Compressors are the highest producers of revenue among driven equipment, in upstream and down-stream industries. They are usually un-spared, and are considered as critical equipment items. Compressor reliability is therefore a high priority, and is directly proportional to company profit. Many compressor trains are continuously condition-monitored for predictive maintenance purposes to obtain optimum reliability – which can exceed 99.7% for compressor types that are properly selected designed, installed and monitored.
Below are some of the best practices for this subject.
Best Practice-1: Use site, company and industry lessons learned to determine if positive displacement or dynamic (turbo) type compressors are required.
Accurately determine all of the required operating conditions. Process conditions not anticipated in the selection phase account for 80% or more of compressor reliability problems. Then use company guidelines for selection of the proper compressor type if available. Whether company guidelines are available or not, determine the type by site and industry lessons learned to avoid plant safety and reliability issues. Finally, indicate the selected compressor type on the data sheet with all best practices dictated by plant, company and industry lesson learned.
Improperly selected compressor type has led to field safety issues, revenue and legal costs as noted by the following examples:
- Lubricated screw compressor in sour gas service – never operated for more than 24 hours continuously
- Reciprocating compressor, used when screw compressor should be used, resulted in extensive maintenance costs and sour gas leakage.
- Rotary lobe compressors, used instead of centrifugal compressor, resulted in the need for multiple compressor when a single centrifugal compressor could have been used.
Benchmark: This best practice has been used since the mid-1980s in Upstream, refining, and chemical plants, giving the following optimum compressor reliabilities:
- Centrifugal greater than 99.5%
- Dry screw greater than 99%
- Lubricated screw greater than 97.5%
- Dry reciprocating greater than 94%
Best Practice-2: Define all process conditions and specific requirements on the compressor data sheet.
Confirm all process conditions and specific requirement on the compressor data sheet, while accurately defining the gas analysis, the highest gas head required and all expected operating points (head and volume flows).
Meet with the EP&C and/or process licensor process engineers to ensure that all operating and upset process conditions are accurately defined.
Inaccurate process have led to scope changes during the project and increased delivery time by 6 months or longer. One recent example was a lubricated screw compressor installation where the compressor could not operate for more than 24 continuous hours due to the actual field gas analysis (sour) being different from the data sheet gas analysis (sweet).
Benchmark: Since the mid-1970s process engineers and senior operators in the process design phase to ensure that all conditions were defined. This best practice has resulted in compressor reliability of the highest levels (greater than 99.7%).
Best Practice-3: Pre-select centrifugal compressor casing type, impeller type, the number of compressor cases and impellers in each casing.
Pre-selection of centrifugal compressor casing type and impellers ensures optimum safety and reliability. Determine if a horizontal split casing or vertical (barrel) type is required based on process conditions and vendor/company/industry guidelines and plant lessons learned.
Determine the impeller type (opened or closed) base on company/industry guidelines and plant lessons learned. Also determine the number of impellers allowed in each casing based on head per impeller stage limits and shaft stiffness
There are many case histories of failure because centrifugal compressors were not selected using the proper case design and limiting the number of impellers per stage.
Benchmark: Since the mid-1970s this practice been used to achieve success in all centrifugal compressor installations resulting in plant installations of greater than 99.7% centrifugal compressor reliability, without:
- – Critical speed issues
- – Gas instabilities (gas whirl and whip)
- – Impeller failures
- – Excessive factory acceptance test (FAT) time.
Best Practice-4: Limit non-lubricated reciprocating compressor piston speed to below 600ft/minute for optimum compressor non-lubricated compressor reliability (greater than 94%).
Non-lubricated reciprocating compressors have the lowest reliabilities; only around 92%. Invest in an extra cylinder if necessary to keep piston speeds below 600 ft/minute for maximum packing, valve and piston ring/rider band life.
Failure to limit piston speed in non-lubricated reciprocating compressor application has led to valve, packing and ring MTBFs of less than 6 months.
Benchmark: This best practice resulted in non-lubricated reciprocating compressor reliabilities greater than 94%. Packing, valve, piston ring and rider band MTBFs have exceeded 24 months.
