Probabilistic Damage Stability

Probabilistic damage stability is the modern statistical framework under SOLAS Chapter II-1 Part B-1 and Resolution MSC.421(98) for assessing a ship’s survivability after damage. Instead of testing specific deterministic damage cases (the older approach: ‘flood any one compartment, must remain afloat’), the probabilistic framework computes an Attained Subdivision Index (A) by summing the probability of survival of all credible damage cases, each weighted by the statistical probability of that damage occurring. The Attained Index must equal or exceed a Required Subdivision Index (R) that scales with vessel type, length and (for passenger ships) number of persons onboard. The framework was effective for new dry cargo ships from January 2009 under MSC.421(98) and progressively extended to other vessel types through subsequent amendments. The probabilistic framework allows more design flexibility than the older deterministic standard (a strong subdivision in one area can compensate for weakness elsewhere, as long as the aggregated Index meets the requirement) and provides a more rational basis for safety assessment because it accounts for the relative probability of different damage scenarios. It is the modern global standard, replacing the SOLAS 90 deterministic framework for the principal ship types. The probabilistic framework supplements (does not replace) the Stockholm Agreement water-on-deck criterion for ro-ro passenger ferries; the two regimes operate in parallel for vessels in the relevant scope. ShipCalculators.com hosts the principal computational tools: the Attained Subdivision Index calculator, the Required Subdivision Index calculator, the damage probability (p factor) calculator, the survival probability (s factor) calculator, the vertical extent factor (v) calculator, the intermediate stage flooding calculator and the GZ curve calculator. A full listing is available in the calculator catalogue.

Contents

Background

Why probabilistic

The probabilistic framework was developed because the older deterministic framework (test specific damage cases, pass or fail each) had several limitations:

Conservative for some scenarios, weak for others: a vessel passing the deterministic test for one or two adjacent compartments could be very strong against the tested damages but weak against untested combinations.
Discourages design innovation: the deterministic framework rewarded compliance with specific tests rather than overall safety.
Inconsistent with other safety frameworks: civil engineering, aerospace and other safety-critical domains had moved to probabilistic / risk-based assessment in the 1980s and 1990s.

The probabilistic framework was developed through the IMO Stability and Load Lines (SLF) Sub-Committee from approximately 1988 onwards, with the foundational papers by Wendel (1960s and 1970s) on damage statistics. The framework was first applied to passenger ships under SOLAS 95 amendments (in force 1996) and progressively extended to dry cargo ships through MSC.421(98) (in force January 2009) and subsequently to tankers and other vessel types.

Conceptual structure

The probabilistic framework calculates the Attained Subdivision Index (A):

$$ A = \sum_i p_i \cdot s_i \cdot v_i $$

where:

$p_i$ is the probability of damage to a specific compartment combination $i$ (computed from statistical damage data).
$s_i$ is the probability of survival given damage to compartment combination $i$ (computed from the damaged stability calculation).
$v_i$ is the vertical extent factor (probability of damage to specific decks).

The summation is across all considered damage cases, typically hundreds to thousands. The Attained Index $A$ must equal or exceed the Required Subdivision Index (R).

For the Stockholm Agreement and SOLAS Chapter II-1 Part B-2 (passenger ships), additional supplementary calculations apply.

Required Subdivision Index

The Required Subdivision Index $R$ scales with vessel type and length:

Cargo ships (dry, MSC.421(98)): $R = 1 - \frac{128}{L_S + 152}$, where $L_S$ is the subdivision length.
Passenger ships (SOLAS Chapter II-1 Part B-2): $R = 1 - \frac{5,000}{L_S \cdot N + 2.5 N^2 + 15,225}$, where $N$ is the number of persons onboard.
Tankers (MARPOL Annex I + SOLAS): separate hierarchy.

The functional form ensures that larger vessels and vessels with more persons onboard face higher required Indexes.

Damage probability factors

p factor

The damage probability $p$ for a specific compartment combination is computed from statistical damage data covering thousands of historical collision and grounding casualties. The data gives the probability distribution of:

Longitudinal damage extent (length of damage along the ship).
Longitudinal damage location (where along the length the damage occurs).
Transverse damage extent (depth of damage from the shell into the hull).
Side affected (port or starboard for collision; bottom for grounding).

