May 2021

Special Focus: Maintenance and Reliability

Restaging/rerating of centrifugal compressors: Fundamentals, practices and challenges

The primary requirement for rerating or revamping an existing centrifugal compressor is to match the new operating requirements, such as increased or decreased flow, increased or decreased head, or a combination of both. Rerating or revamping is necessary where the existing compressor or compressor train is incapable of meeting the new operating requirement efficiently.

The primary requirement for rerating or revamping an existing centrifugal compressor is to match the new operating requirements, such as increased or decreased flow, increased or decreased head, or a combination of both. Rerating or revamping is necessary where the existing compressor or compressor train is incapable of meeting the new operating requirement efficiently.

The rerating/restaging of operational existing compressors is an attractive cost-effective solution to address the latest development in process, changes in gas composition, debottlenecking of the existing plants and production maximization. Successful rerating/restaging plays a significant role in capital cost optimization processes.

Brief descriptions of fundamental concepts, practices, challenges and process checklists involved in the process of rerating/restaging of centrifugal compressors are discussed here.

Thermodynamic analysis

Centrifugal compressors are sensitive to operating conditions, molecular weight, gas properties, etc. However, the new operating environment may call for different operating scenarios, different molecular weights and different gas properties, as described in the subsequent paragraphs. During the restaging study, the adequacy of an existing compressor system will be analyzed and the requirements for modification or opportunities for improvement will be identified.

In centrifugal compressors, a decrease in inlet pressure will shift the operating envelope toward a lower flowrate, and the surge margin will decrease in proportion with the decrease in suction pressure. The compressibility factor (Z) also increases as the suction pressure/inlet pressure decreases, as the reduced pressure decreases while the reduced temperature remains constant. A decrease in inlet pressure decreases the gas density, which may result in higher power consumption.

Lower suction pressure results in a lower Reynolds number, influencing the boundary layer and friction loss. At a lower Reynolds number, friction loss will be higher, resulting in lower performance. As a result, a small fall in the inlet flowrate due to higher pressure at the inlet piping/nozzle often requires reduced rotational speed. On the other hand, higher suction pressure will result in higher discharge pressure at a given speed. As a result of a change in composition, the compressor operating envelope will change and consequently calls for revised operating procedures.

Temperature

An increase in the inlet/suction temperature will result in an increased discharge temperature; however, the influence of suction temperature is important when approaching the stonewall point (≥ mach flow). When the suction temperature is high at choke flow, the pressure ratio is reduced. The molar heat capacity has non-linear propositional relation with temperature, and an increase in temperature will result in a lower specific heat ratio. Additionally, higher inlet temperature will result in higher compressibility and lower gas density.

Lower suction pressure results in a lower Reynolds number because of lower density and influences on the induced boundary layer and frictional loss in the flowing path, leading to lower overall efficiency. When suction pressures increase, discharge pressure will increase and must ensure the sufficient flow/adjust speed to avoid surge. TABLE 1 summarizes the impact of lower suction pressure and composition change.

The high specific heat ratio will result in higher discharge temperature and slightly lower discharge pressure at the same pressure ratio. On the other hand, lower specific ratio gases consume more specific power for a given pressure ratio.

Polytropic head will increase along with specific heat ratio. The head is increasing proportionally with the volume polytropic exponent, which in turn increases as the specific heat ratio goes up. However, the influence of the pressure ratio on the head value is less significant compared with volume polytropic exponent effect. The impact of gas heat ratio on the compressor head becomes more pronounced at high rotational speed. Operating range decreases as the specific heats ratio of the handled gas reduces.

Polytropic head is influenced by gas constant (R), which decreases as molecular weight increases. Compressor efficiency is also affected by gas constant. The high molecular weight value leads to a reduction in the gas constant, yielding a lower volume polytropic exponent, which is inversely related to the stage efficiency. The required polytropic head of low molecular weight tends to increase as the flowrate reduces. This indicates higher frictional losses associated with low gas density and viscosity.

