Every offshore platform, from the first day steel goes into the water until the last day of operation, shares its working environment with other vessels. Supply boats, standby vessels, tankers, workboats the marine traffic around an offshore installation is constant, varied, and largely operating under time pressure. Most of the time, the system works. But vessel collisions with offshore platforms do happen, they have caused casualties and major damage in the past, and they remain a recognized major accident hazard that has to be formally assessed as part of any credible Safety Case.
Elixir Engineering conducted a Ship Collision Study for the Drilling Platform and Living Quarters associated with the FEED phase of the Kalamkas-Sea and Khazar Field Development Project. The client for this study was Metrix LLC. The purpose was straightforward: evaluate the risk that a vessel, during the course of routine or abnormal marine operations around the platforms, collides with the installation and determine what design, operational, and procedural measures are needed to keep that risk within acceptable limits.
The Kalamkas-Sea field sits in the Northern Caspian Sea, approximately 132 km south-west of Kashagan, and Khazar lies roughly 28 km further south-west again. Both fields are in shallow water around 6 to 7 meters which creates its own set of challenges for structural design and marine operations. Between December and March, large parts of the Northern Caspian are ice-bound, which restricts normal marine operations and changes the character of vessel traffic in ways that have to be accounted for when building a realistic collision risk picture.
It is worth being clear about what this type of study is and what it is not. A ship collision risk assessment is not a structural impact analysis it does not calculate whether a specific beam will deform under a specific load. What it does is quantify the risk: how likely is a collision, and how much energy would be involved? Those two numbers together give you the risk picture that determines whether the platform is safe by design, or whether something needs to change.
The study addresses two types of collision scenario, which behave quite differently from each other. The first is a powered collision a vessel that is under propulsion but has lost effective control, typically through a watchkeeping error, a steering failure, or a communications breakdown. In a powered collision, the vessel may be travelling at or near its normal operating speed when it makes contact with the platform. The second type is a drifting collision a vessel that has lost all propulsion and is being carried toward the installation by wind, current, or wave action, with no ability to correct course.
These two scenarios matter in different ways. Powered collisions tend to involve higher impact velocities, which means higher kinetic energy at the point of impact. Drifting collisions generally involve lower velocities, but the absence of any recovery mechanism the vessel cannot stop itself, slow down, or change course makes them a distinct and serious hazard in a region where wind and current conditions can be significant. Both types were assessed for this project.
The technical output the structural and safety engineers actually need from a collision study is the kinetic energy associated with each scenario type. Kinetic energy is the product of the effective mass of the vessel and the square of its approach velocity. Because velocity is squared, even a moderate increase in approach speed produces a significantly higher energy demand on the structure. That number goes directly into the structural design brief and the risk acceptance assessment.
The study had five things it needed to deliver. Each one was a direct input to either the structural design, the marine operations planning, or the Safety Case.
The first was a clear picture of what vessels are operating near the platforms and how. This is not generic it is specific to the Northern Caspian operating environment, the types of vessel that serve this kind of shallow-water development, and the traffic patterns in this part of the sea. You cannot do a credible collision study without first understanding what ships are actually out there.
The second was a defined set of collision scenarios specifically those that are credible given the vessel traffic, the platform layout, and the local conditions. We are not interested in every conceivable collision; we are interested in the ones that could genuinely happen. That distinction matters for producing a study that is useful rather than exhaustive.
The third was a frequency estimate for each scenario: how many times per year, on average, would we expect a collision of this type to occur? This is the probabilistic heart of the assessment. Without it, you have consequence analysis but no risk analysis.
The fourth was an impact energy calculation for each scenario. As noted above, this is the primary technical output that feeds the structural design. It answers the question: if this collision happens, how much energy does the structure have to absorb?
The fifth was a risk screening outcome a determination of whether each scenario falls within or outside acceptable risk criteria, and what risk control measures are needed based on that determination.
The assessment followed a four-step methodology structured in accordance with CMPT (1999) and DNV-RP-F107 the two internationally recognized standards that define best practice for this type of offshore collision risk work.
Before you can assess collision risk, you have to know what collisions are actually possible. This step involved reviewing vessel navigation routes, shipping lanes, and the marine activities taking place in the waters around both platforms. We looked at approach paths for supply vessels, the movement patterns of standby vessels on station, and the traffic from any passing commercial or tanker vessels in the broader Caspian Sea shipping picture.
From that review, we identified the scenarios worth assessing not every theoretical encounter, but the ones that represent genuine operational realities. Both routine marine activities and abnormal conditions were included in the scope. A supply vessel losing power during an alongside approach is a routine marine risk. A passing vessel drifting off course in deteriorating Caspian weather is less frequent but fully credible. Both belong in the study.
