How to Conduct a Production Facility Bottleneck Study

A Bottleneck Is a System Limit

When an oil and gas facility falls short of its production target, the first explanation is often a piece of equipment: the export compressor is at full load, the separator level is unstable, or the produced-water system cannot accept more liquid. That equipment may be where the constraint becomes visible, but it is not necessarily the root bottleneck.

A production facility is a connected hydraulic, thermal, control, and operating system. Increasing well deliverability changes separator pressure, gas liberation, compressor duty, cooling demand, water production, flare load, and export conditions at the same time. Removing one constraint often exposes the next one immediately.

A bottleneck study therefore asks two different questions:

1. What limits production today?
2. What will limit production after that constraint is removed?

The useful answer is not “the compressor is at 100%.” It is a tested constraint hierarchy, a realistic maximum operating rate, and a set of modifications ranked by production gain, cost, shutdown requirement, and risk.

Define the Business Question First

The study boundary should start with the commercial objective rather than the simulation model. Typical objectives include:

• Restore production that has declined below the facility's historical performance
• Accommodate new wells or a satellite tie-in
• Increase oil production without exceeding gas, water, or export limits
• Extend plateau as water cut and gas–oil ratio increase
• Reduce flaring, fuel use, or chemical consumption at the current rate
• Determine whether a proposed production target is achievable with existing equipment
• Identify low-cost operating changes before approving capital modifications

The objective needs a reference case and a time horizon. “Increase production” is too vague. “Determine the modifications required to process 18,000 BPD at 65% water cut and 12 MMscfd associated gas for the next three years” is an engineering basis.

The study should also agree what success means. The maximum instantaneous test rate is rarely the right target. A sustainable production rate must remain within equipment design limits, control margin, environmental permits, product specifications, inspection constraints, and the operator's accepted risk envelope.

Assemble Evidence Before Building the Model

Brownfield facilities are not represented completely by their latest P&IDs. Valves have been throttled, exchanger performance has degraded, instrument ranges have changed, temporary hoses have become semi-permanent, and control logic may no longer match the original cause-and-effect chart.

A credible study begins with a structured data request:

At least one representative period should be reconciled in detail. Stable operation is useful for establishing normal performance; a high-rate test or near-trip period is useful for seeing the constraint. Raw historian data should be checked for bad tags, frozen values, unit errors, calibration drift, and inconsistent time stamps before it is used.

Operator interviews are essential. The control-room team usually knows which valve reaches 95%, which cooler loses performance in the afternoon, and which alarm appears before every production cut. Those observations are hypotheses to test—not substitutes for calculations—but they focus the work quickly.

Establish a Reconciled Base Case

The base case is the technical reference against which every improvement is measured. It should reproduce an actual operating period closely enough that the team understands any remaining mismatch.

Reconciliation normally closes the following balances:

• Total mass and component balance across the facility
• Oil, gas, and water allocation through each separation stage
• Energy balance around heaters, coolers, and heat exchangers
• Fuel gas, flare, vent, drain, and produced-water balances
• Compressor suction and discharge conditions and power
• Pump differential head and hydraulic power

The model should use representative fluid properties, current equipment geometry, actual pressure drops, and measured utility conditions. Calibration factors—such as exchanger fouling resistance or compressor efficiency—must remain physically plausible and be documented. A model that matches the plant only because several arbitrary efficiencies were adjusted is curve-fitting, not reconciliation.

Measurement uncertainty should be visible. If inlet multiphase flow is inferred from allocated well tests while export meters are fiscal quality, forcing both to match exactly can create a false sense of precision. The study should distinguish measured, calculated, assumed, and reconciled values.

Screen Every Constraint Category

The constraint register should cover more than major equipment nameplates. A systematic review usually finds limits in seven categories.

Wells and gathering system

Available wellhead pressure, choke performance, flowline pressure drop, slugging, sand production, and hydrate or wax limits may restrict what reaches the facility. A plant debottleneck cannot recover production that the gathering system cannot deliver.

Separation and liquid handling

Separator gas capacity, liquid residence time, oil–water separation, level-control range, internals condition, and liquid carryover all matter. A vessel can have adequate theoretical gas capacity but poor real performance because of damaged internals, foaming, unstable pressure, or an overloaded liquid outlet.

Compression and gas treatment

Compressor power, speed, discharge temperature, surge margin, driver ambient derating, suction pressure, and cooler performance should be checked against the actual map. Downstream dehydration, sweetening, fuel-gas, and export systems can constrain the same gas rate even where compression margin remains.

Pumps and liquid export

Pump head, NPSH margin, motor load, minimum-flow recycle, control-valve pressure loss, pipeline backpressure, and tank or export constraints form one system. The centrifugal pump sizing and NPSH checks should use current fluid properties and suction conditions rather than the original design case.