The effect of gas molar mass on the compressor head causes the required number of mechanical stages to vary. The low molecular weight gas raises the required head to achieve the desired discharge pressure, leading to greater pressure coefficient on the impeller blades.

During rerating, the probability of formation of deposits/hydrates on the impeller and diffuser surface of the compressor must be evaluated to find the suitable solvent injection requirements to avoid costly, time-consuming offline cleaning. Normally, the amount of solvent required should not be more than 3% of the total flowrate, and excessive flow can lead to serious erosion problems.

Settle-out pressure should be estimated for a rerated/restaged compressor system and the design adequacy of the existing system—including the knockout drum (KOD), inter-stage coolers, piping system, pressure relief valve, blowdown system and associated systems like the flare system—should be checked. Vibration analysis of the piping system is required to avoid acoustic-induced vibration/flow-induced vibration.

Requirements of alarm and trip schedule changes should be addressed based on the rerated or restaged compressor operating conditions and composition. The adequacy and requirements of the lube oil system, sealing system, bearing cooling water system, inter-stage cooler load and KOD must be checked. Brief adequacy check requirements are highlighted in TABLE 1.

While rerating with material of different gas composition, compatibility with respect to the new composition should be studied. As per API 617,1 casing should be radially split when the partial pressure of hydrogen (H2) at maximum allowable working pressure (MAWP) exceeds 200 psig. The presence of hydrogen sulfide (H2S) in the gas composition should be considered to check the material compatibility, as per NACE MR 0103, NACE SP 0472 and NACE MR 0175. Gas service where the partial pressure of H2 exceeds 100 psig or H2concentration exceeds 90 mol% at pressure should be considered as hydrogen service, and the compatibility should be checked during the rerating of compressors.

Compressor discharge temperature should be limited to 150°C (302°F) for natural gas services to avoid polymerization of hydrocarbons (fouling in the impeller), possible decomposition/cracking, auto-ignition hazards and seal/lube system limitations, even though the compressor components can withstand higher temperatures. Discharge temperature should be estimated based on the new composition and operating conditions.

The following rules of thumb have been deduced for ease of understanding and application:

  • Rule of thumb 1: The head of a compressor varies as the square of the tip speed and flow handled by the compressor varies with tip speed and impeller diameter.
  • Rule of thumb 2: A 1% increase in speed corresponds to a 3% increase in flowrate.
  • Rule of thumb 3: A velocity of 38 m/sec–45 m/sec is generally considered as the upper limit for an inlet nozzle operating on air or gases with acoustic velocities like air.
  • Rule of thumb 4: Higher molecular weight or lower temperature may reduce the allowable flow through the nozzle.
  • Rule of thumb 5: Volumetric flow and molecular weight define the size of the compressor.
  • Rule of thumb 6: In general, reduced gas molecular weight reduces pressure ratio surges at lower flowrates.
  • Rule of thumb 7: Lower suction pressure affects the overall efficiency of the compressor.
  • Rule of thumb 8: Excessive fluctuation in the molecular weight causes change in the incidence angle to the entry of the diffuser vane.

Compared to an isentropic process, the discharge temperature at the impeller exit is greater in the polytropic process. Therefore, the provided work by the impeller is higher in the case at constant discharge pressure. Even though the exit pressure of both polytropic and isentropic processes are equal, the discharge temperature and enthalpy differences are greater in the polytropic process. Discharge temperature in both isentropic and polytropic processes is a function of the pressure ratio, suction temperature and process exponent. The polytropic process is a function of inlet temperature, pressure ratio, compressibility factor and molecular weight.

Fouling/scaling

Fouling is normally accompanied by lower efficiency and decreased head due to changes in aerodynamic performance and flow restrictions. One of the early warning symptoms is a decrease in the amount of turndown to surge, i.e., an increase in the minimum flow. This is normally caused by deposits in the diffuser/guide vane, which offers flow restrictions that often result in surging. Online chemical washing/offline removal techniques improve the performance of the compressor. This requirement can be identified during the design stage and appropriate provisions can be provided. Water-based or petroleum-based solvents can be used to dissolve the contaminations.

