MARPOL Annex I Reg 14: oil filtering equipment

Background: MARPOL 73/78 to consolidated 2025 edition

The control of machinery-space bilge water discharge has been a core concern of MARPOL since 1973. The 1954 OILPOL Convention limited oil-in-water discharge to a notional 100 parts per million within 50 nautical miles of land, with no measurement obligation and no equipment standard. The 1973 Convention with its 1978 Protocol, entering into force on 2 October 1983, replaced this regime with a hard equipment-based standard expressed first in former Regulation 16 of Annex I (which at that time covered oil discharge monitoring for tanker machinery spaces, distinct from the cargo-area Regulations that covered segregated ballast) and later, after the 2004 consolidated revision, in the renumbered Regulation 14.

The original 1983 regime required ships of 10,000 GT and above to carry an oil discharge monitoring system linked to an oily water separator at 100 ppm or below, with the separator alone required for ships of 400 GT and above. The 1992 amendments under Resolution MEPC.51(32), entering into force on 6 July 1993, tightened the limit to 15 ppm for all bilge water leaving any ship of 400 GT and above and required the automatic stopping device at the separator outlet that interrupts overboard discharge whenever the oil-content meter exceeds the limit.

The 2004 consolidated revision (Resolution MEPC.117(52), in force 1 January 2007) renumbered former Regulation 16 as Regulation 14 inside the new Chapter 3 “Requirements for machinery spaces of all ships”, alongside Regulation 12 oil residue (sludge) tanks, Regulation 13 standard discharge connection, and Regulation 17 Oil Record Book Part I. The two type-approval resolutions, MEPC.107(49) for the 15 ppm bilge separator and MEPC.108(49) for the oil discharge monitor, were both adopted on 18 July 2003 and applied from 1 January 2005.

MEPC.107(49) superseded Resolution MEPC.60(33) (adopted 28 October 1992) which had set the first-generation 15 ppm type-approval standard under the MEPC.51(32) amendments. The supersession is important for survey purposes: any unit approved under MEPC.60(33) only, without a subsequent MEPC.107(49) certificate, does not meet the current standard and must be replaced at the next IOPP renewal.

The 2025 consolidated edition incorporates refinements that recognise the diversification of marine fuels following the IMO 2020 sulphur cap and FuelEU Maritime. It clarifies that the 15 ppm limit applies to all oil, including bio-blended residual fuel, methanol-derived oil, and synthetic-fuel slip entering the machinery-space drain, and that MEPC.107(49) is to be supplemented with an emerging-fuels test fluid at the next revision.

Regulation 14.1: 15 ppm discharge limit and threshold conditions

Regulation 14.1 sets the headline discharge rule. Any discharge into the sea of oil or oily mixtures from machinery spaces is prohibited except when all of the following conditions are satisfied: the ship is proceeding en route; the oily mixture is processed through oil filtering equipment complying with paragraph 6; the oil content of the effluent without dilution does not exceed 15 parts per million; and the mixture does not originate from cargo pump-room bilges or mix with cargo oil residues.

The 15 ppm limit is expressed as parts of oil per million parts of effluent by volume, measured continuously at the separator outlet without prior dilution. Any arrangement that draws sea or fresh water into the sample line ahead of the meter to lower the apparent oil concentration is a defeat device and a serious deficiency under any port-state regime.

The proceeding en route condition excludes discharge while the ship is stationary at anchor or alongside, and is not satisfied during manoeuvring in a confined channel even at slow speed. The machinery-space origin condition limits the regulation to bilge water from the engine room, pump room (excluding cargo pump room), shaft tunnel, and emergency generator space; galley grey water and swimming pool drains must not be co-mingled with the OWS feed. The cargo pump-room bilge stream on tankers is excluded and addressed in Regulation 34.

The 15 ppm limit applies in all sea areas, including special areas under Annex I (Mediterranean, Baltic, Black, Red, Gulfs, Gulf of Aden, Antarctic, North-West European Waters, Oman, Southern South Africa). Regulation 15 imposes the operational discharge criteria for bilge waters more broadly; within special areas its provisions add procedural conditions but do not relax the 15 ppm number itself.

400 GT and 10,000 GT applicability thresholds

Regulation 14 establishes two equipment thresholds based on gross tonnage.

The 400 GT threshold is the basic OWS threshold. Every ship of 400 gross tonnes and above must be fitted with oil filtering equipment that meets the type-approval standard in MEPC.107(49). That standard, as described in detail below, mandates the 15 ppm alarm (visual and audible, at the engine control station) and the automatic stopping device (fail-safe valve) as integral components of every approved unit. So the alarm and stop are required for all 400 GT+ ships as part of the OWS equipment itself, not as a separate add-on.

The 10,000 GT threshold adds the ODMCS recorder requirement. Every ship of 10,000 gross tonnes and above must carry an oil discharge monitoring and control system compliant with MEPC.108(49) that records continuously the oil content of the effluent, the rate of discharge, and the position of the discharge valve, with a tamper-evident time stamp retained on board for at least three years. This continuous recorder is the obligation that distinguishes the 10,000 GT fleet from the 400-to-9,999 GT fleet.

