Process Discipline FEED for Centrifugal Compressors on Oil & Gas Installations
TABLE OF CONTENTS
1.2 The Impact of Process Design Quality
1.3 Criteria for Successful Process Engineering FEED
1.4 Process Flow Diagram (PFD) and Heat and Material Balance.
1.5 Process and Instrumentation Diagrams (P&IDs).
1.5.1 Process Safeguarding Systems Design
1.6 HAZOP Report and SIL Evaluation
1.7 The Criticality of FEED HAZOP for Project Success
4. Suction Knockout Facilities
6.1 Functions and Configuration of Surge Control System
6.1.2 Compressor Discharge Volume
6.1.4 Compressor Recycle Cooler
7.3 Pressure Equalization (Settle-out)
1. Introduction
This article on centrifugal compressors is one of a series which I am writing to assist process engineers engaged in the front-end engineering design (FEED) of the major process components employed on oil and gas production facilities. These are:
· Wellheads and flowlines.
· Production separators.
· Centrifugal compressors.
· Centrifugal pumps.
· Heat exchangers.
The term “component” is from API RP 14C (2017) and is used for reasons which are explained below.
The inspiration for this work comes from two main sources.
The first was my personal experience.
After graduating with a bachelor’s degree in chemical engineering, I was a process engineer in the hydrocarbon industry for 35 years. It was my fortune to ride the wave of design activity in the North Sea from the ‘80s onwards.
I struggled throughout to resolve the tension between quality and time. There was rarely enough time to do everything that seemed necessary; and most detailed design projects finished late and over budget. This was an existential issue for me since I worked mainly as a freelancer. I became focussed on discerning what data and activities were necessary for a successful process design, and the order in which activities needed to be carried out.
The second inspiration is a book by Merrow (2011) which I happened upon a couple of years before retirement.
Merrow analysed the performance of 130 oil and gas megaprojects (value exceeding $1 billion). The sample was biased toward Europe and North America.
78% of the projects failed to deliver on the estimates that had been made at project sanction. The average cost growth and schedule slippage were 33% and 30% respectively.
The 22% which were successes were at least as challenging technologically as the failures.
It was interesting to see statistical confirmation of my anecdotal experience of dysfunctional projects.
Merrow identified that project outcomes are particularly sensitive to the completeness of FEED, and in his book he recommends improvements (in project management) aimed at achieving satisfactory completion of FEED.
This work recognizes the critical importance of successful FEED. In it I propose the essential activities that process discipline must carry out during FEED, and the most effective sequence.
1.1 What is FEED?
Oil and gas projects are generally matured in two consecutive phases: planning, and execution.
The planning phase of major projects generally comprises a gated sequence of feasibility studies, conceptual design studies and front-end engineering design (FEED).
The execution phase comprises detail design, followed by construction and plant start-up.
During the planning phase the plant design, the estimated cost and schedule for completion of the execution phase, and reliability of attaining production targets, are successively refined. At each gate the decision is taken on whether to proceed to the next stage.
The function of conceptual design is to arrive at a single scope for development in FEED.
Conclusion of FEED is a major milestone, when the operator´s level of confidence is sufficient to decide whether to sanction investment in the execution phase. In the event of a positive conclusion, the operator canvasses the authorities for sanction to proceed.
Barringer & Weber (1996) cite studies which indicate that 2/3 of project life cycle costs (LCC) are fixed during FEED, and that the opportunity to reduce LCC declines sharply thereafter.
The required accuracy of the FEED cost estimate varies among operators, and according to the magnitude of the investment under consideration.
Merrow (2011) cites typical accuracy of – 15 / 25%.
An example in the public domain is found with the Norwegian Petroleum Directorate (NPD), which requires operators to include a cost estimate in licence applications for execution of projects on the Norwegian continental shelf.
The NPD (2020) normally stipulates a cost estimate of / - 20% within an 80% degree of confidence. This means that if such a project were executed many times, in 8 out of 10 cases the costs would be between / - 20% of the estimate.
1.2 The Impact of Process Design Quality
During FEED, process discipline carries out the core activities necessary for design of the oil and gas production facilities and utilities.
Consequences of deficiency in the process discipline FEED are illustrated by the examples below.
These are from a review by the Norwegian Petroleum Directorate of 26 oil and gas projects on the Norwegian continental shelf, each having a value exceeding $1 billion (NPD, 2013).