  • Rule of thumb 9: Generally, petroleum-based solvents are not useful to remove salty deposits.
  • Rule of thumb 10: The amount of injected solvents should not be more than 3% of the total flowrate.

Capacity enhancement

Existing compressor system capacity can be enhanced by the following methods or a combination of the following methods:

  • Utilization of design margin
  • Addition of stages/impellers
  • Optimizing impeller tip speeds
  • Adding a parallel compressor train
  • Optimizing the impeller design
  • Suction boosters
  • Eliminating the compressor losses.

DIMENSIONLESS PARAMETERS ANALYSIS AND FEASIBILITY ANALYSIS

The following dimensional number analysis (TABLE 2) will provide the scope for restaging/rewheeling of the compressor. A rationalized approach to describe the aerodynamic characteristics of compression machinery can be used for analysis.

Polytropic head per impeller

As per Simmon’s method,2 the maximum polytropic head can be approximately 30 kJ/kg (3,058 m) for an operating pressure < 100 bar, whereas the maximum head is limited to 20 kJ/kg (2,038 m) when the discharge pressure exceeds 100 bar (FIG. 1). In low-pressure services (> 100 bar), the maximum head values must be corrected based on molecular weight using the Brown method.

FIG. 1. The maximum polytropic head can be approximately 30 kJ/kg (3,058 m) for an operating pressure < 100 bar, whereas the maximum head is limited to 20 kJ/kg (2,038 m) when the discharge pressure exceeds 100 bar.
FIG. 1. The maximum polytropic head can be approximately 30 kJ/kg (3,058 m) for an operating pressure < 100="" bar,="" whereas="" the="" maximum="" head="" is="" limited="" to="" 20="" kj/kg="" (2,038="" m)="" when="" the="" discharge="" pressure="" exceeds="" 100="">

As per the Brown method, the value of 30 kJ/kg (3058 m) is to be used for molecular weight between 28 and 30. For molecular weight above this value, 0.3 kJ/kg (31 m) is to be subtracted from this head value for every unit increase in molecular weight, whereas 0.6 kJ/kg (61 m) head is to be added for every unit decrease in molecular weight. In general, a single-stage closed impeller compressor can raise the head up to 42 kJ/kg (4,281 m).

The change in molecular weight will have an impact on the number of stages (FIG. 2) of the compressor since a lower molecular weight calls for more required head to achieve the desired pressure, leading to greater pressure coefficient on the impeller.3,4,5,6 It is always better to review the number of impellers/stages required with respect to all possible operating scenarios and molecular weight to avoid unnecessary cost and time delay at a project’s later stages.

FIG. 2. The change in molecular weight will have an impact on the number of stages of the compressor since a lower molecular weight calls for more required head to achieve the desired pressure, leading to greater pressure coefficient on the impeller.
FIG. 2. The change in molecular weight will have an impact on the number of stages of the compressor since a lower molecular weight calls for more required head to achieve the desired pressure, leading to greater pressure coefficient on the impeller.

  • Rule of thumb 11: Polytropic head per impeller and the number of stages of a compressor depend on the molecular weight of the gas being compressed.

Optimum impeller tip speed

Optimum tip speed depends mainly on the type of impeller, molecular weight, material strength and tip mach number. TABLE 3 can be used for preliminary design. However, for corrosive and low-temperature service [below –50°C (–58°F)], maximum tip speed is normally limited to 250 m/sec even though the molecular weight is < 35. Apart from these criteria, rotational stress, critical speed (mechanical resonances) and driver capabilities must be considered.7,8,9

During rerating, tip speed can reach up to 274 m/sec with careful engineering. Gas acoustic velocity is directly related to specific heat ratio and temperature and inversely propositional to molecular weight. Higher molecular weight gases are associated with lower acoustic velocities and lower allowable tip speed. It should be noted that the impeller material stress level is directly proportional to the square tip speed. Material strength often plays a critical role when low molecular weight gas is being handled with high head requirements. Cases exist in which machines are operated at more than machine mach number without any issues—the important factor is the inlet relative mach number, which is proportional to the machine mach number and flow coefficient. Speed shall be optimized with respect to lower/upper critical speeds. The maximum speed is limited by aerodynamic and mechanical limitations.

