Turbine journal run out

We have 1000 Mwe turbine. Due to moisture ingress into lub oil and subsequent bearing damage, which also left marks on journal area, the journals were machined. A run out of 0.05 on the journal area is what we have achieved after grinding. What will be the consequences of high journal run out and what is the permitted run out value?

Original Post

A runout if 0.05 mm is approximately 0.002 inch or 2 mils. Since this a power generation turbine I would assume that the max speed is 3600 rpm.

 

API 670 provides guidelines for the limits for shaft runout at the probe track area if that is what you are referring to.

There is also an API Best Practice (API 684 please check) for machinery rotor repair that may help you also.

In my world, the runout limit at the bearing location is 2 mils. This could be a "simplification" if you are a detailed person. The DTI collects many imperfections in its reading. Some of these factors may not be due to the journal repair you mentioned.

Regards- Ali M. Al-Shurafa

I see a pertinent detail, in that you say this is a 1000 Mw (1,341,022 HP) machine so this is likely a nuclear steam turbine, and at 1500 RPM.  Rotor diameters at the target areas therefore are likely to be quite large, perhaps 700 to 800 mm.   I would therefore not be too overly concerned with 0.05 mm runout.

What is the allowable vibration at the journal in question?

 

I believe it is excessive for noise.  I would not guess the application.  You might consider adjusting your alarm thresholds - depends upon what they are now.

Is this mechanical runout? 

Are the journals centerlines the same as before?  If the journals are 0.05 mm off center then you would have a large imbalance to take care of.

Did  you de-Gauss after grinding?  Not only can this leave magnetic runout (noise) for the probes, but this can affect machinery reliability.  Did you check for residual magnetism?  Did you burnish?

Note:  I am surprised that your company, Ali, allows 2 mils runout.  One of your company members is working on the API 670 update, which is in progress.  The usual limit (simplified) is 0.25 mils pp.

Bill,

The 0.25 mils is also used as a limit.The limits we use (in the operating facilities not for new machines) actually vary based on the case. That is why I said the 2 mils is a simplified limit. So, no more is acceptable regardless of the reasons. But, as you know crazy things happen and you end up with shafts with high runout at the site.

Now, the OP actually did not mention the probe track area, what he mentioned is the journal at bearing location. They could have different runout conditions.

With respect to API 670, if you recall it states something like "what ever greater". So, if the radial vibration setpoint is 4 mils, the limit of 0.25 mil will not hold. In my personal opinion, the API should have a separate limit for runout regardless of the vibration setpoints.

I'm interested to hear from you and the rest about your practice on the machines under repair.

Regards- Ali M. Al-Shurafa 

William_C._Foiles posted:

I believe it is excessive for noise.  I would not guess the application.  You might consider adjusting your alarm thresholds - depends upon what they are now.

Is this mechanical runout? 

Are the journals centerlines the same as before?  If the journals are 0.05 mm off center then you would have a large imbalance to take care of.

Did  you de-Gauss after grinding?  Not only can this leave magnetic runout (noise) for the probes, but this can affect machinery reliability.  Did you check for residual magnetism?  Did you burnish?

Note:  I am surprised that your company, Ali, allows 2 mils runout.  One of your company members is working on the API 670 update, which is in progress.  The usual limit (simplified) is 0.25 mils pp.

Bill, while much of your comment is valid and deserves consideration, I would not think this a guess as to application.  If the OP has correctly stated that the machine is a "1000 Mwe turbine", then it is almost certain to be a nuclear power generation machine.  The equivalent of 1000 Mwe is over 1 million horsepower!  Further, and based on the OP profile that he resides in India, the machine is in 50 cycle service, and likely operates at 1500 RPM.  Lots of torque, large journal diameters, large clearances as a result.

Accepting the physical/electrical size of this machine, I think it a stretch to apply the limits of API.  For example the overall limit of vibration is specified as (12000/N)1/2 or 2.8 mils (0.07mm).  An old GE EM1534 states that under steady state conditions,  high load 6 mils is considered satisfactory in a machine of this probable type.

API 612, the special purpose steam turbine specification states for proximity probe target areas "These areas shall be properly demagnetized to the levels specified in API 670 or otherwise treated so that the combined total electrical and
mechanical runout does not exceed the following.

"For areas to be observed by radial-vibration probes, 25 % of the allowed peak-to-peak vibration amplitude or 6.3 μm (0.25 mil), whichever is greater."

API addresses "new" machines and this would be almost 0.04 mm (1.5 mil) as compared to the OP stated 0.05 mm.  This is (IMO) marginally acceptable for an older machine.

 

John,

Our former employer made this size steam turbine and larger.  They could hold (and would sign up for it on the radial surface - not thrust) 0.5 mil pp, which for the larger machines is practical.  The GT guys would not sign up for this. 

My remark about 0.25 mils was addressed to a fellow API member, where this is a common requirement, not 2 mils as mentioned.  The 0.25 mils should be possible and practical on most API machines; shaft material can be an issue.  One time Ali's company did some work (I participated in this.) to see if metal (powered metal) spray techniques could be used to help with this and other issues.).  