Development operators reported performance data for the topsides portion of each project, which gives an indication of the adequacy of the process engineering.
Norwegian Project A: Greenfield FPSO.
According to the operator of the development, 59% of FEED was completed before execution.
Topsides costs during execution were 55% over budget, and production start-up was delayed by 16 months.
Several equipment packages were ordered early due to long delivery times. However, engineering was not sufficiently completed when the equipment was ordered, which led to many changes along the way.
Norwegian Project B: Brownfield subsea tie-backs to host platform.
According to the operator of the development, 100% of FEED was completed before execution.
There was cost growth of 96% and a 100% increase in equipment weight on the host platform, due to underestimating the scope of the required brownfield modifications.
The operator also cited a lack of understanding of the complexity of implementing modifications concurrently with operation of the host platform (“simops”).
A press release from the engineering contractor at commencement of execution announced the scope for new equipment on the host platform was limited to a manifold module for receiving the subsea well streams and tie-ins to the existing process systems. However, an additional production separator was in fact required.
In a follow-up review covering the years 2007-18 (NPD, 2020), development operators highlighted the topsides as particularly challenging in platform developments, with engineering errors and deficiencies being important causes of cost growth and schedule delays.
Topsides incurred large cost increases even in several projects which ended up with overall costs in line with the FEED estimate.
1.3 Criteria for Successful Process Engineering FEED
This article provides recommendations for the process engineer´s essential tasks during FEED for a centrifugal compressor, so that the documentation is fit for detailed design, laying the basis for a successful facilities design.
A successful facilities design is one in which:
· The actual facilities costs during execution are equal to or less than the cost estimate budgeted at the conclusion of FEED.
· The production plant is started up on time, and performs satisfactorily under start up, operation and planned shutdown scenarios. Maintenance can be carried out safely.
· The process safeguarding systems automatically shut down the plant in a safe manner in the event of deviations from the acceptable range of process conditions.
1.4 Process Flow Diagram (PFD) and Heat and Material Balance.
Creation of a PFD is the first task that must be accomplished in FEED of the oil and gas production facilities. The PFD shows the configuration of the main equipment items and control schemes and establishes the foundation for the process and instrumentation diagrams (P&IDs).
Realistic heat and material balances are critical for generation of specifications and sizing calculations for equipment, piping and control valves and safeguarding elements.
On brownfield projects they are used in assessing the suitability of existing equipment for the projected service.
They are input to the material selection diagram, which is generated by a metallurgy specialist to show the basis for selection of construction materials for equipment and pipework.
The PFD is developed by process engineers using as far as possible standard equipment units, to convert the imported well streams into export products.
The reservoir engineering function must define the well streams, by providing a matrix of fluid properties and operating parameters for each well, encompassing the entire service life of the source reservoir (s).
The data required is indicated by the shaded boxes on the figure below. To proceed with the facilities design absent these data is a fundamental error.
If lift gas is provided from a different facility, the supplier should state its full composition and delivery conditions at project battery limit.
The required product specifications and operating parameters at battery limits should be verified against the supply contract.
These data are crucial to enable development of a suitable process configuration, and generation of heat and material balances for the key operating cases.
There should be available a set of material balance data from conceptual design. The reservoir data which forms the basis of the material balance may however have been updated in the interim. It is therefore important to have reservoir engineering confirm these data at commencement of FEED.
The project sponsor’s process engineer needs to be clear that changes in reservoir data during FEED may adversely impact not only process engineering manhours at the contractor, but potentially affect specification of major equipment, with repercussions for layout arrangements and deck weights.
The wise process design engineer at the engineering contractor will spell out that the process manhour estimates and schedule are based on the reservoir input data supplied by the project sponsor.
1.5 Process and Instrumentation Diagrams (PIDs).
FEED P&IDs are developed based on the process configuration that is shown on the PFD. A brownfield project must also include updated modification P&IDs showing tie-ins to the existing plant.
Generally, each P&ID shows an item of production equipment along with its control system and pipework connecting other production and utilities equipment.
For each item of production equipment, safeguarding systems must be provided.
Particular attention should be accorded battery limit interfaces, where other plants or pipelines may present hazard sources.
Design pressures, temperatures, and construction materials of equipment are stated, and are indicated in coded form for the pipework.