  • Rule of thumb 12: Impeller maximum speed is limited by mach number, material strength (yield stress at maximum continuous speed) and critical speed (resonance).
  • Rule of thumb 13: Some original equipment manufacturers (OEMs) set the overload limit based on the relative Mach number of 0.96 or lower.

Utilization of design margin

Utilizing the existing overload limits is a common method of optimization. However, such limits are based on experience and function of machine mach number limitations, gas composition, number of stages, etc. It is well known that a compressor running at a high machine mach number or handling high molecular weight gases will have less overall flow range than a compressor that runs at a low machine mach number or handles low molecular weight gases. Therefore, the allowable overload margin may vary. For example, the high molecular weight compressor with an overload margin of 120% will have a 140% design margin when it handles low molecular weight.

Adding additional trains

The addition of parallel compressor trains is quite common in the oil and gas industry; however, deciding the parallel configuration option—a series-parallel (each casing will have independent driver) and tandem-parallel (single driver for multiple casings)—plays a key role in plant reliability, turnaround capabilities and efficiency.10,11,12

In general, a tandem-parallel arrangement requires more casing compared to a series-parallel arrangement for the given flowrate, resulting in higher increased fixed capital. More casing calls for more piping, sealing, lube system and accessories, which adds to the cost in a tandem-parallel arrangement. The loss of a compressor in a series-parallel arrangement is associated with variation in the inter-stage pressure, and the impact should be studied during the engineering stage. The loss of driver in the series-parallel will result in flow instabilities in the other stage. Loss of driver in the tandem-parallel will result in capacity reduction. The tandem-parallel represents a simple control system. Careful analysis of the configuration during the revamp/installation of additional trains should be exercised with respect to reliability, availability, safety, capital cost, space constraint and turndown flexibility.

Number of impellers

The number of impellers can be estimated by dividing the total polytropic head by the maximum head per impeller. The inlet flow coefficient can be improved by improving the eye diameter and impeller diameter ratio. Normally, it is preferred to have the same diameter for smooth flow. However, the maximum number of impellers in a single casing is normally limited to 10 (or 10 impeller stages) due to rotodynamic, aerodynamic, operational and design constraints.

Casing

Horizontally split casings are typically used for lower pressure applications (up to approximately 40 bar discharge pressure), while vertically split (barrel type) casings have successfully been used for discharge pressures up to 800 bar. During rerating/restaging, casing will not usually be modified.

Compressor stability

Both operational and aerodynamic stability should be considered during the rewheeling/rerating of compressors. FIG. 3 presented in API 617 outlines compressor stability requirements.

FIG. 3. The diagram presented in API 617 outlines the compressor stability requirements.
FIG. 3. The diagram presented in API 617 outlines the compressor stability requirements.

Suction boosters

Gas density can be increased by many ways, including suction boosting and refrigeration. Simple suction boosting is considered as an additional stage.

Process challenges and the ripple effect on a process system

KODs, piping/nozzles, relief systems, interstate/discharge coolers, drain and vent systems, emergency shutdown and depressurization systems, alarm and trips schedule, area classification and its impact, impact on material corrosion, lube and seal systems, utility requirements, rerouted piping system, operating envelop, driver capabilities, etc., should be revisited during the restaging/revamping of the compressor system, as tabulated in TABLE 4.

Impeller trimming

Impeller trimming is often used to improve energy efficiency to optimize the oversized compressor and alter the exit mach number of an impeller for different molecular weight.

In radial trimming, the outlet of the impeller will be trimmed in the radial direction by reducing the diameter. Normally, radial trimming is preferred where the reduction in pressure rise across the impeller is required. This is simple and does not require modification in shroud; however, it will change the fluid exit angle, which often calls for diffuser redesign. This option is not preferred for high-speed wheels since it will result in lower efficiency, whereas in low-specific speed wheels, it improves the specific speed value closer to optimum value, and radial trimming does not alter the flow coefficient at the inlet.

