Gas Compressor Re-Rate and Process Solutions Through

Engineering and Collaborations

Barry Hall M. Najib Bin A Rahim

Chief Engineer – Compressor Technology Senior Rotating Equipment Engineer
Weir Engineering Services, Ltd Petronas Carigali Sdn Bhd

Ir. Syed Muhammed Afdhal Altug Bilgic

Staff Engineer Engineering Manager - TMD
Petronas Carigali Sdh Bhd Weir Engineering Services, Ltd

1.0 Background
PETRONAS Carigali Sdn Bhd, Peninsular Malaysia Operations (PCSB-PMO) operates a natural
gas processing offshore facility located about 150 km off the coast of Terengganu, Malaysia. The
complex was commissioned with 3 x 50% natural gas compression trains designed to handle 350
million standard cubic feet per day (MMSCFD) export gas. However, due to presence of
contaminants such as hydrocarbon liquid, H2S and CO2 in the natural gas, the compressor trains
experienced multiple issues and failures as soon as the field was commissioned.

The failures on the centrifugal gas compressors were mainly associated with fouling and
resonance. There were occasions where the compressor impellers were cracked due to fatigue
as identified by metallurgical analyses performed. Modal analyses were conducted, which
confirmed resonance issues, where the impeller natural frequencies were within the operating
speed range of the compressors. There were also issues involving material erosion and corrosion
due to presence of H2S and CO2.

In addition to the operational and design related issues, a new gas platform was commissioned in
2009, which required the gas processing complex to handle additional gas capacity. The natural
gas from the new platform would be directed to the existing compressor trains. As a result, the
aerodynamic design of the existing compressors was no longer suitable to handle the combined
gas flow and properties, and PCSB-PMO decided to retrofit the existing compressor aerodynamic
design.

A contract was awarded to Serba Dinamik Sdn Bhd, through a competitive bidding exercise, for
design of a re-rated bundle to be retrofitted into existing compressors and resolve the
rotordynamic issues. As part of the contract, Weir Engineering Services, Ltd. (WEIR) designed,
manufactured the re-rated compressor internals and performed factory testing. Serba Dinamik
Sdn.Bhd provided local personnel, facilities and logistical support to their alliance partner Weir
Engineering Services, Ltd. (WEIR) during the design, manufacturing, testing and commissioning of
the re-rate. Improvement projects were also undertaken by PCSB-PMO, which successfully
mitigated wet gas issues through upgrading of the inlet gas scrubber internals.
This paper discusses the issues faced on a gas processing platform that led to a Compressor
Upgrade project, and aims to demonstrate how various organizations collaborated on solving
operational issues via application of sound engineering solutions and technology know-how to
improve facilities’ reliability and performance. Despite various challenges, this collaboration has
proven to be among the significant contributing factors to successfully turn the platform from a
problematic operation into one of the best performing assets in PCSB-PMO.

2.0 Introduction
Three identical natural gas compression trains (K-2410/40, K-2510/40, K-2710/40) on the PCSB-
PMO platform were commissioned in 2002. The trains comprised of:

Gas Turbine: Speed Range: 2521 – 3308 RPM

LS Coupling:
Gear Box: Gear Ratio: 3.28
HS Coupling:
Compressor: Speed Range: 8273 – 10858 RPM

The units experienced failure due to several reasons soon after commissioning.

Unit 2 (May ‘03):

Low performance and vibration.

Heavy greyish cement-like deposit.
Build up mostly on 2nd impeller.

No broken impellers. Cleaned and put back in service.

Unit 3 (Jun ‘03):

Low performance and vibration.

Lighter deposit, greenish brown, higher density.

Impeller No.2 broken in 5 locations
4 in front, 1 at back.

Sent to OEM for repairs and investigation.

Unit 2 (Aug ’03):

Tripped on high vibration.

Little deposits

Impeller No.1 broken in 1 location

Sent to OEM for repairs and investigation.
 Unit 1 (Nov ’03):

Tripped on high vibration.