Normally the calculated nominal diameters are indicated for major pipework (4” NB and larger).
1.5.1 Process Safeguarding Systems Design
API RP 14C (2017) presents a methodology for designing process safeguarding systems for offshore oil and gas production facilities. It is generally referenced also in design of onshore facilities. It was first issued in 1974 and is currently in its 8th edition.
Grounded in the observation that most threats to safety from the production process involve the release of hydrocarbons, the methodology is focussed on preventing such releases, stopping the flow of hydrocarbons to a leak if it occurs, and minimizing the effects of hydrocarbons that are released.
Based on industry experience, API RP 14C lists for each type of equipment and associated piping (“component” in 14C terminology):
· Undesirable events which potentially lead to release of hydrocarbons.
(“undesirable” is equivalent to “hazardous” in IEC 61511 terminology)
Undesirable events which are applicable to the components considered in this work are overpressure, leak, liquid overflow, gas blow-by, underpressure, excess high temperature, excess low temperature.
· Various causes for occurrence of the undesirable events.
· Detectable abnormal conditions which portend undesirable events at the component.
· Appropriate protective devices, and where they should be located.
Two levels of protection are provided.
The highest order (primary) device acts to prevent escalation of the event, and the next highest order (secondary) device is a mitigation measure in case of failure of the primary device.
The two levels employ functionally different types of safety device to eliminate common mode failure, and they must be completely separate from the equipment control systems.
API Std. 521 (2014) is referenced for guidance on additional causes that should be considered, and evaluation of devices for overpressure protection.
1.5.2 Generic PIDs
Each item of production equipment (i.e. “component” in the terminology of API RP 14C), calls up an exclusive set of safeguarding elements. It is therefore possible to develop for each component a generic P&ID complying with API RP 14C, which can subsequently be finalized with project-specific data.
A generic P&ID for production centrifugal compressors is presented hereunder.
In addition to process safeguarding measures, the generic P&ID illustrates a fit for purpose control scheme and maintenance isolations.
1.6 HAZOP Report and SIL Evaluation
HAZOP is the standard acronym for Hazard and Operability study, which has evolved from pioneering work in the UK chemicals industry in the late 1960s and early 1970s.
The main objective of the HAZOP study is to evaluate the fitness of the plant as represented on the P&IDs, i.e., is the process safely operable? The HAZOP also typically identifies errors and omissions in the P&IDs.
The HAZOP considers the normal operation mode(s) and in addition, the following irregular operation modes:
· Maintenance
· Start-up
· Shutdown
· Commissioning
In cases where brownfield construction is planned concurrent with production operations (“simops”), this mode must also be considered.
The following process engineering documentation is required for the study:
· A set of P&IDs with the development operator´s comments incorporated. In the case of brownfield projects, the tie-in P&IDs must be included.
The P&IDs shall show all production equipment and piping, piping classes, pressure specification breaks and PRV set pressures.
· A PRV schedule, listing the relief cases for each PRV and the determining case.
· A cause and effect matrix to illustrate the design intent of the process safeguarding instrumentation.
· Production equipment and instrument schedules with information on sizes, design temperatures and pressures, and materials of construction.
In preparation for the HAZOP study the production process is subdivided into relatively small sub-systems which are highlighted on the P&IDs and are termed “nodes”.
During the HAZOP study, the process engineer explains for each “node” in turn its design intent.
The study chairperson then applies a standard set of “guidewords” (more, less etc.) to describe positive or negative deviations from normal operating values of “parameters” (temperature, pressure etc).
For each deviation in turn, feasible causes are posited by members of the team.
The team then assesses potential consequences of the deviation without taking credit for safeguarding measures, i.e. assuming all safeguards fail. There may be several consequences which propagate to other sub-systems.
Consequences are considered to be anything that adversely affects:
· Health and safety of offsite populations, operator employees and contractors.
· Environment.
· Operator finances, in respect of equipment, business security and market opportunities.
· Operator reputation.
For each cause / consequence pair, the team lists any safeguarding measures that are already in place on the P&ID.
If the identified safeguards are assessed to be inadequate, then team member(s) are assigned with remedial actions.
1.6.1 SIL Evaluation
A safety integrity level (SIL) evaluation is normally carried out on the process facilities FEED. It may take place concurrently with or following the HAZOP.
