Shop test of centrifugal compressors

Centrifugal compressor purchasers and end-users realize the value and payback of accurate and proper shop performance test and shop mechanical run test. Shop test is a necessary step for centrifugal compressor future reliability and trouble free operation. ASME-PTC-10 type 1 performance test should be done wherever possible. Otherwise arrangement and details of ASME-PTC-10 type 2 performance test, as close to specified operating condition as possible, should be fixed before machine order. Three case studies are discussed.

PERFORMANCE TEST PROCEDURE

Centrifugal compressor performance tests are described in ASME-PTC-10 and API 617. With reference to ASME-PTC-10, there are two different types of performance tests. A type 1 performance test is actually a shop performance test in anticipated site condition. It is conducted with same gas as site (same gas with molecular weight deviation below 2%). Generally pressures, temperatures, compressor speed and capacity permissible deviations are below around 4-8%. Type 2 performance test is completely different. It permits the use of a substitute test gas and accepts extensive deviations between test and specified operating conditions. There are only a few limits on some essential gas dynamic parameters of test conditions (compare to specified operating conditions). Specific volume ratio and flow coefficient should be within around 5% deviations.

There are some limits on machine Mach number and machine Reynolds number. The test speed, capacity, mass flow, pressures, temperatures, power, etc are often totally different from the specified operating condition speed. In Type 2 test, a suitable gas is identified which does not lead to excessive power or discharge temperature and is readily and cheaply available. Substitute gas such as air, nitrogen, CO2, CO2/He mixes, fuel gas, etc are used. Safe operating speed, critical speeds, maximum allowable pressures, allowable temperatures and other machine limits are considered in test condition selection. ASME-PTC-10 Type 2 allows considerable deviations in test conditions.

For example a natural gas (MW=16) centrifugal compressor can be performance tested (Type 2) using CO2 (MW=44) with around half inlet flow, approximately 20% mass flow, around 50% speed, approximately 6% gas power and much less pressures (less than 10%) compare to specified operating conditions. There is always question regarding accuracy and usefulness of type 2 test. The flow patterns of centrifugal compressors are complex. This complex performance is function of main fluid characteristics such as volume ratio, flow coefficient, machine Mach number and machine Reynolds number. Concept behind type 2 performance test is performing a test with different gas and flow details, where as main fluid characteristics are within certain limits (let say volume ratio, flow coefficient within around 5%, machine Mach number within around 0.1 deviation and machine Reynolds number within 0.1 to 100 times) and use available mechanics of fluid knowledge and formulations to estimate machine performance in specified operating conditions. For example the friction aspect of compressor performance is affected by the machine Reynolds number.

In type 2 test, test Reynolds number is different compare to specified operating condition but it is still within certain limits to keep governing friction formulations the same (same model and flow regime). Based on theory, a modification (or let say correction) to the test results is applied based on available gas dynamic knowledge to estimate the friction effects of compressor performance in specified operating condition. All correction formulations are available in ASME-PTC-10 for estimation. Some engineers believe that ASME-PTC-10 type 2 test is not an actual performance test but it is a laboratory-type fluid test on real machine to confirm some fluid dynamic characteristics. ASME-PTC-10 type 1 test is always preferred. If ASME-PTC-10 type 1 is not possible (for example real gas can not be supplied and used in shop), it is necessary to plan for a type 2 test with test conditions (gas molecular weight, speed, capacity, power, pressures, etc) as close as possible to specified operating conditions. Arrangement and details of type 2 teats should be fixed in bidding stage and before machine order. Vendors always prefer simplest and cheapest arrangement for type 2 test (and code ASME-PTC-10 type 2 allows this). Unfortunately type 2 test gas selection and arrangement are generally discussed after order or even near test time. It usually causes considerable change order costs or purely laboratory-type test on real machine.

Type 2 test can be useful if gas mixture that closely approximate the job gas used and pressures, compressor speed, capacity, power, etc matched as close as possible to specified operating conditions (within ASME-PTC-10 limits and as close as practical to operating conditions). This properly arranged type 2 test may give useful prediction on future machine behavior such as operation close to surge point, some types of aerodynamic excitations, effects of seal on dynamic behavior, etc. These effects are of serious concern particularly in high pressure applications.

For performance test whether type 1 or type 2 is used, following limits should be confirmed (in case of type 2, based on estimation):

1- Head and capacity: zero negative tolerance.

