Weld repair Generator cooling air duct


I hope this message finds you all well.

We have 25MW Holec Generator driven by GE Frame 5. There is a crack in the inlet cooling air duct. We want to repair it by TIG weld. Is there any risk in welding without dismantling the ducts ?

Thank you in advance.

This is one of those questions where a picture is worth a thousand words. I don't think many people have ever seen a Holec generator. And, inlet cooling air ducts can take many different shapes. You haven't told if the duct is on the generator, if it's in the generator, if its connected to the generator with a flexible "boot" that is non-metallic. A picture would definitely be extremely useful, if not outright necessary.

Next, as I understand TIG welding, it's done using an inert gas--but it still uses high current electricity that requires a ground/neutral be attached to a common piece of metal (common to the piece being welded). However, in this case (a heavy duty gas turbine-generator with an electronic turbine control system and electronic field devices/instruments), the placement of the welding machine ground/neutral/negative lead is extremely important whenever welding on equipment that has sensitive electronics attached to it. The welding machine ground/neutral/negative lead should be attached to bare metal as CLOSE to the area where the welding will be done as is possible. If this means paint has to be scraped or ground off to make a suitable ground, then that's what has to be done. The welding machine ground/neutral/negative lead cannot be attached to a piece of structural steel or earthing strap 25, 50 or 100 meters away. (Well, it can--but the results probably won't be very much fun, or inexpensive.) If it is, then the current can take many different paths--sometimes through the control system or field devices--when "returning" to the welding machine. This can be very disastrous.

Also, while this goes without saying, but just in case this was one of the questions you were implying but not actually asking (and may ask in another post to this thread): It would NOT be recommended to perform any welding while the turbine and generator was running. Many sites actually power the turbine control system down entirely, though it's not really necessary if the welding machine ground/neutral lead is attached very close the area being welded (so current doesn't have to travel far to "return" to the welding machine).

Hope this helps!

(I think it's been a while since you last posted to Control.com, but now, one can attach all types of files to threads/posts/responses--even photographs!)
Thanks a lot CSA for your support. Really appreciate it.

Below are some images for the crack and the ducting assembly. I still don't understand the risk on the control system. The generator protection we use is electromechanical, not electronic. Do you mean the bearing thermocouples and stator RTDs ? Does this mean that, in general, we must avoid welding on the generator while instrumentations are installed and connected to the control system ?

Of course we will do the job after the turbine stops, the only reason we want to do the weld without dismantling the duct is to reduce the turbine downtime; i.e. dismantling the ducts will take time and the turbine will not be stand-by.

Image  (1).JPG
Image  (2).JPG
Image  (3).JPGImage  (4).JPG

Well, most embedded generator RTDs have one lead grounded (earthed) at the JB on the generator. So, that is a risk. Some bearing T/Cs are also grounded/earthed.

Most generators have lightning suppression which can be affected by welding if the welding ground/neutral/negative lead is not attached close to where the welding is being done on or around the generator.

And, there has also been damage to bearings/bearing journals caused by not attaching the welding ground/neutral/negative lead close to where the welding is being done on or around the generator.

Look, any welder worth his wages knows when welding around high-speed rotating equipment with any kind of instrumentation involved or journal bearings, etc., to attach the welding ground/neutral/negative lead close to where the welding is being done. It is--or should be--taught in any elementary welding or welding certification course.

Just use common sense when analyzing the situation. If the embedded generator RTDs are NOT connected to the Mark*, or the bearing T/Cs are NOT connected to the Mark*, then look at all the other instrumentation that might be connected to the Mark* (load transducers; voltage transducers; PTs; CT's; etc.) as potentially being damaged--including the Mark*. BUT, good welding practices should always have the welding ground/neutral/negative lead attached closely to the area where the welding is being done--especially if there any sensitive electronics involved. ESPECIALLY.

And that's that.
I don't get why any engineered product would ground an RTD lead. The beauty of an RTD is that it be inherently isolated, because the measurement relies on a constant current 'excitation' source so that the resulting voltage drop can be 'read'. The resistance wire, which is wound element around a ceramic core, is insulated from the sheath with mineral oxide (MgO) to ensure isolation. Grounding a lead introduces the extremely high probability of a ground loop which could add or subtract current to measurement circuit. The same issue applies to a 'grounded thermocouple' which has a faster response than an ungrounded junction, but is as likely to be subject to ground loop errors. I find the concept strange. One finds by-the-seat-of-the-pants controls cobbled together with all sorts of strange wiring, but the companies making turbines and generators have a lot resources and experience to put into design.

I do like the photos. Now I have some idea what these beasts look like.

