Research Experimental 8
12-24-2011, 12:40 PM
#1
Research Experimental 8
The Wankel engine in its current form has some inherent flaws. First because of it's design it has a hot side and a cold side. While this allows for better power due to a colder air charge, it also causes a continuous hot spot which can lead to dry channeling in the coolant passages and engine failure. This also limits the heat transferred initially to the fuel which in turn limits the usable amount of fuel because liquid fuel does not burn (only the vapors do).
In an effort to solve these flaws I have created Research Experimental 8. The first thing I changed was using a water free coolant. My choice is close in composition to Evans NPG but widely more available and it has allowed me to use low grade pump gas with no fear, as well as removing the potential for coolant channel dryout and engine boil-over.
The second and more revolutionary change I have made is waste heat recovery and steam injection. By using a copper heat exchanger on the exhaust manifold fed by a water injection system to create steam, and then injection the steam into the intake manifold I have significantly increased my fuel economy while at the same time gaining a wider margin to detonation and the potential to advance timing and lean out the mixture. This may seem like some sort of voodoo, but the concepts are quite clear. Waste heat recovery is simple enough, the exhaust manifold provides more than enough heat to create the steam I need.
Yes that is copper to brass to stainless. Yes i do understand galvanic corrosion. If I was doing this for a production model I would be placing stainless inside the exhaust pipe, since that isn't an option I'm using copper for it's heat transfer and stainless for its lack of heat transfer. The brass was just what I could find at the plumbing store. It's not perfect, it's not permanent, but it does work for now.
The steam is then injected into the intake. Right now it's going in a pvc port I JB welded into place originally for water injection. This is less than optimal, but works for now. My next modification is to inject to the maintenance ports because they prevent water from hiding behind the closed SSV.
Now for the physics that make this work: The steam enters with the fuel air charge. As there are no moisture separators the steam is very low quality (high water concentration). The steam then interacts with the air and some of it condenses while it heats the fuel air charge to around 170 degrees ( saturation temperature depending on vacuum). Now the engine begins to compress the charge. As the charge is compressed the properties of the water (which again is a large part of the steam) change and the water begins to evaporate into the air. I say evaporate and not boil because it has more to do with the dew-point of the air than the boiling point of the water at this point in the cycle. This causes the compression to be nearly isothermal. This near isothermal compression reduces the heat load on the engines cooling system and more importantly prevents hot spots from forming which could cause detonation. Then the fuel air charge is ignited with the spark plugs. The remaining water which is finely dispersed in the charge is then flashed to steam both adding extra power and lowering peak combustion temperature. The end result is lowering internal engine temperatures while adding power. The effect on EGTs is minor because the water slows the burning of the fuel and causes more burning fuel to make it out the exhaust. This is easily modified with timing, but without a tested and reliably system I'm not going to mess with tuning. The effects on coolant temperature are major. So much so that a higher temperature thermostat may be required to help retain the heat that is so useful for both power and fuel economy. So far however the results are most promising.
Currently the system is set up for vacuum drag only. I plan on getting the pump progressive controller tuned in soon. Also there is no shut off for start up/shutdown. This causes me to have steam out the tailpipe for a couple minutes during start-up as water percolates out of the steam generator.
This will be solved by a solenoid which will connect pre throttle body air with the outlet of my water nozzle to remove vacuum and stop the vacuum drag.
It's currently on order along with a permanent catch can.
I'd love to tell you MPG results, but I just got the system set up and do not have enough road time for solid data. What I can tell you is that the indicated OBD2 MPG is incredible and historically only a couple mpg away from actual. A couple tanks of gas will tell the story. I have no control MPG data at this time because I'm sure there will be some heckler who will claim my gains are due to steam cleaning the carbon deposits out of the engine, so I'll take that data after I have completed a few tanks with water injection.
Oh and there is the distinct possibility that I will blow up my engine by either too much water or a bad tune when I get that far, or even loosing water injection with a good tune so as always DO NOT TRY THIS AT HOME. Let me risk my engine so you don't have to! I will happily post results as I get them as well as any additional data. I want this test to be completely transparent and repeatable.
