New Tool for Riser Break and Post-Break Riser Behavior

Gas Release in Water, Shock, and Impulse Driven Movements

When a Gas Riser subject to high Pretension Forces and high internal gas pressure breaks, the gas will start leaving the riser at each side of the break, and suddenly jets will form. A jet may reach very high velocities, we have seen (in simulations) velocities as high as 900 meters per second.
If a riser part emitting gas is kept at a fixed position, the jet may extend as much as 200 - 300 meters in vertical or horizontal direction before the jet "collapses". It is also interesting to study the diameter of the jet, in our simulations the diameter of the jet is not much more than 2 - 3 meters at a distance of 150 meters from the break.

Let us have a brief look at the physics of the problem "Riser Breaking":
 We assume that the riser is pre-stressed, and the gas in the pipe is under (very) high pressure.
 When the pipe breaks, the elastic energy is released, and the separated ends pick up momentum and part. Further, the gas jet from one separated part will - at least in a short period of time - set up forces on the other part (this effect is not included at present).
 Also, the gas streaming through the riser (parts) will exercise forces on the riser due to change of direction of the streaming gas internally in the pipe.
 In addition to these forces, the riser parts are subject to gravity, buoyancy, and drag forces.
 Further complicating the matter, is the fact that the riser may hit obstacles or go out of the water.
 And, final difficulty, the gas release is depending on the length of the riser ("support volume") and the level above the seabed.

For a customer, we have developed a rather revolutionary consept for calculating the post-break behavior of a gas riser. The physics of the problem is extremely complicated (ref. above). However, some simple assumptions make the problem managable, the accompanying technology developed is unique. Below is the procedure briefly described:
  1) Compute the gas release in water utilizing AUTODYN, so getting the gas stream at every point in the riser at every point in time.
  2) Model the riser system as manufactured.
  3) "Install the riser" (and so include prestress).
  4) Separate the riser at break.
  5) Now take the gas stream to the pipe and account for curvature to compute forces (momentum equation).
  6) Compute the drag forces on the pipe according to Morison's Equation
  7) Account for inertia forces.
  8) Follow steps 2 through 7 for a sufficiently extended time.
  9) 

Input Data

In order to better explain the concept, we simply list the input data (parameters etc) below:

  __________GENERAL__________

    Gravity (m/s^2)
    G-accelEration axis e.g. 1 => +x-dir
    Density of displaced fluid (Salt Water)
    Fluid region (Validity region of Buoyancy and Morison)
    Global sea current velocity vx, vy, vz (m/s)
    Cd = Morison normal drag coefficient
    Ct = Morison tangential drag coefficient
    Cm = Morison added mass coefficient
    Number of cylindrical pipe sections

  __________CYLINDERS___________

  ---Cylinder#1
    Number of subgrids
    Starting point x0 of cylider
    Direction of cylider center axis (Has to be straight at t=0)
    Number of Buoyancy and Impulse volume segments for whole cylinder (Uniformly distributed)
    Name of data file with hydrodynamic diameter D(s) for Buoyancy ("None" => no Buoyancy)
    Name of data file with drag diameter D(s) for Morison ("None" => no Morison)
    Name of data file with flow data (P,ro,v) for Impulse ("None" => no Impulse)

  ---Cylinder#2
    Number of subgrids
    Starting point x0 of cylider
    Direction of cylinder center axis (Has to be straight at t=0)
    Number of Buoyancy and Impulse volume segments for whole cylinder (Uniformly distributed)
    Name of data file with hydrodynamic diameter D(s) for Buoyancy ("None" => no Buoyancy)
    Name of data file with drag diameter D(s) for Morison ("None" => no Morison)
    Name of data file with flow data (P,ro,v) for Impulse ("None" => no Impulse)

Verification

To ensure that concept and implementation is sound, we have put heavy emphasis on verification. Below is a figure showing three curves. They represent the displacement response in one node computed in AUTODYN (with Morison's Equations included), in ANSYS (PIPE59), and by integrating the differential equation governing the dynamics of the problem. As one can see, the results are very much the same. To do the implementation correctly is an extremely difficult task, all credits go to Ola Pramm Larsen.

Slugs in Risers

We plan to develop the concept outlined above to include the effects of multiple slugs at uneven distribution on the movement of the risers. We think this is a "relatively easy" task to perform, as we have all numerical tools and procedures in place.