2nd stg diffuser heavily contaminated, some
almost blocked.

Impeller No.1 broken in 2 locations at front and back.

Sent to OEM for repairs.

Unit 2 (Oct 04):

Removed due to high vibration

Impeller No.1 broken in 3 locations

Unit 3 (Jun ‘05):

Tripped on PT overspeed.

Lighter greyish deposit

Shaft sleeve at buffer gas injection heavily
eroded

Coupling broken

Unit 1 (Aug ‘05):

Tripped on compressor excessive vibration.

Grayish deposit.

Impeller No.6 broken in 2 locations (hub).

3.0 Problem Definition

The failures in the compressors were associated with fouling and resonance. Metallurgical
analyses on failed impellers suggested that the broken impellers failed due to fatigue. Modal
analyses confirmed impeller natural frequency inside the speed range of the compressors.

Gas compressors on the platform operate in incredibly harsh environments. The gas component
comprises a problematic combination of CO2, H2S, H2O, and Mercury. However, although traces
of mercury were found inside the gas, no direct mercury related failure was reported in the original
designs.
PCSB-PMO worked on the process system to minimize and prevent fouling as much as possible.
The measures taken included cleaning of the scrubbers, but no design modification was
implemented. However, the resonance issue could not be resolved since it would require new
impeller manufacture. Instead, the impellers were scalloped by the OEM (see Figure 1) and the
maximum running speed (MCS) was reduced to avoid resonance. The excessive scalloping and
the reduced speed range reduced the compression capability of the compressors, while
decreasing aerodynamic efficiency levels.

In addition to the operational and design related issues, PCSB-PMO introduced a new gas
platform. The natural gas from the new platform would be re-directed and the combined gas
would be compressed by the K-2410/40, K-2510/40, K-2710/40 compressor trains. As a result,
the aerodynamic design of the existing compressors would no longer be capable of handling the
combined gas, which would amount to approximately 50% of PCSB-PMO’s production. To make
the issue more complex, the platform would have to operate at existing process conditions until
the new platform was functional. This required the compressors to be able to fulfil their function
under existing as well as future conditions.

Gas production from the two wells was not treated for Mercury removal, not dehydrated, and not
sweetened. Hence, looking forward, process gas was going to remain as an issue for the
platform.

Figure 1. Scalloped impeller on Future Rotor

4.0 Compressor Re-Rate Solution
In order to resolve the design related issues and to account for the additional gas from the new
platform, PCSB-PMO opened a bid for a re-rate to the existing compressors. The re-rate was
required to

Resolve resonance issues

Handle existing production

Handle new production

Handle combined existing and new production

with a maximum available power of 15.5 MW. A factory acceptance test (FAT) and later a site
acceptance test (SAT) was required for verification of the design. The materials used on the re-
rate would have to be NACE compliant as well as mercury resistant.

The WEIR solution to the re-rate was a 7 stage back-to-back design with 19 inch impellers. Table
1 and Figures 2 through 4 show the performance data and the Head-Flow curves of the re-rated
compressor. It can be clearly seen on Figure 4 that the re-rated compressor was able to achieve
the existing performance as well as the higher flow and head requirements of the future
processes. With these changes, the re-rate solution offered by WEIR eliminated the need for a
completely new compressor that PCSB-PMO had believed they would be forced to implement.
The re-rate solution therefore allowed the utilisation of all ancillaries around the compressor
casing and prepare for future operations of the platform with minimal cost and disturbance to the
current operations

The existing driver, gear-box, couplings and ancillaries were retained. The existing dry-gas seals
and seal system were also retained. The existing journal bearings and thrust bearing were to be
utilized upon confirmation by rotordynamic analyses and detail design work.

To be able to fulfil all duty requirements, the compressor polytropic efficiency had to be high. To
achieve the required high efficiency levels, vane diffusers were introduced at each diffuser (see
Table 1 for performance data). To prevent any resonance issues, modal analyses of all impellers
would be performed during the design stage and later verified by ring testing. To delay fouling,
the rotor and stator wetted surfaces would be coated with a Teflon based coating.