In an effort to solve these flaws I have created Research Experimental 8. The first thing I changed was using a water free coolant. My choice is close in composition to Evans NPG but widely more available and it has allowed me to use low grade pump gas with no fear, as well as removing the potential for coolant channel dryout and engine boil-over.
The second and more revolutionary change I have made is waste heat recovery and steam injection. By using a copper heat exchanger on the exhaust manifold fed by a water injection system to create steam, and then injection the steam into the intake manifold I have significantly increased my fuel economy while at the same time gaining a wider margin to detonation and the potential to advance timing and lean out the mixture. This may seem like some sort of voodoo, but the concepts are quite clear. Waste heat recovery is simple enough, the exhaust manifold provides more than enough heat to create the steam I need.
Yes that is copper to brass to stainless. Yes i do understand galvanic corrosion. If I was doing this for a production model I would be placing stainless inside the exhaust pipe, since that isn't an option I'm using copper for it's heat transfer and stainless for its lack of heat transfer. The brass was just what I could find at the plumbing store. It's not perfect, it's not permanent, but it does work for now.
The steam is then injected into the intake. Right now it's going in a pvc port I JB welded into place originally for water injection. This is less than optimal, but works for now. My next modification is to inject to the maintenance ports because they prevent water from hiding behind the closed SSV.
Now for the physics that make this work: The steam enters with the fuel air charge. As there are no moisture separators the steam is very low quality (high water concentration). The steam then interacts with the air and some of it condenses while it heats the fuel air charge to around 170 degrees ( saturation temperature depending on vacuum). Now the engine begins to compress the charge. As the charge is compressed the properties of the water (which again is a large part of the steam) change and the water begins to evaporate into the air. I say evaporate and not boil because it has more to do with the dew-point of the air than the boiling point of the water at this point in the cycle. This causes the compression to be nearly isothermal. This near isothermal compression reduces the heat load on the engines cooling system and more importantly prevents hot spots from forming which could cause detonation. Then the fuel air charge is ignited with the spark plugs. The remaining water which is finely dispersed in the charge is then flashed to steam both adding extra power and lowering peak combustion temperature. The end result is lowering internal engine temperatures while adding power. The effect on EGTs is minor because the water slows the burning of the fuel and causes more burning fuel to make it out the exhaust. This is easily modified with timing, but without a tested and reliably system I'm not going to mess with tuning. The effects on coolant temperature are major. So much so that a higher temperature thermostat may be required to help retain the heat that is so useful for both power and fuel economy. So far however the results are most promising.
Currently the system is set up for vacuum drag only. I plan on getting the pump progressive controller tuned in soon. Also there is no shut off for start up/shutdown. This causes me to have steam out the tailpipe for a couple minutes during start-up as water percolates out of the steam generator.
This will be solved by a solenoid which will connect pre throttle body air with the outlet of my water nozzle to remove vacuum and stop the vacuum drag.
It's currently on order along with a permanent catch can.
I'd love to tell you MPG results, but I just got the system set up and do not have enough road time for solid data. What I can tell you is that the indicated OBD2 MPG is incredible and historically only a couple mpg away from actual. A couple tanks of gas will tell the story. I have no control MPG data at this time because I'm sure there will be some heckler who will claim my gains are due to steam cleaning the carbon deposits out of the engine, so I'll take that data after I have completed a few tanks with water injection.
Oh and there is the distinct possibility that I will blow up my engine by either too much water or a bad tune when I get that far, or even loosing water injection with a good tune so as always DO NOT TRY THIS AT HOME. Let me risk my engine so you don't have to! I will happily post results as I get them as well as any additional data. I want this test to be completely transparent and repeatable.
Last edited by Harlan; 12-24-2011 at 12:43 PM. Reason: Applied emphasis to "Do not try this at home." ps. Please don't!
Good luck and keep us posted as you get more data
12-24-2011, 01:11 PM
#4
Registered
iTrader: (2)
Join Date: Oct 2007
Location: Pacific Northwest
Posts: 16,684
Likes: 0
Received 239 Likes
on
109 Posts
What is the actual benefit of steam injection vs normal water injection? Your description sounds like the same thermal reaction of water injection (assuming it's a fine enough spray).
Is this benefit of steam vs proper fine water mist worth the complication of running water to the exhaust for the steam creation?
