Skip to content

Drawing No. EH–CA–003 // Engineering Case Study

Deepwater Horizon — Well Control Simulator

Reviewed August 2026

On 20 April 2010, the Macondo well lost control during temporary-abandonment operations. A failed bottom-hole cement barrier, a misinterpreted negative-pressure test, displacement of dense drilling mud with seawater, delayed kick detection, overwhelmed surface handling, and ultimately unsuccessful emergency well-sealing systems combined into a catastrophic blowout. This page reconstructs the barrier sequence and lets users explore the underlying hydrostatic logic without pretending to reproduce the actual multiphase transient.

Prepared by EngineerHub · CSB / BSEE / National Commission-based educational reconstruction

Accident: 20 Apr 2010Water depth: just under 5,000 ftFinal well depth: 18,360 ft subseaCrew onboard: 126Fatalities: 11Oil release: 87 days

01 // Interactive well-control failure replay

Select a stage or run the complete sequence.

Macondo well — barrier and flow reconstructionDocumented sequence, depths and barrier states are shown schematically. Vertical proportions and gas-bubble motion are not to scale.
Deepwater Horizon Macondo well-control replaySchematic showing drilling rig, riser, subsea blowout preventer, well casing, bottom-hole cement barrier, reservoir, mud and seawater columns, kick migration and surface diversion. MGSoverboard SUBSEA BOPAnnular preventerPipe ramBlind shear ram BOTTOM CEMENThydrocarbon-bearing reservoir seawater displacement ↓ hydrostatic headhydrocarbon influx ↑ cased wellborebottom barrierANNULARPIPE RAMBSR SEA SURFACE · 0 ftBOP / SEAFLOOR · ~5,000 ftRESERVOIR / TD · ~18,360 ft BARRIER STATUSstate changes with replay stage NEGATIVE-PRESSURE TESTtest interpretationoverall statusnot yet testedbleed-down result pendingflow / returns not yet assessedacceptance decision pendingMismatch unresolved = STOP.
Well-control state
Visual encoding
Dense drilling mud
Seawater / reduced hydrostatic head
Hydrocarbon influx / gas migration
Failed / overwhelmed barrier
Barrier verificationBlowout / lessons
Barrier / operation
Well-control indication
Risk / missed barrier
Consequence / lesson
Hydrostatic / flow state
Barrier state
Controlled wellUncontrolled blowout

Key case facts

Water depthjust under 5,000 ft
Final well depth18,360 ft subsea
BOP size≈57 ft / ≈400 tons
Kick undetectedalmost 1 hour
People onboard126
Fatalities / hospitalized11 / 17

Operations console — latest stage first

02 // Hydrostatic barrier explorer

See why replacing dense mud with seawater can remove well-control margin.

18,360 ft
14.17 ppg
8,367 ft
11,850 psi
All-mud hydrostatic
Mixed-column hydrostatic
Static pressure margin
Static state
STATIC HYDROSTATIC MODEL — NOT A TRANSIENT WELL-CONTROL CALCULATIONFORMATION BARRIERformation pressureseawater replacementdrilling mudPRESSURE BALANCEhydrostaticformationmargin
Model boundary. Hydrostatic pressure is calculated in oilfield units with the standard relation P = 0.052 × mud weight (ppg) × vertical depth (ft). With two fluids, the pressure contributions are summed by column length. This ignores transient friction, equivalent circulating density, multiphase flow, gas compressibility/expansion, temperature, well geometry and influx dynamics. The “Macondo reference preset” combines documented or published reference values only to demonstrate static pressure logic; it is not a forensic re-simulation.

03 // Barrier cascade

Major accidents require multiple layers to fail or become ineffective.

01 / BOTTOM CEMENTThe cement barrier intended to isolate the hydrocarbon-bearing zone was ineffective. This removed a primary physical barrier before temporary abandonment.
02 / INTEGRITY TESTThe negative-pressure test produced unexpected results, but the crew accepted the test as a pass and proceeded.
03 / HYDROSTATIC MUDDense drilling mud was displaced with seawater, reducing hydrostatic pressure and allowing influx through the failed bottom barrier.
04 / KICK DETECTION & RESPONSEHydrocarbons flowed for almost an hour without effective human intervention or automated closure. By the time the kick was recognized, gas had migrated far up the riser.
05 / BOP + DIVERTERSurface diversion to the mud-gas separator was overwhelmed; the BOP temporarily shut in the well but the emergency shear system ultimately did not seal it.

04 // Engineering figures

Historical milestones and barrier logic, not invented transient curves.

Static hydrostatic comparison

All-mud referenceAfter reference displacementReference formation pressure
View reference values
InputReference valueSource / role
Total fluid column18,360 ftCSB final well depth below sea level
Mud weight14.17 ppgPublished Macondo drilling reference used here only as educational input
Seawater replacement length8,367 ftCSB description of temporary-abandonment configuration
Reference reservoir pressure11,850 psiUSGS initial reservoir pressure used in post-blowout modeling
The chart demonstrates why a large reduction in fluid-column density can change a well from overbalanced to underbalanced. It does not reproduce the exact pressure state at every point in the historical operation.

Barrier availability through the sequence

effective / availableconditional / human dependentfailed / overwhelmed
View interpretation
This is a qualitative barrier-health visualization. It intentionally avoids assigning invented reliability probabilities to cement, human response, the BOP or diverter.
As the cement and hydrostatic barriers were lost, the system increasingly depended on timely human detection and emergency hardware. CSB emphasizes that humans themselves were being relied upon as safety-critical barriers.

05 // What actually happened?

The blowout began as a loss of well barriers, not as a sudden BOP malfunction. A critical cement barrier at the bottom of the well had not effectively isolated the hydrocarbon-bearing zone. The negative-pressure test intended to verify that barrier produced unexpected results and was nevertheless accepted as a pass.

