Gluhareff 130R power failure issue still open ... Any Idear?

[Viv]

Hi all

I cant post the last CFD report to the forum and I am in a rush to go off to Grahams for the day so if you go on to my .Mac site, then look in the pressure jet folder , then at the bottom of the list is supersonic nozzle report in .doc and .pdf

www.homepage.mac.com/vivcollins/FileSharing6.html

Viv

[Dave]

I have gone back to scan through the various CDF reports, but could not draw any clear conclusions from them. Could someone enlighten me as to where things go from here.
Dave

[luc]

Hi Dave,

Basically, what you have to do when looking a the document is put your attention mostly on the images as far as plume chape and flow patern is concerned.

If you want to do calculations, you will need to put your attention in the last pages, wich include the surface area data (Flow, speed, temp... ect.).

But, the Supersonic Nozzle CFD @ 225 psig. is only a aftway process, were I am doing the CFD of the Gluhareff Nozzle @ 700 psig. At the end, we will all be able to visualize flow paterns, plume shape, mach numbers and data, and come to a conclusion by comparing both nozzles.

I hope this can put some light on your question.

Bye the way .... Were is Raymond? ... I taugh he would have been the first one to be back with a coment about his Nozzle... He he he.

Hoooo ... One more thing. I don't know want Gluhareff did when he design his nozzle, but it is hard like hell to simulate. My CFD program is having a hard time to reach all the same goals as the Supersonic Nozzle. CPU time is now 77 hours, symmetric calculation and still 2 goals to achieve.

Okeyyy .... Cya guys.

Luc ...................... Out

[Raymond G]

Luc Wrote:

Bye the way .... Were is Raymond? ... I taugh he would have been the first one to be back with a coment about his Nozzle... He he he.

I've been away for awhile. Give me a little time to digest the report, and then I will comment. Looks good and interesting on the surface though.

Later,
Raymond

P.S. It was nice of you to credit me in the report. Thank you.

[Dave]

Luc
Thanks for the feedback. I will take another look through the reports then chill while others draw their conclusions.
Dave

[luc]

Okey Guys,

CFD simulation of the Gluey Nozzle in its own intake stack @ 700 psig./1200 °F is over. I will build the repport as soon as I get Viv feed back about what should be included in the repport and what should be left aside.

Then.... we will analyse and compare.

Cya,

Luc ........................... Out

[Raymond G]

Well guys,

We've invented perpetual motion! My initial analysis of the supersonic nozzle's CFD simulation is that the results are highly suspect. At first, many of the results just didn't seem right. These included:

1. Mass flow from the fuel nozzle is too low. The value of .039 lbm/sec seems to be one forth of what it should be. (p.22 of the .doc file)

2. The flow velocity of 689 ft/sec from the 2nd stage ejector seems much too high. (p.25)

3. One would expect too see some shock and expansion waves in the nozzle exhaust plume, but they are almost non-existent. I see an oblique shock at the fuel nozzle exit (p.20), and a normal shock just downstream of the 1st stage ejector exit (p. 21). I would normally expect to see some reflected shock and expansion waves (i.e. shock diamonds) in the image on page 20. Their absence, along with the other items above suggests that the solution we are looking at is not converged.

4. Flow amblification is startlingly high. The ratio of 2.55 to .039 lbm/sec is 65.4 to 1, which is huge, and highly unlikely.

5. Finally, the kinetic energies don't match. The kinetic energy at the 2nd stage ejector exit is about 3.5 times greater than at the fuel nozzle. Generally speaking, we cannot create energy ;-) I havn't done a complete energy analysis to include thermal and pressure energies to see if they all balance, but my experience is that you never get good enough thermal mixing in an ejector to actually increase kinetic energy. Remember that the manual gives examples where the measured KE efficiency was around 67%. We are looking at 350%!!!

Comments and corrections would be greatly appreciated.

Regards,
Raymond

[luc]

Hummmmm.....

The only thing I could do is double check the boundary conditions, make sure the data is correct.