Best Practice-5: Restrict reciprocating compressor field pulsation limits to +/-2% of line pressure for safe and reliable operation.
Take care during the design phase that an accurate isometric piping arrangement is used for the digital pulsation analysis. Check field pulsation values immediately after start up and take immediate corrective action (installation of orifices, pulsation bottle corrections, etc.) to ensure optimum levels of safety and reliability.
Failure to address pulsation problems as indicated during the design and/or field operation phase gives rise to a number of issues. Examples or these are:
– Instrument breakage and gas release
– Piston rod breakage
– crank case/distance piece bolt breakage
– Packing failure
– Valve failure
Benchmarks: Since the 1980s, this best practice been used to correct many field pulsation issues that had cause safety problems and had significantly reduced compressor MTBFs.
Best Practice-6: Do not use medium or high speed reciprocating compressors (greater than 400 rpm) for plant duty.
Justify low speed reciprocating compressor use based on site, company and industry experience during the early phases (pre-FEED) of the project.
The use of medium or high compressors in plant service has resulted in $MM in lost revenue and maintenance costs.
One plant’s maintenance cost for high speed reciprocating compressors (1200 rpm), in sour gas service, was five times the cost of low sour gas as the high speed compressors.
Benchmarks: This practice used since 2000, in all project which use lubricated reciprocating compressors in plants. This action has resulted in reliabilities for lubricated reciprocating compressors in excess of 96%.
Best Practice-7: DO not use lubricated screw compressors in sour gas services or when C6+ components are present in the gas analysis.
Accurately determine the gas sample in the pre- FEED phase and note if the gas is sour (contains any H2S, CO2 and/or has carbon components C6+). If either of the two characteristic mentioned before are present, do not use a lubricated screw compressor. Lube oil in these services will deteriorate (mineral or synthetic will be equally affected), hence reducing the life of bearings, and quickly plugging the lube oil filters, requiring frequent attention (often hence per day).
The use of lubricated screw compressor types in sour gas services and/or where C6+ components are present has results in the following consequences:
- – Need for daily lube oil filter changeover
- Shutdown of compressor weekly for bearing change and complete oil system cleaning (one week duration)
- – Legal suits costing $MM.
Benchmarks: This best practice has been implemented since 2000 when multiple lubricated screw compressor field issues were experienced.
Best Practice – 8: Centrifugal horizontal split casing compressor nozzles should be located on the bottom of the casing for lowest mean time to repair (MTTR).
Be sure to justify either horizontal split with bottom nozzles or radial split casing with top nozzles to minimize time to repair (MTTR) during the pre- feed phase of the project:
- – The time to disassemble a horizontal split case with bottom nozzles is 1 – 2 days maintenance time.
- – The time to remove the inner casing from a radial split compressor is 2 – 3 days maintenance time.
- – The time to disassemble a horizontal split case with top nozzles considering piping removal and re-assembly without exceeding external piping forces and moment limits is 7 days or more.
Benchmarks: This best practice has been used since 1990s. It saves 4 – 5 days maintenance time and therefore increase daily revenue for this time, potentially in excess 10MM additional revenue. If the driver is condensing steam turbine use bottom nozzles since the height of the compressor platform will be set by the condenser size and condensate pump NPSH requirements.
– For all type of driver that are planned to be installed at ground level, use barrel type compressors with top nozzles – the capital cost compared to horizontal compressor is approximately 10% increase.
Best Practice – 9: During pre- bid project phase (before priced vendor bids are received), screen for proven impeller or blade row flow, head coefficient and tip speed experience on similar gas compositions.
Centrifugal and axial compressors are custom designed, but should incorporate proven impellers or blade and gas path parts. The success of the factory acceptance test (FAT), field start-up and process life safe and reliable compressor reliability is directly related to impeller, blade and gas path component integrity.