Axial trimming involves the reduction of impeller blade height at the outlet of the impeller without changes in the inducer. This is preferred for high-specific speed impellers. Reducing the outlet height will lower the specific speed of the impeller. Trimming to control the losses due to tip gap (clearance between impeller and shroud) should be checked and revisited. It should be noted that the tip gap for the given operating speed is generally based on material distortion, which is greatest at the outer diameter where centrifugal forces have the strongest effect. It is normally expected that axial trimming will result in an efficiency decrease and narrow choke margin but will have no major effect on the flow range of the compressor in the initial stages.

Control of losses

The following losses are normally encountered in the compressors, affecting performance. By controlling these losses, the compressor system can be optimized.

  • Disk friction loss due to adhesive forces between rotating disk and fluid. The shear force acting between the impeller back face and the stationary surface is to be overcome by cost of power. Generally, it increases with rotational speed and impeller exit radius.
  • Skin friction loss due to adhesive forces between the channel surfaces and the fluid. Channel surfaces includes the hub, blades and shroud.
  • Incidence loss caused by the direction of the gas flow diffusing from blade angle. The deviation between the relative inlet angle of the gas and the actual blade angle causes the gas to change its direction, resulting in (energy loss) incidence loss. Incidence loss occurs when the relative fluid flow angle of fluid entering the impeller deviates from the actual blade inlet angle. This loss can be minimized by adjusting the blade angle closer to the relative flow angle, curving of the blade in the direction of the entering flow, and by improving the inducer (initial part of the impeller) design.
  • Blade loading loss due to a pressure difference between blade to blade. Blade loading loss is due to the growth of the boundary layer in the impeller and is highly dependent on the diffusion of working fluid to the impeller. Blade loading loss is a function of the diffusion factor and the tangential impeller velocity. It can be optimized by increasing the rotational speed or by increasing the flowrate. Changing the relative flows (inlet and exit) and the tangential impeller exit speed will result in a change in blade loading loss.
  • Recirculation loss caused by the back flow of fluid to the impeller, which requires additional power to overcome the backflow.
  • Clearance loss due to significant flow leakage through the clearance between the impeller and casing due to pressure difference. This will form a small vortex on the suction side (low-pressure side) of the impeller vane outer tip. It affects both the suction side and discharge side.
  • Leakage loss caused by the leakage of fluid through compressor seals. Seal loss decreases the energy available to convert into pressure head due to internal recirculation inside the compressor.
  • Vaneless diffuser loss in the vaneless diffuser space as a result of friction and the absolute flow angle. Whether the compressor has a diffuser or not, a vaneless space is always directly flowing through the impeller where the flow is diffused. In this space, primary and secondary flows mix, which results in friction loss and other losses referred to as vaneless diffuser loss.
  • Vaned diffuser loss depends on diffuser shape, blade loading factors, incidence angle and surface friction of the vane.
  • Mixing loss is also defined as slip of the flow field against the direction of rotation from pressure to suction side. After the flow has left the impeller, the primary and secondary flows mix. This is not ideal and energy loss and more recirculation will take place.
  • Wake mixing loss due to mixing with jet flow right after impeller exit.
  • Super critical mach number loss due to shock wave loss.
  • Chocking loss due to the relative mach number at the throat.
  • Incidence loss due to the difference between the inlet blade angle and flow angle.
  • Entrance diffusion loss due to the diffusion inlet to throat.
  • Mechanical losses include power dissipated through bearings, seals, shaft-driven lube pumps and gear boxes.

The role of OEMs

OEMs play a critical role in the restaging/rewheeling of compressors. Normally, all mechanical, rotary modifications, rewheeling, internals modification and restaging are carried out by OEMs due to their expertise, experience, facilities and tools. OEMs will conduct the thermodynamic, mechanical, rotary, aerodynamic and vibration studies, and will establish the safe operating envelop and provide performance guarantees. An engineering contractor or in-house engineering team can study equipment and processes upstream and downstream of the compressor unit.