What I could do also... is to replace the actual Nozzle inlet data, wich is : Propane gas under 225 psig. @ 1200 °F, by an inlet flow condition wich would be, Propane gas Flow, temp. and pressure. But I need to know first, if the .78 lbs/lb/Hr is exact.

Also, you need to know that the fluid media is a choice available in the CFD program itself, therefore, I take for granted that the computer knows what gasious propane is.

Just tell me the gasious flow I should input or if I can take the .78 Lbs/Lb/Hr and I will rerun the simulation.

I think this is our problem.

Regards,

Luc

[Raymond G]

Luc wrote:

Just tell me the gasious flow I should input or if I can take the .78 Lbs/Lb/Hr and I will rerun the simulation.

You lost me there. Was this a reply to my post about the supersonic CFD simulation discrepancies? If so, I don't understand where the .78lbs/lb/Hr came from. This looks like the claimed SFC of the G2-8-130, which has little to do at this point with the inputs to the CFD analysis.

Regards,
Raymond

[luc]

Yes Raymond,

My previous posting was in answer to your discrepancy.

What I am saying, is that the previous simulation was done using only the following conditions :"Gasious Propane", "225 psig." and "1200 °F" as Nozzle Inlet parameters, assuming that the CFD program knew what would be the flow of gasious propane under those conditions.

I guess I was wrong assuming that... Therefore, what I need is to add another inlet condition parameter, wich would the flow (CfC of .78 Lbs/Lb/Hr = 14.55 Liter/Lb/Hr or 3.84 Gal./Lb/Hr), wich represent the quatity of fuel this engine swallows per pounds of thrust per hour.

What I am asking is : can we use the CFC of .78 Lbs/lb/hr (14.55 Liters or 3.84 gal.) at a pressure of 165 psig. wich is supose to be the CFC consumption of the 130R engine and rerun the simulation.

I think this is were the CFD simulation error come from.

Cya,

Luc

and set the pressure at 165 psig. wich is the .

[luc]

Come on guys,

I need an answer .... I need a flow figure. I have looked into the boundary conditions of the supersonic nozzle CFD analysis and realise that the nozzle inlet conditions were establish by only 3 parameters (Media, Pressure and temp.)

But, I can change that for "Inlet Volume Flow" or "Inlet Mass Flow". But I am not sure what I should use or if I can use the Gluhareff of .78 Lbs/Lb/hr as specified in its manual and convert it. So...here are the figure I need confirme with you guys.

Media = Is Gas Propane Ok?
Pressure = Is 165 psig. Ok? (As stated by Gluhareff in its manual)
Temp. = 1200 °F
Flow = ??????????????? (.078 Lbs/Lb/Hr as stated in the Gluey manual)

I can enter Mass Flow rate (Lb/s) or Volume Flow (Cubic Feet Min.)

Need your answers...Guys,

Cya,

Luc ...... Out

using

[Raymond G]

I guess I was wrong assuming that... Therefore, what I need is to add another inlet condition parameter, wich would the flow (CfC of .78 Lbs/Lb/Hr = 14.55 Liter/Lb/Hr or 3.84 Gal./Lb/Hr), wich represent the quatity of fuel this engine swallows per pounds of thrust per hour.

Luc,
I would not do that. The fluid properties, pressure and temp should be all you need, and the throat area of the fuel nozzle will determine your mdot(propane). If you add flow rate you will overconstrain, and the program will probably balk at that. Furthermore, the .78 SFC figure is highly suspect, and should not be used to estimate flow rate. Incidently, to get mdot from SFC, multiply thrust times SFC and divide by 3600 (sec/hr). So you would get: (130lbf)*(.78lbm/lbf*hr)/(3600sec/hr)=.0282 lbm/sec.

As I alluded above, your CFD solution does not look converged. This is not inconsistant with the program telling you it has found a solution. It may just mean that you need to use the current solution as input to another run, with tighter convergence tolerances.

