Objectives of Centrifugal Compressor Mechanical Tests

It is a good idea for the user to have a witness to check that someone has inspected and verified operation of all safety and warning instrumen tation used in the shop. Granted, the compressor is in the vendor's shop and repairs are on the manufacturing account, but why risk damage and late shipment just because the witness was bashful? While he's at it, have him ask if the oil is being filtered to a level of at least 10 microns. It probably isn't feasible to open the filter and look, but at least ask.

For this test, as much of the contract equipment as is practicable should be installed on the compressor, such as the vibration probes, temperature sensors, even the job coupling if a match-up to the driver is possible. If the contract coupling isn't used on the test, the coupling used should duplicate the mass-moment of the contract coupling. Recall that the ideal test would be to duplicate all the field conditions. This picture should be in the user's mind, bounded only by contract scope and practical shop limitations. A little creative imagination can make a routine test a meaningful test for the user.

One objective is to verify oil flow, friction horsepower, and heat rejection to the lube oil. Very often these tests are carried out under vacuum conditions resulting in very low aerodynamic thrust loads. It should be noted that the thrust friction horsepower constitutes the greatest loss and, therefore, the results will have to be adjusted by calculation from the test thrust to full design thrust. The error found from the test compared to design is of little significance to compressor efficiency, but is serious to oil cooling capacity. It should be noted that no-load to full-load friction loss is a linear relationship for thrust bearings and for gears. Errors in predicted loss are due to variations in running clearance and oil flow that affect the churning loss. It is important that tests are carried out at sitedesign temperatures with the correct viscosity of oil.

If oil buffered seals are used on the compressors, the seal leakage toward the process side of the compressor must be carefully measured, as it is (and should be) a small value. While five gallons per day doesn't sound too small, in a four-hour run, this is less than two pints, making the hold-up time at the inner seal chamber and in the lines to the drain pots a significant value. This makes exact measurement quite difficult.

On compressors with gas seals, it is very desirable to use the contract buffer gas skid. Use of the skid provides the contract filtration level as well as having the regulators sized for contract conditions and checks the buffer gas skid functioning. Most shop systems are built of surplus parts and tend to use manual regulation, which may lead to serious test stand problems. If the skid is not available, it is important to have the vendor provide a comprehensive plan on how the shop buffer gas will be set up.

On those compressors where rotor dynamics can be a problem, which is on all but the standard units (even them sometimes) and reciprocating compressors, this is the point where the acceptance of the compressor for vibration and critical speed criteria should take place. Finding these in the field later is what the user and vendor want to avoid, and this is why all the elaborate and careful work is done at the running test time.

Suitable instrumentation has to be installed to monitor rotor response. This must be taken into consideration during the design stage and the required finish, concentricity, and physical properties of the instrument target areas monitored as a quality control operation during manufacture. Installation and calibration of probes during shop assembly in conjunction with manufacturing quality control avoids costly test stand delays. If magnetic tape records are obtained of the shaft vibration during shop test, these can then be used for field comparison during operation.

Shop test facilities should include instrumentation with the capability of continuously monitoring and plotting rpm, peak-to-peak displacement, and phase angle (X-Y-Y'). Presentation of vibration displacement and phase marker by use of an oscilloscope makes visualization easier.

The vibration characteristics, determined by use of the instrumentation, will serve as the basis for acceptance or rejection of the machine. API standards generally require that the equipment be operated at speed increments of approximately 10% from zero to the maximum continuous speed and run at the maximum continuous speed until bearings, lube-oil temperatures, and shaft vibrations have stabilized. Next, the speed should be increased to trip speed and the equipment run for a minimum of 15 minutes. Finally, the speed should be reduced to the maximum continuous speed and the equipment should be run for four hours. API does not require that the four hours be uninterrupted; however, it is generally interpreted that way. The interpretation is one of the many test criteria to be discussed. It would seem that a break in the test at the midpoint is not the same as having it cut short five minutes from the end because the vendor's boiler took an upset that was not related to the compressor test. The vibration during the shop test is normally specified as the API limit of 1.0 mils peak to peak, or the value from Equation 10.1, unfiltered. whichever is lower.


A = peak to peak amplitude, mils

N = maximum continuous speed, rpm

The limits given are for the centrifugal compressor and for the steam turbine. API 619 should be consulted for the limits of vibration of the helical lobe compressor.

The parameters to be measured during the test will be speed and shaft vibration amplitudes with corresponding phase angles. The vibration amplitudes and phase angles from each pair of X-Y vibration probes shall be vectorially summed at each response peak to determine the maximum amplitude of vibration. The major-axis amplitude of each response peak shall meet the limits specified and set by the specification. At the end of the four-hour test, the unit should be decelerated and recording measurements taken. The unit should then be accelerated to full speed and similar recordings made. The gain of the recording instrumentation used shall be predetermined and pre-set before the test so that the highest response peak is within 60-100% of the recorder's full scale on the test unit coast-down (deceleration).

Vectorial subtraction of slow-roll (300-600 rpm) total electrical and mechanical runout is permitted by the API rules. Vectorial subtraction of bearing-housing motion may be justified if it can be demonstrated to be of significance.

If the previously agreed-on test acceptance criteria are not met, then additional testing will be required.

Living Off The Grid

Living Off The Grid

Get All The Support And Guidance You Need To Be A Success At Living Off The Grid. This Book Is One Of The Most Valuable Resources In The World When It Comes To When Living Within The Grid Is Not Making Sense Anymore.

Get My Free Ebook

Post a comment