"Fat Boy" 12/118 thrust engine

[finiteparts]

Ron,

Yes, I agree with your points. We would reduce the pressure rise in the stage also which still goes along with the main idea...to reduce the work in the compressor stage as a means to try to get the compressor and turbine to better "match". I haven't worked through the calculations, so I am not sure if the system balances out any better...just thought that IGVs do offer a means to "control" the compressor operation and maybe they could help here. Just thought that I would throw it out there as food for thought.

I guess I didn't actually specify that the pre-whirl should be in the same direction as the rotor rotation, I just assumed that by saying that I reduced the work in the stage it could be inferred from the vector diagram, sorry about that.

[racket]

Hi Chris and Ron

Yep , familiar with IGVs , I made a stator for my TV84 back in ~July 2000 to try and improve thrust .

At the time, due to the mounting of the turbo in the bike , I was experiencing very large vortexes of air entering the comp inlet during static testing , now here in the Southern Hemisphere they were rotating in the "wrong" direction , opposite to wheel rotation , so the IGV was to both "straighten" air going into the comp , and because I gave the vane a 10 degree twist it should have produced a slight rotation in the direction of wheel rotation to potentially improve the inlet velocity triangle.

Reading through my journal for test runs around that time there doesn't seem to be any mention of "improvements" ................LOL, understandable really as I was already producing 46 kgs - 102 lbs of thrust which was about all the engine was capable of , I did see 50 kgs on a few occasions when running some pretty high TOTs .............I spent a couple of fruitless years chasing more thrust before learning to do the maths and finally realising there wasn't any more to be had

A coupla pics






Cheers
John

[turboron]

John, beautiful work. How did you fabricated your IGV?

Thanks, Ron

[racket]

Hi Ron

LOL........with difficulty :-(

Some wooden jigs to hold bits in place for welding , and lotsa work with a file shaping the vanes to knife edges.

Cheers
John

[CH3NO2]

Nov 15, 2016 19:48:16 GMT -5 racket said:

...

What the test did show up was the big jump in fuel pressures once past that 1.0 bar P2 , it was indicating a fuel pressure of 23 psi at 1 bar , so only a pressure drop of ~8 psi across the injectors .

When I attempted to give her a "gut full" of fuel to try and reach a higher P2 the fuel pressure ended up at 76 psi for only a P2 of 26 psi , a 50 psi fuel pressure drop , this is excessive and probably >50% more fuel than its possible to burn "correctly" with the airflow available, TOTs were >1,000 C for several seconds , P4t was just under 8 psit.

There appears to be a combustion problem , maybe the flametube is simply too short to complete combustion before the Tertiary Zone .

I'll drop the engine out and check the flametube and evaps , I might have to resort to burning liquid LPG gas to obtain the speed of combustion necessary.

Cheers
John

Hi John,

When I read this it gives me the feeling you could be experiencing boiling of fuel within the fuel manifold.

Calculations of fuel mass flow from the inlet pressure can be misleading if there is boiling somewhere within the manifold.   The pressure is increased to the point where there should be >50% more fuel than possible to burn correctly.    If you were pumping in that much excess fuel, you would likely see, smell and measure it in a dramatic way.   Unless fuel mass flow wasn't actually going up >50%.

Can you put a thermocouple on the fuel manifold?
Does the manifold show signs of being strongly heated?

Another way of knowing if there is boiling is to measure the change in fuel consumption rate at a given pressure.  Does "measured" fuel mass flow rate agree with the fuel mass flow "calculations" based on inlet pressure?

If there is fuel boiling in the manifold, it will create numerous other anomalies that could be interpreted as flow obstructions, offsets in flow distribution or even coking in the lines.

When you found the debris in your manifold, if you still have it, you can potentially test it in various ways to see if its metallic or organic in nature.
Can the debris be dissolved in a polar or non-polar solvent.
Can it be dissolved in a acid or base? 
Can it be volatilized on a hot plate?
Smell & texture?

If the fuel is coking even a little, its boiling... and that could explain a good portion of the operational difficulties. 


Just throwing the idea out.  
Tony

[CH3NO2]

Dec 10, 2016 16:13:32 GMT -5 racket said:

...
I'm pretty certain that the vibrations I was feeling through the test stand and car trailer it was mounted on is combustion instability considering the "impossible" amount of fuel I was attempting to burn .

