Page 287 - Practical Design Ships and Floating Structures

262

1 INTRODUCTION
One of the first decisions in the preliminary ship design phase concerns the degree of watertight subdi-
vision needed to ensure a satisfactory stability of the vessel following a collision. Since the sinking of
the ‘TITANIC’ the international community has issued and updated rules and regulations with the aim
to avoid such disasters. The naval architect has to take these demands into account early in the design
phase in order to reduce the adverse effects such measures usually have on the cargo handling and pay-
load capacity of the ship. For passenger ships most national maritime authorities require that the mar-
gin line is not submerged in the case of a side shell damage of a predefined length. This demand can be
transformed into so-called floodable length curves giving the naval architect immediate information on
the necessary number and optimal positions of transverse bulkheads. The required damage stability is
most often obtained by use of a deterministic approach where the residual stability for any location of
the rule damage extent is checked against regulations laid down by the maritime authorities, notably
the IMO. With only transverse bulkheads this can be done simultaneously with the floodable length
calculations. If the residual stability is too low, longitudinal bulkheads can be inserted usually in a lo-
cation at least one fifth of the ship’s breadth inboard from the shell plating in way of the deepest sub-
division load line, as this is the maximum penetration of a rule damage.

However, it has unfortunately been demonstrated in several accidents that the concept of rule damages
of a predefined size does not always capture real life. This has led to the development of the probabil-
istic damage stability regulations, first issued in the seventies in IMO Resolution A.265 as an alterna-
tive to the deterministic approach for passenger vessels. Due to the heavier demands than those of the
deterministic approach and the need for consideration of more damage cases, A.265 is, however, sel-
dom used.

Restrictions are imposed on tankers and chemical carrier concerning the subdivision and residual sta-
bility as prescribed by IMO through the MARPOL and IBC codes, aiming mainly at reduction of the
pollution in case of collision or grounding. For dry cargo vessels no general requirements existed be-
fore 1992 except for the position of an afterpeak and a collision bulkhead.

With the development of the probabilistic damage stability as introduced in SOLAS Part B-1 in 1992
for dry cargo ships and the ongoing harmonisation process towards uniform rules for all vessels, it has
become very urgent for the naval architect to be able to filfil these requirements in a rational manner.
An Attained Index A now measures the residual stability of a vessel, taking into account all possible
sizes of damage. Each damage is weighted by the possibility that such damage can be expected. The
survivability of a given damage is measured in tenns of a factor s, calculated from the properties of the
associated residual stability curve. For a real vessel literally thousands of damage cases must be con-
sidered, cf Koelman (1995), and only extensive use of dedicated computer programs, cf. Jensen,
Baatrup and Andersen (1999, makes the procedure tractable.

Since only a single measure, the Attained Index A, obtained as a weighted average of contributions for
all possible damages is used to assess the damage stability, it is not obvious how to use the regulations
in the conceptual design phase to obtain the most appropriate subdivision. An infinite number of sub-
division layouts might satisfy the regulations. Sen and Gerigk (1992) proposed the use of local attained
indices, aiming at a uniform contribution to the Attained Index A along the length of the hull. This
works well if only transverse bulkheads are considered and no constraints are made on the length of
the compartments, but when longitudinal and horizontal bulkheads are included it is not straightfor-
ward to apply this procedure. Then it is more important in the initial design phase to be able to esti-
mate rapidly the Attained Index 4 based on a simplified layout of the subdivision, and to determine
the sensitivity of A with respect to design parameters as bulkhead coordinates and centre of gravity.
These two topics are addressed in the present paper.