GE Frame 3 Starter

[kahlilj]

Has anyone ever replaced the starter for a GE Frame 3? We have several of these units at a couple of facilities. They were initially installed with Elliott Steam turbine starters (circa 1960's). Some of those steam starters were replaced with Coppus gas turbine starters in the 90's.

We are currently looking at replacement designs (for the remaining steam turbines). One of the designs is a pneumatic (positive displacement) starter. The manufacturer has installed these on other Frame 3's similar to ours, but the starting torque design used for those others was significantly less than what our turbine starters use. The pneumatic used 550 ft-lbs of torque. Our turbine starters are spec'd to 1650 ft-lbs of torque.

I contacted G.E. directly & they sent me a graph showing the torque requirements for the Frame 3 (attached). At 0 rpm the required torque is only 700 ft-lbs. And at light-off, (~25% rated speed), the needed torque is ~550 ft-lbs. I think this corresponds to the Frame 3's 1750 rpm point in which the turbine starter's required torque is 450 ft-lbs.

Anyway I'm a little miffed at why there is such a huge difference. Does anyone know? Has anyone ever used a pneumatic starter on a gas turbine?

 

[rmw]

Is yours a 2 shaft or a single shaft Frame 3? I think Frame 3's were all 2 shaft machines but it has been too long ago to remember. I know that Frame 5's came in both two and single shaft versions.

I have not used a pneumatic starter on a gas turbine but I have on Diesel engines. I have used steam and gas expander turbine starters as you have, electric motors and diesel engines (for black start). The only starters I ever replaced were like for like.

I think that yours may be a case where the steam turbine that was picked for the application had a higher rating than was required by the turbine since they used a basic 'off the shelf' item. I never really paid attention to the rating of those I used, just whether or not they did their thing.

What is the breakaway torque of the pneumatic starter? The 550 rating may just be the full speed torque or the minimum torque. Do they provide a torque curve? There is no reason why a pneumatic device can't do it and there isn't a great deal of difference in the basic configuration of the old Frame 3, so maybe they just need to do some more convincing to assure you that it will work.

To me, the breakaway torque requirement is the determining factor. If it can't break it away, what torque it produces at light off speed is academic.

Also, check to see that the starter gear box ratios between your application and what they are offering as their experience are the same. Torque rating is worthless if it isn't at the speed at which that rating is good for.

rmw

 

[j79guy]

I've never been a fan of positive displacement pneumatic starters. Unless the supply air is very clean, they tend to have high wear rates, and generally require lubricators placed in-line with the air supply. (Vane type.)(Read messy, and high maintenance.) Frame-3 units hang on the starter for quite a while, thus if I was to spec a pneumatic starter, it would be a turbine type, with a non-contact primary rotating element. As RMW asks, if your Frame-3 unit is two spool, the lower torque rating of the proposed new starter shouldn't really be an issue. You may have to extend your crank timers, or other mods to make the control logic happy, but the unit should still light off and ramp ok.
Some of the less expensive positive displacement starters, I recall rebuilding them once a month, to get them to survive in a less than perfectly clean supply air environment. Finally replaced with a non-contact, air turbine type, problems solved.

Robin Sipe.

 

[kahlilj]

Thanks RMW & Robin,
The unit is a 2 shaft Frame F-Models. The pneumatic medium will be compressed natural gas & it will be "clean".

If by breakaway torque you mean the torque required to start-up the unit, then the pneumatic starter is rated for 550 ft-lbs. According to GE the required start-up (breakaway)torque for Frame 3's is ~700 ft-lbs.

 

[rmw]

Breakaway torque is the torque required to roll it off from a dead stop. Once it is broken away and rolling, the torque required to keep it rolling is less as GE's curve shows.

rmw

 

[kahlilj]

I found some job book data on this frame 3 unit (F-model) which give some history or design basis. The original starter was an Elliot steam expander turbine rated for 400 HP at 4200 rpm (@600# & 750°F). I calculated this to be ~500 lb-ft of torque at that speed.

That still seems a bit high based on the manufacturer's torque curve. This unit is rated for 6900 rpm for the 1st stage turbine. 4200 rpm is ~61% of the rated speed. according to the manufacturers curve at ~61% the required torque is only ~360 lb-ft. Is it typical to have this much cushion (40% more) from the required torque vs the actual/potential torque??