This is the first time I've had a combustion problem with any of my engines , usually they just work .

Cheers
John

Yeah.   I'm starting to think there is a cascade of interrelated issues here.  

The vibrations you noted are a symptom of the combustion instability you describe... Your description also reminds me a lot of what is called "chuffing" in the rocket propulsion community.   If you have chuff your combustion chamber is having correspondingly large oscillations in chamber pressure.   If the pressure oscillations are not caused by compressor surge, the pressure oscillations will certainly create compressor surge at the same fundamental frequency.  A potential positive feedback mechanism for reinforcing pressure oscillations.

Additionally, the chamber pressure oscillations will cause an oscillation of fuel Mdot if the injector pressure drop is low...   When you described the pressure drop across your injector needles it struck me as low.  Very low.  

Low enough for fuel Mdot to be susceptible to the local environmental pressure oscillations.   Fuel Mdot oscillations will create combustion oscillations... another positive feedback mechanism for combustion instability.    

Furthermore, chamber pressure or deflagration oscillations could pump combustion products in and out of the combustion chamber through the flame can orifices.  It could also explain how you noted fuel back flowing from the evaporators.  

In your photo of the fuel manifold and injectors, it looks well coated with soot.  There can be a number of reasons why it has soot on it but a combustion instability sure could be one cause.   A  combustion instability, at the vibration frequency you noted, could be back firing flames all over the injectors and manifold, splashingback evaporator fuel onto the manifold/injectors.... which could cause boiling of the fuel in the injectors creating even more instability.

I recommend to increase the fuel injector delta P.  
Ideally 2X+ chamber pressure for a more rigid flow rate.  *No boiling!*  If there is boiling in the manifold or injectors they will be squishy no matter how much delta P there is.  If the root cause is compressor surge this certainly wont fix it... but it may bring everything closer to stability. 

Tony


postimg.org/image/d7rioe4wr/

[CH3NO2]

One other thing you could try is to significantly increase the fuel manifold flow velocity by decreasing its flow area.
Decrease the manifold tube diameter such that it increases velocity. It will decrease the time the fuel is being exposed to heat.

You can reduce the manifold flow area down to about 4X the total injector orifice flow area. This will avoid excessive manifold pressure drop yet keep the velocity high enough to reduce fuel heat transfer and boiling.

[racket]

Hi Tony

Mmmmm, I'll have to "digest" this and have a good think :-)

The original Schreckling micro engine used a vapourising coil within the flametube for a fuel system and suffered from oscillations , but I haven't heard of it happening with any other "micro" engines that have the fuel manifold only exposed to hot compressed air .

It'd take some serious heat transfer to boil 2.5 litres/min of kero passing thru an ~8 mm ID steel tube of ~900 mm length .

Cheers
John

[CH3NO2]

Hi John,

You did a really good job of documenting your observations in this thread. In the documentation is a wealth of clues that may not seem like much on their own but when the pieces are put together its starting to build a picture.

Consider that the total flow area of the 18 0.5mm ID injectors is 3.53^2mm.
The total flow area of the 5/16" (8mm) manifold will be about 31.7^2mm after accounting for wall thickness... but because its flowing in two directions at the inlet that brings the manifold total flow area to ~63.3^2mm.

This makes the manifold flow area 18X the total injector flow area. (much too high)
Then if you consider that as the fuel flows, from both directions, to the furthest end of the manifold to the last injector the manifold flow area is 322X the flow area of the last injector. (now we are really looking at the possibility of boiling.)

Then consider the fuel flow rate of 2.5 liter/min. Is that measured or calculated? Remember the huge combustion inefficiency you calculated at that flow rate? It's possible that you dont have a massive combustion inefficiency. It may be that the pressure calculated fuel flow isn't really 2.5 liters/min.
Measure it.  If fuel flow rate calculations and measurements are in agreement, there is no boiling.   If they are significantly different, you got boiling in there.

All this in combination with all the other details you posted:
Suspected flow obstruction.
Particles of "rubbish" in the manifold (coke?)
Uneven temperature distribution seen on the thermocouples.
Vibrations in test stand.
Little change in performance despite jacking the fuel pressure way up.
The problem getting worse when you switched to larger diameter injectors.
Witness marks of fuel/soot where there shouldn't be any.
Silver solder joint melting at the vaporizer inlet.

