Mech'l Running Test of multistage Cent Comp as per API 617 1

[CaracasEC]

Hi All, one objective of mechanical running test is to verify oil flow, friction horsepower, and heat rejection to the lube oil.Since the test is carried out under vacuum conditions, resulting in very low aerodynamic thrust loads. It should be noted that thrust friction horsepower constitutes the greatest bearing loss and, therefore how is the calculation from the thrust to full design thrust? Thnks in advance...

 

[jbattershell]

CaracasEC,
first, I am not aware of many tests being conducted on a centrifugal compressor at vacuum pressure. are you positive on this aspect of the test?

most compressors are aerodynamically tested and mechanically tested, often both occuring at the same time to speed up the process. during the aerodynamic test, all mechanical properties are also being monitored, especially while the compressor is under load. in fact, a mechanincal test is not valic unless the compressor is loaded, near design power.

also, while thrust bearings do contribute a fair share to the overall mechanical losses, they are not the majority of losses. each bearing has a significant horsepower loss and if the compressor is a geared unit, the gearing losses and bull gear windage losses are significant also.

if the compressor OEM is claiming that the mechanical test will be conducted while the compressor is under vacuum conditions, notifying them that a true mechanical test must be done while the unit is loaded.

regards,
John

 

[CaracasEC]

According to the Mechanical Running Test Procedure by the OEM, compressor will be tested under vacuum, i'll ask your question prior to our Mechanical Running Test. API 617 standard on which the compressor will being purchase under this international code has not stated whether mechanical running test will be done under vacuum or not. As you mention that a mechanical test is not valid unless the compressor is loaded, near design power? what part of the code that i can find that paragraph.thanks for the sharing...

 

[pmover]

jbattershell,

fyi & clarification, centrifugal compressor mechanical tests are conducted with a vacuum as stated by the OP.

"if the compressor OEM is claiming that the mechanical test will be conducted while the compressor is under vacuum conditions, notifying them that a true mechanical test must be done while the unit is loaded."

does not make much sense . . .

the performance test is conducted with either the actual gas or a test gas (type I or II) per ASME PTC-10; hence, the performance test is under a "load". not aware of any mechanical tests conducted "under a load". the two tests are separate and distinct.

-pmover

 

[CaracasEC]

Hi primemover,
As per OEM procedure, type two test will be conducted as per asme ptc 10-1997 (performance test) while mechanical running test will be as per api 617. However, the code does not elaborate whether mechanical running test will be conducted on vacuum or not. How is your actual experience on mechanical running test?

 

[centeng11]

Mechanical run tests are not required to be carried out under vacuum. Mechanical run tests in a vacuum can create problems, for example, the operation of oil-film type seals cannot be tested (if applicable), the lack of significant gas density can influence rotor dynamics, and even the casing can heat up abnormally from the churning of residual gas (given the vacuum is not perfect).

 

[jbattershell]

CaracasEC, you are right, the API-617 code does not specifically state that the mechanical test shall be conducted under load. However, section 4.3.8.6 does specify that a full pressure, full load test can be conducted by the OEM and that this test procedure should be agreed upon between the vendor and the customer. as this paragraph is marked with a dot, it is extra and not manadatory.

But I will state this: if i were the customer, I would surely want my compressor's mechanical performance to be evaluated WHILE the compressor is LOADED and at or near design pressure. this test can be conducted with a test gas like N2 or it can be done with a hydrocarbon gas. during the mechanical testing, the seals performance is evaluated and flow of the seal vent line is monitored. vibration is monitored and axial displacement is monitored. if all of these parameters are monitored and measured while the compressor is under a vacuum and hence, under very low loads, i would not believe the compressor is really being tested and proven to be fit for function. how would i know with some degree of certainty that the compressor will run good in the field at full load conditions if it isn't verified during the factory test?

unless the contract states that the compressor is to be tested under full load, then the OEM can do a test at part load. But for my compressor, I would not be satisfied with that test. many things ("gremlins") things, tend to show up only if the compressor is fully loaded. seals are not being tested much if the sealing pressure is quite low. axial displacement not being checked unless near design thrust loads are being imposed on the thrust bearings.

so, while API 617 doesn't force the compressor mechanical test to be done loaded, I am not off base with wanting to test the mechanical properties of a compressor while loaded. an integrally geared compressor may warrant a loaded test more so than a single shaft machine since the gearing, once loaded, exhibits radial loads on the bearings that would otherwise not be present unloaded.

for the majority of my career thus far with integrally geared compressors, the OEM would perform the mechanical test at the same time the performance test was being conducted. therefore, parameters were measured and checked while the compressor was loaded at or near design conditions. and in addition, the compressor was unloaded to check vibration at that point as well. occasionally, compressors would be fussy while unloaded and that was checked.

the testing procedures with other OEM's can certainly be different. and typically, the test procedures are sent to the customer 6 weeks or so prior to testing for the client to review. at that time, deviations or alterations to the test procedure can be discussed.

