"Fat Boy" 12/118 thrust engine

[racket]

Hi Smithy Hi Ian

Yep , I'm working on it , ........I had a long discussion with Andy M this morning , love our phone chats Andy, and we've worked out some possible causes and solutions.

Will get back to you on it .

On a side note, I checked the fuel flow through the manifold at 5, 10 , 15 and 20 psi , with ~1,000 ml/min at 5 psi pressure drop , 1500 ml/min at 10 psi , 1900 ml/min at 15 psi and 2400 ml/min at 20 psi ...............currently trying to guesstimate actual mass flow of air using the turb temps and am getting flows a bit either side of 3 lbs/sec depending on burn efficiency .

More wet towel around head calculating this arvo :-)

Cheers
John

[Deleted]

Oct 20, 2016 22:20:16 GMT -5 racket said:

Hi Smithy Hi Ian

Yep , I'm working on it , ........I had a long discussion with Andy M this morning , love our phone chats Andy, and we've worked out some possible causes and solutions.

Will get back to you on it .

On a side note, I checked the fuel flow through the manifold at 5, 10 , 15 and 20 psi , with ~1,000 ml/min at 5 psi pressure drop , 1500 ml/min at 10 psi , 1900 ml/min at 15 psi and 2400 ml/min at 20 psi ...............currently trying to guesstimate actual mass flow of air using the turb temps and am getting flows a bit either side of 3 lbs/sec depending on burn efficiency .

More wet towel around head calculating this arvo :-)

Cheers
John

Hi John

I don't know if i was much help :-) but yes it's good to chat...

Am sure before long you will have got her running at full power and starting on the freepower unit :-) it is going to be one powerful beasty :-)

All The Best

Andy

[racket]

Hi Guys

Well, I've turned the numbers every which way but can't seem to find a reason for the high temps other than the rather poor comp efficiency of 68% , it should be 10% better , which would mean a lot less horsepower required so lower turb temps .

I've removed the diffusing exhaust and will construct a water injection system for the comp inlet to try first as it won't require disassembling the engine to fit an internal spray manifold , I'll design for an injection rate of ~4-5 litres per minute and see what happens , at the very least it should cool the turb temps .

I've redone the comp diffuser calcs and they still look OK , certainly not enough of a problem to cause a 10% reduction in efficiency especially as I have a reasonably generous vaneless space outboard of the comp wheel , it should provide a tad more "flexibility" .

Just ordered a water pump , www.ebay.com.au/itm/ws/eBayISAPI.dll?ViewItem&item=401093424479 hopefully this will quench "the Fat Boys's" thirst ;-)

Cheers
John

[jetjeff]

Hey John,

The G.E. J47-GE-25 used water injection halfway down the burner cans. I understand complete dissassembly sucks (been there done that), but it may be best to position there.

Regards

Jeff

[racket]

Hi Jeff

Yep , internal injection would , and probably will , be the be method as inlet injection doesn't do a lot to improve compressor stage performance with our low pressure/temperature compression , I'm looking more for benefits with regards what'll happen once the water is inside , the increased mass flow and lower temps might shift air flow rates to a more efficient region of the map .

Once I have the engine running at a steady state I can start feeding water in and note any changes , good or bad , as it'll hopefully indicate where I need to proceed.

I really have no idea whats going on at present , all I can do is start trying things , LOL, I might just fluke a good result :-)

Cheers
John

[turboron]

John, my experience in rotating machinery is that most problems look silly when you find the root cause. For example, perhaps you just have a leak between the compressor discharge and the turbine inlet flange.

Thanks, Ron

[racket]

Hi Ron

Yep , I dare say I'll have a good laugh when I find the answer

Cheers
John

[racket]

Hi Chris

I've been thinking about your venturi gas velocities in the Solar engines fuel delivery system with my own strange problems .

Could it be possible to produce a venturi throat just downstream of the turbine wheel exducer when a diffusing exhaust is fitted to a turbine stage designed for little gas deflections .

Currently at the 2.5:1 PR the NGV throats should be running choked with an open exhaust , but because my NGV angles are a bit "higher" than normal at ~33-34 degrees , the "radial" gas speed going into the wheel with a choked NGV and high temps could be approaching 1,200 ft/sec, this produces a good start to the continuing expansion through the ever decreasing flow area up to the next "throat" just downstream of the exducer where my thermos and pitot are situated , the cutback exducer is also producing more axial gas flows than a conventional exducer , the gases are just "slipping through" with minimal deflections

Theres a roughly 2:1 area ratio between the diffusing exhaust outlet and "the throat" , and plenty of time for diffusion over the roughly 3 inlet diameter length of the diffuser , ~300 mm long.

