[MFGT]

Hello,

i wanted to create a topic about the surrogate fuels, because i came across a few difficulties and would like to exchange knowledge and experience regarding this. My main fuel is gasoline with RON of around 92 and MON about 84, simulated in a Gasoline DI engine. Target OP is 8 bar bmep at 2000 rpm.

Previously i was using the simple approach with Iso-Octane and n-Heptane, blending both by volume according to the Octane number. However, that blend does not have the right LHV and also the H/C ratio and LSt are too high.

Adding Toluene and using the the method described here (http://www.sciencedirect.com/science...10218010000337) i got a three component fuel, which better matches the LHV and LSt, although H/C is a bit too low now.

Ive set up the same simulation using PRF and TRF each, with corresponding Jia mechanisms included in the CONVERGE example cases. At the IN- and OUTFLOW boundaries, i use pressure and temperature profiles of a GT-Power model, which is matched measurement data.
Due to the different fuel density, my discharge coefficent slighty changes. Furthermore, I have added another mutlizone dimension to account for the third fuel component.

Beside that the simulation duration was increased by 25% (multizone dimension and bigger reaction mechanism), i noticed the following points:

• As previously described, the LSt had changed and therefore my Lambda is different. With nearly the same fuel and oxidizer masses, i receive a Lambda of 0.96 (PRF) compared to 1.02 (TRF). I have also calculated these values on my own for validation. Here, TRF seems to be better suited, as LSt was closer to original fuel.
• Due to the different mixture composition, i have a difference in kappa and cp. Compared to GT Power, the PRF matches these values better.
• A difference in kappa means also a difference during the compression stroke. At spark timing, the pressure of 15 bar is increased by 0.3 bar simply due to the TRF. PRF matches the compression stroke profile very good.
• Finally, there is a big difference in pressure profile/heat release during combustion. PRF matches my pressure profile of GT-Power at the second cycle. TRF has a slower burn rate and slower pressure rise. Difference in peak pressure is around 8 bar (44.6 vs 37 bar).

From the points mentioned above, i have concluded that i will use PRF in future, due to the simplicity and better results (at least in this case). I am not looking at emissions or knock phenomena yet.



What are the experiences that other users have with this? Does someone has investigated this as well?

[SauravMitra]

Hi Tobias,

This is a great thread and therefore a great initiative. I agree with you, all the users can share their own experiences and expertise with respect to surrogate fuels. This thread has a huge scope. Let us keep in mind that we might get limited by proprietary/IP issues.

Coming back to the original discussion, I have observed the same trends with PRF and TRF for RON95. TRF lags in pressure rise and burn rate. One of the factors that affect burn rate, is the different spray behavior for the two surrogates. Have a look at the evaporation rates, penetration, smd etc. Maybe there is room for tuning there.

Due to presence of toluene, soot tends to be more in TRF.
I have not checked knock, but am guessing toluene will act as a knock-suppressant most probably, due to its strong negative temperature coefficient. Correct me, if I am wrong.

By the way Tobias, I did not understand what you meant by LSt. Break it up for me.

Let us keep the discussions going.

Thanks

[MFGT]

Hi Saurav,

with LSt, i meant the stoichiometric air requirement, calculated by:

Lst=1/yO2*(2.664*c+7.937*h+0.998*s-o)

with yO2 mass fraction of the oxidizer, and c, h, s and o being the mass fractions of C, H, S and O within the fuel.

At the moment, i calculate lambda and reaction lamda within my clinder before IGN timing on my own, to doublecheck results.

[MFGT]

Originally Posted by SauravMitra

One of the factors that affect burn rate, is the different spray behavior for the two surrogates. Have a look at the evaporation rates, penetration, smd etc. Maybe there is room for tuning there.

The evaporation rate is slower with TRF, because i chose different mass diffusivity constants. While it had been Iso-Octane defaults for PRF, i took Toluene defaults for TRF as its my main component.

SMD is different due to different discharge coefficent of the nozzle. Same timing, pressure and mass, but the mixtures have different densities and therefore i had to adjust the discharge coefficent. If i remember correctly, the TRF is slightly smaller initially (blob injection model).

[SauravMitra]

Hi Tobias,

I think it will be a worthwhile exercise, if you only allow the mechanism to be different between the two and keep the same properties for both the surrogates. My proposal is to use the same LHV, same diffusivity constant (use Gasoline defaults), same injection pressure and timing and then see what happens.
Others, pour in your ideas and suggestions for exploring surrogate fuels.

Thanks

[MFGT]

I would really like to do that, but I only have limited licenses available. And therefore a full cycle simulation takes ~30-40 hours.

Regarding the PRF, the thing that I dislike most is the discrepancy with respect to air/fuel ratio. With the given fuel masses from test bench, my mixture is getting too rich and I am not sure how this affects my simulation/comparison. Despite the fact, that its difficult to get the target air/fuel ratio, as fuel can escape into the intake port.

[MFGT]

I recalculated my fuel masses in reference to the oxidizer mass, so its around 95% of the original fuel mass. This allows me to achieve the target AFR (i hope) and the injected fuel energy (mass*LHV) is similar to the experiment. Now i have to wait for simulation results.

[SauravMitra]

Keep us posted Tobias with your progress and findings.

Again, this thread is for general purpose and you can think of solving a research problem in an Open forum. Others, feel free to contribute your ideas and thoughts.

[MFGT]

Today, while browsing through the Liquid Database, i accidentally came across the Liquid "CSI_Gasoline_v1".

If using this Liquid, into what species does it evaporate then? And what RON/MON numbers does it refer to?

My original intention was to look for a surrogate for lubricant oil, where i have Diesel2 or Hexadecane (C16H34) in mind.

[dprobst]

This liquid CSI_Gasoline_v1 is documented in the manual on pg 247. It is a surrogate with 50% isooctane, 35% decane, 10% pentane, and 5% dodecane.

You have different options for evaporation from a liquid species like this. It could be evaporated to isooctane if a detailed chemistry mechanism were being used that was developed for isooctane. If a mechanism with isooctane and heptane was being used, and you were simulating a RON 92 fuel, you could evaporate to 92% isooctane and 8% heptane. Evaporating into a species named SI_Gasoline_v1 would technically be possible, but may not make much sense since we do not have a therm.dat / mech.dat with this species.

[MFGT]

Thanks for the answer Dan.

Originally Posted by dprobst

If a mechanism with isooctane and heptane was being used, and you were simulating a RON 92 fuel, you could evaporate to 92% isooctane and 8% heptane.

How do i do that? I tried to figure this out already but didnt came to any conclusion.

If I set evap_source_flag=0, everything would evaporate to one selected species (e.g. IC8H18)
If I set evap_source_flag=1 or 2 (which should be the same, as composites are not available via SAGE), where do i define my target evaporation composition?

[dprobst]

If I set evap_source_flag=0, everything would evaporate to one selected species (e.g. IC8H18)

You are correct, the option to evaporate into another species only allows you to evaporate into a single species.

I suggest you create two liquids with the same liquid.dat properties of CSI_Gasoline_v1 but name them as the two fuels in your mechanism (i.e. c8h18 & c7h16) and then evap directly.

[MFGT]

This is similar to what i did with my lubricant oil surrogate.
I took C16H34 as liquid, but wanted to evaporate it into C8KET and simply renamed it in liquid.dat.