SIL evaluation is an exhaustive procedure which is normally led by electrical or PACO discipline and is therefore not detailed here. A simplified description follows.
Each consequence identified in the HAZOP is assessed according to the magnitude of its impact. The likely frequency of its occurrence is also assessed on an order of magnitude basis, without taking credit for any safeguarding measures in place. The inherent risk is then evaluated as the product of impact and frequency of occurrence.
To mitigate the risk to a value that is tolerable, requires a safety instrumented function (SIF) with a defined SIL, which operates independent of the process control systems.
For example, if the risk needs to be reduced by a factor of 100, a SIL 2 SIF is required; for reduction by a factor of 1000, a SIL 3 SIF is required.
1.7 The Criticality of FEED HAZOP for Project Success
It is important to conduct a HAZOP towards the end of FEED, to eliminate the need for major changes during detail design and construction. FEED HAZOP studies typically uncover a significant number of safety deficiencies in the process design which require redesign.
If the process design is otherwise incomplete at the time of HAZOP, the situation is greatly exacerbated.
In an ideal world, engineering solutions would be developed and incorporated in a final revision of P&IDs issued for detail design. However, at this late stage of FEED the manhour and schedule budgets are often exhausted, and a list of deficiencies is deferred to detail design for resolution.
Detail design commences with a set of P&IDs which are frozen procedurally; modifications which follow from redesign need to be justified through the project management of change system.
Often several unplanned issues of P&IDs are necessary as detail design gets under way. It becomes very difficult to synchronize the work of mechanical, piping and other disciplines which require process input. This results in cost growth and schedule delay.
Clearly it is very favourable to minimize the number of process design deficiencies discovered during HAZOP.
The present work proposes a generic P&ID for centrifugal compressors.
The methodologies described in API RP 14C and API 521 are applied during development of the P&ID so that the number of deficiencies discovered in the HAZOP study is diminished, the potential for re-work during detail design is reduced, and the likelihood of a successful project is enhanced.
2. Centrifugal Compressor Principles of Operation
A centrifugal compressor boosts the kinetic energy (i.e., dynamic pressure) of gas flowing into the eye of the impeller by increasing its radial and tangential velocity. The accelerated gas then passes through the static diffuser, where the increased flow area causes controlled deceleration of the gas, resulting in conversion of the dynamic pressure to static pressure. The compressed gas then flows to the outlet piping.
The performance of the compressor at constant speed is demonstrated by a characteristic curve, which is usually expressed in terms of static head versus inlet volume flowrate.
The piping system into which the compressor discharges generates a static pressure curve which varies according to gas flowrate. The intersection of the piping system curve and compressor performance curve is defined as an operating point. See GPSA (2017) for illustration.
The operating point can be altered either by:
· changing the location of the characteristic curve, by changing the properties of the inlet gas or the speed of the compressor; or
· changing the piping system curve, by redesigning a control valve for example
The compressor requires smooth, undisturbed flow into the eye of the first impeller, otherwise performance will be reduced. This requirement may be achieved with a straight inlet piping length equivalent to two flange diameters.
3. Process Specifications for Centrifugal Compressor
Customization of compressor internals such as impeller, diffuser, etc., is required for optimal performance. This results in long fabrication times relative to other equipment items; consequently, compressors are often ordered before completion of FEED.
The following list of process variables needs to be provided to the manufacturer to enable the custom design:
· Gas mass flow
· Complete gas analysis
· Toxicity assessment of gas
· Inlet pressure
· Inlet temperature
· Discharge pressure
It is important to provide the manufacturer with the full range of expected compressor operating cases; each case is a discreet data set encompassing all the above variables.
As a minimum the cases considered should include the following from the overall installation PFD: Maximum Oil: Maximum Water: Maximum Gas.
The frequency and duration of sub-cases should be indicated, for example, summer / winter operation.
At the front end of oil and gas process plants, gas properties vary short-term depending on concurrent well selection, and changing well properties in the medium to long term. These variations in addition to episodes of surging flow from the wells impose challenging conditions on compressor control and surge prevention systems.
Particular attention is therefore warranted when defining flow cases at the front end of the plant.
The presence of corrosive components such as water, hydrogen sulfide, carbon dioxide, chlorides (due to carryover of saline liquid from production separator), should be highlighted to the manufacturer to enable selection of appropriate materials of construction for compressor internals and casing.