As a result, Vendor requirements should be noted in the invitation to bid (ITB), to provide the following parameters for each stage (impeller or blade row), This will ensure optimum FAT results and field safety/reliability:
- – Flow coefficient experience
- – Head coefficient experience
- – Tip speed experience
- – Impeller inlet (eye) Mach number experience
- – Above parameters for similar gas compositions
Once impellers or blade rows are designed and operating, field changes are difficult, time consuming. Will produce revenue loss and difficult to confirm, since field instrumentation does not have the same accuracy as FAT instrumentation.
Benchmarks: This best practice used since 1970s to achieve problem free FATs, smooth start-up and optimum centrifugal and axial compressor safety and reliability.
Best Practice – 10: Limit closed centrifugal compressor head per impeller values for optimum impeller performance and reliability in the project specification.
Limiting centrifugal compressor impeller head per stage will ensure the optimum flow range for each impeller and the overall compressor performance flow range. In addition, it will ensure that all impeller performance stresses are within safe, proven limits that will result in failure-free impeller life. Limit heavy gas applications (molecular weights above 40) to 24 KJ/Kg or 8000 Ft-lbf per impeller. Limit other services below 40 molecular weight to 3000 meters or 30 KJ/Kg or 10,000 lbm per impeller.
In applications where the use of these head limits forces the compressor into 2 cases FT-lbf/lbm, be prepared to justify this requirement based on the loss of daily revenue. Reducing the head per impeller in the field will either require reduce operating speeds and proportional lower revenue or a new compressor train design. Do not accept impeller heads above these values even if vendor experience is proven.
Failure to meet predicted performance curve flow limits beyond the design point and field impeller failure have occurred when the above head limits were exceeded compressor design. Centrifugal compressor head vs flow performance curve fell off sharply in head and flow beyond the rated point, but had to be accepted since the centrifugal curve shape is not guaranteed. This reduction in flow means a lower production rate for the life of the compressor, which amounts to be significant loss of revenue. Closed impeller failures in the impeller cover, hub side plate and scallop failure (between the impeller vanes at the outside diameter on the cover or hub) were experienced when the head per impeller limits were exceeded.
Benchmarks: This best practice used since 1970 and it results in:
– Meeting or exceeding predicted compressor predicted curve shape beyond the rated point.
– Fault-free closed impeller performance (zero compressor impeller failures)
Best Practice – 11: Require centrifugal compressor head rise to be a minimum of 5% in order to prevent control and protection system issues in the field.
Centrifugal compressor head rise is defined as the head at surge condition divided by the head at rated point. The lower is the head rise, the more rapid a change in flow for change in the head required by the process. Review each proposed impeller head rise during the pre- bid phase, and require the vendor to re- select any impellers that’s have a head rise that is less than 5%. Note that for heavy gas applications, greater than 40 mol. Weigh, this may be difficult and in those cases, acceptance of the highest impeller head rise available from the vendor will have to be considered, based on stated vendor experience and possible discussion with end users.
Benchmarks: This best practice has been used since 2000, resulted in optimum compressor safety and reliability.
Best Practice – 12: Select individual impellers or blade rows to operate as close to best efficiency point (BEP) as possible.
Select individual impellers or blade rows to operate as close to best efficiency point (BEP) as possible. This will ensure:
– Maximum overall compressor curve flow range
– Trouble free heavy gas impeller operation for gas molecular weights above 40.
Ensure that each proposed impeller, when operating at the rated condition is a maximum of +/-10% from its best efficiency point and +/-5% from its best efficiency point for gases above 40 molecular weight. Agree to review and not retain the individual vendor individual performance curves since they are proprietary. Many compressor vendors can presently (2010) manufacture each impeller to operate at its best efficiency point when operating at the specified compressor rated operating point.
Benchmarks: This best practice has been used since 1980 and resulted in trouble free factory acceptance test (FAT) and compressor field reliabilities in excess of 99.7%
Best Practice – 13: Gas density changes of more than +/-20% affect centrifugal compressor performance curve (head vs actual flow) shape. Require the vendors to re-issue the performance curve in these instance.