(Serious questions)
Is this benefit of steam vs proper fine water mist worth the complication of running water to the exhaust for the steam creation?
(Serious questions)
12-24-2011, 02:30 PM
#5
What is the actual benefit of steam injection vs normal water injection? Your description sounds like the same thermal reaction of water injection (assuming it's a fine enough spray).
Is this benefit of steam vs proper fine water mist worth the complication of running water to the exhaust for the steam creation?
(Serious questions)
Is this benefit of steam vs proper fine water mist worth the complication of running water to the exhaust for the steam creation?
(Serious questions)
12-24-2011, 03:21 PM
#9
will not start up be a problem?
what about deceleration--assuming this is working of a vacuum injection?
what about water in the oil?
not being critical--interesting concept of steam over water injection. maybe like the old smokey theory of pre heating the gas?
what about deceleration--assuming this is working of a vacuum injection?
what about water in the oil?
not being critical--interesting concept of steam over water injection. maybe like the old smokey theory of pre heating the gas?
12-24-2011, 03:47 PM
#10
Deceleration will collect water, but the engine itself will still be hot and most of it should carry on through. I have not seen any stumbling during or after deceleration. Of course with the clutch in or the transmission in neutral this is the same as idling, which is also not problem.
Water in the oil is a consequence of a rotary naturally. The existing oil system should remove water from the oil so long as it gets up to normal operating temperature and has a vacuum on the oil sump. I will monitor this closer though, thanks for the heads up.
I'll post more pictures as I get the rest of it presentable. Right now it's raining outside, so no progress or pictures.
12-24-2011, 05:59 PM
#14
You can get 24mpg out of a stock rx8. Perhaps even more.
What are your goals? How are you going to tune it? How will you deal with the performance loss? What will you do to maintain a good oil film inside the engine?
It's a curious project but I'm skeptic it will be of any sue.
What are your goals? How are you going to tune it? How will you deal with the performance loss? What will you do to maintain a good oil film inside the engine?
It's a curious project but I'm skeptic it will be of any sue.
12-24-2011, 06:18 PM
#15
You can get 24mpg out of a stock rx8. Perhaps even more.
What are your goals? How are you going to tune it? How will you deal with the performance loss? What will you do to maintain a good oil film inside the engine?
It's a curious project but I'm skeptic it will be of any sue.
What are your goals? How are you going to tune it? How will you deal with the performance loss? What will you do to maintain a good oil film inside the engine?
It's a curious project but I'm skeptic it will be of any sue.
I have an accessport, and racetuner, but I don't know how I'm going to tune it yet. I need the system to be completely bulletproof before I tune around it, and then I will still need a margin for error. So that is more a long term goal.
So far I have noticed no performance loss. I have actually seen a performance gain because the engine/coolant system does not heat soak, so it's fresh and ready to go all drive long.
Oil film is a really tricky matter. Folks running water injection seem to have no problem, perhaps it's due to the relatively small amounts of water being added. I premix, but conventional oils do not perform well as steam oils. I may end up going for a sohn adaptor with a steam oil. Again this is long term if the gains I'm seeing pan out after a couple of fill ups it will be time to think about the long term.
12-24-2011, 07:42 PM
#16
Just so I understand your baseline setup, are you using the OEM tune? It seems to be what you are saying but I don't want to assume.
I had been considering a fuel preheat system based on the experiments of the late Smokey Yunick but hadn't worked out how to incorporate the needed heat exchangers without melting the upper manifold.
I had been considering a fuel preheat system based on the experiments of the late Smokey Yunick but hadn't worked out how to incorporate the needed heat exchangers without melting the upper manifold.
12-24-2011, 08:00 PM
#17
The tune is a slightly modified Cobb stage 1. The only things I changed were the CAT CEL and lowered the fan temps.
A simple plate heat exchanger should give you some preheat I had pondered the idea myself. You could tie into the coolant bottle return. Sure it wouldn't vaporize all the fuel (actually it shouldn't vaporize any in the lines) but it would allow more of it to flash boil as it is injected. Since the components of gas boil at 90-450F the objective shouldn't be to vaporize all the fuel, but instead to get enough to vaporize to spread the rest into a fine mist which can fully vaporize late in combustion instead of in the catalytic. That in itself should increase fuel economy, but would lead to hotter combustion temps (as you have less fuel to boil and remove heat during combustion) less margin to detonation and burn your side seals/springs to slag. If you heat the fuel you must also remove the heat the fuel would have removed. This is why running lean mixtures produces so much heat! All the fuel is burning, none of it is simply boiling without burning.