Any anticipated variation in the internal operating environment due to off-design operation, process upsets or chemical cleaning during shut down should also be defined.
The manufacturer should provide the following calculated data:
· Inlet gas Cp/Cv (Note 1)
· Inlet gas compressibility (Note 1)
· Inlet gas volumetric flow (Note 1)
· Outlet temperature
· Outlet gas Cp/Cv
· Outlet gas compressibility
· Polytropic head
· Brake power including losses, at performance guarantee point(s)
Note 1
These key variables are calculated from equations of state based on gas composition, pressure, and temperature, and are used by the manufacturer to design the compressor internals.
The manufacturer should therefore be responsible for evaluating them, using their preferred equations of state.
Prior to manufacturer’s data becoming available, it will be necessary for the process engineer to estimate compressor power requirements and the outlet gas properties. Usually, a computer simulation tool will be available to carry out these calculations. The underlying calculation principles and manual methods can be found in the GPSA Data Book (2017).
The process engineer shall indicate ambient conditions. For facilities located offshore and coastal locations onshore, NORSOK M-001 (2014) requires that the atmospheric environment be considered wet, with the condensed liquid saturated with chloride salts, when selecting materials and coatings to prevent chloride induced corrosion of equipment.
4. Suction Knockout Facilities
Suction knockout facilities are provided upstream of the compressor inlet. The facilities comprise a scrubber vessel, and optionally a cooler which shares the duty of recycle gas cooler as discussed in 6.1.4 below.
The purpose of these facilities is to remove droplets of condensed water and liquid hydrocarbons carried by the gas. Typically, the scrubber vessel is specified to deliver gas containing maximum 13.4 ml per 103 Sm3 (0.1 US gallons per 106 Sft3).
The inlet scrubber design should consider the compressor recycle case.
A vane pack anti-entrainment device fitted in the scrubber is preferred over a wire mesh demister, since the latter introduces hazards of loose wire entering the compressor; demister pads are also subject to progressive clogging, which restricts flow and may impair the effectiveness of anti-surge control.
The line between the knock-out vessel and the compressor inlet should be as short as practicable, be free of pockets, and should slope back towards the knock-out vessel.
The inlet line should be ensured free of condensate during operation and shut down, by insulating and heat tracing the line.
A single (not spared) PSV sized for pool fire should be installed. The inlet line to the PSV should be connected on the scrubber vessel below the level of the vane pack, so as not to dislodge it if called upon to operate.
5. Throughput Control
Depending on process objectives, throughput control can be achieved by regulating either suction flow or discharge pressure.
On oil and gas installations the compressor throughput control is more commonly targeted at meeting downstream pressure conditions, and discharge pressure control is employed.
5.1 Fixed Speed Drive
Where a fixed speed drive is installed, throughput control may in principle be achieved using inlet guide vanes, discharge throttling or suction throttling.
Guide vanes are only suitable for clean gases, and therefore should not be used in hydrocarbon gas service.
Suction throttling is usually selected for throughput control of fixed speed compressors because it is more efficient and offers better turndown capability than discharge throttling.
The suction throttling valve is located downstream of the compressor suction scrubber.
A butterfly valve is commonly used in throttling applications.
A pressure controller downstream of the compressor regulates throughput by modulating the suction throttling valve.
5.2 Variable Speed Drive
When a variable speed drive is used, a pressure controller downstream of the compressor regulates throughput by varying the rotational speed of the compressor driver.
The pressure controller provides the set point of the speed controller. Depending on whether the driver is a gas turbine or an electric motor, the output of the speed controller either regulates the gas flow to a turbine, or it varies the electrical frequency of the power to the motor.
6. Surge Control
There exists on the performance curve for a given compressor speed, a (low) flowrate known as the surge limit, where the static pressure generated by the compressor is not sufficient to balance the pressure in the discharge piping system. Momentary reverse flow to the compressor results, followed rapidly by recommencement of forward flow as pressure balance returns. The rapid oscillation of reverse and forward flow results in vibration which may damage the compressor and pipework. The potential for severity of damage increases with compressor discharge pressure, so gas injection compressors are particularly vulnerable.
It is necessary to develop an effective surge control philosophy during FEED. A specialist vendor will finalize the surge control system during detailed design, and the system will be tuned during offshore commissioning.