Mud displacement removed a major pressure barrier

After the test was accepted, the crew continued temporary-abandonment operations and replaced dense drilling mud with seawater. Drilling mud controls a well by exerting hydrostatic pressure against formation pressure. Removing it reduced bottom-hole pressure while the surface plug had not yet been installed.

The kick developed without timely intervention

CSB concluded that hydrocarbons flowed from the reservoir for almost an hour without human intervention or automated closure. As influx continued, hydrocarbons accelerated upward through the well and the roughly mile-long riser. Once large volumes of gas were above the subsea BOP, even successful shut-in could not remove the hazardous inventory already traveling toward the rig.

The surface response could not safely handle the flow

The diverter was configured to send flow to the mud-gas separator rather than directly overboard. The separator was rapidly overwhelmed, and hydrocarbons discharged from multiple points onto the rig. They found an ignition source, causing explosions and fire.

The BOP story is more complicated than “it did not work”

Pressure evidence indicates that a pipe ram successfully shut in the well for a period. The later emergency blind shear function likely actuated, but CSB concluded that effective compression had buckled the drillpipe off-center so the blind shear ram could not fully close and seal the well. Post-incident examination also revealed latent control-system deficiencies in the redundant emergency pods.

08 // Human and organizational factors

The CSB treats Macondo as a process-safety and barrier-management failure involving hardware, people, procedures, organizations and regulation.

No clear negative-test procedureNeither BP nor Transocean management supplied a test procedure with explicit acceptance criteria and required actions for deviations.
Temporary-abandonment plan changed repeatedlyCSB found numerous plan variations and no explicit risk assessment for the final procedure used at Macondo.
Humans were relied upon as safety barriersAfter other barriers were gone, successful control depended on recognizing the kick and manually operating BOP equipment before conditions escalated.
Warning information was difficult to interpretUnexpected pressure and flow behavior, changing operations and multiple simultaneous tasks created a demanding human-performance environment.
Operator–contractor responsibilities were diffuseCSB concluded that BP and Transocean had policies for risk management but neither assumed effective responsibility for ensuring their implementation at Macondo.
Emergency equipment assurance was incompleteRoutine BOP tests did not expose latent failures in emergency functions such as the AMF/deadman control system.
09 // Defense in depth

Controls that should prevent a kick from becoming a major accident.

Well designIndependent primary barriers

Maintain multiple verified barriers between the reservoir and surface. Do not remove a hydrostatic barrier until replacement barriers are demonstrably effective.

TestingExplicit pass/fail criteria

Integrity tests need written procedures, instrument expectations, acceptance limits, actions for anomalous results and independent challenge.

MonitoringKick indicators protected from operational noise

Flow-out, pit volume, pressure and displacement data should remain visible and interpretable during changing transfer and displacement operations.

ResponseEarly shut-in philosophy

Well-control procedures and drills should favor timely conservative action while the influx remains below the BOP rather than waiting for unambiguous surface evidence.

Emergency handlingRoute large gas releases away from occupied areas

Diverter configuration and surface gas-handling capacity should be aligned with credible blowout scenarios, not routine small-gas handling.

Safety-critical equipmentVerify full emergency functionality

Testing must challenge redundant BOP emergency systems and latent failure modes, not only components used in everyday drilling.

07 // Chronology

Date / phaseEvent
Oct 2009Drilling at Macondo begins with the Marianas rig.
31 Jan 2010Deepwater Horizon arrives after Marianas is damaged by a hurricane.
9 Apr 2010Drilling stops at 18,360 ft below sea level; temporary abandonment preparations follow.
20 Apr — cement verificationPositive-pressure test is accepted; negative-pressure testing then produces unexpected results over several attempts.
20 Apr — test acceptedThe final negative-pressure test is accepted as a pass even though cement integrity has not been demonstrated.
20 Apr — displacementDense drilling mud is replaced with seawater as the crew prepares to set the surface plug.
~following hourHydrocarbons enter and rise through the well without timely intervention or automatic closure.
Late eveningMud and hydrocarbons reach the rig; annular/diverter actions are taken and the mud-gas separator is overwhelmed.
9:47 p.m. contextPressure evidence indicates a pipe ram successfully shuts in the well and drillpipe pressure builds above 5,000 psig.
~2 min laterFirst explosions occur and data transmission to shore ceases.
22 AprDeepwater Horizon sinks.
15 JulA capping stack stops the release after 87 days.
12 // Engineering lessons learned

Macondo demonstrates how quickly a complex system can become dependent on its least reliable remaining barrier.

Barrier status must be explicitTeams need a shared, continuously updated picture of which barriers are verified, conditional, degraded, removed or dependent on human action.
An unexplained test is not a passed testUnexpected pressure or flow behavior should stop the operation until the mechanism is understood and integrity is demonstrated.
Hydrostatic head is an active safety barrierChanging fluid density or column height is not just a logistical step—it changes the pressure balance controlling reservoir influx.
Early kick recognition is decisiveOnce gas has migrated far above a subsea BOP, closing the well cannot remove the hazardous inventory already in the riser.
Diverters are consequence-management systemsRouting decisions must account for credible flow magnitude and gas release location; routine equipment may be overwhelmed in a blowout.
Redundancy must be tested as redundancyTwo control pods do not provide two reliable layers if latent wiring or battery failures can remain undetected.
Human factors belong in well designProcedures, displays, alarm logic, concurrent tasks, team communication and decision authority shape whether human barriers can succeed.
Process safety is different from personal safetyGood injury statistics cannot demonstrate that major-accident barriers are healthy. Barrier condition and leading process-safety indicators need separate attention.

09 // Frequently asked questions

Common misconceptions and training points.