A simple plate heat exchanger should give you some preheat I had pondered the idea myself. You could tie into the coolant bottle return. Sure it wouldn't vaporize all the fuel (actually it shouldn't vaporize any in the lines) but it would allow more of it to flash boil as it is injected. Since the components of gas boil at 90-450F the objective shouldn't be to vaporize all the fuel, but instead to get enough to vaporize to spread the rest into a fine mist which can fully vaporize late in combustion instead of in the catalytic. That in itself should increase fuel economy, but would lead to hotter combustion temps (as you have less fuel to boil and remove heat during combustion) less margin to detonation and burn your side seals/springs to slag. If you heat the fuel you must also remove the heat the fuel would have removed. This is why running lean mixtures produces so much heat! All the fuel is burning, none of it is simply boiling without burning.
12-25-2011, 09:39 AM
#19
For what it's worth, I do know that some stationary gas turbines used in power generation use steam injection in order to inhibit production of oxides of nitrogen.
Are you using the OEM cat? You mentioned that your tune changed the CAT CEL, presumably disabling it. Was your catalytic convertor having trouble before your experiment?
Are you using the OEM cat? You mentioned that your tune changed the CAT CEL, presumably disabling it. Was your catalytic convertor having trouble before your experiment?
12-25-2011, 10:19 AM
#20
most w/m injection companies have an electronic solenoid valve that could be used in conjunction with a coolant temp gauge that you can set to signal the injection set up.
So start up would not be a problem.
couple things:
1- How do you insure equal distribution to the rotors? (Its harder than you may think)
2- W/m guys usually do not run with the w/m injecting all the time--that may be a difference when thinking about the oil film? If its steam going in then it may not really be a problem though? I really dont know.
3-It would be better to have a system that would be rpm activated ( can roughly do this with maF signal) and not injecting during deceleration.
4- you will loose power--about 10hp at least.
Interesting idea, but I am thinking you are going to find that heating the water before hand will be of limited benefit over conventional water injection? Water injection into the combustion chamber turns to steam almost instantly.
I get 25 mpg now--with my sc set up. Just lower the rpms and you are there.
So start up would not be a problem.
couple things:
1- How do you insure equal distribution to the rotors? (Its harder than you may think)
2- W/m guys usually do not run with the w/m injecting all the time--that may be a difference when thinking about the oil film? If its steam going in then it may not really be a problem though? I really dont know.
3-It would be better to have a system that would be rpm activated ( can roughly do this with maF signal) and not injecting during deceleration.
4- you will loose power--about 10hp at least.
Interesting idea, but I am thinking you are going to find that heating the water before hand will be of limited benefit over conventional water injection? Water injection into the combustion chamber turns to steam almost instantly.
I get 25 mpg now--with my sc set up. Just lower the rpms and you are there.
12-25-2011, 04:46 PM
#21
couple things:
1- How do you insure equal distribution to the rotors? (Its harder than you may think)
2- W/m guys usually do not run with the w/m injecting all the time--that may be a difference when thinking about the oil film? If its steam going in then it may not really be a problem though? I really dont know.
3-It would be better to have a system that would be rpm activated ( can roughly do this with maF signal) and not injecting during deceleration.
4- you will loose power--about 10hp at least.
1- How do you insure equal distribution to the rotors? (Its harder than you may think)
2- W/m guys usually do not run with the w/m injecting all the time--that may be a difference when thinking about the oil film? If its steam going in then it may not really be a problem though? I really dont know.
3-It would be better to have a system that would be rpm activated ( can roughly do this with maF signal) and not injecting during deceleration.
4- you will loose power--about 10hp at least.
2.Steam can really be a problem. Do a google search on "steam oil" and you will see why. I'm just expecting that in small amounts it will be ok.