On a map of pressure rise versus inlet flowrate, the objective of surge control is to keep the operating point to the right of the surge limit at a given compressor speed. This is achieved by recirculating a portion of the flow from the compressor discharge to suction by means of a pipework loop incorporating a recycle valve. The surge control system opens the recycle valve at a set point which gives the desired offset of the operating point from the surge limit.
This surge prevention strategy requires flow measurement at the compressor inlet, with temperature and pressure measurement at the compressor suction and discharge.
In the case of a variable speed compressor, there is a range of set points for the surge controller which generates a surge control line parallel with and to the right of the surge curve. This prevents surge conditions for all operating points, whilst minimizing unnecessary recycle flow, thereby promoting higher operating efficiency. See Staroselsky and Ladin (1979) and McMillan (2010) for development details.
6.1 Functions and Configuration of Surge Control System
The recycle / surge control system is required to operate satisfactorily in three different circumstances:
· Compressor start-up.
· Gradual approach to surge due to incremental changes in suction or discharge conditions.
· Precipitous approach to surge due to rapid process changes, such as emergency shutdown of the driver or closure of a downstream block valve.
These requirements unfortunately entail conflicting recycle valve functionalities.
In some cases, a separate recycle loop must be provided to cater for the precipitous approach to surge, though in oil and gas installations it is usually possible to use a single recycle loop. Botros (2011) discusses considerations that facilitate selection of single or dual recycle loops.
The elements of a surge control system are:
· Surge controller
· Compressor inlet flow meter
· Compressor discharge FSV
· Recycle line incorporating recycle valve.
Where compressors are installed in series or parallel, each compressor should have a dedicated surge control system. The discharge FSVs and the recycle take off and return lines are configured to allow independent operation of the separate recycle loops.
6.1.1 Inlet Flow Meter
Various flow measurement methods are available for the surge control system, such as orifice, venturi, and ultrasonic meters. Each has relative strengths and weaknesses in respect of cost, reliability, and accuracy. PACO discipline is responsible for selection of the optimum technology.
To function properly, each type of meter requires certain straight pipe lengths upstream and downstream, expressed in multiples of pipe diameter. The process engineer ensures that the appropriate lengths are marked on the P&ID, so that they will be properly incorporated on the Piping Layout drawings.
It is important that the meter technology is selected during FEED, because the pipe diameters involved in compression systems are relatively large and therefore have a significant impact on piping and equipment layout.
6.1.2 Compressor Discharge Volume
The compressor discharge volume should be minimized, to maximize the required system response time of the surge control system.
The boundaries of the compressor discharge volume are the discharge FSV, and the recycle valve inlet.
The discharge FSV should be located as close as possible to the compressor outlet, with the recycle line connected immediately upstream of the FSV.
The FSV should be of the damped type.
Paragraph A.9.3.2 of API RP 14C (2017) recommends the FSV be located outside of the compressor module.
6.1.3 Recycle Valve
Ghaisas (2016) recommends that the recycle valve be located within five pipe diameters of the recycle takeoff point if it is feasible.
It should be located at a high point, so that the piping on either side drains away from the valve.
During FEED there may be insufficient data available to specify completely the recycle valve.
For costing purposes, the valve maximum Cv may be estimated by applying a factor of 1.8 to the Cv that is obtained from the compressor operating data at 100% speed. This approach has been found to take care of the most severe surge condition, according to Singleton (2014).
An alternative approach, recommended by McMillan (2010), is to size the recycle valve for full compressor flow with only 70% of the discharge pressure.
McMillan (2010) further recommends:
· A globe valve with linear characteristics should be specified except where the recycle line is too large.
· The total full-scale stroking time of the recycle valve should be less than 1 second.
The recycle valve opens on air failure.
6.1.4 Compressor Recycle Cooler
Compressors on production installations require considerable time to start up from cold, resulting in deferred production and potential penalties for non-delivery of product. It normally makes economic sense therefore to design for the compressor to continue to run (“hot standby”) during outages of upstream or downstream equipment.
During hot standby mode, and during compressor start-up, the recycle valve is open and the discharge block valve is closed, so all compression power serves as heat input to the recycled gas, exacerbating the risk of trips due to excessive discharge temperature.
