Does anyone know the vibration limts for a Carrier 17DA centrifugal chiller? The end user has alarms set at 1.5 mils pp which seems unusually tight for a 4500 RPM machine.

I can't claim to know Carriers limits, but 1.5 Mils at 4500RPM is around 0.35 in/sec. That amplitude is NOT something I would like to see on new or newly rebuilt chiller. In my opinion, 1.5 Mils should be no problem to meet.

Take Care,

Chris, not sure I totally agree with the concept of converting to velocity and then applying the criteria. API 617 allows the sq. root of (12000/N) where N is RPM. That is for a "new" machine and represents the typical acceptance criteria used at the OEM test stand. That would be approximately 1.6 mils pp.

I agree with Chris. I'd like to see about half that level for normally operating machine, so 1.5 mils pp is a pretty good alarm limit.

I would question monitoring displacement for alarms as it totally ignores all the high frequency components of the vibration signature, such as the blade rates of the compressor stages.

Of course, with vibration guys: if you lock 10 of them in a room, they'll come out with 15 opinions!

Jon
Spintelligent Labs

John,

The responses sound like you entered the wrong forum. As you indicated, shaft relative vibration is not the same as casing vibration.

Example: 2 mils pp on a 3600 rpm generator would be approximately 0.38 ips 0p in velocity. Would this machine be in danger? On your typical generator this would be good.

I think the bearing dimensions, geometry (Some may have a partial arc bearing.) and loading would play a part in the alarm setting. Any criticals below running speed? In general API is shop acceptance levels and not shutdown levels. Generally, one expects the shutdown levels to be greater than the acceptance levels. Does the OEM manuals have a recommendation?

What blade pass tells us most of the time is that the compressor is operating off design, which can be determined more directly. Other issues can certainly influence this, but the typical failure modes are best monitored using shaft relative on most compressors. If you have a flexible foundation like the axial compressor on most gas turbines, one should look at both seismic and shaft relative. John you understand this, but not everyone does.

Bill, didn't realize that "vibration/alignment/balance" limited one to rolling element bearings. But perhaps I should have been more clear up front that this is a proximity probe measurement on a bearing with about 9 mils diametral clearance.

John from PA (specifically Media PA)

P. S. How you doing by the way?

Without having a stake in this or without getting a steak out of this, I can offer the following comments. I trust that they are worth the price (free). Of course they come with no warrantee.

Your company has a part of it that makes compressors. They may have a design practice that gives guidance with regard to alarm points. You could investigate this.

It sounds like around 2 X API would make a good alarm point. Shutdown or danger might be 2.5 X or even 3X API; at the high alarm setting one should not operate for extended periods. However, this could be modified by several factors. Since the speed is 4500, does this have a gear that loads a bearing significantly? The bearings are not partial arc bearings are they?

How do operations treat alarms? Does an alarm mean that action is required and will be scheduled? Is the first alarm only a nuisance waiting for a shutdown or high high alarm?

Is the equipment spared? Risk tolerance for operations would include the value of the product if this unit goes out of service. How easy is it to get spares?

What is the delay on the alarms? Any voting for shutdown? These combine with the risk tolerance investigation. Modify as needed.