Excellent documentation. These small individual observations are fitting together.

Could waisting of the vaporizer tube be working as an internal flame holder?

[CH3NO2]

When troubleshooting, throw out all the possibilities you can think of then go through a process of elimination.

Lets use waisting of the vaporizer tube acting as an internal flame holder.
How could this happen?
How could we eliminate it as a possibility?

If we consider the possibility of the waisting in the vaporizer causing flow detachment from the tube wall how far down will the flow go before it reattaches to the wall? 1mm? 3 cm? It's hard to say but its probably somewhere within this range.
There will be a re-circulation zone up to the point of flow reattachment.

The calculated fuel/air ratio in the vaporizer may initially say its too rich to burn... but its only too rich if the fuel is completely vaporized. So what if the fuel isn't fully vaporized in the first 3cm of the waisting and is only partially vaporized... now we can have a potentially combustible vapor in the re-circulation zone behind the waisted rim.

So what if that combustible vapor catches fire?... all kinds of trouble and wild instability inside the vaporizer tube.
Once it catches fire it will quickly self extinguish because so much more fuel is vaporized by the added heat. Then once it extinguishes its self it will cycle back to a cooler/leaner condition and go right back to a combustible condition and start the cycle all over again. Combustion instability and flow instability working in unison.

Can the possibility of a recirculation zone inside of the vaporizers be eliminated? Potentially. Change out the vaporizer tubes for tubes with no waisting and a little less demming?

Another potential scenario...

How can fuel boiling be part of the system instability? If frothy vapor bubbles are being passed through the far side injectors, the local equivalence ratio inside those vaporizers will be out of whack. A lot leaner and erratic. Potentially in the flammable range and causing more trouble.
All that soot on the fuel manifold/injectors suggests it's being washed in fire.

Where did the melted silver solder joint happen in relation to the fuel manifold inlet? The far side? On the side where bubbles in the manifold could rise to? If it did either one or both, you have a strong suspect.

Just throwing out some possibilities.

[racket]

Hi Tony

Had a look at the fuel manifold and the syringe needles are still "silver" under the oily soot , so no heat there ................the evaps are also "silver" for a goodly length , only the last 25% has serious heating as they should , so all good there as well.

Air speeds through the evap is several times ignition speed of kero .

As for fuel pressures , I need to run fairly low pressure drops across the injectors if I want to use an automotive EFI fuel pump , with 40 psi of P2 and some line friction , I'm only left with a max of ~25 psi for drop across the injectors , with flow varying as the square of pressure I soon run into problems if I start off with any more than a few psi pressure drop at idle , manifold testing gave 1,000 cc at 5 psi , 1,500 cc at 10 psi 1,900 cc at 15 psi and 2,400 cc at 20 psi , with a mass flow of 3.6 lbs/sec of air and a ~50 :1 A/F ratio I need ~4.4 lbs/min of fuel or ~2.6 litres/min .

When I suffered silver solder melting , it was due to some blocked injectors , the working ones were "overfueling" sending a hot streak back to the flametube rear wall , overheating it and the silver solder holding the evaps in place .

LOL, better get back to my tiling in the kitchen ;-)

Cheers
John

[finiteparts]

Hi Tony,

I really doubt that John was experiencing any combustion instability, the simplest and more likely cause would be a flow instability in the compressor system.

The reason that I say this is because conventional combustors (this means most combustors that are not designed for low emissions) are not very susceptible to acoustic coupling due to the large number of cooling holes in the liners which give very high acoustic damping. Rocket engines are a whole different animal! They have one big hole (the exhaust!) and it is usually acoustically blocked by a shock wave, so the acoustic waves (pressure waves) reflect all over the place and cause all kinds of havoc. So for a small combustor like this to suffer from some acoustic coupling is very, very unlikely. The same thing goes for the fuel pressure. In a rocket, the acoustic coupling drives the requirement for a high pressure drop across the injector so the injector flow rate doesn't strongly couple to the acoustic pressure waves, but that is not an issue in small GT combustors. The acoustic pressure waves are attenuated through the liner and thus don't reflect and set up standing waves in the combustor.