It should therefore be considered to install a recycle gas cooler, designed for full continuous recycle.
Placement of the recycle cooler within the compressor discharge volume would reduce the required system response time of the surge control system, therefore the cooler should be placed downstream of the recycle valve.
The outlet temperature of the recycle cooler should be specified equal to the normal compressor inlet temperature. Credit should be taken for Joule-Thomson cooling across the recycle valve when considering the cooler inlet temperature.
In case there is a series compressor upstream, the recycle cooler duty can be shared with the outlet cooler duty of the upstream compressor, with due consideration given to the thermal and hydraulic design of the shared cooler for all operating scenarios.
The recycle gas tie-in to the cooler would be located downstream of the discharge FSV of the series compressor, to prevent potential backflow of the recycle gas towards it.
7. Process Safeguarding
7.1 Shutdown Valves
Shutdown valves (SDVs) should be located:
· on the compressor discharge line, immediately downstream of the discharge FSV, and
· on the compressor inlet line, upstream of the recycle line return connection.
The discharge SDV, discharge FSV, recycle valve, compressor casing and connecting pipework should be rated for the maximum discharge pressure that can be generated by the compressor, defined as maximum suction pressure plus the head at surge at maximum compressor speed. Maximum suction pressure is taken as the upstream trip pressure.
The inlet SDV should be rated according to settle-out conditions (see 7.3 below).
The SDVs should fail to the closed position.
This design strategy when combined with an adequate Safety Instrumented Function (SIF) on the compressor discharge SDV (see 7.2 below), eliminates the requirement for PSVs on the compressor suction and discharge lines, as discussed in paragraph A.9.2.2.2 of API RP 14C (2017).
Paragraph A.9.3.5 of API RP 14C (2017) recommends that the SDVs be located outside of the compressor module.
The pressure system volume between the compressor inlet and discharge SDVs will normally require an emergency depressurization facility, as discussed in Section 4 of API Standard 521 (2014). The inlet line to the emergency depressurization (EDP) valve should be connected on the compressor discharge line, immediately upstream of the discharge FSV. The EDP valve should discharge to flare.
7.2 Reverse Flow Protection
Sudden shutdown of the compressor or its driver introduces a reverse flow hazard. Potential consequences include damage of the compressor internals and loss of containment from the compressor casing and upstream scrubber and cooler.
Layers of Protection Analysis (LOPA) would assign Safety Integrity Level (SIL) 2 or 3 to this hazardous situation, depending on assessment of its frequency of occurrence and consequences. The compressor discharge SDV may be designated as the final element in a SIF, to reduce the reverse flow risk to a tolerable level.
7.3 Pressure Equalization (Settle-out)
During emergency shutdown of the compressor, the suction and discharge valves will close, and the recycle valve will go fully open. Via the recycle line, the gas pressure and temperature in the suction side equipment and pipework equalize with the discharge side, at conditions intermediate between the steady state suction and discharge values, known as settle-out. The settle-out conditions are calculated with reference to the total system volume and the relative quantities of gas in the suction and discharge segments.
The minimum design pressure of the suction side equipment can then be evaluated at 1.05 times the settle-out pressure, as discussed in Annex B.3 of API Standard 521 (2014); API 521 does not however advise how to determine maximum discharge pressure.
Section 4.2 of NORSOK P-001 (2006), recommends calculating settle-out pressure based on coincident high-pressure trip on the suction and the discharge sides of the compressor, then adding a 10% margin to determine the design pressure or PSV set pressure of the suction side equipment.
7.4 Seal Gas
Dry gas seals are usually installed on both ends of the compressor rotor shaft where it penetrates the case, to prevent the escape of process gas to the environment and machine lubrication system. A thrust balance line is installed between the suction and discharge seal chambers to equalize the seal pressures at the suction value.
Stahley (2001) discusses the configuration and operating principles of dry gas seals. The primary seal is located outboard of the labyrinth seal and consists of a pair of mechanical seal faces which are forced together by coil springs when the compressor is shut down, and which separate by approximately 1 micron during compressor operation. This miniscule gap gives rise to the superb performance of dry gas seals. The primary seal vents to a hydrocarbon vent or flare system.
