Jump to content
RA004E

dual fuel system setup (using two sets of fuel rails)

Recommended Posts

Hello folks,

Sicne the gas stations here don't sell E85 so I am trying to build a Subaru STi with a dual fuel system. 

The main goal is to have good street drive-ability (by using ID1050x injectors as main injectors) and E98 power on demand, anytime I want.
HP goal is around 800~850whp. Turbo is going to be either EFR9180 or Precision 6870... 

Here is my plan; 
There are going to be 2 sets of fuel rails, total 8 injectors, one fuel system for gasoline and the other for pure ethanol.
Primary fuel system -  ID1050x *4, one rail for each bank, feed by an AEM320 pump thru stock gas tank, fuel is RON98
Secondary fuel system - ID2000 *4, one rail for each bank, feed by a huge in-tank brushless fuel pump thru a E100 fuel cell. (or a lift pump + in-line pump)

I wanted to adjust the E% by different boost level. EX;
below 1.5bar of boost - RON98 - using primary fuel system only
1.5~2.0bar of boost - E50 - using both primary and secondary fuel system
above 2.0bar of boost - E100 - using secondary fuel system only

In the mean time, I want to still retain the wideband closed loop lamdba control in auto mode.
With two different VE & IGN maps, ramping by a multi fuel blending ratio, and using staged injection.

Is it possible by using the G4+ Thunder? If I use "modled - multi fuel" equation mode I can't set the parameters of the secondary fuel system. 
The "multi fuel" injection parameters seems only correspond to the one(main) fueling system.

Since I am a local Link ECU dealer I really want to finish this project by using LinkG4+. Don't really want to switch to syvecs or motec. 
Please help!

Thanks, 
Mike
Taipei, Taiwan

Share this post


Link to post
Share on other sites

How are you planning on calculating the %Ethanol in the fuel? 

Would be easier it one tank was 98 and the other was a ethanol blend. Trying to blend both on the run would be a nightmare.

I would adjust the boost based on E% not E% based on boost level.

 

Share this post


Link to post
Share on other sites

Couple of ideas come to mind but there would still be some issues to work through with both.

1) have the 2 tanks and 2 fuel pumps, but then put check valves in both lines and bring them together with some kind of selectable valve then through an ethanol sensor before going into a single fuel rail with just the ID2000's. This makes it so the ECU sees it as just a traditional flex fuel setup with an ethanol percentage.

Issues you will need to sort out though are

a) finding an electronically controllable valve (and then hook it up to an aux output/pwm output) - maybe you need 2x fuel pump controllers, one for each line instead of an actual valve?, 

b) you will need to run it semi returnless to avoid the mixed fuel getting back into what should be pure ethanol or fuel tanks. Maybe because you have a ethanol sensor you can live with a bit of mixture in each tank and so you could decide to run a return which dumps all fuel into one or the other (or set up a similar valve/dual controller system to switch where the returned fuel goes as well - you will get a little be of impurity because of "wrong" fuel in the line when it switches but at least it wont mean the whole tank quickly ends up in the wrong place) Your ethanol sensor will also need a big hose size for that kind of HP if its on the inlet line, most the ones i've seen have been smaller than you will want on a feed line at that kind of HP.

 

2) have it physically as you describe with 2 completely separate fuel systems, and configure an aux output to control a couple of relays that provide/block the 12v side of each injector bank, but wire both "injector 1" to the same ecu output. This way the ecu wouldnt see it as multi fuel though, so you will need to treat it like a cam switched engine and just have 2x fuel tables, 2x ignition tables etc, and have it so the ECU switches to the other set of tables based on the same input/output as you are using to control the relay that switches injector banks. You then set this "fuel control" output to be based on TPS/KPA/whatever you want.

Issues:

a) because the ecu only allows one set of deadtimes and injector size info, you will have to use the info for whichever set it idles on. The deadtimes will be wrong and the flow volume will be completely wrong so you will need to "fix" this by setting your second fuel table to some overly high numbers, or setting the fuel pressure on the second rail lower (which may partly defeat the point of having larger injectors anyway).

b) any fuel pressure sensor you feed into the ecu will only be accurate for one of the fuel systems, you may not be able to use this as an input because of this

c) You would only be able to get one fuel type or the other like this though, not the 50/50 mix you describe (unless you set up something really crazy like energize both sets of fuel injectors at once, and use a 4d fuel table to drop 50-60% out of the fuel value when in "mixed mode", or some really low values above a certain KPA/RPM so that the fuel from both rails together adds up to the amount you need to run well). You would be fighting against the calculations within the ECU by doing this though so might be a lot of work to get right. 