2- Power: not exceed 104% of predicted power (sometimes when plant efficiency is vey important, lower limits, as low as 2%, may be adopted and agreed before machine order).

3- Surge: stable operation at near calculated surge (let say around 8% above calculated surge flow).

Based API 617 and refer to type 2 test procedures, test speed may be completely different with specified machine speed. It is not permitted for ASME-PTC-10 type 1 test (in type 1 test only 2% speed deviation is allowed). Complete unit test is always recommended (if practical). All train components including compressor casings, gear units, driver, and all auxiliaries are tested all together. If complete unit test is not possible, a tandem test is useful. In tandem test, usually shop driver, and shop oil system are used. A separate auxiliary test may be performed. Torsional vibration measurements are recommended to evaluate torsional calculations. It is necessary for complex multi-machine casings including gear units with variable speed driver.

MECHANICAL RUN TEST

Job seals and bearings are generally employed in mechanical run test. Sometimes additional heat dissipation elements (such as specially designed fins) are required to avoid overheating particularly in low pressure tests. Pressurized run test is always preferred, except special cases (such as cases when under vacuum tests are required). Extensive measurements (speeds, pressures, temperatures, oil and seal flows, bearing metal temperatures, vibration data, seal gas data, etc) are required for mechanical running test. Generally as far as practically possible, all contract equipment and systems are used for the test. If using job coupling is not practical, test can be implemented with simulator. Overhung moments should be simulated within suitable margin of job coupling (preferable below 5%, maximum deviation 10%). Test facilitates should be capable of continuously monitoring, displaying, and recording vibration, vibration spectra, Bode plots, and shaft orbits. Shop mechanical run test procedure is straightforward. The machine is started and speed increases to the maximum continuous speed. Operation continues until bearing metal temperatures, oil temperatures and shaft vibrations are stabilized. Then machine is operated 15 minute at trip speed. After that, 4 hours operation at maximum continuous speed. Main focus is on vibration. Usually vibration measurement covers a range of 0.25 – 8 times of operated speeds.

Based on experience, high frequency ranges (usually above 1500 Hz or 90,000 rpm) do not contribute to the judgement of the mechanical performance. The frequency analysis recorded must not show significant amplitudes at frequencies other than running speed or twice running speed. Shop verification of the unbalanced response analysis is usually required. Each spare rotor needs a mechanical running test. Other important goal is verification of lateral critical speeds. The locations of all critical speeds below the trip speed are confirmed on the test stand. Actual critical speeds are expected within 5% of analytical values. Peak responses amplitudes should not exceed the analytical values. During test the lube oil temperature rise through the bearing is not expected to exceeding 30°C.

Vibration readings and bearing temperatures at the end of the four-hour run should be essentially the same as those recorded at the beginning of the four-hour test. Real-time vibration data (from startup to shutdown) are recorded and delivered to end-user. All hydrodynamic bearings are removed, inspected, and reassembled after the mechanical running test.

Removal and inspection of dry gas seal is not recommended. Dry seal gases (particularly cartridge type) may require that the seal be returned to the seal manufacturer if removed for inspection. It is recommended to remove oil seal (mechanical oil film type shaft end seals which very rarely used today) for inspection after test. Minor scuffs and scratches may occur on the bearings. Subsequent minor cosmetic repairs of these parts do not justify repetition of the test. If melting or smearing, overheating or distinct wear occurs in the babbit of bearing shoes, these parts should be replaced. The cause of the defect must be investigated and eliminated, and the mechanical run test should be repeated. After run test, compressor casing is gas leakage tested to evaluate joints and seals. Assembled compressor is tested to maximum operating pressure for minimum of 30 minute (inert gas MW - Molecular Weight, less than job gas MW, helium for low MW gas or nitrogen or refrigerant gas for high MW gas). Please consider following recommendations regarding optional tests:

1- Varying lube oil conditions (oil pressure and temperature at minimum and maximum values) is an API 617 optional test but it is strongly recommended.

2- Noise level test is API 617 optional test. It is only required for large machine when high noise is expected.

3- Post test inspection of casing internal is not recommended due to advantage of delivery of proven run and pressure tested compressor.

SITE CONSIDERATIONS

Care should be taken when designing the team that represents equipment purchaser for shop tests. Some of individuals, who attend shop test, need to be worked at site for commissioning and at least one for long term operation.