Heat transfer and utilities

Cooling medium temperature, air-cooler ambient conditions, exchanger fouling, heating duty, electrical generation, instrument air, fuel gas, chemicals, and seawater systems often set seasonal limits. Utility constraints are commonly missed because they sit outside the main hydrocarbon process model.

Produced water, flare, and environmental limits

Mature facilities frequently become water-limited rather than hydrocarbon-limited. Hydrocyclones, flotation units, disposal pumps, injection wells, and oil-in-water compliance must be reviewed as an integrated train. Flare capacity, emissions permits, discharge specifications, and waste-handling limits can also cap sustainable production.

Controls, alarms, and operating limits

A control valve continuously above 90% open, a controller kept in manual, or a narrow gap between normal operation and a trip set point may reveal a practical bottleneck before a mechanical limit is reached. Instrument ranges, valve characteristics, alarm settings, and control interactions belong in the study.

Calculate Capacity with Margin

Each item in the constraint register should have four values:

1. Current operating load
2. Credible maximum capacity
3. Required operating margin
4. Predicted load at the target production case

A simple utilisation measure is:

Utilisation (%) = predicted load / allowable capacity × 100

The difficult term is allowable capacity. It is not always the datasheet design value. The allowable rate may be lower because of fouling, degraded performance, a regulatory limit, uncertainty in the calculation, or the control margin required to handle normal disturbances.

Margin should be functional rather than arbitrary. A compressor needs distance from surge and driver overload. A separator needs enough working volume to absorb inlet fluctuations. A control valve needs travel in both directions. A produced-water system needs margin for changing water quality, not just average flow.

Results are clearest in a constraint table:

This makes the constraint hierarchy visible and shows which limits appear together.

Use the Right Analysis for Each Question

Steady-state process simulation is the main tool for a facility-wide bottleneck study. It closes the heat and material balance, tests rate and composition changes, and predicts loads throughout the process. But it is only one part of the analysis.

Additional methods may include:

• Compressor performance-map review and driver derating
• Vessel gas-capacity and separation calculations
• Hydraulic network calculations for gathering, flare, utilities, and export
• Heat-exchanger rating using current fouling and utility conditions
• Pump curve and system-curve analysis
• Relief and flare load review where modifications change credible scenarios
• Electrical load-flow or generation-capacity assessment
• Dynamic simulation for control interactions, slug response, trips, and start-up
• Field performance tests where data cannot resolve the real capacity

Dynamic process simulation should be reserved for time-dependent questions. It is valuable when increasing rate reduces separator surge volume, changes compressor-control behaviour, or brings normal operation too close to a trip. It is not necessary for every equipment capacity check.

Sensitivity analysis is usually more valuable than excessive precision in one case. Water cut, GOR, ambient temperature, export pressure, fluid viscosity, exchanger fouling, and equipment availability should be varied to identify the assumptions that change the recommendation.

Test Changes in the Plant Before Buying Equipment

The fastest production gain is often an operating change, but any test must be planned and controlled. A structured performance test can confirm whether a suspected constraint is real and provide better data for the model.

A test plan should define:

• Objective and success criteria
• Starting conditions and required stable period
• Rate-change steps and hold times
• Parameters to trend and sampling frequency
• Alarm, trip, vibration, quality, and environmental stop criteria
• Required personnel and decision authority
• Method for returning to the original condition

Possible low-capital interventions include controller retuning, pressure-set-point optimisation, redistributing flow between parallel trains, restoring exchanger cleaning, repairing damaged internals, reducing unnecessary recycle, changing well routing, or adjusting chemical treatment.

Operating closer to a protective trip is not debottlenecking. Any set-point change requires confirmation of the equipment design basis, alarm philosophy, relief protection, and Management of Change process. The study must preserve the safeguards that make the facility safe to operate.

Develop Options as Packages

Options should be assembled into coherent packages rather than issued as an unranked list of ideas.

Package A—Operational recovery might include instrument calibration, exchanger cleaning, controller tuning, revised well routing, and compressor wash. It is fast and low cost but may recover only part of the target.

Package B—Targeted modifications might replace one control valve, re-wheel a pump, upgrade separator internals, and add cooling area. It requires engineering and a short shutdown but preserves most existing equipment.

Package C—Strategic expansion might add parallel produced-water treatment, a compressor stage, or a new processing train. It delivers greater future capacity but carries more capital, schedule, and interface risk.

Each package should state:

• Sustainable production gain across the forecast range
• CAPEX accuracy class and OPEX impact
• Production deferral and shutdown duration
• Required engineering and safety reviews
• Implementation schedule and long-lead items
• Residual bottleneck after completion
• Sensitivity to forecast uncertainty

The residual bottleneck is critical. If a compressor upgrade increases capacity by 15% but the produced-water system reaches its limit after 5%, the standalone compressor project has only 5% usable benefit.