[edit] you could also use the dual fuel table in "switchover" mode rather than blend to acheive the same thing for switching between fuel tables. Still doesnt help that the ECU isnt expecting the injector changes.

Both systems require an additional fuel tank which could be 50-100kg of weight when full and both seem like a lot of work. Are you sure it wont be cheaper to just buy a barrel of E85/E100 and just mix your own fuel?

Good luck

Share this post


Link to post
Share on other sites

You can do the blend using staged injection function but I dont see a good way to blend between 2 VE and Ign tables using staged injection as the control variable.  CLL should still work fine.

Why would you even need two VE & Ignition tables if your eth content is based on boost - as long as your main fuel and ign tables have MAP on one axis then that is going to be directly rated to eth content and take care of itself.  Wouldnt you  only need 2 tables if you were going to vary the eth content by some other parameter that was not already referenced in the table?

To clarify, I would not use modelled-multifuel mode for this, I would set it up just like a single fuel, staged setup.  Primaries one fuel, secondaries are the 2nd fuel, using the flow ratio method.

Share this post


Link to post
Share on other sites
2 hours ago, Ducie54 said:

How are you planning on calculating the %Ethanol in the fuel? 

Would be easier it one tank was 98 and the other was a ethanol blend. Trying to blend both on the run would be a nightmare.

 I would adjust the boost based on E% not E% based on boost level.

  

By setting up the secondarry injection staging table, using Sec/Pri flow ratio in staged injection control you get a pre-calcuated E%. It would be the same if I put the E85 in the second tank...
Unless the second fuel has the same stoich ratio as the main fuel.

Share this post


Link to post
Share on other sites
2 hours ago, cj said:

Couple of ideas come to mind but there would still be some issues to work through with both.

1) have the 2 tanks and 2 fuel pumps, but then put check valves in both lines and bring them together with some kind of selectable valve then through an ethanol sensor before going into a single fuel rail with just the ID2000's. This makes it so the ECU sees it as just a traditional flex fuel setup with an ethanol percentage.

Issues you will need to sort out though are

a) finding an electronically controllable valve (and then hook it up to an aux output/pwm output) - maybe you need 2x fuel pump controllers, one for each line instead of an actual valve?, 

b) you will need to run it semi returnless to avoid the mixed fuel getting back into what should be pure ethanol or fuel tanks. Maybe because you have a ethanol sensor you can live with a bit of mixture in each tank and so you could decide to run a return which dumps all fuel into one or the other (or set up a similar valve/dual controller system to switch where the returned fuel goes as well - you will get a little be of impurity because of "wrong" fuel in the line when it switches but at least it wont mean the whole tank quickly ends up in the wrong place) Your ethanol sensor will also need a big hose size for that kind of HP if its on the inlet line, most the ones i've seen have been smaller than you will want on a feed line at that kind of HP.

 

2) have it physically as you describe with 2 completely separate fuel systems, and configure an aux output to control a couple of relays that provide/block the 12v side of each injector bank, but wire both "injector 1" to the same ecu output. This way the ecu wouldnt see it as multi fuel though, so you will need to treat it like a cam switched engine and just have 2x fuel tables, 2x ignition tables etc, and have it so the ECU switches to the other set of tables based on the same input/output as you are using to control the relay that switches injector banks. You then set this "fuel control" output to be based on TPS/KPA/whatever you want.

