Steam turbine speed sensor spikes while at 3 rpm. 4

[hlh1]

We have a steam turbine/generator set that operates at 3600rpm. No problems when operating. When off line and on turning gear at 3 rpm we sometimes get false spikes to 350 rpm. This is not real of course. I have six speed probes in the front standard. They are connected back to an OEM turbine control system. All of the probes are produced by Honeywell, we have used these models since 2002, no real problem until this has occurred. These probes are VRS type passive probes. Sometimes this spiking will throw off or block our startup sequence. It appears and disappears only when at 3 rpm with a medium warm turbine.
I have replaced sensors and tested the used ones, all tested good. I pulled in new cables, checked grounding and shielding. Checked air gaps! Made sure they were at OEM recommendations.

These probes are in our front standard portion of the turbine. The lid to this area covers the number one bearing. It takes three or four days to cool down enough to take out unit off oil to the bearings.
So it is not easy to get access to these probes. Our six probes are divided into two groups of three. The first group is for speed control and is wired back to a control function circuit board. The second group of probes are used for protection monitoring and are wired to a overspeed protection system. Seperate control boards!!

And yet all six sensors show the same spikes. Sometimes the spikes are one minute apart and sometimes they are about 7 seconds apart. And sometimes they just quit for no apparent reason.
These probes are aimed at a gear wheel with 60 teeth. 180 pulses a min at 3 rpm. This problem only happens when on turning gear at 3 RPM.
The OEM has been unsuccessful in solving our problem at this time. Lost in Midwest.

 

[3DDave]

Sounds like a software problem that might have a hardware interaction; like the rate is high enough to be detected as a valid rate, but there is a timeout that is allowing double sensing of a single sensor and it's calculating that this must be the next tooth over some tiny interval. I bet that it is very sensitive - you say 3 rpm, but is is actually 3.00000000000000 rpm or is it 2.98 rpm some times and 3.02 other times? That may be the cause of the variability in occurrence.

I think this will require putting an oscilloscope on the sensor outputs to see how the sensors are reporting. You need to look for transitions in state for all 3 sensors on each control board. If the sensors are doing the thing they should be doing, then it's off to the controller software.

The fact they all do this at the same time leads me to believe it is software, unless these boards are all different designs.

Fun software diversion - was at a company that had a complex controller board. Almost always worked great except when it didn't. Looked for defective components. Found none. Looked for defective soldering. Found none. Chased it back and forth and finally called back a retired guy. He looked at the microprocessor programming guide and found that the default was for the microprocessor to wait exactly 0 clock cycles before reading data. Sometimes it took more than 0 clock cycles for the data bus to present stable values. Programming guide said - use a minimum of 2 or 3. Voila! Set the correct wait time and controller board was flawless.

 

[LittleInch]

A good point - a turndown of 1000:1 is challenging for anything and you might find that the internal clock speed of the sensor or the counting mechanism just can't manage the low speed without the occasional glitch. If it's sensing at full speed 216000 pulses a minute or 3,600 a second and now its only doing 3, it might just record one pulse less than the number of cycles by one or two and then calcualte wrongly.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[hlh1]

Thank you for your help. The OEM has repeatedly referred to the air gap of the sensors to the toothed wheel. I will follow up with them as to software sensitivity for turn down on pulse monitoring. My O-scope died a few years ago and because of lack of use I have not replaced.
Well I guess you know what I am going out to buy or rent. hahaha!!! We had a software up grade a few years back. We had an OEM engineer here for that and he may have changed some settings. No accusations here just info. We had this just every now and then before the up grade. No I/O cards
changed. Just a software tune up. Again no accusations. I will be doing a turbine outage in three weeks, you bet I will be playing in this area deeply. And I will have a OEM software engineer right by my side.

Again thank you to you guys, I do appreciate your input. Live long and prosper!!

 

[JJPellin]

What is the tooth profile. In my experience, turbine manufacturers make toothed wheels with a square tooth design. The manufacturers of the governors and overspeed trip systems expect a gear tooth profile. This makes the system extremely sensitive to air gap.

Johnny Pellin

 

[hlh1]

Our Mechanical department have no drawings of said wheel. I will find out during this upcoming outage. I will request this info from OEM also. This is the same "gear" we have had on our turbine since the 2002 spring overhaul. We did not have this problem at that time that I am aware of. We did not start having any problems until after 2015, and then only intermittently until 2020. Then it started visiting more frequently. Thank you. I will use this idea, along with the others offered. The OEM may have these drawings, however they are having trouble finding other requested prints for our upcoming outage. So who knows.

 

[IRstuff]

3 rpm is pretty slow; perhaps there's some mechanical issue with, say, the bearing lubrication, that's causing a power spike that affects the sensors. As the turbine/generator spins faster, the issue is masked by, say, sufficient lubricity.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[JJPellin]

There is a setting within the governor / tachometer / OST system for the probes that are looking at that toothed wheel. I think it might be called gain. We changed this setting and altered the probe air gap to solve a buggy speed indication problem on a large steam turbine. All square toothed gears are not the same. A very slight chamfer on the corners makes a large difference. As noted above, looking at the signal on an oscilloscope will show the problem clearly.

Johnny Pellin