Turbine Cold Start Instrumentation

[kayaker1]

When warming the turbines, cold start or warm, we have manufacturers recommended procedures for temperature soak times. Sometimes these are longer than necessary. Does anyone know of other methods, perhaps monitor Differential Expansion or strain gauges, to speed up startup?

Allis CHalmers 1960's impulse reaction cross compound type 370MW.

 

[ilan7]

Slaw heating the turbine is very important.
I would recommand to strickly follow the manufacturer instructions to prevent an un-expected failures.
If you need to buy some start up time - consulte the manufacturer. I work for a small turbines manufacturer and do not like to see an operator changing the operation instructions without getting permition to do it.

 

[byrdj]

My experance is with GE LSGT. Which pressurizes the HP section while on turning gear to achieve most of the warming, then speed holds sub synchronis for RH/LP sections.

The importance of prewarming and speed soak holds is to allow the rotor bore tempertures to exceed the FTT (transiton from brittle to ductile). The manufacture curves provided with older units that relied on operators recording a TC and keeping up with time has and can be updated with cumputer programs that calculate the bore tempurature. The advantage to this programs is they can take into account the amount of warming that was achieved prior to reaching soak temp and any elevation in soak. The new units Say since mid 80s introduced such a computer called RSI (rotor stress indicator). I have seen a few utilites write thier own, but one must be carefull. I've seen the utilites program give the OK to roll way too soon.

While the manufactor i worked with thought the times given in the cold soak (For beginning 70F turbine, TG pressurized soak at 300F for 4 hours) were more than aduaquate to achieve HP rotor bore FTT, It did not allow sufficient time for shell expansion, and rotor long and rubbs were normally encountered.

Therefore units that I am resopsible to restart following an outage, besides being sure that the soak times are meet, I also verify that Shell expansion has leveled out and agrees with my expected expansion for the soak temp

 

[byrdj]

In order to determine if your starting and loading (SAL) procedure can be enhance, your existing procedure and instrumentation needs to be defined. For example, the vintage of your unit may not have had the essential instrumentation at installation in the 1960s.

I have no experience with AC (I have walked by one). For this vintage GE, without provisions for pre-warming nor the needed TCs to monitor, the SAL instructions was three timelines for cold, warm, and hot, with the definition for each being duration off line. Warming was achieved by slow accelerations (60rpm/m) and a couple sub synchronize speed holds. This method did not provide sufficient warming and a lot of failures occurred.

Most of this vintage was updated with changes to the control system to allow pre warming prior to rolling. The SAL instructions were change to a series of flow charts with metal temperatures from TCs determining when warming was sufficient for varies steps in the SAL ramp. However, the addition of the essential TCs to monitor was not always completed.

As stated in the previous post, the next evolution was to use a computer to monitor the TCs and control the SAL ramp.

Another consideration in reducing the SAL ramp times would be how many such ramps will be made. All temperature changes are cyclic stress on the turbine. If there will only be a few cold starts (once per year), then the stress (temperature ramp rates) can be higher. If the there will be many warm starts, then the stress should be kept lower.

The condition of the shell and rotors should also be a factor in SAL procedure. Turbine steel has a tendency to become brittle with age. The shells and rotors may have minor cracks that higher SAL stress will propagate.

I would think the Company that took over for AC (

http://www.regencoservices.com)

would be a good source to assist you. However, there are many turbine consultant companies that would also (like the one I work for).

A word of warning, there are companies that promote a rapid cold startup procedure that will cause damage.

 

[kayaker1]

Thanks for the advice.
We are looking at ways to improve our availability. Current cold start procedures are 48hrs. I have been told there are ways to shorten start times and didn't know how this could be accomplished.

Currently we allow the steam chest to warm for a minimum of 1.5 hrs. and until the chest is 150F of main steam temperature. Steam chest > 185F.

HP & IP turbine inner cylinder temperature at 190 f minimum.

Increase speed to 600 RPM in 10 minutes.

Maintain 600 RPM for 30 minutes.

Increase HP turbine speed to 2500 RPM in 60 minutes.

NOTE: MAINTAIN 2500 RPM UNTIL HP & IP (TOP & BOTTOM) INNER CYLINDER TEMP REACH 400F.

 

[byrdj]

If I make an anologe to GE, It sounds like you have the original configuaration, ie no provisions for prewarming before rolling.

Again not familair with all the intricaces of AC, but you should be able to pressurize the HP and reheater boiler section and achieve HP inner greater than 350F before rolling.