The design would be verified by an ASME PTC-10 Type II and API617 factory test and a site
acceptance test (SAT) following commissioning.

All design, manufacture and commissioning work was to be carried out by WEIR. Project
management and regional logistical and field support was to be provided by Serba-Dinamik.

Dimensions of existing equipment (e.g. casing, bundle, rotor, housings, etc.) were measured by
WEIR personnel at PCSB-PMO facility in Kerteh. Amendments to O&M Manuals were provided
by Weir.
 CASE 1 CASE 2 - RECYCLE CASE 3
(ALL DATA ON PER UNIT BASIS) LP HP LP HP LP HP
GAS HANDLED Case 1 Case 1 Case 2 Case 2 Case 3 Case 3
STANDART FLOW, MMSCFD 175 175 119 132 255 255
WEIGHT FLOW, kg/h 197477 197477 146846 162888 314633 314633
INLET CONDITIONS
PRESSURE (kPaA) 1800 5199 1500 4661 3000 6706
TEMPERATURE (°C) 26.59 42.95 42.61 42.92 42.61 42.92
MOLECULAR WEIGHT (%) 22.59 22.59 24.69 24.69 24.68 24.68
COMPRESSIBILITY (Z1) 0.954 0.890 0.964 0.886 0.927 0.835
INLET VOLUME, (m³/h) 11543 3934 10029 3294 10333 4172
DISCHARGE CONDITIONS
PRESSURE (kPaA) 5349 12500 4811 12500 6856 12500
TEMPERATURE (°C) 119.3 121.5 142.7 131.5 113.7 98.2
COMPRESSIBILITY (Z1) 0.957 0.922 0.965 0.916 0.931 0.868
FRICTION POWER (KW) 50 49 30
kW REQUIRED (ALL LOSSES INCLUDED) 15484 12699 16207
SPEED (RPM) 10341 10211 8780
POLYTROPIC HEAD (kJ/kg) 131.8 103.2 137.7 106.9 90.5 61.1
POLYTROPIC EFFICIENCY (%) 83.9 82.9 83.5 81.5 83.3 79.8
CERTIFIED POINT YES NO NO

Table 1. Re-Rate Performance Data

65000 45000

40000
55000

Case 2 Case 2 - Recycle

35000 Case 1
Polytropic Head (ft_lbf/lbm)

Polytropic Head (ft_lbf/lbm)

Case 2 Case 2 - Recycle

45000
Case 1

30000 10858 RPM

35000
10341 RPM
Case 3 10858 RPM 25000

25000 10341 RPM

Case 3 9307 RPM
20000

9307 RPM
15000 8273 RPM
15000
8273 RPM
7239 RPM
7239 RPM
5000 10000
3000 4000 5000 6000 7000 8000 9000 10000 1000 1500 2000 2500 3000 3500
Inlet Volume Flow (acfm) Inlet Volume Flow (acfm)

Figure 2. Low Pressure (LP) section performance map Figure 3. High Pressure (HP) section performance map
55000

45000
Case 1
Polytropic Head (ft_lbf/lbm)

35000

10858 RPM

25000

15000

7239 RPM

5000
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Inlet Volume Flow (acfm)

 Figure 4. Existing vs Re-Rate LP Performance Map

5.0 Re-Rate Project Execution

5.1 Design
A joint project team was formed by WEIR and Serba Dinamik. All design work was carried out by
WEIR personnel. An initial design was carried out using approximated dimensions of the existing
equipment from user-supplied drawings. This design provided approximate rough stock
dimensions for critical components and allowed material purchase while detail design work was
being carried out. Later, the Design Engineer and the Drafter took detailed dimensions off a spare
rotor and bundle of the existing compressor, which was made available by PCSB-PMO.