There are many notes about necessity of vendor representative involvement in compressor piping check and alignment. Based on experience it is useful that vendor representative observes a check of the piping but it is unnecessary for the vendor representative to be present during the initial alignment check or to check alignment at the operating temperature.

As soon as possible a site performance test should be conducted on each compressor (after installation and pre-commissioning). The performance is checked based on ASME-PTC-10 type 1 test. It should be planned in advance. Test procedure is necessary and some extra instruments and temporary provisions may be required. The test should confirm, as a minimum, that the compressor can meet the guarantees of flow, pressure and power and that the other points on the curve are within +/- 5% of the shop performance test curve.

Site performance tests may also be implemented during operation (for example after several years of operation), to identified extent of degradation.

5    CASE STUDIES

The first case study is presented for type 2 performance test of a large centrifugal compressor for hydrogen service. Table 1 shows performance test data for this centrifugal compressor.

Table 1 Type 2 Performance Test Data for Hydrogen Centrifugal Compressor (first case study).

 

Specified Operating Condition

Type 2 Test Condition

Gas Molecular Weight

~4

~6

Inlet Volume Flow (m3/h)

6200

5000

Volume Ratio

~1.12

~1.12

Machine Mach Number

0.241

0.245

Machine Reynolds Number

~8x106

~106

Inlet Pressure (Barg)

150

23

Discharge Pressure (Barg)

185

28

Power (MW)

~6

~1

Speed (rpm)

9950

7800

Based on experience, proposed test conditions of this case study are better than many type 2 tests performed in industry. Test gas (gas mixture with molecular weight 6) and test speed are not so far from operating conditions. But improvements can be done for more realistic type 2 performance test such as using closely matched gas mixture (molecular weight~4), increasing test gas pressures (while keeping parameters within ASME-PTC-10 limits), closely match isentropic index, etc. These modifications can help to better match test power, capacity and compressor dynamic behavior to specified operating conditions. In this case, because it was not negotiated in bidding stage, agreement could not be reached for improvements and performance test was done based on data of Table 1. This case study shows importance of negotiation of performance test details before machine order.

The second case study is presented for type 2 performance test of medium pressure natural gas compressors. Table 2 shows proposed performance test plan for this machine. It is a good example of considerable differences between test and specified operating conditions. As a rule of thumb, test speed within 20% limits of machine speed is always preferred. For this test, test speed is around 60% of machine operating speed. But for this purchase order, several identical compressors are required and these type 2 performance tests are applied for subsequent units after first unit successful type 1 performance test. This arrangement is optimum and acceptable.

Table 2 Type 2 Performance Test Data for Medium Pressure Natural Gas Centrifugal Compressor (second case study).

 

Specified Operating Condition, Natural Gas

Type 2 Test Condition with CO2

Gas Molecular Weight

~16.3

~44

Inlet Volume Flow (m3/h)

~20,000

~12,000

Volume Ratio

~2.6

~2.6

Machine Mach Number

0.614

0.605

Machine Reynolds Number

~6.4x106

~9.8x106

Inlet Pressure (Barg)

12

0.7

Discharge Pressure (Barg)

46

4.8

Power (MW)

~13

~1

Speed (rpm)

~10,000

~6,000

Third case study is presented for a high pressure natural gas compressor. Test speed is around 80% of specified operating speed. Improvement still can be done for a more realistic type 2 shop test using gas mixture to closely match gas molecular weight. Again modification proposals were not successful because of commercial issues. Test was implemented based on data of Table 3. Comparison with site performance test shows acceptable results.

Table 3 Type 2 Performance Test Data for High Pressure Natural Gas Centrifugal Compressor (third case study).

 

Specified Operating Condition, Natural Gas

Type 2 Test Condition with Nitrogen (N2)

Gas Molecular Weight

~16

~28

Inlet Volume Flow (m3/h)

~5500

~4500

Volume Ratio

~1.8

~1.8

Machine Mach Number

0.47

0.48

Machine Reynolds Number

~10x106

~1.3x106

Inlet Pressure (Barg)

40

8

Discharge Pressure (Barg)

100

16

Power (MW)

~10

~2

Speed (rpm)

11,800

9,400

Notation

MW    Molecular Weight

ASME                American Society of Mechanical Engineers

 

(Amin Almasi is a rotating machine consultant based in Brisbane, Australia)