The damage probability for a specific compartment combination is the integral of the damage probability density function over the damage extents that would breach the boundaries of those compartments. The framework provides standardised formulae and tables for this calculation.

v factor

The vertical extent factor $v$ accounts for the probability that damage extends upward through specific decks. The factor is calculated based on the vertical extent of the damage relative to the deck heights, integrated over the vertical damage probability distribution.

s factor

The survival probability $s$ for a specific damage case is computed from the damaged stability calculation:

Apply the damage to the vessel (flood the damaged compartments at the appropriate permeability).
Calculate the resulting equilibrium trim, heel and waterline.
Calculate the residual GZ curve.
Compare the GZ curve characteristics to specified survival criteria.
The survival probability $s$ is between 0 (does not survive) and 1 (fully survives), based on how the GZ curve characteristics compare to the criteria.

The principal survival criteria are:

Maximum GZ value ($GZ_{max}$): minimum required value, typically 0.05 to 0.15 m depending on the damage scenario.
Range of stability (angle from equilibrium heel to vanishing stability): minimum required range, typically 7 to 30 degrees.
Equilibrium heel ($\theta_e$): maximum permitted, typically 7 to 25 degrees.
Margin line is not exceeded by the damaged waterline.

Each criterion contributes a partial $s$ factor; the lowest determines the case-specific $s$.

Damage cases enumeration

Single, two and three compartment cases

The probabilistic framework requires consideration of:

Single-compartment damage: damage to each compartment individually.
Two-compartment damage: damage to each pair of adjacent compartments.
Three-compartment damage (for some cases): damage to three adjacent compartments.

For a vessel with N transverse compartments, the number of single-compartment cases is N, two-compartment cases is approximately N-1, three-compartment cases is approximately N-2. The total number of damage cases for a typical merchant ship is approximately 50 to 200, depending on subdivision complexity.

Side and bottom damage

Both side damage (collision-type) and bottom damage (grounding-type) must be considered. The damage probability functions differ for the two types.

Intermediate stage flooding

For some damage cases, the vessel may pass through intermediate stages of flooding before reaching final equilibrium. For example, water flowing through the damaged compartment may take several minutes to fill the compartment fully; during this time the vessel may pass through transient states with reduced stability. The probabilistic framework requires consideration of intermediate stages where these may have lower stability than the final equilibrium.

Calculation methodology

Software tools

Probabilistic damage stability calculation is computationally intensive. The principal commercial tools are:

NAPA Damage Stability (NAPA Ltd, the dominant commercial tool, used by approximately 80% of major merchant ships).
AVEVA Marine Damage Stability.
DNV NAUTICUS Damage Stability.
Lloyd’s Register IntelliShip Damage.
NK PrimeShip Damage.
Bureau Veritas BV-LCM Damage.

A typical probabilistic damage stability calculation for a large container ship considers approximately 500 to 1,000 damage cases and takes approximately 4 to 8 hours of computer time.

Process flow

The standard process:

Compartment definition: detailed definition of every watertight compartment in the vessel, including permeabilities.
Damage case enumeration: generation of all single, two, three-compartment damage cases per the framework.
Damage probability calculation: $p_i$ for each case from the statistical formulae.
Damage stability calculation: equilibrium trim/heel and GZ curve for each case.
Survival probability calculation: $s_i$ for each case from the GZ curve and survival criteria.
Attained Index calculation: sum $\sum_i p_i \cdot s_i \cdot v_i$.
Required Index calculation: from the formulae for the vessel type.
Compliance check: $A \geq R$.

Class society review

The full calculation package is submitted to the Class society for review. The review typically takes 4 to 12 weeks for a major newbuild, less for a sister vessel of an already-approved design.

Comparison with deterministic framework

Feature	Deterministic	Probabilistic
Approach	Specific damage cases	All credible cases weighted by probability
Output	Pass/fail per case	Aggregated Index A vs Required R
Conservatism	Often very conservative	More balanced, allows trade-offs
Computational effort	Modest	Substantial
Design flexibility	Limited	Greater
Use as of 2024	Some legacy applications	Modern global standard

The deterministic framework remains in use for some specific cases (tanker double hull under MARPOL Annex I, bulk carrier single-hold flooding under SOLAS XII, some passenger ship requirements). The probabilistic framework is the modern standard for most merchant ship subdivision design.

Implementation in service

Trim and stability booklet

The vessel’s trim and stability booklet includes:

The Attained Subdivision Index $A$ value.
The Required Subdivision Index $R$ value.
Representative damage case results.
Loading and operational restrictions arising from the damage stability assessment.

The booklet is reviewed at every periodic survey and revalidated for any vessel modification.