The impeller modal analysis was carried out after all aerodynamic and stress analyses were
completed. The modal analyses assured that there were no impeller structural modal frequencies
inside the running speed range of the compressor. The calculated frequencies were later verified
by ring testing one of the impellers at WEIR’s Bedford facility. There was good agreement
between the calculated and tested natural frequencies. The mode shape leading to the failures of
original impellers is shown for the LP 1st stage impeller in Figure 6.

Figure 5. Re-Rate Design

Inlet scoops (commonly known as baby’s bottoms) were added to each section inlet to better
guide the flow into the first stage. Low solidity vane diffusers were designed for each diffuser to
improve efficiency levels.

The labyrinth seals were designed with swirl-breaks to reduce destabilizing effects. The labyrinth
material was chosen as Arlon to protect against mercury attack, as agreed with the customer.
The centre-seal was designed as rotating teeth against honeycomb design. The original TAM
seal was eliminated. The rotordynamic analysis was carried out with the new rotor and stability
analysis was carried out during design stage. The existing deflection-pad type bearings were not
re-used due to unavailability of stiffness and damping coefficients. Instead, new 4-Pad LBP
configuration tilting-pad type bearings were designed and manufactured.

The train torsional analysis was performed and results suggested that the existing coupling could
be re-used.

Figure 6. LP 1st Stage Modal Analysis – 9th Mode

Thrust analysis suggested that the existing thrust bearing could be re-used. In addition, the
existing dry-gas seals were re-used.

Concurrently with the design calculations, detail drawings of components and assembly drawings
were completed. All drawings were made with SI units as agreed before the start of the project.

5.2 Manufacturing and Assembly

Manufacturing was undertaken at WEIR facilities and at WEIR-approved vendors. The impellers
were milled as two pieces; the blades were milled into the hub and a shroud with weld slots was
manufactured. The blade leading edges were welded with stellite weld rods to protect against
erosion in the field. Later, the shroud was welded on to the hub. The welded impeller went
through heat treatment to comply with NACE hardness requirements.

After heat treatment, the impeller was finish machined, NDT inspected (dye penetrant and
magnetic particle), balanced and spun to 115% of the Maximum Continuous Speed per API617 7th
Edition requirements.
 Hub CNC Milling Shroud CNC Milling

Hub and Shroud Weld Prep Post Weld

Figure 7. Impeller Manufacturing Steps

The rotor was assembled and sequentially balanced at WEIR facility in Barton-on-Humber. After
low speed balance, the rotor was coated with a Teflon coating to protect against corrosion and
provide anti-fouling properties. The rotor was then high-speed balanced.
 Figure 8. Rotor with Anti-Fouling Coating

Figure 9. Manufactured Bundle and Rotor

5.3 Testing
ASME PTC-10 Type II aerodynamic performance and API617 mechanical testing (FAT) were
performed at the WEIR facility in Barton-on-Humber, UK. FAT was carried out by assembling two
purpose built loops, one for the LP and one for the HP section. Each loop contained a flow control
valve, cooler and flow meter, together with the instrumentation to record the suction and discharge
pressures and temperatures. The compressor was driven by a variable speed DC motor driving
through two gearboxes to achieve the required rotational speed. The performance testing was
carried out using Nitrogen at equivalent conditions and the field performance calculated in
accordance with the procedures in PTC10.

Both aerodynamic and mechanical tests were successfully completed for both the machines that
were tested. Figures 10 through 12 show the test setup and the resulting test performance curves
for both sections of the compressor.
 Figure 10. Test Setup – Compressor

70000 1.0

65000 0.9

60000 0.8
Polytropic Head (ft_lbf/lbm)

55000 0.7
Polytropic Efficiency

50000 0.6

45000 0.5

40000 0.4

35000 0.3

30000 0.2
Specified Performance
25000 Corrected Tested Performance 0.1

20000 0.0
4500 5000 5500 6000 6500 7000 7500 8000 8500 9000
Inlet Volume Flow (acfm)

Figure 11. LP Section FAT Performance

 50000 0.90

45000 0.75
Polytropic Head (ft_lbf/lbm)

Polytropic Efficiency
40000 0.60

35000 0.45

30000 0.30

25000 0.15
Specified Performance
Corrected Tested Performance

20000 0.00
1500 1700 1900 2100 2300 2500 2700 2900 3100
Inlet Volume Flow (acfm)

Figure 12. HP Section FAT Performance

6.0 Process System Upgrade Project Execution

The gas compressors on the gas processing platform are operating in an incredibly harsh
environment. The gas component comprises a problematic combination of CO2, H2S, H2O,
and Mercury. The gas which originated from existing and new wells is not treated for
Mercury removal, not dehydrated, and not sweetened.