The thresholds are expressed in gross tonnage under the 1969 Tonnage Measurement Convention, not in deadweight tonnes; the surveyor confirms the applicable threshold from the ship’s International Tonnage Certificate. A modification that increases gross tonnage across either threshold triggers re-examination at the next intermediate or renewal survey.

The 400 GT cutoff captures essentially the entire commercial fleet, including most fishing vessels, offshore supply vessels, and mid-sized passenger ferries. Below 400 GT the discharge prohibition still applies; smaller vessels comply through bilge holding tanks and shore disposal. The 10,000 GT cutoff captures the bulk-carrier, tanker, container, gas-carrier and large cruise fleets. Between 400 GT and 10,000 GT the OWS with its alarm and stop is required but the continuous recorder is not; the operator records discharges manually in the Oil Record Book Part I under Code D. Harbour tugs and pilot launches that do not proceed en route are exempted in practice through a trading-pattern declaration endorsed by the recognised organisation at the IOPP renewal survey.

Type-approval framework: MEPC.107(49) and MEPC.108(49)

The type-approval regime that supports Regulation 14 rests on two parallel resolutions adopted at MEPC 49 on 18 July 2003, both of which entered general application on 1 January 2005 and have since been the binding technical baseline for every approved 15 ppm separator and oil discharge monitor.

Resolution MEPC.107(49) governs the 15 ppm bilge separator itself, replacing the earlier MEPC.60(33) standard. It defines the unit as the combination of a separator stage and an oil-content meter that together deliver effluent at or below 15 ppm under the test conditions of the resolution. The unit must include the automatic stopping device that interrupts overboard discharge when the meter alarms, the 15 ppm alarm with visual and audible indication at the engine control station, the sample line drawing the test stream after the separator stage and without dilution, and a tamper-evident approach to all calibration controls. The resolution prescribes the test fluids, test sequences, alarm and stop response times, and the documentation the manufacturer must submit for the type-approval certificate.

Resolution MEPC.108(49) governs the oil discharge monitoring and control system (ODMCS) required for ships of 10,000 gross tonnes and above. The system must continuously measure the oil content, record the measurement together with the discharge rate and valve position on a tamper-evident medium retained on board for at least three years, and control the overboard discharge through an automatic stopping device interlocked with the meter. MEPC.108(49) also extends to oil tanker cargo discharge monitoring under Regulation 31, but in the machinery-space context it provides the recorder and control logic that complements MEPC.107(49). On many ships both functions are built into a single combined unit.

The type-approval certificate lists the unit’s rated flow capacity in cubic metres per hour, the applicable test fluids, the alarm and stop response times measured during the type test, the calibration interval, and the expected meter cell service life. The certificate is held at the engine control room with the identification plate of the unit and the manufacturer’s installation drawings. The recognised organisations of the International Association of Classification Societies (IACS) operate mutual recognition of type-approval certificates: a separator approved by ABS is accepted by DNV, ClassNK, LR, BV and KR for installation on ships flying any IACS-recognised flag, subject only to verification of the installation arrangement at the IOPP commissioning survey.

Test protocol: Type C and D fluids, transient performance

The MEPC.107(49) test protocol is a sequence of laboratory and full-scale tests at an accredited test facility, witnessed by a representative of the recognised organisation. Each unit type is tested at the rated flow capacity with four standard test fluids:

• Test Fluid A: Marine Diesel Oil (DMA grade) plus distilled water at 100 ppm dispersion, simulating the typical dilute bilge stream from a clean engine room.
• Test Fluid B: the same DMA fluid at 1,000 ppm, simulating a heavily contaminated stream after a fuel-oil leak event.
• Test Fluid C: a stable emulsion of DMA, surfactant (Tween 80 or equivalent) at 0.5 percent, and iron oxide particulate at 100 milligrams per litre, simulating realistic bilge composition with engine-room cleaners, lube-oil leaks and rust. Type C is the most demanding because the surfactant stabilises droplets against gravity coalescence and the iron oxide loads the meter cell.
• Test Fluid D: pure water, used to confirm the zero baseline of the meter.

The unit must process each fluid at rated flow without exceeding 15 ppm at the outlet, with the alarm and automatic stop operating within 15 seconds of the meter reaching threshold and a further 5 seconds before the discharge valve closes fully. Transient performance is tested by feeding alternating slugs of clean water, Type A and Type C fluids in a 30-minute sequence simulating variable bilge composition.

The type-test report documents the measured alarm and stop times, the meter reading versus reference laboratory analysis, and the mass balance between inlet, outlet and reject streams. The report supports the type-approval certificate and remains on file with the recognised organisation for the life of the unit type.

OWS technologies: coalescing, centrifugal, membrane, dual-stage

The 15 ppm separator approved under MEPC.107(49) is not a single technology but a family of approaches that all meet the type-test outcome. The four principal technology classes installed on the global fleet are gravity coalescing, centrifugal, membrane, and dual-stage hybrid units.

