Testing KJ-66 for research
 

Hello, I am a masters' student in mechanical engineering. I'm investigating small scale electrical generation, specifically at using gas turbines for small (>10 kW) generator units. I'm experimenting with a KJ-66 engine to determine how much shaft power I could extract from the engine. To convert from a jet engine to a gas turbine I replaced the propelling nozzle with a diffuser that I had made.

Last week after a long time of setting up and testing problems (I have completely got rid of the manufacturer's control system and have my own, as well as custom sensors throughout the engine) I finally got the engine up to 126 kRPM. Unfortunately the engine failed at this speed (but I did have enough data for that run to be useful). I think it was actually a rear diffuser guide vane that failed and then hit the turbine which ripped off a turbine blade and generally caused havoc all through the engine. I do have a backup engine worth of parts so its not catastrophic or anything, just annoying to have to reassemble everything (and also not have a backup from now on).

In assembling the engine I found that the shaft tunnel and the NGV piece had a sliding fit. I know the previous ones I had used were pretty tight so I was surprised. Now I looked in the KJ-66 assembly manual and it didn't say anything about the fit. I also have a JG-100 manual which says it should be a sliding fit (for the JG-100 engine which is very similar to the KJ-66). So essentially I was wondering, why would a sliding fit be good? How does the NGVs stay in place? Is the outer casing the only thing holding it from moving axially?

If you guys are interested I can put up some pictures later (and give updates on how the research is going)

Paul

RE: Testing KJ-66 for research
 

What was the EGT when it failed? Please post pics

RE: Testing KJ-66 for research
 

126K is way beyond the max rpm for a KJ66. 100K was pushing the limit if I recall. Chances are the turbine blades streched due to heat and centrifical loads and fouled in the NGV.
Be carefull, when they blow it's like a hand grenade going off.

Been there (see below)!

RE: Testing KJ-66 for research
 

I'm very interested in hearing what data you collected from the engine. Can you please post your results on the run you did?

RE: Testing KJ-66 for research
 

Did you mention 126k on a KJ66[X(] I think you did well to keep it together for as long as you did. What RPM were you hoping for?


But anyway, the slide fit of the NGV to the shaft tunnel keeps the shroud concentric to the turbine wheel tips and the case sets the distance of the NGV guide vanes to the back of the wheel. Some engines had the NGV bolted to the end of the shaft tunnel but I think maybe only AMT still use this.
For whatever reason, after the engine has done some running the fit of the NGV to the shaft tunnel seems to close up and they can be really tight to get apart.

It's unlikely the NGV guide vanes failed and at 126k I would say it was most defiantly the turbine wheel that let go first.

If its a proper set of parts you are using and you have the facility to balance the assembly correctly then I would be running the engine at 117k.

I'd like to see some pics.

Jason

RE: Testing KJ-66 for research
 

I have a Wren turbine wheel in my KJ and it is rated at 120K RPM and I run my engine at 118K RPM with no problems to date.

During the testing of the turbine, as the turbine was accelerated, the turbine wheel expanded with exhaust temperature and this casued the turbine wheel to rub against the NGV, stopping the turbine. After the turbine wheel cooled down, it became free to rotate again and inspection had shown no damage to the NGV or turbine wheel so we had to skim the NGV because the clearance between the turbine wheel and the NGV was too tight.

duthie.

RE: Testing KJ-66 for research
 

Thanks guys for your comments. I didn't know that 126K was too high for the KJ66 - I saw the rated max speed was 128K (just from the old artesjet website, probably not that great a resource), and I knew that one of my predecessors had run a JG-100 to 126K consistently. Though I guess the JG100 is a better design than the KJ66.. but it has the same turbine wheel and NGVs does it not? Anyways, I will slow it down, probably stay to max 118K or lower.

This is a picture of the diffuser that I replaced the propelling nozzle with. The guide vanes are only spot welded on which was why I thought it failed first. But yeah, it could have been the turbine. I'll have pictures of the damage tomorrow since I don't have the ones I took with me today.

RE: Testing KJ-66 for research
 

Here's an operating point at 102.5 kRPM. m in kg/s, P in kPa (abs), T in k, C in m/s (velocity).