Loading computer integration

Modern loading computers (NAPA, Aveva, etc.) include damage stability checking modules: for any actual loaded condition, the system can verify that the vessel still satisfies the probabilistic damage stability framework. This is operationally useful for unusual loaded conditions that may not have been explicitly tested in the design submission.

Modifications

Major hull modifications (lengthening, bulbous bow retrofit, conversion) require recalculation of the probabilistic damage stability assessment. The Class society review the recalculation and approve the modification.

Specific applications

Cargo ships (MSC.421(98))

For new dry cargo ships above 80 m length, MSC.421(98) is the principal framework. The Required Index increases with length per the formula $R = 1 - 128/(L_S + 152)$.

Passenger ships

Passenger ships under SOLAS Chapter II-1 Part B-2 face higher Required Indexes:

Length factor: longer ships face higher Required Indexes.
Person factor: more persons onboard face higher Required Indexes.
Two-compartment standard for cruise ships: typically required, providing margin against multi-compartment damage.

The probabilistic framework was first applied to passenger ships in 1996 (SOLAS 95 amendments) and has been progressively refined.

Tankers

Tankers under SOLAS plus MARPOL Annex I face hybrid requirements: the probabilistic framework for general damage stability plus the deterministic double hull requirement for oil cargo containment. The two frameworks are complementary.

Bulk carriers

Bulk carriers above 150 m length face the SOLAS Chapter XII single-hold flooding deterministic requirement in addition to the probabilistic framework. The two requirements are complementary.

Ro-ro passenger ferries

Ro-ro passenger ferries face the probabilistic framework plus the Stockholm Agreement water-on-deck criterion for vessels operating in NW European waters.

Limitations and ongoing development

Damage statistics

The damage probability functions are derived from historical casualty data. The data is necessarily backward-looking; if the actual collision and grounding patterns change (e.g. due to autonomous shipping, climate-driven Arctic operations, route changes), the probability functions may need to be updated. The IMO SLF Sub-Committee periodically reviews the statistics.

Survival criteria

The survival criteria (minimum GZ, minimum range, maximum heel) are partly empirical. Subsequent casualty experience may suggest refinements. The criteria are reviewed periodically.

Time-domain effects

The current probabilistic framework is largely steady-state: it assumes the vessel reaches equilibrium after each damage scenario. Time-domain effects (transient flooding rates, dynamic motion in waves during flooding) are partially captured through intermediate stage flooding and the Stockholm Agreement water-on-deck criterion, but are not fully integrated. The IMO is working on enhanced time-domain damage stability assessment, expected to enter the framework over the 2025 to 2030 period.

Computational scaling

For very large and complex vessels (cruise ships, mega-container ships, FPSOs), the number of damage cases can exceed 10,000, requiring substantial computational resources. Modern cloud computing and parallelised solvers have largely addressed this constraint.

Future developments

Enhanced time-domain assessment

The IMO is developing enhanced time-domain assessment methods that simulate the actual flooding dynamics. The work is ongoing through the SLF and SDS Sub-Committees; commercial implementation is expected from approximately 2027 to 2030.

Climate and operating area considerations

Climate-driven changes in storm frequency and intensity may require re-examination of the underlying assumptions. The IMO is monitoring this through ongoing review.

Integration with autonomous shipping

Autonomous shipping introduces new failure modes (system failures, cyberattacks) that may need to be integrated into the damage probability framework. Initial work has begun at the IMO.

References

IMO Resolution MSC.421(98): Amendments to the International Convention for the Safety of Life at Sea, 1974, as amended (SOLAS Chapter II-1 Subdivision). International Maritime Organization, 2017.
IMO Resolution MSC.281(85): Explanatory Notes to the SOLAS Chapter II-1 Subdivision and Damage Stability Regulations. International Maritime Organization, 2008.
SOLAS Chapter II-1 Parts B-1 (probabilistic subdivision for cargo ships) and B-2 (passenger ship subdivision). International Maritime Organization, 1974 with subsequent amendments.
IMO. SLF Sub-Committee historical reports. International Maritime Organization, ongoing.
Wendel, K. Subdivision of Ships. SNAME, various papers 1960s and 1970s.
IACS. Common Structural Rules for Bulk Carriers and Oil Tankers (CSR BC and OT). International Association of Classification Societies, 2024 edition.
DNV. DNV Rules for Classification of Ships, Pt 5 Ch 4 Damage Stability. DNV, 2024 edition.
Lloyd’s Register. Probabilistic Damage Stability: Methodology Guide. Lloyd’s Register Group, 2022.