The report contains a concise description and discussion of three incidents involving gas discharges under water on the Norwegian shelf in recent years. These incidents were selected with the intention of illuminating the characteristic and important features of discharges under water, and the problems related to these incidents.
The starting point for the project was the chain of events that leads to leaks and further to the loss of assets such as life and health, loss of/damage to the environment, as well as material/financial assets.
Efforts to mitigate the risk can be aimed at all stages of the chain of events, with efforts being implemented in those areas where the benefit/utility value is greatest in relation to disadvantages/costs.
This requires knowledge about the individual stages in the chain of events, how they can be influenced/guided, and the interplay between the various factors. The report discusses some of the factors that affect the chain of events as regards:
The work was gradually concentrated on the actual discharge scenario, from leak/rupture of a pressurized pipeline on the seabed, the flow/dispersion of the gas in the water masses up through the water column and subsequent dissipation in the air above the sea surface.
Good technical insight into this scenario is essential for understanding the risk picture and the ability to influence the risk associated with gas discharges under water.
A central part of the project has been mapping the methods and models developed to describe and calculate the progression of a gas discharge, the movement of the gas up through the water column and further dissipation in air.
Based on a more precisely defined test case, four companies; DNV, Scandpower, Safetec and Lilleaker, performed independent calculations of a number of different discharge scenarios using their respective models.
The results provided a picture of the extent to which the industry has a mastery of this technical area, and would be a good starting point for a potential improvement process.
The project was concluded with a seminar on 15 November 2006, where 75 people attended. The results of the project were presented and discussed in the seminar.
The conclusion is that the mutual deviations are too great, and that there is a significant potential for improvement. Continuation of the work has now been initiated.
The project participants want to cooperate in 2007 to identify the underlying causes of the very different results. The sources of these differences can be divided roughly into the following areas:
Factors which must be clarified or worked on in the future:
1) Understanding and modeling of the discharge under water, from the actual discharge on the seabed up to and including the transition from sea to air (when the gas breaks through the sea surface)
The challenge here is to establish the basic understanding of the physical phenomena, and the ability to translate this into models that are sufficiently correct and precise for the various purposes.
It is particularly in connection with the large, massive discharges that today's models are inadequate. We realize that there is a need for a more fundamental method of attacking the problem, and probably also a need to develop new models.
2) Dissipation of gas in the air above the sea
In this area, the problem can be solved by making new calculations where the problem (the scenario) will start with a set of precisely defined "discharge profiles" on the sea surface, and a more detailed and differentiated description of the results of the calculations. The results will then be compared with a view towards clarifying whether the dissipation calculations in air are the source of the observed differences.
3) Physical experiments on a large or medium scale
Today's models are based on small-scale experiments, all less than 1 kg/s. A complete rupture of one of the large pipelines that transport gas from the Norwegian shelf will result in initial discharge rates of 20,000-30,000 kg/s. This means that we have no possibility of verifying the models used in realistic experiments. Large-scale experiments will be complex and costly. However, there is a great deal of interest in this, and as a first step there should be a "study" of:
Contact person in the PSA:
Odd Tjelta, project leader