The ~5 psi measured on my pitot is above ambient , but if the static pressure at the throat was substantially below ambient then the pressure ratio between the NGV throat and exducer throat might be high enough to choke things .

If the Solar guys could several hundred feet per second using just the small pressure drop across a flametube wall whilst starting from a virtual standstill , whats to say I couldn't get a couple of thousand ft/sec starting from 1,200 ft/sec .

Your thoughts on this would be greatly appreciated .

Cheers
John

[finiteparts]

Hi John,

I think that is really what you are doing with the exhaust diffuser, except the throat is occurring at the natural location in the turbine stage.

If I am reading this correctly, you are proposing to put a reduced cross-section just downstream of the turbine exit plane? I don't think that would help since you would need to supply an increased static pressure to "push" the flow through the new exhaust throat area, which would push the static pressure at the turbine exit plane up and thus reduce the total to static pressure ratio across the turbine stage. I think working on getting the best diffuser that you can for the exhaust is your best bet. Perhaps even adding a centerbody to reduce the dump loss at the turbine hub as the flow dumps to the center.

With a 5 psi total pressure residual in the exhaust, from the isentropic relations we see that your turbine exhaust Mach number is around 0.66. That is a huge amount of residual kinetic energy in your exhaust flow. I am concerned that due to the severe clipping, you might have shifted torque capability of the turbine too far off to keep up with the compressor torque demand.

If you dig into Whitfield and Baine's "Design of Radial Turbomachinery" book, they develop a method for specifying the relative discharge flow angle for a minimum relative and absolute discharge Mach numbers. The reason they are driving to minimize the relative discharge Mach number is that high passage Mach numbers are very inefficient due to frictional losses. Similarly, the desire to minimize the absolute discharge Mach number is because the kinetic energy loss is becomes very significant.

So as always in engineering problems, there is a balance where both are minimized. The minimum loss occurs with relative discharge angles of between -55 to -70 degrees from the axial direction (thus 20 to 35 degrees from the tangential). If there is an ability to diffuse the exhaust flow efficiently, then having a larger absolute discharge Mach number is acceptable and the -55 degree from axial relative flow angle is ok. But if there is not a good means to recover the kinetic energy in the exhaust, then the turbine design should target to have lower exit blade angles (closer to the -70 degrees to the axial) to reduce the absolute discharge flow speed at the expense of increased passage velocities (relative flow velocities).

If we look at what happens with reduce exit swirl angle, you have essentially reduced the enthalpy drop through the stage at a given rotational speed. The reduced enthalpy drop means the rotor has a higher specific speed which means that it incurs higher exit kinetic energy losses. See the plot on page 13 of:

ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19680006474.pdf

If you look further in at page 15, you can see that efficient recovery of the exhaust kinetic energy allows peak static efficiency to be moved to higher specific speed turbines. They reference another document where they did the diffuser testing, here:

ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19680002452.pdf

This diffuser doesn't look very easy to build for the homebuilder, but the document does show that there can be a "large" gain in efficiency due to the addition of an efficient diffuser. There is some good diffuser maps here that you might use to "check" your diffuser sizing and make an estimate of your diffuser performance and if there might be any way to extract more recovery.

ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19730022195.pdf

Another bit of information that might be helpful is the residual swirl...you might try is to hook your total pressure probe to a water manometer and then while the engine is running, if you could devise a way to rotate it, you could find the swirl direction by watching for the peak total pressure. You might also to put some oil or ink in the compressor and exhaust duct to see if you can get some "flow streaks" as a means to understand the residual swirl and diffuser incidence angles.

I think you are on the right path though...if you can get the static pressure recovery of the exhaust diffuser up, you will be able to get the velocities increased through your turbine minimum cross section and thus increase the turbine work.

Good luck,

Chris

[racket]

Hi Chris

Thanks for your thoughts and especially the Links , a bit of "digestion time" required with those .

Ok, bear with me here , if my current "throat" is at the exducer outlet plane , the flow area increasing a bit after that due to the lack of blading , but my exducer shroud curve extends radially inwards a couple of millimetres further due to material thickness and design requirements , this should roughly maintain the exducer flow area up to my pitot and thermos , so I can assume I'm reading the throat conditions .

Now assuming I currently have a "venturi" between turb inducer and diffuser exit , there being a reduction in flow area up to the throat and an increase afterwards , then the static pressure at the throat will be substantially below ambient , this would increase the gas velocities up to the throat , but, if I remember correctly , density drops faster than velocity increases, so there could be some "problems" there , not to mention the losses from the very high gas speeds coming out of the turb wheel and into the diffuser , with undoubtedly more losses as those high speed gases try to diffuse , 2 lots of losses ...........mmmmm :-(

The 5 psit I'm measuring is above "exterior" ambient , the actual static pressure at the throat would be below ambient , so the dynamic component would be greater than the 5 psi indicates , I really need a total minus static pressure gauge to get a true indication of the velocity component.

