Compressor rotor balance quirks.

[j79guy]

A couple engines come to mind, the GE J79/LM1500 family and the RR Avon. These engines, or more specifically, these engine's compressor rotors share a common first design critical, in the 5700-6400 rpm range.
It seems that no matter how well you balance these rotors, it is still a crap shoot as to how severe the FDC will be. Some engines won't even cause the vibe meter to wiggle going through the FDC, others will shake your teeth fillings out. Rolls' "fix" is to numb down the vibration metering in this RPM range, and/or get creative as to where they place the vibration probe on the engine. No rhyme or reason, every installation is slightly different, often within the same facility that has multiple Avon units operating side-by-side. GE toughs it out by sticking to it's alarm and shutdown values through the whole RPM range, which leads to some perhaps unecessary shutdowns. During balance of these engines' compressor rotors we try to reduce the residual imbalance to minimal levels, but sometimes this is not effective. Over the years we have found that of more importance is balancing the rotors "in the middle", by removing approximately 60% of the imbalance in the middle of the rotors, and then performing a final trim balance on the rotor ends. With the J79/LM1500's, this is easy, as there are specific balance locations to accomodate this. With the Avon, this is a bit more tough. Also, we pay particular attention as to where any residual imbalances lie, as they cannot be in couple.
To top it all off, we leave a significant imbalance in the turbine rotor, and specifically do not 0-balance this assembly. If upon testing of the engines on the cell, midframe vibration is high through the FDC, by re-indexing the turbine rotor 180 degrees, we have a good chance of reducing the FDC.
I have a copy of a report GE made up in the mid 1970's, of a study done over a period of 11 years of J79 production. The report is a good inch thick, and covers a lot of ground. It can all however be boiled down to the last paragraph, on the last page. Basically, GE concluded that they really cannot answer what is going on with the compressor rotors, and no matter what, all Airforce engines had a 1.9 percent failure rate on the test cell due to high vibration, and the Navy engines 1.8 percent.
Other recommendations? If you have a GE engine with a significant FDC, and upon accelerating through it, vibration levels are acceptable, do what Rolls does, and dampen down the vibe monitoring through the FDC. It works for Rolls, it will work on a GE engine. Avons?, well they all shake through the FDC, some more than others. One thing I did run into is in high ambient temperature conditions, the FDC shifts up the RPM range. I was in Iran last August, commissioning a Mk-1533 Avon, and it was +46 degrees C. outside. The FDC was pronounced in the 6200-6400 rpm range, and their vibration monitoring returned to full sensitivity at 6400. It was touch and go as to whether the engine would come on-line or not, and at exactly 6400 the vibes would drop off. Too close to call, and I recommended they raise their vibration dampning to 6450 RPM.

Robin.

[UtahTurbineMan]

Now would you consider vibration dampening and moving probes around to be a form of "lie,cheat and steal" in order to sell an engine? What exactly would be the FAA's take on the process? Its not that i think its wrong, im just ignorant as to the practices of other shops in regards to the balance procedures. How much would the compressors weight be in ratio to the turbines, i would imagine the turbines would weigh alot more and have more of a probability to counteract the vibe data from the compressors pickup. However i would think that leaving significant imbalance the turbine would also spike vibs on the front fan pickups. Does this not seem to be the case?
Oh, PS. How would one get their hands on such a report from GE?

[j79guy]

Well, the OEM does this all the time. The FAA takes their leads from the OEM, so all is ok. Or is it? Does the engine operate reliably, and meet warranty period? If so, no problem then. However though, industrial turbine operators expect their engines to run smooth for tens of thousands of hours between major overhauls, so the observed vibration levels better be real good. Rolls will "move" vibe probes around to find a nice, quiet spot on the engine, and call it a day. Not so much on the GE units.

Every engine design is different, and in your case, you need to look at where the mass is. Which rotor is heavier, the compressor or turbine? Which has a larger radius of the rotating centre of gravity? Do the rotors have a bearing on each end, or overhung? Is the turbine hard coupled to the compressor, or soft coupled? The rotor with the most rotating mass will control the vibration signature of the engine. (The tail wagging the dog.) If the turbine has the most rotating mass, and the compressor is light and long, the turbine can whip the compressor through the FDC and SDC. I'm used to working on the older military aircraft engines where the compressor has the most mass and is long and flexible. Turbine 2/3 the weight of the compressor, short and stiff. For these designs very specific procedures are used in balancing and engine build to try and reduce the FDC/SDC as much as possible. For your specific application, you need to let us know details of the construction and general layout.

Front fan indicated vibes are to me to be treated as a separate issue, apart from the other rotors as it clearly has the most rotating inertia, and usually overhung. I cringe every time I'm on a wibe-body, and hear the fan "growl" on take off. If balanced properly, no high bypass fan engine will growl. Occasionally, and only very occasionally, I will hear a good one. About 10:1
When I mention leaving in a "high" residual imbalance in a turbine rotor, this is relative (Still under the maximum imbalabce allowed for an overhaul.), and specific to single-spool engines where the turbine rotor is hard coupled to the compressor. In multi-spool engines, or flexible copling engines, this goes out the window, and we will 0-balance the turbine rotor. However, leaving in a specific amount of imbalance on the rotors with less rotating mass is a tool that can be used to get a troublesome engine to pass test. Remember, there are no hard and fast rules, and manuals are a guideline, written by people who in their best estimates think that these are the best procedures. The best manuals are the ones that evolve over time, as better procedures are brought to light.

The report no longer exists. I made a copy of the last few pages while digging through GE's J79/LM1500 archives way back in 1996, and they were destroyed shortly thereafter. The last few pages are where the good stuff is anyways and I can tell you about it.

Robin.

[UtahTurbineMan]

With our talk on Friday and your repost as well, yes i am a little stunned that Rolls Royce would do this practice. It defenitely makes sense that military aircraft would have a much larger/longer compressor than turbines, so thats out as well. I have tried to knowingly leave imbalance in the HP turbine and "line up" corrections in the build and with luck in doing so. I havent had the same luck on the LP sides of things. When you said you do like your weights within a 120 degree spectrum that made me a little happy, last year i implemented the exact same measurement into our build planning here. I like to know that im on the same track as the smart people are. Im constantly trying what i can with these engines in order to solve vibration issues, as im sure we will talk again about it. Thanks!

[j79guy]

No problem, you learn, I learn, we all gain from it. Ultimately, a turbine is just a man made mechanical device that does not know what name is on the data tag. The basic principles of rotor balancing are universal to all makes, so "special" problems just need to be worked out.

Robin Sipe.