Gravity coalescing separators are the most common technology, installed on roughly 70 to 80 percent of the global fleet by unit count. The separator vessel is a vertical or horizontal pressure tank fitted internally with plate packs or fibre coalescing media that promote droplet growth as the oily water passes through. The coalesced droplets rise to the surface and drain by gravity to the oil residue (sludge) tank; the cleaned water leaves through the meter cell and the discharge valve. Coalescing separators handle Type A and Type B test fluids reliably but struggle with stable Type C emulsions, and many older designs from before 2005 cannot meet the Type C transient test without retrofit.

Centrifugal separators use mechanical rotation at typically 5,000 to 8,000 revolutions per minute to drive oil and water apart by density. Centrifugal units handle emulsions better than coalescing units and have a smaller footprint, but they consume more electrical power and require more maintenance attention to bowl wear.

Membrane separators use ultrafiltration or nanofiltration modules with pore sizes of typically 0.01 to 0.1 micrometres. Membrane units offer the most consistent 15 ppm or lower performance across all test fluids and are increasingly specified for passenger ships and environmentally sensitive trades. Membrane modules are sensitive to fouling and require pre-filtration and periodic chemical cleaning; module replacement intervals run to 12 to 36 months.

Dual-stage hybrid units combine a first-stage gravity or coalescing separator with a second-stage polishing module (membrane, adsorbent bed, or biological treatment). The first stage reduces concentration from typically 100,000 ppm to 100 ppm; the second stage polishes from 100 ppm to below 15 ppm and provides the alarm-stop interlock. Dual-stage units are the standard for new-build cruise ships and large containerships from approximately 2010 onward, and are the preferred retrofit solution for older units that fail the Type C transient test.

The rated flow capacity of the installed unit must be matched to the expected daily bilge generation; typical practice is to size the unit for two to three times the daily bilge generation to allow batch operation over a fraction of the day. The OWS 15 ppm alarm threshold checker can help verify compliance conditions before a planned discharge.

ODM interlocks and the automatic discharge stop

The automatic discharge stop is the regulatory and physical core of Regulation 14: the hard interlock that forces the unit to refuse discharge when the meter alarm is active.

The interlock is implemented as a fail-safe valve in the overboard discharge line, typically a pneumatically actuated ball or globe valve with a spring-return-to-closed characteristic. Loss of compressed air, electrical signal or meter signal all cause the valve to close. The valve is energised to open only when the 15 ppm meter reads below threshold and the system control logic has confirmed the discharge permissive, which includes the proceeding en route signal where hardwired, the flow-rate sensor confirming that the sample stream represents the actual discharge, and the tamper detection circuits on the meter housing and sample-line shutoff.

The alarm response time required by MEPC.107(49) is 15 seconds between the meter reading exceeding 15 ppm and the visual and audible alarm activating at the engine control station. The automatic stop response time is a further 5 seconds to the discharge valve reaching the fully closed position. The combined response of 20 seconds is the regulatory ceiling; modern units achieve total response under 10 seconds.

The alarm panel indicates meter reading, alarm status (normal, warning, alarm), valve position (open, closed, transit), and sample line flow. The panel is lockable against unauthorised adjustment with a tamper-evident seal on the calibration controls. Any seal break is a mandatory entry in the Oil Record Book Part I under Code I (failure of the oil filtering equipment).

The recorder function for ships of 10,000 gross tonnes and above logs meter reading, discharge rate, valve position, date and time, and ship’s position continuously while the discharge valve is open. The record is retained on tamper-evident media for at least three years. The recorder is the principal evidentiary instrument at PSC inspection: a recorder showing gaps in coverage, flat-line traces, or post-event gaps correlated with port-state intervention is a leading detention trigger.

Bypass valve seal-and-record requirements (Reg 14.4)

Regulation 14.4 acknowledges that the OWS may fail in service during a voyage. The regulation does not permit discharge of bilge water through a temporary bypass arrangement: the bilge water must be retained in a bilge holding tank until reception facilities are available. Any bypass valve that could physically route bilge water around the OWS must be physically sealed in the closed position, with any seal break recorded immediately in the Oil Record Book Part I.

The seal arrangement is a lead-and-wire or plastic-tag seal at the manifold of any valve capable of bypass. The seal is identified by a unique number logged in the OWS commissioning record at the IOPP initial survey and reproduced in the ORB at every renewal. The PSC surveyor confirms the seal is intact, the number matches the record book, and no broken-and-replaced seal is present without a corresponding ORB Code I entry.

The emergency bypass scenario contemplated by the drafters was a flooding casualty where bilge pumping flow rates exceed OWS capacity. The master is then permitted under Regulation 4 to discharge oily bilge water without OWS treatment to save the ship, with reporting to the flag administration and the nearest coastal state and an ORB Code L entry. In normal service the expected number of seal breaks across an IOPP renewal cycle is zero; an unexplained seal break is treated as systematic non-compliance and may support detention under Paris MoU code 30.