With my current fuel burn rate of ~2,400 ml/min and temp rise in the flametube the mass flow is ~3 lbs/sec which gives me a velocity at the exit of the diffuser of ~750 ft/sec , it'd normally require a minimum doubling of speeds to just satisfy the area ratio between in and out of the diffuser , but the density changes would have to increase that speed change, LOL.... I'm in uncharted water here.

Yep , I'll need some good luck if I'm to get this miss match of components to work .

Thanks again for your thoughts , sooner or later something will gel in my grey matter if I keep throwing ideas around :-)

Cheers
John

[turboron]

John, when we added a boost compressor to the 501KB5 to create the 501KB7 we used the original T56 turbine with only a 4th stage turbine blade change to reduce the swirl. The turbine exit Mach Number was around .6 at the 4th stage turbine blade exit. The original configuration released used the necked down aircraft nozzle which increased the velocity into the diffuser. We used a simple conical diffuser with a 5 degree included angle with no problems. The 5 degree angle results in a fairly long diffuser. We later changed the configuration to remove the T56 exhaust nozzle and had a straight section out of the turbine before entering the diffuser. I believe that the literature suggested that a 7 degree angle could also be used. Also, we injected steam into the 501KB7 giving a .6 turbine exit Mach Number with no problems using the same type of diffuser.

Thanks, Ron

[finiteparts]

Hi John,

Ok, I think we are both saying the same thing. Yes, I agree...if you get a high static pressure recovery in the diffuser, your turbine exit plane will be substantially below ambient. Hopefully that will make your wheel behave more like a lower specific speed turbine.

I for some reason thought that your total pressure measurement was near the discharge, so I assumed the static pressure was close to ambient...sorry about that...so wow, your exhaust Mach number will be even higher.

As for the static pressure measurement, have you tried just drilling a static tap in the wall of the diffuser in the same plane as the total pressure measurement? As long as the tap is flush with the wall, you should be able to get a good measurement. There is a nice NASA document that goes over how to "correctly" install wall static pressure taps, but I can't seem to find it right now. This presentation seems to have a lot of the pictures from that NASA document, but no references to it.

ocw.metu.edu.tr/pluginfile.php/1865/mod_resource/content/0/AE547/AE547_3_Pressuremeasurements.pdf


Good luck,

Chris

[racket]

Hi Ron

Yep , the diffuser shouldn't really be a problem on my engine, its probably a combination of factors that don't normally occur in a well designed engine with matched comp and turb , my combination is probably borderline operational ...............LOL, its sorta like when I first started this hobby and I didn't have any idea what I was doing or whether it could even been made to work , in 1990 there wasn't much info around on turning a turbo into a jet engine, I guess I just like to try the unknown , it makes for an interesting journey :-)

I'll do some tests with the diffuser removed to get an idea of the "difference??" it makes .

Cheers
John

[racket]

Hi Chris

An interesting article on pressure taps , it helps explain why I had such "misleading" results when I first tried it on my TV84 back in 1998 , in 2000 I was running a P2 of 40 psi with 7 psi static and 10 psi total in my jetpipe , but when I did the calcs the flow was too great for the temps/density and flow area , in hindsight the distance between the two taps was probably the cause , there most likely being a faster central core in the jetpipe which affected the mass flow calcs.

LOL, sometimes I feel the easiest way for us DIYers is simply to measure the thrust and if its "sufficient", don't get too hung up on the data that could be less than accurate unless extreme care is taken in its positioning .............ah , its a complicated business at times :-)

Cheers
John

[racket]

Hi Guys

Test conducted without diffuser exhaust , at both 0.5 and 1.0 bar P2 the temps were higher than with the diffuser attached , the diffuser was undoubtedly doing its job of increasing the pressure drop across the turbine wheel and reducing "wasted" energy in a high speed exhaust .

But, as with previous tests once up to a 1.5 bar P2 , the engine "hit the wall", and temps were unacceptably high .

I can rule out the diffuser being the problem ...........thats one thing I can tick off the list.

The 0.5 and 1.0 bar P2 temps were in the mid to high 600s C , not good , but could be accommodated .

Water injection inside the engine is now the only option to cool things down , so time to pull her apart and fit a water spray manifold .

LOL, removing the diffuser didn't change the noise , she's still loud :-)

Cheers
John