 

[j79guy]

To answer your question; Is it typical? No. Is it desireable? Sure. Having a starter that potentially has more torque available than that required, is a good thing, and leads to consistent lightoffs, and starts. Short of breaking the N1 starting drive shaft, there isn't such a thing as too much starting torque, and nothing says you have to use all the available torque anyways.
However, the older Frame type engines generally have ultra conservative lightoff and ramp rates, that can be shortened, to reduce fugitive methane emissions. (If your starting turbine is venting methane to atmosphere as most are here in Northern Canada.) On my test cell, we have only three starters, that we use on all turbines we run. (7 different engine models.) Sometimes the starters are a bit low on torque for the particular engine we're testing, but if you are carefull with the lightoff, and ramp, this generally doesn't cause any trouble. There is a lot of leeway either way, as far as required starting torque is concerned. If your starter can get the engine reliably to lightoff speed, and assist on up, it will work fine.

Robin Sipe

 

[rmw]

Any horsepower rating has to be given at a specific speed or it is meaningless. So this is the particular RPM that this turbine was rated at, not where it might or might not operate. And, the rated torque (HP) is based on the work being delivered by steam at this particular temperature and pressure (and exhaust pressure has to be considered too.

All that said, this turbine would have an entirely different torque and HP rating for gas at some given pressure and back pressure.

The 6900 speed that you mention, is it the operating speed or the light off speed? All the starting device has to do is to get it to the ignition point and sustain the speed until the fire gets lit and accelerates the turbine away from the starting device.

rmw

 

[kahlilj]

thanks robin & rmw.

I am aware that the rated speed of the starter does not have to exactly match the required torque of the turbine. i just thought the difference was a bit large, but then I have never designed a gas turbine nor its componentry.

6900 rpm is the rated speed for full load (operating speed?). I think it lights off at 25% rated speed or ~1700 rpm?? Not completely sure as I not in the office today & do not have my data files with me.

Robin, what is this test cell you referred to? & What kind of starters are being used? We actually had this Elliott steam turbine expander starter on 8 units Frame 3 originally. Since then 4 of those starters were replaced with Coppus turbine expanders which could accommodate both steam & natural gas. So I'm curious to know what you are testing & what works best.

 

[j79guy]

kahlilj; 6900 Rpm will be full rated speed for a Frame-3 engine. For liquid fuel, you can light off the engine at a minimum of 10% speed, or 690 Rpm. For Natural Gas fuel, light off minimum speed is 14% or about 1000 Rpm.
The test cell I refer to, is my test cell, where we run aeroderivitive gas turbine engines, post overhaul or repairs. The GE LM1500 industrial engine is similar to the Frame-3 in terms of light off, and full power speeds, thus I was extrapolating a bit to your application. Some of the older Westinghouse heavy frame turbine units here in Northern Canada use steam turbines as starters as well, with natural gas as the driving medium. These were put into service in the late 1960's, and are still functioning well. What are we using? On our test cell, on the aeroderivitive engines, we're using Garrett AiResearch, TDI and Bendix air-turbine starters. All are working well, but I would give the highest rating to the units that have constant lubrication, (Garrett.) as opposed to self contained oil sump. (Bendix and TDI) Positive displacement startes, such as Powerquik, and Ingersoll Rand, were disasters. The Hilliard units though, seem to be robust.

Robin Sipe.

 

[kahlilj]

Thanks for sharing Robin. Our G.E. Frame 3 units are actually from that same 60's era & might be that Westinghouse design as well?.. I'm unsure.

I am hopeful about the Hilliard product for both cost savings & reduced emissions. We have a site visit scheduled for next week. Any thoughts/suggestions on things to ask Hilliard about their product in this application?

One of the things I hope to do is to start one of the GE turbines & actually measure it's starting (breakaway) torque. Does anyone know of a way to do this?

Thanks again.
Kahlil

 

[j79guy]

The Westinghouse turbines I mentioned are the old W-92 units, (9000 shp, 2-spool.)a bit bigger than the Frame-3.
Questions for Hilliard? Well, first you will need to overlook the sales pitch, and ask for real empirical data as to average starts between rebuilds, past failure modes, and tolerence to impurities in your starting medium. Hilliard make an acceptable product, but tend to be associated with Rolls Royce, thus suffer from a superiority complex that isn't necessarily warranted.

I can't help you with determining breakaway torque, other than suggesting physically measuring the torque it takes to turn the N1 from a dead stop, with a breaker bar. Do you have an access port to the N1 spool, to bar the engine over, as during inspections? If so, then you can stick a torque wrench in and actually measure the breakaway torque.

Robin Sipe

 

[kahlilj]

Robin,
Thanks for your valuable input & insight. I will certainly bring those issues up. As for access port on the N1 side, I am not sure?? I will check with the station to find out if there is one or what they do during inspections. But i like your idea it seems a simple enough approach.

Kahlil