Magic-pipe fraud cases: Carnival, NCL, Princess

The phrase magic pipe entered the maritime regulatory vocabulary in the 1990s through US Coast Guard investigations of operators who had bypassed the OWS with hidden piping, hose, or valve arrangements that routed bilge water directly to the overboard discharge while the ORB was falsified to show normal OWS operation. The cases that followed were the largest pollution-prosecution series in any merchant-marine regulatory regime and shaped the current enforcement posture of the United States Coast Guard, the Department of Justice, and several European port-state authorities.

The Carnival 2002 case was the first headline prosecution. Carnival Corporation pleaded guilty in April 2002 to falsifying records on six cruise ships and paid an 18 million US dollar criminal fine plus a five-year probation requirement with fleet-wide environmental compliance plans and independent audit by court-appointed monitors. The investigation established that crews had used flexible hoses to bypass the OWS during routine bilge discharge, with falsified ORB entries throughout the period.

The Norwegian Cruise Line 2007 case followed a similar pattern. NCL pleaded guilty to charges arising from magic-pipe operation on the Norwegian Sky and paid a 750,000 US dollar fine plus probation. The case established that junior engineers acting on instruction from senior officers could be considered agents of the company for purposes of corporate criminal liability under the Act to Prevent Pollution from Ships (APPS), even where corporate written policies prohibited the conduct.

The Princess Cruises case is the largest single penalty in the series. Princess Cruise Lines, a subsidiary of Carnival, pleaded guilty in 2016 and was sentenced in 2018 to a 40 million US dollar criminal fine plus five-year probation that placed the entire Carnival cruise fleet under court supervision. The case arose from a 2013 inspection of the Caribbean Princess in the United Kingdom, where a junior engineer disclosed that a magic pipe had been used for years to discharge oil-contaminated water without OWS treatment. The 2016 guilty plea and the 2018 sentencing are distinct events and should not be conflated.

The cases produced two structural responses. First, the MEPC.1/Circ.643 guidance of 2008 strengthened the interpretation of MEPC.107(49) to require tamper-evident sample lines, alarm setpoint controls, and meter housings, with explicit acknowledgement that unidentified piping near the OWS skid is to be treated as a presumptive bypass arrangement at PSC inspection. Second, the United States Coast Guard and the Paris MoU intensified the use of whistleblower-incentive provisions under APPS Section 1908(a), under which a crew member reporting a documented violation may be awarded up to half of any criminal fine. The provision has produced a steady flow of cases since 2010 and is credited with the substantial decline in magic-pipe prosecutions through the 2020s.

2025 amendments: emerging-fuels bilge composition

The 2025 consolidated edition of MARPOL Annex I addresses the bilge-composition implications of fuel diversification under the IMO 2020 sulphur cap and FuelEU Maritime. Three operational realities drive the update.

First, bio-blended residual fuels at B24 to B30 produce bilge composition with elevated fatty-acid methyl ester (FAME) content that some legacy oil-content meters do not detect reliably; calibration must be verified against the actual bilge stream and the dominant fuel grade declared in the OWS commissioning data sheet. Second, methanol-derived oil in dual-fuel installations behaves differently from petroleum-derived oil under gravity coalescence, and a polishing-stage retrofit is recommended where the existing unit was approved before 2024. Third, synthetic-fuel slip from low-carbon installations including ammonia and synthetic methane contributes minor lubricant-derived hydrocarbon to the bilge; the alarm response time under transient conditions must be re-verified at the next intermediate survey for ships converted to dual-fuel after 2024.

The next revision of MEPC.107(49) is on the MEPC working agenda with proposals from Japan, the Republic of Korea and the European Commission, with the central proposed change being the addition of an emerging-fuels test fluid simulating B30 bilge composition. The IMO had not adopted a revised resolution as of mid-2026.

OWS skid construction: sample port, alarm panel, recorder

The OWS unit is delivered as a factory-built skid combining the separator vessel, the meter cell, the sample line, the discharge valve, the alarm panel, and the local control electronics on a single steel base plate. The skid carries the certification plate, the type-approval certificate holder, the manufacturer’s identification, and the connection points for inlet bilge water, outlet to overboard, reject to the oil residue (sludge) tank, and electrical power and signals.

Inlet piping comes from the bilge holding tank through a dedicated bilge transfer pump, typically a screw-type or rotary-vane positive-displacement pump sized to the rated flow. The inlet line carries a strainer for particulates above approximately 5 millimetres and a non-return valve to prevent back-flow during shutdown.

Outlet piping passes through the 15 ppm meter cell, the automatic stop valve, and a dedicated overboard hull penetration clearly labelled in English and the working language. The penetration carries a screw-down non-return valve in accordance with SOLAS damage-stability rules and serves no other discharge stream.

Sample line draws from the outlet line after the separator stage and without dilution, passes through a conditioning section (degassing chamber and flow-rate stabiliser) before the meter cell, and returns the sampled water to the bilge holding tank or to the separator inlet. The line carries a lockable isolation valve labelled “SAMPLE LINE - OWS”.