The ice-bound winter season introduces a genuine complication here. Between December and March, the operational envelope for marine activities in this part of the Caspian narrows considerably. Vessel types change, traffic volumes change, and the probability profile of control failures changes. These seasonal factors were built into the scenario identification from the start, not added as an afterthought.
With the scenarios defined, the next step was to estimate how often each one could be expected to occur. The frequency evaluation used an event based approach, which breaks each collision scenario into three probability components: how often does a vessel operate in a position where a collision is physically possible; given that, how likely is it to lose control; and given that control loss, how likely is it that no recovery action by the vessel's crew, by other vessels, or by warning systems prevents contact with the installation.
Environmental conditions were explicitly factored into the analysis. Wind conditions in the Northern Caspian were characterised using representative Beaufort scale ranges, because wind strength directly affects both the probability of a vessel losing control and the probability of successful recovery. In ice conditions, where manoeuvrability is restricted and visibility can be poor, both probabilities shift in ways that alter the frequency estimate compared to open water assumptions.
These three probability components were combined using the structured framework from CMPT (1999) and DNV-RP-F107 to produce annual collision frequency estimates for each scenario expressed as the expected number of collision events per year.
The consequence side of the risk picture comes down to impact energy. For each collision scenario, the kinetic energy transferred to the platform was estimated based on the effective mass of the vessel and its approach velocity at the point of contact.
For powered collisions, the approach velocity is essentially the vessel's operating speed because in most powered collision scenarios, the navigator is unaware of the hazard until it is too late to take meaningful corrective action. For drifting collisions, the approach velocity is set by environmental conditions rather than engine power, and is generally lower but the absence of any active braking or steering means that what velocity there is gets delivered fully to the structure.
The resulting impact energy values gave the structural engineering team the specific accidental load cases they needed to design against. They also provided the basis for the risk screening comparison that followed.
The last step brought together the frequency and energy results. Using impact energy versus frequency curves a graphical framework that plots each scenario against applicable acceptance criteria the assessment produced a clear picture of where each collision scenario sits in terms of risk. Scenarios that fall comfortably within the acceptance region confirm that the design is adequate for those hazard levels. Scenarios that approach or exceed the criteria prompt a review of what risk control measures are needed.
This comparison framework also provided the basis for the risk control recommendations the specific design, operational, and procedural measures developed to manage the collision risk that remains after the primary structural design decisions have been made.
The Ship Collision Study was conducted in accordance with two internationally recognised frameworks:
| Document No | Standard | How It Was Applied |
| CMPT 1999 | A Guide to Quantitative Risk Assessment for Offshore Installations | Provided the event-based probabilistic framework for estimating collision frequency across powered and drifting scenarios — including how vessel traffic data, control failure probabilities, and recovery failure probabilities are combined into annual frequency estimates. |
| DNV-RP-F107 | Risk Assessment of Pipeline Protection (Det Norske Veritas, 2010) | Governed the methodology for impact energy calculation, collision risk screening, and assessment of risk reduction measures for risers, pipelines, and structural components within the platform safety zone. |
These two standards were chosen because they are complementary. CMPT (1999) gives you the probabilistic methodology for collision frequency the event tree framework, the data on vessel control failure probabilities, and the approach for combining environmental influences into a frequency estimate. DNV-RP-F107 gives you the consequence methodology how to calculate impact energy, how to screen the risk against criteria, and what protection and risk reduction measures to consider. Together they cover the full scope of a collision risk assessment from hazard to recommendation.
The Ship Collision Assessment produced a documented risk picture for the Drilling Platform and Living Quarter platforms, covering both the frequency and energy dimensions of the collision hazard. The findings directly informed structural design decisions and marine operations planning for the FEED deliverables.
The scenario identification confirmed that the primary collision risks for these platforms come from two distinct categories of vessel. The first is infield traffic supply vessels, standby vessels, and workboats that are operating specifically in support of the platforms and interact with the installation on a regular basis. These are the vessels most likely to be involved in a powered collision during an approach or departure manoeuvre. The second is passing traffic merchant vessels and tankers transiting through the broader Caspian Sea that have no intended interaction with the platforms but could drift into the exclusion zone following a propulsion failure.
Both categories, and both powered and drifting collision types within each category, were included in the assessment. The winter ice season was treated as a distinct operational condition with its own traffic profile and collision probability characteristics.
The collision frequency analysis produced annual estimates for each scenario. The Northern Caspian environment shaped these estimates in several ways. The ice restriction between December and March reduces certain vessel traffic categories during that period, which affects the overall annual frequency. But deteriorating weather conditions including strong winds in the transitional seasons increase the probability of control failure per vessel movement, which works in the other direction. The net effect varied by scenario type and was not simply a function of traffic volume.
