Page 90 - Practical Ship Design

60 Chapter 3

The line for bulk carriers and tankers seems quite satisfactory with few data
points deviating by more than 0.025 from the line. If this deviation is related to a
“line” value of say 0.825, the percentage error in a displacement estimated in this
way will not exceed about 3%. If this error is distributed pro-rata between the three
dimensions of length, beam and draft, the error in each of these at this earliest step
in the design process will be less than 1%. In fact it may be even less than this,
since it seems likely that divergences from the mean line may be due more to the
ships in question having block coefficients differing from that given by the mean
line in Fig. 3.12 than because their lightship weights differ markedly from the
‘‘line’’ value.
If at the next step in the design process a block coefficient corresponding to the
mean line is used when deriving the main dimensions, there should be the happy
result of two deviations at least partially cancelling one another, leaving only a
small adjustment to be made to C, when final design weights become available.
Although the deadweight/displacement ratio of container ships is not of much
practical use since, as already noted in 32.9.2, these vessels are volume or stability
critical rather than weight based, an approximate line for this type of ship is given
on Fig. 3.3 but not surprisingly there is quite a large scatter of data points. The
same applies to refrigerated cargo ships which are also volume-based designs.
A deadweight/displacement ratio for either of these ship types and indeed for a
tanker should be corrected using Fig. 3.2 before being used in association with the
design deadweight.
It may seem strange that the base used for the plot in Fig. 3.3 should be the
deadweight itself, but in this case it is used primarily as an indicator of size and
because at the time when a designer is still trying to establish Kd it is the only
parameter available to him.
A number of other lessons can be learnt from Fig. 3.3. Firstly, it seems clear that
the biggest factor in determining the deadweight/displacement ratio is not the ship
type as might have been expected but the block coefficient, whilst next in order of
importance seems to be ship size, with ship type coming third, suggesting that a
regression analysis of this type of data might produce an interesting result.
When drawing Fig. 3.3, only a limited number of data points were used and even
so a number of the points diverge significantly from the suggested lines. Without
plans of the ships concerned it was impossible to establish whether there are good
explanations for these divergences but as the figure is only intended to provide a
quick approach to initial dimensions, it is probably accurate enough and a reason-
ably good example of “the art of drawing sufficient conclusions from insufficient
premises” quoted on the title page. It was interesting to note that points relating to
double skin tankers and bulk carriers plotted quite close to the line indicating that
with good design the weight penalty - if not the first cost penalty - for this type
of improved construction is remarkably small.