Alarm panel at the engine control station indicates meter reading, alarm status, valve position, sample line flow, and recorder status. The panel is lockable against unauthorised adjustment, with a tamper-evident seal on the calibration controls.

Recorder for ships of 10,000 gross tonnes and above is mounted near the alarm panel and records continuously on a tamper-evident medium. Modern recorders are typically integrated into the integrated alarm and monitoring system (IAMS), with a tamper-evident archive retained for at least three years.

Electrical and pneumatic supply must include redundant power feeds from the main and emergency switchboards, with automatic changeover that does not interrupt the alarm or stop function. Loss of power drives the discharge valve to the closed position by spring return. The skid identification plate carries the manufacturer name, model designation, type-approval certificate number, rated flow capacity, year of manufacture, and serial number on a steel plate bolted to the skid frame.

Record keeping under Regulation 17 ORB Part I

Regulation 17 requires every ship of 400 gross tonnes and above (excluding oil tankers) and every oil tanker of 150 gross tonnes and above to maintain an Oil Record Book Part I in which every operation involving oil residues, bilge water and the OWS must be entered. The physical equipment is governed by Regulation 14; every operational record relating to it is governed by Regulation 17.

The operation codes relevant to OWS entries are:

• Code D: non-automatic starting of discharge of bilge water and oily mixtures from machinery spaces, with volume, start and stop times, ship’s position, and meter reading throughout.
• Code E: automatic starting of discharge through the OWS, with operating period, volume, and any alarm events.
• Code I: failure of the oil filtering equipment, including alarm event, action taken, recorder verification, and any seal break under Regulation 14.4.
• Code L: discharge in the case of damage or other exceptional circumstances under Regulation 4, with full report to the flag administration and coastal state.

The surveyor at IOPP renewal cross-references the recorder traces against ORB Code D, E, I and L entries for the previous five years; any discrepancy is flagged as a serious finding. Missing entries, post-dated entries, erasures or correction-fluid use, and handwriting changes between consecutive entries are recognised indicators of falsification and escalated to the flag administration. The ORB is retained on board for at least three years after the last entry and is available for inspection at any reasonable time by the flag administration, any MARPOL party PSC officer, and the US Coast Guard for ships calling at US ports. Falsification of the ORB is a criminal offence under most flag-state laws and under APPS, with historical convictions including imprisonment of senior officers in addition to the corporate penalty.

IOPP renewal-survey verification

The International Oil Pollution Prevention (IOPP) certificate is renewed every five years following a renewal survey in accordance with Regulation 6 and the HSSC harmonised survey scheme. The renewal survey is the principal occasion at which Regulation 14 compliance is verified in detail.

Type-approval certificate verification is the first step: the surveyor confirms that the OWS carries a current MEPC.107(49) certificate (not merely a MEPC.60(33) certificate from the pre-2005 era), that the ODM unit where required carries a current MEPC.108(49) certificate, that each certificate identifies the actual installed serial number, and that the original or a certified copy is held at the engine control room.

Functional test of the alarm and stop is the second step. The surveyor witnesses an introduction-of-oil test in which a calibration fluid is injected upstream of the meter cell and the alarm and stop are observed to operate within the type-approval response times. Modern installations have a built-in test feature that injects a simulant fluid; the surveyor accepts the simulant test where the flag administration permits.

Calibration verification is the third step. The OWS meter cell must be calibrated to the manufacturer’s recommended interval, typically annually or every 8,000 operating hours. The surveyor confirms the calibration certificate is current and that the calibration was performed by an authorised service technician.

Recorder review is the fourth step. For ships of 10,000 gross tonnes and above, the surveyor reviews the recorder traces for the previous five years, looking for gaps, flat lines, post-event gaps, and consistency with the ORB Code D, E, I and L entries. Any anomaly triggers a more detailed inquiry.

Bypass-valve seal verification confirms that any bypass valve in the OWS area is sealed closed with the documented seal number and that no broken-and-replaced seal is present without a corresponding ORB Code I entry. Sample-line verification confirms that the sample line draws from the correct point in the discharge stream, that no dilution arrangement is present, and that the sample-line isolation valve is locked open with the documented seal.

Common PSC findings and detention triggers

Port State Control statistics under the Paris MoU, the Tokyo MoU and the United States Coast Guard Certificate of Compliance regime track Regulation 14 deficiencies under specific codes: 01215 (oil filtering equipment 15 ppm) and 01220 (oil discharge monitoring and control system) under the Paris MoU classification, with broadly equivalent codes elsewhere.

Paris MoU annual reports for the 2020 to 2024 period record approximately 800 to 1,200 Annex I deficiencies per year, of which roughly 20 to 30 percent relate directly to Regulation 14. The breakdown within that subset is approximately:

• 35 percent: alarm bypass jumpers, defeat devices on the meter cell, evidence of alarm-setpoint adjustment.
• 25 percent: calibration drift, expired calibration certificates, failure of the introduction-of-oil functional test.
• 15 percent: bypass-valve tampering, broken seals without corresponding ORB Code I entries, unidentified piping in the OWS area.
• 15 percent: ORB Code D, E, I and L discrepancies relative to recorder traces.
• 10 percent: type-approval documentation missing or out of date.

