of water to that of fuel.  I am thinking of using a BASIC Stamp II to monitor the RPM of the engine and control the water injector (either monitor passively or actively, possibly via the fuel injector voltage connection, not sure though yet).  There may be a simpler way that using a computer to monitor and maintain the proper ratio.

Modeling: I still have no idea what the 2.2 L EFI's flow rate is (lbs per hour - pph), I am guessing that it may be around 50 pph or higher (else the Shadow couldn't make 93 hp stock), but it could be as high as 75-100 pph (based on the fact that my Shadow is now pushing a little over 200 hp, though that could be from increased engine and system efficiency and as such the table below will not be correct).  Here is the approximate flow rates for a typical NA engines:

The above assumes fuel weight of 6 lbs per gallon and a typical brake specific fuel consumption (BSFC) of 0.5. 

I will update you when/if anything changes (ideas, experiments, models, etc.), last update 2/20/2002.

Status: Have a working prototype running on the Shadow (update 05/17/2002)

This is the actually the second model I am using, the first one was used just for experiments and to figure out what I really would need to build a good working prototype.  This second model is powered by a windshield washer pump and uses a McMaster-Carr water jet nozzle injector.  It is a real simple set up, the pump is electronically regulated by the TPS voltage (Throttle Position Sensor), though I have played around with using the MAP (Manifold Absolute Pressure) Sensor's voltage.  I placed the water injector directly over the throttle body, on top of the air box, and used a siphon block above the water injector to keep water from dripping down into the throttle body when the car is off.  The Shadow's new motor eats only distilled water through a stainless steel inline water filter. 

The electronics try to keep the fuel to water ratio around 14:7:1 (though 5 to 20:1 all work, and that's about the total range that this prototype produces).  I also made a manual on / off switch for the prototype, so that I could turn the water injection off (since it is on all the time) 5 to 10 seconds before I shut off the engine (to burn out any moisture). 

Results:  All I can say is WOW!  There is a noticeable gain in accelerating (at any speed) and a great difference in fuel economy.  The water also acts as octane booster (in effect, since the water causes the ignition flame front to travel slower) and I have been able to raise the ignition advance to 15 degrees without any knock on normal cheap 87 octane fuel (and higher, but with no noticeable gains).  The ignition advance along yields a very noticeable increase in performance (5 horsepower, maybe closer to 10 on a modified engine).  I also picked up another 2 to 4 MPG increase on an engine that is heavily modified and already getting well above stock fuel consumption (getting closer to 40 MPG now, probably could get it if I laid off the throttle when on the street).  I have played around with using 10:1 and 14.7 fuel:water ratio be my base (optimal) ratio, and finally decided on 14.7:1 because it seems to give me a smoother temperature reading and the best performance at both high and low rev. 

Status: Latest results and more data from water injection (update 07/01/2002)

After using and tweaking the first prototype (and the electronics), all I can say is WOW x 2.  One thing I didn't mention much in my last update (and side note * you need to post more updates on this project *) was the incredible difference in intake temperature.  With about a 22:1 fuel:water ratio the intake air temperature is the same as the water temperature.  Another neat side effect is that the water actually cools the engine from inside the combustion chamber (like an internal coolant), and here in Mesa, AZ it takes a lot of strain off the somewhat overwhelmed cooling system when you really push the car.  This is great in the summer time here in Mesa, AZ (110 degree days are the normal temperature days....), I really, really have noticed the difference with water injection this summer.  The only problem is that it is so hot here, that the water heats up in the water tank, so the intake is not as cool as it could be.  But a bag of ice has solved that problem a few times.  I've experimented more with this prototype trying to get as much data as I can to figure out the best way to make a permanent water injection system.

This is real data from running the Shadow (not driving it since the ram air could through off the temperatures, plus some reading would be at different points in the fuel:water curve).  I recorded results with a digital lab grade thermometer, the temperature of the water was 92° F and the ambient temperature in front of the car was 108° F:

For every 20° F decrease in intake temperature you will gain about 1% horsepower.  So a 91° F or 49.7% decrease in intake temperature should net about 4.55% more horsepower alone.  This by itself already makes the water injection system a good mod to the Dodge Shadow.

I was able to find the limit that this setup will allow you to advance your timing.  This current prototype will allow 6° ignition advance with the same low 87 octane fuel without any knock.  Though I have stuck around 15°  to 16° advance, but this may be more useful for people with larger engines or different camshafts. 

Over the past few months I have discovered some disadvantages to using water injection.  If the fuel:water ratio is too high, some excess water will remain in the combustion chamber after it should have been changed to steam.  If this situation arises it will thin the oil slightly.  I was experiencing this slight (very, very slight) oil thinning with the 10:1 fuel:water base ratio, changing over to 14.7:1 solved the problem.  I did worry about ring wear from using the water injection (steam cleaning effect in the combustion chamber stripping the walls of oil, increasing friction on the rings), so much so, that I was checking my compression nearly every day.  After no real changes in compression tests over a month period, I decided to switch to a thicker oil (10w30 and later 15w50 over 5w30 Mobil 1 Synthetic) and use Engine Restore (which I have seen work good in the past on my engine and others) as precautionary measures, and still after about 5 months of use, no changes real changes in compression on any cylinder.  All in all, I would have to say the positives much outweigh the negatives with water injection, I have seen a good gain in performance with the only major problem being all the constant worrying I had to live with originally.

Copyright © 2004 Karlsnet.com. All Rights Reserved.

[an error occurred while processing this directive]