External Organisations

  • DONG Energy A/S, Denmark

View graph of relations

The purpose of the project is to demonstrate the feasibility of developing a tool to assess the risk on offshore platforms due to releases of hydrocarbons. This risk is to be understood as the risk to personnel, expressed in terms of Individual Fatality Risk (IR), Potential Loss of Life (PLL), Fatal Accident Rate (FAR, at platform and workplace level), and group risk (distribution of number of simultaneous fatalities).
The tool will be based on simulation of the interaction between concurrent phenomena following loss of containment, specifically:
•The physical processes (outflow, dispersion, ignition, heat radiation, explosion)
•Detection, alarming and emergency shutdown
•Escape and evacuation
•Impact on persons, escalation and impairment of safety functions
The interaction between these event trees depend on the stochastic outcome of times needed from one event to the other, this set of event trees will be referred to as the dynamic event trees. The simulation technique will be based on Discrete Event Simulation .
The tool will provide an overall framework to describe and simulate the interactions between the above mentioned concurrent chains of events. The final tool is intended to offer a platform for offshore QRA offering flexibility in choosing data and models to describe these single events.
Focus will be on the documentation of the input that users will use to apply the tool, i.e. the data and models to describe the events in the simulation. The QRA process is considered as combining two sets of knowledge:
1) Assumptions – this consists of the majority of the input to the QRA, such as failure data of components, consequence models, expected responses from systems and people, ambient conditions, the interaction between physical lay-out and impacts, and many other simplifications of reality;
2) Probabilistic reasoning – this consists of the logic reasoning that generates all possible combinations of events in the dynamic event trees– according to the assumptions – into final outcomes.
In principle the probabilistic reasoning will include some level of assumption (viz. a limited number of events in the dynamic event trees), but the ideas is that the principle of the probabilistic reasoning can be justified and verified (i.e. it is possible to ensure that a set of input correctly leads to some output) independent of the correctness of the assumptions, so this part can be verified “once and for all”, while the other input, the assumptions, is presented completely and transparently, so that these assumptions can be verified or justified for each study (or consequence models can be verified or validated independently).

The feasibility study will include the following activities:
The probabilistic reasoning: Establishment of event trees
Generic event trees will be developed for the 4 concurrent processes, and links defined between these processes. The event trees will be expressed with “time from release” as the independent variable. The development of the processes will all be expressed with reference to this time.
For demonstration purposes the event tree framework will be implemented in ARENA® Discrete Event Simulation (DES) software. ARENA will allow single simulations to be visualized for verification and testing. The ARENA implementation will be demonstrated to the project partners and its review group. The ARENA implementation is not intended for release to other parties.
Rules for input and documentation
Formal requirements for initial data (information on the platform layout, modules, escape routes, etc.) and the deterministic (for describing physical effects) and stochastic (for describing random events) models contained in the “event boxes” will be established, i.e. dependent parameters and the output of each model will be defined. For example, the model for “jet flame” will depend on release rate, hole size, direction, and deliver as output: flame size, location and radiation contour. Input can be provided either by on-line models, which, for the sake of creating statistical significant results, need to be very simple, or off-line data, such as results from CFD explosion modeling, which requires some processing and interpolation of results in order to transfer results from few scenarios into the DES model.
An important consideration will be to develop criteria or a format for documentation of the assumptions with respect to completeness, transparency and traceability.
For the feasibility study, very simple models for the events will be selected, mainly based on the “Yellow Book” or other similar, relevant sources. Only models necessary to create sufficient credibility of the feasibility study will be included, which e.g. means that the easiest type of release (single phase gas release) will be included.
DTU will deliver to the project partners by the end of the feasibility study:
• A proposal for transparent and complete documentation of QRA assumptions (input), to be delivered by means of a technical report describing structure, and examplification using the demo case study;
• A description of a framework for a QRA technique based on concurrent (dynamic) event trees, evaluated using Discrete Event Simulation, to be delivered by measn of a technical report;
• A demonstration of that framework with a simple set of models and data, to be delivered as a workshop demonstration and output by means of screen dumps with explanatory text.
StatusCompleted
Period01/12/201230/06/2013

    Keywords

  • Risk Analysis, Oil and Gas, Offshore, Discrete-Event Simulation

Activities

Download as:
Download as PDF
Select render style:
ShortLong
PDF
Download as HTML
Select render style:
ShortLong
HTML
Download as Word
Select render style:
ShortLong
Word

ID: 52966646