I believe you could benifit from contacting a consulting company

 

[byrdj]

Making a comparison of GE modern and your current cold start up times (I am not an operator and it has been other a year since involve with cold start, so not to detailed)

GE post 1970
Choose a temperature Ramp rates:
200F/hr max stress (2,000 cycle life)
130F/hr normal (10,000 cycle life)

Soak time on turning gear with HP turbine pressurized with 65 psig steam (saturation temp 300F). HP metal temperature prior to any heating
Less than 100F-4hours
Less than 200F-3hours
Less than 250F-2 hours
Less than 275F-1 hour
Greater than 300F none

Acceleration rates
>100F 1st stage to throttled steam mismatch MS conditions 180rpm/m
>300F (hot boiler)………………………………………….120rpm/m
<200…………………………………………………………360rpm/m

Hold 3000 rpm
Till Xover (IP exhaust) > 200F for 60 minutes

Hold 3600 no load 30 minutes

Hold 2% load till Xover > 350 for 60 minutes

The HP and IP temperatures ramped from roll to minimum load 300 to 700, about 400F, thus the fastest roll time with 130F/hr ramp should be 3 hours, plus 4 hours pre-warm.

Load ramp not to exceed temperature ramp rate. Expected low load to full load total temperature change to be 300F, thus load ramp about 1%/min (about 2 hours to full load)



Yours (similar to GE pre 1970s)
Steam chest >185 for 1.5 hours
MS – SC < 150F
HP&IP>200F

60 rpm/m (10 minutes)
Hold 600 rpm for 30 minutes
30 rpm/m (60 minutes)
Hold 2500 rpm till HP&IP > 400F

Same expected low load HP temperature of 700F
Rolling with 200F turbine to minimum load at 700F would be a 500f change, thus could take just under 4 hours

After review your timeline for cold turbine start, the duration sounds as quick as the “modern TG pre-warming” I’m familiar with.

A couple questions
What type of shaft seals does the unit have, i.e. gland sealing steam or water seals? If water seals, then, TG pre-warming by pressurizing won’t work.

Is the main steam conditions limited during roll to slow down the temperature ramp rate?
After HP&IP TC are > 400F, can the unit be accelerated to rated and loaded, or is there a “no load, rated speed” hold and a minimum load hold (like hold 2% load for another TC to exceed) before load ramping.

While your stated concern was to reduce the cold start time, you must still be concerned with the cyclic stresses being applied. GE did not believe your type of start warmed the HP rotor enough for the metal to transition before the centrifugal stresses at 3000 rpm were applied.

There would be a lot of other questions that some one more familiar with AC should be asking if you are planning on using a unit designed for base loading to now be an occasional peaker

Does your admission valve arrangement allow for 360 degree steam flow in the first stage nozzle, or are you controlling speed with only a partial arc segment, like with a control (governor) valve? Has the nozzle gap to bucket distance been increased to lessen once per rev wheel excitation at low load, especially if rolling on an arc segment?

While I didn’t answer your question, I would hope these comments will be helpful.

 

[kayaker1]

>>What type of shaft seals does the unit have, i.e. gland sealing steam or water seals?

Gland Seals.

>>Is the main steam conditions limited during roll to slow down the temperature ramp rate?

Yes the boiler may have a few hundred pounds of pressure (2300 psig normally). If there is pressure in the boiler, the stop valves allow steam to warm but not enough to roll the unit off turning gear.

>>After HP&IP TC are > 400F, can the unit be accelerated to rated and loaded, or is there a “no load, rated speed” hold and a minimum load hold (like hold 2% load for another TC to exceed) before load ramping.

After HP&IP inner cylinder reach 400F, the speed is kept at 2500 rpm a minimum of 2 hrs to obtain expected transition temperatures of 250F at the shaft bore.

>>Does your admission valve arrangement allow for 360 degree steam flow in the first stage nozzle, or are you controlling speed with only a partial arc segment, like with a control (governor) valve?

We have full arc control.

>>Has the nozzle gap to bucket distance been increased to lessen once per rev wheel excitation at low load, especially if rolling on an arc segment?

I'm not sure I know what you mean by this. Can you explain further?

Thanks for the comments.

 

[byrdj]

1st stage nozzle gap

If a unit is design to be a base load, the wheel tip axial clearance (nozzle to bucket) is close for efficency, Assuming 4 CVs, all opening sequentially and admit steam in a 90 degree segment of the nozzle. If loaded with only one CV open the bending steam force will applied to the row of buckets for 90 degrees, then relaxed for 270, giving a once per rev bending. Peakers will open CV1 and 2 together and have the clearance from nozzle to bucket increased, lossing stage efficency, but not stressing the 1st stage buckets as much.

A more radical change would be to allow a CV supply to the entire nozzle

The design of 1st stage bucket and wheel fit is more difficult due to this

 

[byrdj]

 

[marada93]

Geeze...48 hours?

That seems extremely excessive. The heat soak hold speeds and times seem appropriate. What are the durations of load holds?

Of course differential expansion is the key. This would tell you if you are warming up the rotor faster than the shell.

Is the 48 hour cold start straight from the O&M manual?