1 is in the inlet pipe on my rig, 2 is right before the compressor, 3 is right after the compressor, 4 is before the NGVs, 5 is the outlet of the turbine, 6 is at the exit of the diffuser. The thrust is small because the diffuser recovers the kinetic energy from the exhaust.

RE: Testing KJ-66 for research
 

what type of pressure sensors are you using for inside the engine? Also, what are you using to measure velocity; pitot tube with a dp sensor?

RE: Testing KJ-66 for research
 

I have stagnation pressure probes inside the engine. All the velocities, densities, static pressures (except for P2 and P5) are calculated using the pressure and temperature, mass flow rate and flow areas.

RE: Testing KJ-66 for research
 

Here are some pictures...

first is the turbine after it was taken out of the engine. You can see the missing blade, as well as others that are bent. The reason I thought it was the diffuser vanes (from the diffuser pictured above) was that the blades before ANDafter the missing blade are bent. You can see the broken diffuser vane circled in red and the place where it is supposed to be. These vanes are just tack welded on and thus are relatively weak. But it could have easily been the turbine. Another thing it could have been is the rear bearing since it was really worn down (on the inner race it was almost worn down to the shaft - sorry about the bad quality picture for that one!). Anyone want to hazzard a guess as to which one came first - the turbine causing the bearing to fail or the bearing fail which caused the turbine to rub and fail?

Do you guys have any experience using a top flite magnetic balancer for turbines? Is it sensitive enough? I'm thinking of getting one as the Du-Bro prop balancer I have doesn't seem good enough.

RE: Testing KJ-66 for research
 

You are right there!


<div style="margin: 0cm 0cm 10pt"><span style="font-size: 9pt; line-height: 115%">Either bearing failure caused the wheel to touch and rip off and bend the blades or theturbine failed and the massive imbalance caused the bearing failure. Which ever way it was most likely caused by over speed at your 126k rm and probably helped along by your Dubro prop balancer,balanced shaft assembly...http://www.rcuniverse.com/forum/js/f...used_smile.gif

Jason</span></div>

RE: Testing KJ-66 for research
 

panzer32

It looks like your bearing failed and a part of it damaged your turbine wheel, any pic of your ngv? it will be good if you post some pictures of all the parts you have used.
My kj66 is running at 125.000 rpm at around 680c with no problem so far hope i will not have any problem in the future when i use it in an airframe, for the moment was tested after reconstructing it with full autostart, inconel 12 stick cc new bearings newer turbine and ngv.
One of the best turbine wheels and ngv you can find are from Dieter Albisser if you are looking for ,that's the one i have used and many well known turbine manufacturers. his website is http://jetmax.ch/.



Eftychios

RE: Testing KJ-66 for research
 

Hy,

126k RPMdefinetly should not be a problem for a standard KJ66 setup, given that it is a good turbine wheel, good bearings and a standard internal setup with a perfect balance.
If it was not a bearing failure i'd think you pobably had it off-balance....a du-bro will NOTdo! You need to have it dynamically balanced...the equipment you need for that is pretty complex, your best bet is to send it to somebody to balance it for you. Rescue Turbine Service can do that for you, so can Wren.....
They will also mark the position in which it was balanced, because if it is reassembled just a few degrees different from the last time, balance will be off again!
Just a tip: if you start it up to idle RPM, hold jour fingernail to the housing. If you feel more than just the tiniest of vibration, shut her down and have her balanced! That is the first thing i do when i start up a homebuilt, and i recheck every RPMincrease for vibrations.....

By the way, the sliding fit between the shaft tunnel and the NVG is to allow for temperature expansion. Steel and Aluminum have different temp coefficients, and therefore will grow and shrink differently. The sliding seat will allow it to adjust without putting unneccesary loads on any parts.
You can also use a steel shaft housing and bolt it to the NVG, but then you will need a sliding seat to the housing. Both variants work, both have their advantages and disadvantages.....

best regards
Hank

RE: Testing KJ-66 for research
 

I thought them NVG's were for seeing at night :) or was that the Nozzle Guide Vanes (NGV)?