Soooo, there's a reason GE grounds the compensation leg of 3- or 4-wire RTDs embedded in the generator stator windings and cooling gas temperature measurements. And, it's pretty simple....

Three-conductor, twisted, shielded cables (sometimes called ”triad cables”) cost more than two-conductor, twisted, shielded pairs (sometimes called ”TSP-” or ”Belden cables). Since the length of the cables/conductors is virtually identical (especially when the compensation leads are grounded in the same junction box on the generator), it's only necessary to use two or three conductors as the compensation wires between the Mark* and the generator junction box or the Generator Control- or Protection Panel and the generator junction box. The compensation inputs to the monitoring circuits can then be jumpered together and only two or three conductors from the generator junction box need to be used for the compensation lead measurement.

It saved wire and terminations. And because that's how the original interconnection diagrams were drawn, GE used the practice for decades.

It frustrates most people who have never seen it done like that before, myself included the first time I encountered it. I just couldn't understand how it worked (because it wasn't working when I first saw it!--because the generator factory personnel never connected the compensation leads to ground, it was left to Field Engineering personnel to tell the electricians to do it, but nobody told the Field Engineers it needed to be done....)

In today's industrial world with Protective Earth and Functional Earth, I believe this practice is outdated and responsible for many nuisance and intermittent turbine control system problems and even failures, especially if the site is prone to lightning strikes and electrical storms.

Anyway, the bottom photo shows a GE-design Frame 5 heavy duty simple cycle gas turbine-generator. These machines are nominally rated at around 20-25 MW (depending on age). The gas turbine exhaust is usually around 1040-1100 deg F depending on ambient conditions and the age of the machine. A simple cycle machine exhausts to atmosphere (that's the large vertical duct/structure in the middle of the photo). So, all that exhaust heat is lost and gone forever (when it could be used to heat water in a Heat Recovery Steam Generator, HRSG) to drive a steam turbine and possibly the steam, at a lower pressure, could be used for a process, like drying paper stock at a paper mill, or powering a line drive turbine at a paper mill or drying vegetables to be used in spices and food production or heating large vats of soup or beer).

The generator is shown at the left of the photo, with a small exciter on the end of the generator shaft (probably brushless). The generator probably turns at 3000 RPM (50 Hz), which is derived from a Reduction- or Load Gearbox assembly that changes the turbine speed from approximately 5100 RPM (which is the optimum speed for the axial compressor, and the turbine section).

The generator is air-cooled, with air being drawn in through filters on the other side of the generator by a fan on each end of the generator shaft which blows the air through the generator and it also exhausts back to atmosphere through the right-angle duct on top of the generator.

The axial compressor inlet air filter system is at the far right of the picture, and what appear to be the ends of dust-and rain hoods can be partially seen on the far end of the large white vertical filter house structure.

They are pretty impressive machines. The output shaft of this type of unit passes through the exhaust diffuser, to the Load Gear, which is coupled to the generator shaft. These types of machines are called hot-end drive units. Newer designs have the turbine shaft directly coupled to the generator at the axial compressor end of the turbine--or, the "cold” end of the machine (a cold-end drive). These newer machines can produce as much as 400 MW--by itself! In combined cycle application, they can produce enough steam with the gas turbine exhaust heat to produce another 200+ MW. Very impressive machines with very high efficiencies (as much as 60+%), compared to coal-fired power plants which could never run at more than 35+% efficiency even when they were brand new.

And the emissions are so much lower for these new machines than any coal- or gas-fired steam plant even with state-of-the-art emissions reduction systems (which are extremely expensive to install, operate and maintain--and which usually reduce overall plant thermal efficiency).

Anyway, the unit in the bottom photo is one of the most popular and ubiquitous around the world. They are pretty economical to build and operate (though the simple-cycle units are only about 35% efficient), can be built fairly quickly, can burn gaseous and liquid fuels (some units burn very lightly treated crude oil practically right out of the ground; others burn the very dregs of oil refinery leftovers (called, rightfully so, residual fuels which if spilled on your leather shoes will eat a hole through the leather)). Some units burn naphtha, another nasty liquid. Some units can even burn coke gases from steel mills--these are very versatile machines in this respect.

But, GE-design Frame 5 heavy duty gas turbines are very robust machines that can--and have--withstood untold abuse and misuse and heavy use--all ”in a day”s work” as they say. I've seen Frame 5s shutdown for normal maintenance only to find half of the length of nearly every second-stage turbine bucket (blade) missing, with less than 3% performance degradation before the outage and only a very slight increase in machine vibration levels. Truly, they are simply workhorses built with significant safety and performance margins (something that can't really be said about some recent advanced gas turbine designs).

Hope this helps!!!