 

[CaracasEC]

Hi Jbattershell, thanks to your informative comments. We recently have finished already performing the Performance Test and Mechanical Running Test. Performance Test result shows that as per ASME PTC 10-1997 the test is successful. As per Mechanical Running Test as per API 617 what the OEM has done as per contract is running the compressor at vacuum. Vibration was good, dry gas seal leak test has been performed for reference as it was already tested at the vendors shop, oil was varied from minimum temperature and maximum temperature including minimum and maximum pressure for the bearing performance. I could say that the Mechanical Running Test as per API 617 is successful. What we will conducting next on the compressor is the gas leak test and dismantling of bearing. Rotor will not be dismantled as no abnormal sound has been noted during the test./.

 

[ortabe]

The purpose of a mech running test is not to test the mechanical losses. It would be a meaningless number since "vacuum" will actually be ~0.1 bar causing the compressor to do a few (10-100) kW of gas power.

Varying the oil temperature is done mainly to demonstrate rotordynamic stability over a given temperature range and here the radial bearings are relevant, not so much the thrust bearing. I've never heard of it in relation to "pressure for bearing performance".

 

[ortabe]

Btw, while centrifugal gas compressors are typically mechanically tested at vacuum conditions, I'll have to agree with jbattershell because things like load and density sometimes have profound effects of the rotordynamics.

 

[CaracasEC]

Thanks Ortabe. Yes the compressor will not experience aerodynamic thrust during the mechanical running test. Thanks for the informative sharing..

 

[turb1]

Jbattershell, I agree that a test run on the proper MW gas at full power and full pressure is the most optimum test since it is the most complete test of everything already mentioned (aerodynamic forces on the rotor, etc).

Working for a centrifugal compressor OEM, we rarely test this way due to cost. If a customer is willing to absorb the additional testing cost (which in some cases is as much as the capitol cost of the piece of equipment) then we will do it.

When we design, we base it on experience of bearing/rotor combinations that we have had successful field operation with. Every now and then we have to step outside of these bounds and we sometimes have problems. This is the reality for all rotating equipment manufacturers and in the end there needs to be some level of trust between the vendor and the purchaser. It is not in the OEM's best interest to put a piece of equipment in the field that is going to have warranty issues.

In my experience the times when you have issues is when you are forced to step outside the normal bounds to reduce cost (ie less stages, smaller frame, etc).

 

[raphatss]

As mentioned by turb1, a PTC type 1 or a full power-full pressure test is very expensive. I've seen OEM that don't even have capacity in their test bench to do these kind of tests depending on the equipment, due to limitations on power, heat exchangers...sometimes hydrocarbon is not permitted...

jbattershell,
Those compressors that you were talking, were they air compressors?
If they were, then it is easier to perform the mechanical running test at the same time as the performance test...
But if you have a hydrocarbon compressor, the OEM would need to have the same gas to proceed with the test. This would be a very expensive ptc type 1 test...
Instead, they prefer to do the mechanical running test with vacuum or gas at a very low pressure and don't consume much power.
Then they proceed to the PTC type 2 test, where they use a different gas (CO2, He, etc.) with a different speed, which normaly is not the same of the MRT, to achive similarity with site conditions (Vin/Vout - Reynolds - Mach). With this test is possible to verify head, efficiency, etc. and convert them to site conditions.

For the mechanical losses, PTC 10 says that you have to measure the oil temperature and convert to site conditions using a relation between test and site speed.

I believe that the API mechanical running test and the PTC type 2 test covers the majority of problems that can occur... But if you have a critical compressor, with high pressure, high power and you are not so confident on the rotordynamic, a PTC type 1 or a full power-full pressure would help. And why not a complete unit test to check controls etc.

Have seen some bizarre things during type 1 tests...sometimes it's worthy the money.

 

[CaracasEC]

Hi Raphatss, thank you very much for sharing informative comments...

 

[jbattershell]

Raphatss,
testing was done with both air compressors and gas compressors. most tests were type II. Gas compressor, type II tests were conducted with inert gas. speed was similar to full speed during the aerodynamic testing, but not right on. most times, the inlet temperature could be adjusted to meet mach number requirements with a test running speed very close to actual running speed in field. either way, the compressor was at or very near full load much of the test and not in a vacuum with very little load. final vibration measurements were taken at contract speed while loaded. rpm was simply increased slightly (or decreased slightly) to reach contract speed.

you are quite correct, many things can "pop" up during a full load test, especially if you have the capability of doing a type I test with contract gas.