Performance test

A centrifugal compressor’s performance estimation can be done by three methods: the adiabatic method, N-method and Mollier method. In the adiabatic method, the compression process is completely reversible (isentropic), meaning that the input power is completely converted into pressure energy, which is not the practical case. The N-method is named based on a polytropic exponent (n) that is used to estimate discharge conditions and is independent of the state of compressed gas, and the polytropic head is the sum of each stage head, which are not true in adiabatic processes. Mollier diagrams are generally preferred for pure gases.

A performance test run of the rerated/restaged compressor can be carried out as prescribed in ASME PTC 10 and API 617. However, an agreement with the client and the OEM on PTC 10 test class and methodology is precedent. A complete unit test—including casing, gear, driver and auxiliary units—for evaluating performance is always recommended. Shop performance test limitations must be considered.

Decision-making process

A payback period of up to 5 yr can be considered as a potential opportunity for improvement. If the payback period is less than 2 yr, it requires urgent rerating/restaging. However, TABLE 5 details some of the major factors considered during the decision-making process.

Takeaway

Gas molecular weight, gas property, suction operating conditions, upstream and downstream equipment limitations, and existing compressor data are playing a vital role in the analysis of rerating/restaging. Deep analysis on compressor stability is often required during the rerating study. However, rerating/restaging is an attractive and cost-effective solution to maximize the utilization of available resources (optimization) and address the changes in process and plant requirements. Rerating is often used to reduce the capital cost and improve the viability of a project.

Rerating or restaging often calls for multi-disciplinary input and knowledge and should, in general, be carried out in consultation with the OEM. The ripple effect on upstream and downstream equipment and units should be a part of any study. The outlined guidelines and limitations will help designers to carry out a feasibility study of rerating/restaging. HP

LITERATURE CITED

  1.  API Standard 617, “Axial and centrifugal compressors and expander-compressors,” 8th Ed., September 2014.
  2.  Simmons, P., B. Nesbitt and D. Searle, Guide to European compressors and their applications: The complete practical reference guide to compressors design, operation and applications, Vol. 2, Professional Engineering Publishing, London, U.K., 2003.
  3.  Khan, M. O., “Basic practices in compressors selection,” International Compressor Engineering Conference, Purdue University, West Lafayette, Indiana, 1984.
  4.  Ludtke, K., “Rerate of centrifugal process compressors—Wider impellers or higher speed or suction side boosting?” 26th Turbomachinery Symposium, College Station, Texas, 1997.
  5.  Blahovec, J. F., et al., “Guidelines for specifying and evaluating new and rerated multistage centrifugal compressors,” 27th Turbomachinery Symposium, College Station, Texas, 1998.
  6.  Ludtke, K., “Twenty years of experience with a modular design system for centrifugal process compressors,” 21st Turbomachinery Symposium, College Station, Texas, 1992.
  7.  Sorokes, J. M., E. A. Memmott and S. T. Kaulius, “Revamp/rerate design considerations,” 42nd Turbomachinery & 29th Pump Symposium, Houston, Texas, 2013.
  8.  Garcia, D., et al., “Restage of centrifugal gas compressors for changing pipeline landscapes,” IAGT 2015 Symposium, Banff, Alberta, Canada, 2015.
  9.  Gutierrez Velasquez, E. I., “Determination of a suitable set of loss models for centrifugal compressor performance prediction,” Chinese Journal of Aeronautics, October 2017.
  10. Li, P.-Y, C-W Gu and S. Yin, “A new optimization method for centrifugal compressors based on 1D calculations and analyses,” Energies, Department of Thermal Engineering, Tsinghua University, Beijing, China, 2015.
  11. Albusaidi, W. and P. Pilidis, “An iterative method to drive the equivalent centrifugal compressor performance at various operating conditions: Part II Modelling of gas properties impact,” Energies, School of Aerospace, Transport and Manufacturing, Cranfield University, Bedfordshire, U.K., 2015.
  12. Botha, B. W. and A. Moolman, “Determining the impact of the different losses on centrifugal compressor design,” R & D Journal, 2005.

The Authors

Related Articles

From the Archive

Comments

Comments

{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}