Issues:

a) because the ecu only allows one set of deadtimes and injector size info, you will have to use the info for whichever set it idles on. The deadtimes will be wrong and the flow volume will be completely wrong so you will need to "fix" this by setting your second fuel table to some overly high numbers, or setting the fuel pressure on the second rail lower (which may partly defeat the point of having larger injectors anyway).

b) any fuel pressure sensor you feed into the ecu will only be accurate for one of the fuel systems, you may not be able to use this as an input because of this

c) You would only be able to get one fuel type or the other like this though, not the 50/50 mix you describe (unless you set up something really crazy like energize both sets of fuel injectors at once, and use a 4d fuel table to drop 50-60% out of the fuel value when in "mixed mode", or some really low values above a certain KPA/RPM so that the fuel from both rails together adds up to the amount you need to run well). You would be fighting against the calculations within the ECU by doing this though so might be a lot of work to get right. 

[edit] you could also use the dual fuel table in "switchover" mode rather than blend to acheive the same thing for switching between fuel tables. Still doesnt help that the ECU isnt expecting the injector changes.

Both systems require an additional fuel tank which could be 50-100kg of weight when full and both seem like a lot of work. Are you sure it wont be cheaper to just buy a barrel of E85/E100 and just mix your own fuel?

Good luck

Hello CJ,
I really appreciate you opinions! Will go deep thinking about what you said. I will probably stay with the method 2) and hopefully not fighting against the ECU fueling logics too much.

method 1) was my original thought... 2 months ago I came out this idea and finally decide to give it up. Setting up two fuel system by using one fuel rail will eventually have some reliable issue, like the fuel flow control solenoid. And like you said, since you have to install the flex fuel sensor before entering the rail, it might be a restriction of the fuel system unless you run parallel.
 

21248205_10154916676463034_5308553764735568721_o.jpg

Share this post


Link to post
Share on other sites

Reading other people's answers about the ECU side of it, Adam's solution seems like a lot less "trying to make it do something it doesn't want to" than mine. His solution will only have a single "normal" fuel table and the numbers in it will simply drop to ~half what they should be above 150KPA and to 0 above 200KPA, while using the aux injection table for injectors 5-8 with numbers that are 0 below 150, "half" at 150-200kpa, and "normal" above 200KPA. Only slight downside is you have to tune the aux table the traditional way (ms or duty cycle) rather than VE. There certainly seems less chance of strange things happening when the config is this straight forward. You would then use MAP over 150KPA/200 as a condition to turn on your aux outputs that control the fuel pumps/anything else in the system that switches when the fuel changes.

I still think its easier and cheaper to just mix your own E85 at home from a big barrel of it....

Share this post


Link to post
Share on other sites

Hello Adam, and CJ

First of all really thanks for the help. I am trying to implement the ideas you guys said, and set it up in PClink. Please let me know what do you guys think.
Am I getting it right? By using this method I think I can still retain the closed loop lambda auto mode - even in high boost.

Fuel Main
Injection mode - sequential/staged
Fuel equation mode - traditional
Stoich ratio - 14.7 (gasoline)

Staged Injection
Stage mode - Sec/Pri flow ratio
Sec/Pri ratio - 2.089

AFR_target_dualFuel.png

StagedINJ&E%lambda.png

Share this post


Link to post
Share on other sites
7 hours ago, RA004E said:

By setting up the secondarry injection staging table, using Sec/Pri flow ratio in staged injection control you get a pre-calcuated E%. It would be the same if I put the E85 in the second tank...
Unless the second fuel has the same stoich ratio as the main fuel.

 

Unless you run a dead head fuel system with no return line you would not achieve accurate results. As the returning blended fuel would affect the pw ratio of E% to 98%. 

My way of doing it is slightly different. I would use 8 1050cc injectors. Run a open loop fuel system with both fuels into a PWM 3 way valve. From there into a ethanol sensor (run in parallel off main fuel line) into both fuel rails.

Sound like a lot of problems for no gain. HPA have a good webinar on blending fuels.

Share this post


Link to post
Share on other sites
17 hours ago, RA004E said:

Hello Adam, and CJ

First of all really thanks for the help. I am trying to implement the ideas you guys said, and set it up in PClink. Please let me know what do you guys think.
Am I getting it right? By using this method I think I can still retain the closed loop lambda auto mode - even in high boost.

Your math looks ok. 

However, for the AFR/Lambda target table, the ECU will not know the stoichometric ratio of the fuel is changing so you will need to treat this table in "Lambda" - set units to lambda initially and fill in the table using Lambda values (i.e 0.800 lambda).  

CLL auto mode should still work correctly.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×