Second, if there were any flame events in the vaporizer tubes, there would be a really high likelihood of burnt vaporizer tubes. When a flame moves into a enclosed passage, there is a strong positive flame stretching that leads to local flow deceleration. This can be so severe that the flame can hold on to this local flow disturbance even if the bulk flow speed is much higher than the flame speed. Usually, this would occur near the wall boundary layer due to the lower flow speeds (by definition, the flow speed is zero at the wall and reaches the mean flow speed at the edge of the boundary layer). The presence of the flame and its back pressure also leads to the formation of an adverse pressure gradient in the near wall region and this trips the boundary layer and forms separated flow pockets. This is commonly called flashback in the combustion field. Premixed systems can flash back and damage hardware, but it is rare that they flashback and then blow that flame back out. There exists a hysteresis behavior around flashback and then clearing that internal flame because of the above mentioned flame curvature and boundary layer separation. Once the flashback happens you can't just go back to the same condition it was at before flashback occurred and expect it to clear, you have to go back to a much stronger flow speed to push the flame out of the tube. So by virtue of John still having vaporizer tubes, it suggests that it is unlikely that he was burning inside of the tubes.

The soot on the casing is likely due to fuel leakage burning off after shutdown, but if the compressor or diffuser was in a weak surge, the flow actually reverses and comes out of the liner. The compressor system can surge due to the physical hardware such as the diffuser vanes and the downstream collector, the coupling between the parts (volumes) and even downstream hardware (as long as the diffuser isn't choked). Since there are vaned diffuser, they are just as susceptible to off incidence flows as is the compressor inducer and this is why compressor maps for vaned diffusers are narrower than those for vaneless diffusers. The fact that the compressor exducer flow speeds are some of the highest flow speeds in the engine means that this area is very sensitive to design issues and can lead to significant losses. If the engine is not reaching the design point, by definition, it is operating off design and there could be very large off incidence flows across the diffuser leading edges and high speed flows are very sensitive to very small off incidence flows leading to diffuser stalls.

John, the comment that the fuel flows are higher than expected and the EGTs are high does suggest lower than expected combustion efficiencies, but that is really confounded by the lack of any real flow turning in the exducer, which could lead to the outer region of the turbine not really extracting as much work as you expected in the mean line calculations and thus higher EGTs. That combustor was pretty short. If it was longer I think a water injection system in the combustor like they do on the Rolls Royce Pegasus engines might help...it would put a larger mass flow through the turbine at a lower entry temp....but I am not sure you could fit it in there.

Good luck,

Chris

[turboron]

Chris/John, injecting water or steam in the outer combustion case between the compressor and the turbine requires a compressor with good surge margin.

Thanks, Ron

[finiteparts]

Ron,

I was actually referring to injecting water into the aft end of the combustor to reduce the combustor discharge temperature while increasing mass flow. The Pegasus engine injects water into the back dilution holes for this purpose. You are right, if you are using water injection for power augmentation, you do suppress the compressor mass flow, so if you were sitting on a constant speed line due to a choked turbine, you would push to a higher compressor pressure ratio and lower flow, i.e., closer to the surge line.

My thought was to not use it for power augmentation at a fixed point, I was trying to use it at a lower flow to "make up the difference", kind of letting John sneak up on the choked exducer condition. Assuming that you have a constant turbine flow function (mdot sqrt(T) / P), the idea was to decrease the combustor discharge temperature by say 6%, so that you could push a 3% higher mass flow with no pressure penalty. The enthalpy of the combustion is hopefully not too suppressed, because it is captured into the flow by phase changing to steam, so increased mass flow and hopefully not too large of a drop in enthalpy (remember, turbine power is mass flow times the change in enthalpy across the rotor) keeps the turbine power in the same ball park, hopefully. The local acoustic speed is reduced so for a fixed through-flow area, I can push more mass flow before choke, which hopefully isn't just the extra mass flow due to the steam. It was more of a play to try to help the turbine match the compressor flows better.

I am still poking around with the numbers, which is a bit tougher because I am no good working with psychometric charts. I realize that is will likely be a problem in that this works well when your speed lines are steep, but if John is running down in the lower range, near surge, the speed lines are nearly flat and so it likely would push it into surge. I will let you know more when I get through the calculations.

~ Chris

[turboron]

Chris, I understand your logic. My experience is with steam injection on an axial compressor with a steep rise to surge and a lot of surge margin.

Thanks, Ron