The frequency estimates were expressed as events per year for each scenario and fed directly into the risk screening comparison.
The kinetic energy calculations produced a range of impact energy values across the vessel types and collision modes assessed. The variation is significant. A supply vessel making contact during a routine alongside approach at low speed represents a very different structural demand from a fully loaded tanker drifting into the platform under storm conditions. The assessment made these distinctions explicit, giving the structural team a differentiated set of design load cases rather than a single conservative envelope that would have been difficult to use in practice.
When frequency and impact energy were plotted together using the energy frequency comparison framework, the assessment produced clear guidance on where each scenario sat relative to acceptance criteria. For the supply vessel scenarios, the risk picture was broadly consistent with expectations for a facility of this type and location. The passing vessel drifting scenario, as is typical in such assessments, warranted closer attention not because the frequency was high, but because the potential energy associated with a large drifting vessel is in a different category from infield traffic, and the consequences of structural impact at that energy level need to be explicitly managed.
Specific design, operational, and procedural recommendations were developed from this screening to ensure that all identified risks were managed to an acceptable level.
Based on the collision frequency and impact energy findings, the assessment produced a set of recommendations covering three categories of risk control. These were delivered as part of the FEED package at the point in the project where they could be built into design decisions rather than retrofitted later.
Structural design measures: The impact energy values from the assessment provided the structural engineering team with specific accidental load cases for the platform members most exposed to collision loading legs, risers, and lower deck structural components. Recommendations addressed the specification of structural capacity for critical members, consideration of energy-absorbing detailing where appropriate, and the management of structural redundancy in areas where the loss of a member under collision loading could lead to progressive failure.
Operational measures: Marine operational recommendations addressed the approach corridors and speed restrictions for vessels operating within the platform safety zone, exclusion zone requirements for different vessel categories, and mandatory operational procedures during periods of adverse weather or restricted visibility. The ice season was addressed specifically, with recommendations on vessel attendance and standby arrangements during the December to March period when normal marine response capability is reduced.
Procedural measures: Emergency response recommendations covered escalation protocols for collision events, communication requirements between vessel crews and platform control, and contingency arrangements for managing the immediate aftermath of a collision that affects structural or process systems. In a remote location with limited external rescue response capability, the quality of these procedures has a direct bearing on how well personnel are protected in the event of a serious incident.
There is a practical argument and a technical argument for doing a ship collision study at FEED rather than later in the project.
The practical argument is cost. Structural changes that are straightforward to incorporate during FEED adjusting a member specification, repositioning a component, adding a protective element to a riser become expensive retrofits once detailed engineering has progressed. The window where collision study findings can genuinely influence design without cost penalty is relatively short, and FEED is in the middle of that window.
The technical argument is quality of input. A FEED stage study uses real project data: the actual vessel traffic picture for the Northern Caspian, the actual platform layout emerging from the FEED design, the actual environmental data for the site. The kinetic energy values it produces are grounded in the specifics of this project, not in generic offshore assumptions that may or may not apply. The structural team gets load cases that are actually calibrated to their platform which is the kind of input that produces defensible design decisions rather than conservative guesses.
There is also a Safety Case argument. A credible Safety Case has to demonstrate that the major accident hazards have been identified and that the risks have been reduced to an acceptable level. Ship collision is on that list. A FEED stage study ensures that the Safety Case can make that demonstration based on documented analysis specific to the project not a placeholder that needs to be revisited at a later stage.
Ship collision is one of those hazards that is easy to underestimate because it feels remote a vessel accidentally striking an offshore platform is not something that happens every day at any individual installation. But it does happen across the offshore industry, it has resulted in serious structural damage and personnel casualties, and in shallow water environments with active marine traffic the frequency is not negligible. It belongs in the Safety Case, and it belongs there with real numbers behind it.
The Ship Collision Risk Assessment conducted for the FEED phase of the Kalamkas-Sea and Khazar Field Development Project gave the project team those numbers. It identified the credible collision scenarios relevant to this site, quantified their likelihood using a standards based probabilistic methodology, calculated the structural demand they would place on the platforms, and screened the results against acceptance criteria to identify where risk control measures were needed.
The structural, operational, and procedural recommendations developed from that work were incorporated into the FEED deliverables at the stage where they could shape design decisions, not just document them after the fact. That is what a good technical safety study at FEED stage should do.
If you need a Ship Collision Study, Quantitative Risk Assessment, or other technical safety services for an offshore project, Contact us to discuss how we can support your work.