A single grave deficiency such as a defeat device on the meter cell, evidence of a magic pipe, or a falsified ORB is detention-triggering on its own and is also a prosecution referral to the flag administration and to the United States Coast Guard for ships in US waters. Multiple correlated deficiencies can support detention under Paris MoU code 30 for systematic non-compliance.

The United States Coast Guard applies the Annex I standards through 33 CFR Part 151 and, since 2010, through the National Strike Force specialised oil-pollution investigation team. USCG detentions involving the OWS have historically resulted in APPS prosecutions with fines of 750,000 to 40 million US dollars and senior officer convictions where ORB falsification has been established. The Australian Maritime Safety Authority (AMSA) is the most active enforcer in the Asia-Pacific region, with a particular focus on the eastern Australian iron-ore export corridor and the Great Barrier Reef World Heritage area; AMSA detention rates for Annex I deficiencies have run at 2 to 4 percent of ships inspected. The Norwegian, Danish and Swedish flag administrations under the Paris MoU operate a focused inspection campaign on cruise and ferry traffic in the Baltic special area.

SEEMP Part I overlap

The Ship Energy Efficiency Management Plan (SEEMP) Part I required under MARPOL Annex VI Regulation 22 is principally concerned with carbon-intensity and fuel-efficiency operation, but overlaps with Regulation 14 at two points.

First, the SEEMP includes operational measures for bilge generation reduction: reduction of fresh-water leakage from accommodation, control of cooling-water leakage from auxiliary engines, and minimisation of deck-wash and machinery-space cleaning water entering the bilge. A ship that reduces bilge generation reduces OWS operating frequency, meter-cell wear, and reject sludge volume. The SEEMP narrative is verified at the SEEMP audit cycle and cross-referenced to ORB Code D and E records.

Second, under the 2024 to 2025 amendments the SEEMP addresses the emerging-fuels transition and the operational implications for bilge composition. Operators bunkering bio-blended residuals, methanol or LNG dual-fuel installations are required to document the expected bilge composition and the OWS calibration verification at every fuel-grade transition. The SEEMP Part III for the Carbon Intensity Indicator (CII) does not directly engage Regulation 14 but is part of the broader environmental compliance framework reviewed by the surveyor and the PSC officer.

Commercial implications: retrofit costs in offshore and CSO sectors

The new-build cost of an OWS and its installation runs from approximately 30,000 US dollars for a 0.25 cubic metre per hour gravity coalescing unit on a small offshore supply vessel, to approximately 150,000 to 400,000 US dollars for a 5 to 10 cubic metre per hour dual-stage hybrid on a large containership or cruise vessel. The ODMCS recorder and integration adds 20,000 to 80,000 US dollars. Total OWS and ODMCS compliance at new-build is typically 0.05 to 0.20 percent of the hull contract price.

The retrofit cost for ships originally fitted with pre-2005 OWS units that fail the Type C transient test is the more substantial commercial issue. The offshore sector is affected because the typical offshore supply vessel and Construction Support Operations (CSO) vessel operates an aged OWS at a duty cycle well above the cargo and bulk-carrier average. Retrofit with a current-generation dual-stage unit including meter cell, recorder, discharge valve and IAMS integration runs to 80,000 to 250,000 US dollars per vessel with 5 to 14 days of downtime. Across the global offshore fleet of roughly 3,000 vessels, the cumulative retrofit obligation is in the range of 600 million to 1.2 billion US dollars through the late 2020s and early 2030s.

The cruise sector carries the highest scrutiny because of the magic-pipe history; the major lines operate a fleet-wide standard of dual-stage hybrid units with margin below 5 ppm at the meter and a recorder-and-archive regime well in excess of the MEPC.108(49) minimum, as a direct response to the US-court probation regimes. The bulk-carrier and tanker sector carries the new-build cost at delivery, defers retrofit to the next dry-docking after a unit failure, and operates with calibration service contracts of typically 5,000 to 15,000 US dollars per ship per year. The OWS reject stream drains to the oil residue (sludge) tank and is delivered ashore under the Regulation 12 commercial framework rather than independently.

The 15 ppm discharge limit: technical basis

Regulation 14.1 expresses the discharge ceiling as a single concentration threshold at the meter cell, $c_{\text{discharge}} \leq 15$ ppm, where $c_{\text{discharge}}$ is the oil concentration measured in parts per million by volume after the separator stage and without prior dilution. The number is not arbitrary. It derives from the detection limit and reproducibility envelope of the standard NDIR and UV-fluorescence oil-content meters available at the 1992 amendment under Resolution MEPC.51(32). Practical detection across the Type A, B, C and D test fluids was 5 to 10 ppm with measurement uncertainty of plus or minus 3 to 5 ppm. The 15 ppm figure sits above that detection floor and below the visible-sheen threshold of roughly 50 ppm in unstirred sea water. The IMO has not re-opened the threshold since 1992.