3. Working on it. See below.
4. Have not felt any decrease in hp at low water injection rates.
Today I used my new low profile jack (my wife loves me!) and made some changes to my system. My injection is now going into the maintenance ports. There is now a metered air vacuum break line to prevent vacuum drag. I powered up my proportional controller (linked to MAF) and pump to supply since I had eliminated vacuum drag and this would allow me to turn on and off injection as well as control rates.
The results were bad. With the previous setup although I could not start or stop injection it was limited by intake vacuum. My little 1gph nozzle was putting out about .4gph max. This had little effect on vacuum (however it did increase) and a dramatic indicated improvement in fuel economy. I was indicating 31-35mpg, which I thought was way too high to be real, and was probably due to skewing the MAF (see below) as it was hooked up in a weird configuration to give me the option of starting the pump. Today I was using a pump to supply pressure, and with no recirculation as soon as the controler started the pump at lowest power I'm pretty sure it went right to the cutoff switch. That put greater than 1gph at 150 psi. I drove 50 miles back and forth on a 10 mile stretch of flat road with no stops. Water injection was on for the majority of the time, although I did have it off for about 10 miles as I tried to find differences in engine behavior.
Indicate MPG was around 28. Actual MPG is 18. This is around my normal MPG. Much of this difference can be attributed to the MAF value being skewed by the water injection controller. I noticed this because every time I stopped water injection I pulled power to the controller and it changed my MAF reading. Devils Own claims their design minimizes this. I guess what they mean is that it doesn't cause a CEL. Another possibility is that the controller was skewing MAF voltage while it was without power. Perhaps more investigation is worthwhile.
There was no difference major difference in engine power or AVG MPG indicated with water injection on or off. There was however a noticeable change in vacuum. With water injection on I was reading 16-17" Vac while driving at 65mph. Normally I get about 18", my previous steam injection test was giving me 18-19". This was even more apparent when I pulled into the service station and was only getting about 19" vac idle and the engine was putting and steaming out the tailpipe. After a couple minutes it went back to 24" and the engine was dried out. This is not a situation I intend to repeat, as the drop in vacuum was probably due to inadequate lubrication.
This shows that 1gph is way too much. The limiting factor is probably the amount of contact between the copper tubing and the exhaust manifold as well as the short length of the copper tubing. However since it was a big pain in the butt to make the steam generator and I don't want to drag around 15 gal of water in my trunk I'm going to leave that part of the design as is.
Next step is to try it again with a lower water flow-rate. Since vacuum drag worked so well initially I'm going to revert back to that and break vacuum manually until I have a solenoid.
Last edited by Harlan; 12-25-2011 at 04:48 PM. Reason: gph not gpm... doh
12-25-2011, 08:40 PM
#23
1 gallon per hour may be too much. At wide open throttle on the race track with just a 60 cc nozzle i will use about 1/2 gallon in 30 minutes.
that is at a much higher average rpm rate than you are probably doing.
I dont understand the way you have your lines hooked up. Where is the water being injected.
that is at a much higher average rpm rate than you are probably doing.
I dont understand the way you have your lines hooked up. Where is the water being injected.
12-25-2011, 09:00 PM
#24
The rubber hose and brass fitting on the end is the water injector. It's a 1gph misting nozzle acting as an orifice only. It then goes through a rubber hose which connects to a brass Tee. The Tee ties in with pre throttle intake air then goes into a stainless line which goes to the copper steam generator under the car. After the steam generator it goes into another rubber hose which Tees into the lower intake maintenance ports. The abortion on my accordion tube is simply where I routed in this system plus the Oil sump breather which was originally to the maintenance ports. Waiting on a catch can to make this look slightly better.
12-26-2011, 12:48 PM
#25
Oh ok--so the jet air tubes are now the steam pipes?
That may be part of the difficulty you are seeing. If you are using engine vacuum to pull the steam in---then it is not metered by the maf. I dont understand it any other way? If it is vacuum after the maf then there is a delay between it being read and it entering the CC?
What are your a/f's? Are they steady?
That may be part of the difficulty you are seeing. If you are using engine vacuum to pull the steam in---then it is not metered by the maf. I dont understand it any other way? If it is vacuum after the maf then there is a delay between it being read and it entering the CC?
What are your a/f's? Are they steady?