The primary seal requires injection of seal gas, which has two functions: to prevent contamination of the seal by process gas from the compressor casing, and to provide a working fluid for the gap between the seal faces. According to Dwyer (2020), approximately 90% of the seal gas flows across the labyrinth seal into the compressor casing, and the residual seal gas discharges via the dry gas seal into the primary vent thence to the hydrocarbon vent or flare.
These functions require supply of clean dry seal gas at circa 3.5 bar above the pressure of the compressor suction under all operating conditions, including start-up, shutdown, and hot standby when the compressor discharge gas is returned to the suction side via the recycle line.
The compressor manufacturer supplies a filter package to remove residual solids or liquid droplets immediately upstream of the dry gas seal; however, the owner is responsible for:
· Selection of a suitable source of seal gas supply.
Often compressor discharge gas is selected, but hot standby operation would require supplementary pressure via a booster unit.
The owner must check that the seal gas does not undergo partial condensation due to Joule-Thomson cooling as the seal gas pressure reduces across the primary seal. It is recommended to have 20 degrees of superheat in the gas.
· Performance of a vent study, as recommended by API 692 (2018).
The study assumes that the dry gas seal has been destroyed and pressurized seal gas escapes via the primary vent. The study should be carried out during FEED since it may result in significant vent line size(4” or larger), and/ or installation of a PSV.
REFERENCES
American Petroleum Institute (2017). Analysis, Design, Installation, and Testing of Safety Systems for Offshore Production Facilities - API Recommended Practice 14C. (8th Ed.). Washington, DC: API Publishing Services
American Petroleum Institute (2014). Pressure-relieving and Depressuring Systems - API Standard 521 (6th Ed.). Washington, DC: API Publishing Services
American Petroleum Institute (2018). Dry Gas Sealing Systems for Axial, Centrifugal, and Rotary Screw Compressors and Expanders - API Standard 692 (1st Ed.). Washington, DC: API Publishing Services
Barringer, H.P., & Weber, D.P., Life Cycle Cost Tutorial. Fifth International Conference on Process Plant Reliability, October 1996.
Botros, K.K. (2011) Single vs. Dual Recycle System Requirements in the Design of High Pressure Ratio, Low Inertia Centrifugal Compressor Stations. ASME GT2011-45002
Dwyer, K. (2020) Vent Study. Why, how to do it and how to Interpret the results. Plant Engineering Workshop 20. International Institute of Plant Engineering and Design (InIPED)
Ghaisas and Reitsma (2016) Surge Detection and Surge Control Systems for Centrifugal Compressors - Part 2. In COMPRESSORtech2. Waukesha, WI: Diesel & Gas Turbine Publication Group
GPSA (2017) Engineering Data Book (14th Ed.). Tulsa, OK: Gas Processors Association.
McMillan, G.K. (2010). Centrifugal and Axial Compressor Control (1st Ed.). New York, NY: Momentum Press, LLC
Merrow, E.W., (2011). Industrial Megaprojects: Concepts, Strategies, and Practices for Success (1st Ed.). Hoboken, NJ: John Wiley & Sons, Inc.
NORSOK (2006). NORSOK Standard P-001, Process Design (5th Ed.). N-1326 Lysaker: Standards Norway
NORSOK (2014). NORSOK Standard M-001, Materials Selection (4th Ed.). N-1326 Lysaker: Standards Norway
Norwegian Petroleum Directorate. Evaluation of Projects Implemented on the Norwegian Shelf. October 2013.
Norwegian Petroleum Directorate. Project Execution on the Norwegian Continental Shelf. Report No. OD-04-20, January 2020.
Singleton, E. (2014) Control Valves in Turbo-Compressor Anti-Surge Systems. Brighouse, W. Yorkshire: Koso Kent Introl Limited
Stahley, J.S. (2001) Design, Operation, and Maintenance Considerations for Improved Dry Gas Seal Reliability in Centrifugal Compressors. Proceedings of the 30th Turbomachinery Symposium, 203–208, Turbomachinery Laboratory, Texas A&M University.
Staroselsky, N. & Ladin, L. (1979) Improved Surge Control for Centrifugal Compressors, Chemical Engineering
ABOUT THE AUTHOR
The author was a process engineer in the Oil and Gas sector from 1980 – 2015, working on multidiscipline design projects and in operations support.
He graduated with a Bachelor’s Degree in Chemical and Process Engineering from Heriot Watt University, Edinburgh, UK, in 1980.