Rank by Value, Not Capacity Alone

The preferred option is not necessarily the one with the largest nameplate increase. Ranking should combine economic value, deliverability, and risk.

Useful measures include:

• Incremental oil or gas production over time
• Net production after planned and risk-weighted shutdown losses
• Capital cost and operating cost
• Simple payout or discounted value
• Production gain per unit of capital
• Schedule to first benefit
• Technical maturity and uncertainty
• Safety, environmental, and operability impact

Production forecasts should be time-based. A modification completed after reservoir decline has removed the available upside can be technically successful and commercially worthless. Conversely, a modest operating improvement delivered in four weeks may create more value than a large project delivered in two years.

Cost estimates should match project maturity. Early options can be compared using consistent screening estimates; the selected package can then progress to a more developed estimate and cost-risk assessment. The principles in greenfield versus brownfield CAPEX estimating apply directly.

Worked Example: A Mature Gas-Condensate Facility

Consider an onshore facility designed for 70 MMscfd of gas and 12,000 BPD of liquids. It currently processes 54 MMscfd. Operations reports that the export compressor is power-limited and the inlet separator reaches high level during rate increases. The requested target is 65 MMscfd from two new wells.

Historian review shows that the compressor driver reaches 98% load on hot afternoons, but only 88% at night. The aftercooler outlet temperature is 9°C above its design value, increasing compressor power. The separator level problem correlates with a partially open bypass around the produced-water control valve and a controller in manual—not with vessel capacity.

The reconciled model and field checks identify this hierarchy:

1. Current bottleneck: compressor power under summer ambient conditions.
2. Recovered constraint: aftercooler fouling and degraded fan performance.
3. Next bottleneck: produced-water disposal at the forecast high water cut.
4. Future constraint: gas dehydration at 68 MMscfd.

The first package cleans the aftercooler, restores two fan blades, repairs the water-control loop, and revises the compressor wash programme. A controlled test demonstrates 61 MMscfd without exceeding driver or separator limits.

The second package adds a produced-water disposal pump and re-rates the dehydration contactor using updated operating data. It supports 65 MMscfd with agreed margin. A compressor replacement is rejected because restored cooling performance recovers the required capacity at a fraction of the cost and schedule.

The important result is not that one exchanger was dirty. The study explains how the observed constraints interact, proves the sustainable rate, identifies the next limit, and prevents investment in the wrong equipment.

Deliverables That Lead to Action

A useful bottleneck study should leave the operator with an executable plan. Typical deliverables are:

• Study design basis, data register, and assumptions
• Reconciled current-operation heat and material balance
• Validated process model and case register
• Equipment and system capacity calculations
• Constraint register with utilisation, margin, and evidence
• Field-test plans and test results
• Production-envelope and sensitivity assessment
• Options register and modification packages
• CAPEX, OPEX, production-deferral, and schedule comparison
• Ranked recommendations and implementation roadmap
• Updated model files and monitoring recommendations

The final presentation should distinguish verified facts, model predictions, and items requiring confirmation. Recommendations should identify owners, dependencies, and the next engineering action. “Review compressor performance” is not an action plan; “complete an OEM map review using the attached operating cases before selecting the cooler modification” is.

Common Failure Modes

• Starting with the suspected equipment. This confirms the original theory instead of testing the full system.
• Using design data as operating data. Nameplate capacity does not describe a fouled, aged, or modified facility.
• Ignoring controls and operator workarounds. Many practical constraints are visible first in valve travel, controller mode, alarms, and recurring manual intervention.
• Optimising one discipline in isolation. More compressor capacity is useless if water treatment or export becomes the next immediate bottleneck.
• Treating maximum rate as sustainable rate. A short test does not prove seasonal performance, product quality, or adequate operating margin.
• Changing set points without Management of Change. Moving an alarm or trip does not create physical capacity and may remove protection.
• Ranking by CAPEX alone. Schedule, shutdown loss, reservoir timing, and residual constraints can dominate project value.
• Failing to update the model. The study should become a living facility model as wells and fluid conditions change.

Conclusion

A good bottleneck study combines operating evidence, process engineering, simulation, equipment calculations, and commercial judgement. It begins with a reconciled picture of the facility as it operates today, not as it was originally designed. It then identifies the complete constraint hierarchy, tests improvements safely, and ranks coherent modification packages by sustainable production value.

The most valuable result is often not a larger item of equipment. It may be restored performance, a corrected control problem, better use of parallel capacity, or a small modification that unlocks an existing asset. The discipline is in proving that gain—and knowing which constraint will appear next—before the operator commits capital or production downtime.