The harsh operating environment on the platform has caused heavy fouling (Figure 13) in the
original as well as the re-rated compressor internals. Since this issue stems predominantly
from flows upstream of the compressor, the scrubber design was revisited by PCSB-PMO for
improvement.

Figure 13. Heavy fouling found at the compressor internals

 6.1 Scrubber Internals Upgrade
The 1st and 2nd Stage Suction Scrubbers are vertical two phase separators. They originally
comprised a “Schoenpentoeter” vertical vane type inlet device along with “Vane Pack” and
“Swirl Deck” gas/liquid separation devices. These internals have been modified using CFD
techniques to improve gas/liquid separation whilst retaining the same basic design
requirements.

An internal drain takes liquid from the Swirl Deck back to the liquid filled Scrubber Sump. The
liquids are removed under gap level control.

Figure 14. Upgraded Vane Pack and Swirl Deck of the Scrubber

6.2 Higher Efficiency Scrubber

After the scrubber internals upgrade, the 1st and 2nd Stage Suction Scrubbers were able to
remove wet gas with an efficiency of well above 99% and 95% respectively, which is higher
than before upgrade. Even though the liquid knock-out is not 100%, the upgrade has
significantly improved the gas quality entering the compressors.

Coalescer Contaminants Measurement (CCM)

1st Stage Suction Scrubber
PPM (wt) Gals/day
UPSTREAM GAS 5,824.1 11,102.081
DOWNSTREAM 3.2 6.288
GAS
Efficiency 99.9 %

Table 2. 1st Stage Suction Scrubber CCM result

 Coalescer Contaminants Measurement (CCM)
2nd Stage Suction Scrubber
PPM (wt) Gals/day
UPSTREAM GAS 123 243.749
DOWNSTREAM 5.1 9.921
GAS
Efficiency 95.9 %

Table 3. 2nd Stage Suction Scrubber CCM result

7.0 Conclusions
PETRONAS Carigali Sdn Bhd, Peninsular Malaysia Operations (PCSB-PMO) operates a
natural gas platform, which was commissioned in 2000. The compressors installed on the
platform experienced multiple failures soon after commissioning. The failures were due to
compressor design as well as the process conditions.

In an effort to prevent any further failures and also to accommodate the extra gas to be
supplied from the new platform, PCSB-PMO placed a purchase order to Serba Dinamik Sdn
Bhd for the re-rate of the existing compressors.

WEIR designed, manufactured, tested, installed and commissioned the re-rates to the
compressors in 2009. In the meantime, PCSB-PMO analysed the performance of the
existing upstream process system and upon identification of scrubber performance shortfall,
upgraded the scrubbers for the LP and HP sections.

Together with the re-rate and the scrubber internals upgrade, the compressor Mean Time
Between Failure (MTBF) has improved significantly from 6 months to 2+ years. In addition,
the platform has, for the first time in its lifetime, operated all three compressors
simultaneously, accounting for roughly 50% of PCSB-PMO natural gas production.

This paper discussed the operational problems that could be faced at a gas processing
facility such as an offshore platform and how these problems can be resolved by partnership
of individual companies, utilising sound engineering techniques.

For more information please contact Jon Sharp:

Head of Turbomachinery
Weir Oil & Gas Services - Europe
T +44 (0)1652 632702
F +44 (0)1652 631401
M +44 (0)7768 614482
jon.sharp@weirgroup.com
www.global.weir