0.25 mil on a large turbine can be difficult.  I have seen other manufacturers who would not agree to 0.5 mil as a target.  The owner of the company to which you worked prior to its acquisition, fought hard for low runout levels with good cause.  Nonetheless 2 mils is too large.

Fossil units can be  > 1000 MW, too.  ISO addresses the acceptable vibration limits for 1500, 1800, 3000, and 3600 large turbines.  2 mils eats into what is acceptable and reduces the monitoring capability.  A 1000 MW turbine is a valuable asset. 

---------------------

Ali,

If I recall correctly, API older editions had whichever is less - leaving 0.25 mil as the maximum.  My company keeps that provision internally.  For high speed compressors, one cannot always meet the formula - say a compressor running in excess of 25,000 rpm.  One may need to compromise.  However, your alarm levels will be low as well, and runout affects the reliability of the monitoring/alarm functions.  

I assumed, as should be the case, that the probe track area is very close to the journals and that they would have been affected similarly.

Runout in the journal can be a mechanical issue.  If electrical in the sense of causing prox probe runout, who would care unless there is magnetism. 

An interesting case would be that the journals were machined to a 1 mil center offset.  The amount of imbalance cause by changing the center of rotation would be enormous.

Large turbines are not going to have an API  shop acceptance level vibration as the field acceptable vibration value. The acceptable levels will be far greater, and this has been demonstrated to be valid (exceptions possible).

Without looking up ISO documents, as I recall, a 3600 rpm machine would have acceptable vibration of 5.9 mils pp, and I believe you are correct to say that GE used 6 mils pp.  I know that there is experience with manufacturers of large turbines of greater vibration sustained with no failures reported, but one should use caution - bearing loading is important

Should the turbine be 1800 rpm or 1500 rpm, as in the case of a nuclear application,  the acceptable vibration would be larger, yet.  Would you want to have to reduce your vibration limits acceptable to running, just because of the noise?  Would a plant do this for a 1000 MW machine?

=============

In balancing, one question that should be addressed is whether to use prox probe runout in the balance or not.  Technically it is better to include the use of runout (subtract it from the 1X) - runout can change, especially something like residual magnetism and beware of temporary bows and their runout. The customer only sees the apparent vibration, which includes the runout noise.  So, what is done often - usually even - is to ignore the runout; this will work well if the runout is small, not so well for large runout.  Similar concerns exist around monitoring.

 

William_C._Foiles posted:

 

Ali,

If I recall correctly, API older editions had whichever is less - leaving 0.25 mil as the maximum.

Your comment on the 0.25 mil limit made me confused. The editions I saw states whichever is greater. But, you know I'm much younger than you and maybe you are referring very old editions.

For those who are interested in the exact wording of this paragraph from the standard, here it is.

API STANDARD 670: FIFTH EDITION, NOVEMBER 2014

"6.1.2.2 The surface areas to be observed by the probes (probe areas):

.....

d) shall be properly demagnetized or otherwise treated so that the combined total electrical and mechanical runout does not exceed 25 % of the maximum allowed peak-to-peak vibration amplitude or 6 μm (0.25 mil), whichever is greater."

 

Regards- Ali M. Al-Shurafa

Shurafa posted:
William_C._Foiles posted:

 

Ali,

If I recall correctly, API older editions had whichever is less - leaving 0.25 mil as the maximum.

Your comment on the 0.25 mil limit made me confused. The editions I saw states whichever is greater. But, you know I'm much younger than you and maybe you are referring very old editions.

For those who are interested in the exact wording of this paragraph from the standard, here it is.

API STANDARD 670: FIFTH EDITION, NOVEMBER 2014

"6.1.2.2 The surface areas to be observed by the probes (probe areas):

.....

d) shall be properly demagnetized or otherwise treated so that the combined total electrical and mechanical runout does not exceed 25 % of the maximum allowed peak-to-peak vibration amplitude or 6 μm (0.25 mil), whichever is greater."

 

Regards- Ali M. Al-Shurafa

At odds with that is the common requirement in many of the API "Special Purpose" specifications that state the vibration "...shall not exceed the mechanical test vibration limit of 25.4 μm (1.0 mil) or Equation (5), whichever is less".  Equation 5 being the typical API that involves (12000/N)1/2.  

Keep in mind this applies to the mechanical test.not necessarily the field operation.

 

The 1 mil pp did not used to be in older versions as I recall.  I have not been in that debate, but one could ask, what is needed.  Some of the larger compressors may be running at 3600 rpm or even 3000 rpm in API world.  

Residual magnetism can harm more than the probe measurement.  Grinding can leave residual magnetism, and this should be checked and corrected as needed.

As far as standard's language goes, 'properly demagnetized' is not very good.  It is not measurable. One needs a limit to which to judge - I thought I saw a limit recently with which to measure.  

Add Reply

Likes (0)
×
×
×
×