Impeller head is produced by the velocity of the impeller tip and the tangential velocity of gas through the impeller (Euler’s turbo equation). Unlike pump impeller head, where the liquid is in-compressible, compressor head is affected by the gas density. This is the reason that a centrifugal pump can be tested on water and, provided the field liquid is not viscous, will yield exactly the same performance head vs flow curve in the field. By Euler’s turbo equation, the tangential velocity of the liquid through a pump impeller is the same regardless geometry (flow area) is unchanged. However, if the gas density in a centrifugal impeller or axial blade changes by more than 20%, the centrifugal velocity of the gas will significantly change and the individual impeller performance curve shape (head vs flow) will change.
There are many cases where the field performance checks, carried out where a compressor was thought to be in need of maintenance, have showed that it was the performance curve that was in error, because the gas density had changed by more than 20%. Only the vendor can re-issue a new overall performance curve because only the vendor is in possession of the individual impeller performance curve.
Benchmarks: This best practice has been used successfully since 1990s, for field testing prior to turnarounds. In some cases, field performance where gas density has changed by 20% or more was in fact acceptable, when compared to the reissued curves for the correct gas density. In these cases, time-consuming compressor disassembly was averted, thus saving turnaround time and allowing additional production revenue (3 – 7 days depending on the type of compressor case and nozzle orientation – top nozzle orientation on horizontal split compressor cases being the longest time).
Best Practice – 14: Positively define compressor fouling (debris accumulation on impellers and gas passages) by performance calculations on impellers and phase angle changes (heavy spot of unbalance).
Performance deterioration (efficiency and head reduction) can be caused by:
– Gas density changes
– Internal stage seal wear
– Assembly error
Fouling will have the greatest effect on compressor efficiency, reducing its value by 10% or more. Fouling can also be determined by monitoring phase angle change on the rotor since fouling on the impeller will break off irregularly causing unbalance and change of the heavy spot (phase angle change). If fouling is the cause of the performance deterioration, it can be usually corrected without compressor disassembly, thus saving downtime and revenue losses.
Failure to monitor performance and phase angle of vibration has caused many compressor disassembles when the fouling could have been removed by cleaning the compressor internals without disassembly. Performance deterioration usually dictates that the compressor must be disassembled. This may in fact be the case, but if the rotor phase angle is not monitored, a fouled compressor may be disassembled unnecessarily, causing significant amount of downtime (3 -7 days) and corresponding loss of revenue.
Benchmarks: This best practice has been used since 1980s. The best practice of using on line and off line washing in upstream, downstream chemical and refinery compressor applications has been used since the 1980 to prevent compressor disassembly during a turnaround.
Best Practice – 15: Restrict the use of low solidity diffuser (LSD) to positively eliminate gas disturbances which can result in reduced centrifugal compressor safety and reliability.
Low solidity diffusers are used to direct the gas out of the impeller in a radial direction and not in log spiral, as vaneless diffusers do. The results is reduced friction in a LSD, producing increased head and efficiency, but less flow range when compared to a vaneless diffuser. If the fixed angle of LSD vanes do not match the impeller exit gas angle, gas disturbances can occur (stall cells) at frequencies less than approximately 40% of operating speed, which can excite the rotor to produce large external forces and shaft vibration. The forces exerted on the rotor by LSD stall increase with the molecular weight and density of the gas, and can be large if the molecular weight of the gas is greater than 40.
Benchmarks: This best practice has been used since 1990s.
Best Practice – 16: Use Integral geared centrifugal compressors only for spared compressor applications.
We do not recommend that un-spared integral geared compressors are used for un-spared process duty because:
– Integral gear centrifugal compressor were originally designed for plant and instrument air service, which would use spared compressors.
– They are maintenance intensive since they use multiple bearings, seals, gear meshes and operate at high speeds (above 50,000 RPM) for the last stage)
– Because the gas is cooled after each impeller stage, their performance characteristics and reliability are dependent on intercooler condition.
– Since they are typically supplied with only an overall surge protection system and not individual impeller stage protection, they are prone to surging if intercoolers do not attain design heat removal requirement.
Benchmarks: This best practice has been used since 1990s and has successfully convinced project teams not to use integral geared compressors for un-spared process applications.
To be continued..