The interlock turns that ceiling into a hard discharge gate. The permitted overboard flow is a step function of the measured concentration:

where $Q_{\text{actual}}$ is the actual overboard discharge rate in cubic metres per hour. Below 15 ppm the valve passes the full flow; at or above the threshold the meter alarms and the automatic stop drives the valve to zero. There’s no proportional throttling band and no soft limit: the design is binary by regulation, which is why a meter reading just over threshold shuts the line entirely rather than trimming it. The MARPOL Annex I/14 compliance checker works through the Regulation 14.1 conditions in sequence.

Sizing the oily-water separator

Picking the rated flow capacity of the installed unit starts from an estimate of how much bilge the ship makes in a day. The class societies (ABS, DNV, Lloyd’s Register, ClassNK and BV) assembled shipyard and operator survey data through the 1990s and 2000s into an empirical first-pass model:

where $V_{\text{daily bilge}}$ is the daily bilge generation in cubic metres per day, $\alpha$ is the gross-tonnage coefficient (0.0005 to 0.001 cubic metres per day per gross tonne), $\text{GT}$ is the ship’s gross tonnage, $\beta$ is the age coefficient (0.05 to 0.2 cubic metres per day per year), and $\text{age}$ is the ship’s age in years from delivery. The gross-tonnage term captures the leak-and-condensate load scaling with engine-room and habitable-space volume; the age term captures the drift over service life as gland packings and pipe joints loosen.

The rated flow then follows from the daily volume and how long the operator intends to run the unit each day:

with $Q_{\text{rated}}$ the OWS rated flow capacity in cubic metres per hour and $t_{\text{operating}}$ the daily operating time in hours, typically 4 to 8. Common practice sizes the unit for two to three times the estimated daily bilge generation so it runs in batches over a fraction of the day rather than continuously. The OWS sizing tool runs this arithmetic against the standard catalogue model sizes. The model is a design starting point, not a regulatory requirement; the type-approval certificate fixes the actual rated capacity once a unit is selected.

Worked example

Take a 30,000 gross-tonne supramax bulk carrier, 6 years old, with the empirical coefficients $\alpha = 0.0008$ and $\beta = 0.15$:

Choose a daily operating time of 5 hours, enough to run the unit through the morning watch and once during night manoeuvring. The required rated capacity is:

So the supramax needs a Regulation 14 OWS rated at least 5 cubic metres per hour, which lands on a standard catalogue size (the usual steps are 1, 2.5, 5 and 10 cubic metres per hour). A 5 cubic metre per hour dual-stage hybrid runs roughly 80,000 to 150,000 US dollars at new-build with the ODMCS recorder integrated.

A 60,000 gross-tonne panamax containership of the same age scales up:

at a 6-hour operating window, which pushes the selection to the 10 cubic metre per hour model. The arithmetic assumes the meter is calibrated and the sample line draws without dilution from after the separator; the rated capacity should be at least equal to the bilge transfer pump capacity at maximum head, or the pump will overrun the separator and push it into a transient regime outside its type-approval envelope. Both examples assume no special-area restrictions; Polar Code Part II-A imposes a complete discharge prohibition for Arctic operations that removes the sizing trade-off entirely.

Regulation 14 and the former Regulation 16 distinction

A recurring confusion on tankers is the relationship between Regulation 14 and the former Regulation 16, and what the pre-consolidation numbering meant. Before the 2004 revision under MEPC.117(52), MARPOL Annex I used a different chapter structure. The former Regulation 16 covered oil filtering equipment for machinery spaces (what is now Regulation 14), while separate regulations covered segregated ballast tanks and clean-ballast requirements. The renumbering assigned the current Regulation 14 to the former Regulation 16 content. The present Regulation 16 in the consolidated text covers oil/water interface detectors and slop tanks, which is a distinct subject from the OWS equipment in Regulation 14.

The practical consequence for surveyors: deficiency codes and survey items for oil filtering equipment reference Regulation 14 in all IOPP certificates issued from 1 January 2007 onward. An older ORB or IOPP certificate referencing the former Regulation 16 for OWS equipment is correctly read as now being Regulation 14; no reissuance is required solely because of the renumbering.

AEO quick-reference answers

What ppm limit applies to overboard discharge from machinery-space bilges? 15 ppm, measured without dilution at the OWS outlet, by Regulation 14.1 of MARPOL Annex I.

Which GT thresholds trigger which equipment? 400 GT and above: oil filtering equipment (OWS) with 15 ppm alarm and automatic stopping device, type-approved to MEPC.107(49). 10,000 GT and above: additionally an oil discharge monitoring and control system with continuous recorder, type-approved to MEPC.108(49).

What resolution governs type approval of the 15 ppm bilge separator? Resolution MEPC.107(49), adopted 18 July 2003, which replaced MEPC.60(33) and entered general application 1 January 2005.

How quickly must the alarm and stop respond? The alarm must activate within 15 seconds of the 15 ppm threshold being exceeded. The discharge valve must reach fully closed within a further 5 seconds, giving a combined ceiling of 20 seconds.