It may also mean that your CFD is not up to solving this particular flow field :-( The problem with CFD as compared to other FEA methods is that the solutions are very mathematically "stiff". This means that the actual numbers used in the solution matrix vary by more than 6-10 orders of magnitude (powers of 10). Physically this can be visuallized by the size of a boundary layer compared to the size of the freestream flow. Boundary layers can be very small. They can also be quite large. And the difference in scale can easily exceed 6-10 orders of magnitude. Now add to that physicall geometry that varies in size (i.e. the size of the nozzle throat compared to the length of the induction stack), and all hell can break loose.

Part of what all this comes down to is my recommending using a pie slice of the flow domain to reduce the computational power needed. This will help to make the learning curve a lot less steep since each run will not then take dozens of hours, and more tries can be made untill you find a converged solution of the actual flow field of interest.

Hope This Helps,
Raymond

[Raymond G]

What I am asking is : can we use the CFC of .78 Lbs/lb/hr (14.55 Liters or 3.84 gal.) at a pressure of 165 psig. wich is supose to be the CFC consumption of the 130R engine and rerun the simulation.

OHHHHHH Criky!!!! This question got me wondering why you were so on about the flow rate this SFC would calc. Then it hit me, I had completely forgotten about the flow restrictors! Boy do I have a lot of egg on my face now. So I went back and scanned the previous posts, and sure enought you guys talked about them awhile ago.

I am sorry.

So now Luc, I have to ask a question I asked a while back again: Can you MEASURE the ACTUAL mdot when the engine is running? Back on Jan 16. you posted about three running conditions that you measured thrusts, temps, and pressures, but I see no mdots. This is a very important measurement. Can you do it?

Regards,
Raymond

[Raymond G]

OK. I did a little sanity check and the restrictors will not act as sonic valves as long as the propane is truly still liquid in them. The velocity through them in liquid state is only 100-200 ft/sec. But if ANYTHING in your plumbing causes the propane to vaporize before the restrictors, then your thrust is shot. I tried looking at the picture of the engine at full thrust again, but couldn't see if the restrictors are glowing along with most the rest of the engine. That would be a problem. So again, I wish we could get mass flow rate data.

I also noticed that the exhaust plume looks funny. It balloons downstream of the nozzle implying underexpansion.

Regards,
Raymond

[luc]

Hi Raymond,

Okey....Someone will have to help me learn how to make those nice rectangular when you quote someone previous paragraph or posting.

Raymond ... was is "mdot"? ... As presume your are talking about mass flow rate or volume flow rate. On our side, we never measure this data because Flow turbines and meters are very expenssive and just for measuring the thrust, we busted up 5 meters with the engine vibrations. It seem that electronics really don't shock wave. So we never spent the money for the turbines.

Also, we have a flow meter on the pumps, provided by the gas company but we found out that this thing was highly inacurate and that after 5 minutes of running, propane in the line was suck back in the reservoir. Remember, the reservoir become under vaccum after to much propane is sucked out (Winter condition, when almost no evaporation in the tank). Therefore, we could see reverse reading on the flow meter, screwing up at the same time, our reading.

So... Now you know why I used the .78 figure. With restrictors, you CAN'T
push more propane in, even if you wish for. You are limited by the size, the pressure, the media and the temperature ... That it. What ever the propane become in the coils or after the coils, the propane you push trough the orifices is the propane you get trough the nozzle, vapor or liquid.

Furtheremore, it should be easy for you brainiacs, to calculate what is the quantity of liquid propane going through one orifice of .078" (.00477 area). I can't, I am not that wizard... He he he.

As for taking a slice for the computationnal domaine, I have no problem with that but remember, if I take a slice in the middle of the stack, the program won't consider the sides, and the stack is a circular shape, therefore, not constent from middle to side. My question ... Will it affect the results ???.

1) So ... Do I change the boundary condition and add Flow figure? or Do I live it as it is, wich is 225 psig. @ 1200 °F.

2) If I had a flow, what will it be?

3) Do I take a slice or not?

I need answers ....... HELPPPPPPPPPPP

He he he .... Cya,

Luc