How long must the ODMCS recorder trace be retained? At least three years, on tamper-evident media.

What is the magic-pipe provision under APPS? Under APPS Section 1908(a), a crew member who reports a documented violation to the US Coast Guard may receive up to half of any resulting criminal fine as a whistleblower award.

Limitations

This article describes the regulation and the standard equipment; several practical caveats limit how cleanly the 15 ppm regime works in service, and a surveyor or chief engineer should treat each as a known weak point.

The first limitation is meter tolerance at the threshold. The 15 ppm number carries the plus or minus 3 to 5 ppm measurement uncertainty of NDIR and UV-fluorescence cells, so a unit reading 13 ppm on its own meter may be running closer to 17 ppm against a reference laboratory analysis. The alarm and automatic stop respond within the MEPC.107(49) ceilings (15 seconds to alarm, a further 5 seconds to full valve closure), but a slug of stable Type C emulsion can spike past the cell faster than the laboratory reference would confirm. The interlock is binary, which protects against sustained over-limit discharge but says nothing about brief transients within the response window.

The second limitation is type-approval drift. A unit that passed the Type C transient test on the bench at commissioning is not guaranteed to hold that performance after years of bowl wear, membrane fouling, or coalescing-media degradation. The IOPP introduction-of-oil functional test confirms that the alarm and stop fire, not that the separator still delivers below 15 ppm against the full Type C emulsion. Many pre-2005 gravity coalescing units cannot meet the Type C transient test at all without a polishing-stage retrofit, and the field condition of an in-service unit routinely lags its certificate.

The third limitation is the sludge-and-bilge interface. The reject stream from the separator drains to the oil residue (sludge) tank, and the assumption that bilge composition stays within the engine-room cleaner manufacturer’s recommended detergent dose breaks down whenever crews over-dose degreasers or co-mingle galley grey water with the OWS feed. Surfactant above the test envelope stabilises oil droplets against gravity coalescence and defeats older units even when nothing is physically tampered. The regulation assumes machinery-space-origin water only, proceeding en route, with no dilution; each assumption is an operational discipline, not a hardware guarantee.

The fourth limitation is enforcement reality against the magic-pipe history. The interlock and the tamper-evident seals raise the cost of bypass but do not make it impossible. The Carnival, Norwegian Cruise Line and Princess prosecutions all involved hidden piping or hoses that routed bilge water around a fully compliant OWS while the Oil Record Book was falsified. Type approval certifies the equipment, not the crew operating it, so the strongest practical control remains the whistleblower-incentive provision under APPS Section 1908(a) rather than any feature of the hardware.

The fifth limitation is the sizing estimate itself. The bilge-generation model is empirical and regional; the $\alpha$ and $\beta$ coefficients span a factor of two each, so the daily-volume estimate can be off by 50 percent or more for an unusual trade or an aged, leak-prone engine room. The model is a starting point for capacity selection, not a compliance figure, and the worked-example outputs above should be read as design guidance rather than a regulatory minimum. Short-sea ships with daily reception calls, long Antarctic voyages under the Polar Code Part II-A discharge prohibition, and dual-fuel ships burning LNG with a pilot fuel each break the simple scaling in a different direction; any of them warrants a duty-cycle review rather than a catalogue lookup.

A sixth limitation applies from 2025 onward: MEPC.107(49) predates the fuel-diversity shift by two decades. Its Type C test fluid uses petroleum-derived DMA with Tween 80 surfactant. Bio-blended and methanol-derived bilge compositions are not in the current test matrix, and a unit that passes the 2003 protocol is not guaranteed to handle the 2026 bilge stream from a multi-fuel ship. The MEPC working-group proposals for an emerging-fuels test fluid address this gap, but adoption had not occurred as of mid-2026. Operators of dual-fuel and biofuel-bunkering vessels should treat the calibration interval as a hard annual requirement and commission a reference-laboratory bilge analysis at each fuel-grade transition.

See also

• MARPOL Annex I: oil pollution prevention
• MARPOL Annex I Regulation 12: oil residue (sludge) tanks
• MARPOL Annex I Regulation 12A: oil fuel tank protection
• MARPOL Convention overview
• MARPOL Annex VI: air pollution prevention
• IMO 2020 sulphur cap
• Polar Code
• Marine oily water separators and bilge water treatment
• Marine bilge and ballast systems
• Calculator catalogue

Related calculators

• MARPOL Annex I/14 - Oil filtering equipment
• MARPOL Annex I - OWS Sizing (15 ppm Bilge Separator)
• OWS - 15 ppm Alarm Threshold Check
• MARPOL - Fuel Oil Sampling
• MARPOL Annex VI/18 - Fuel oil quality
• MARPOL Annex I/4 - Oil discharge
• MARPOL Annex I/38 - Reception facilities oil
• MARPOL Annex I/37 - Shipboard Oil Pollution Emergency Plan
• MARPOL Annex I/36 - Oil Record Book Part II