Beruflich Dokumente
Kultur Dokumente
The Use of Transport Factor as a Very Simple Model for Ship Design Exploration
Chris B. McKesson
School of Naval Architecture and Marine Engineering, University of New Orleans, New Orleans Louisiana, USA
ABSTRACT In the present paper the author explores the use of Kennell's
The author has been a proponent of Very Simple Models Transport Factor "TF" as a Very Simple Model,
(VSMs) for ship design. VSMs offer the ability to explore a demonstrating good success in using this method to
design space rapidly and with a minimum of input, normalize a family of dissimilar ships, causing their relative
frequently saving the design team many weeks of effort merits to become more obvious.
exploring fruitless corners of the space. Of course, VSMs Transport Factor is Kennell’s 1998 expansion of von
are limited in fidelity and do not replace detailed design Karman’s 1950 investigations into transportation
models. Instead they should be viewed as complementary, effectiveness. Kennell has published several papers in
in the same way that a high-powered telescope is often fitted which he has used TF as a tool for design analysis and
with a small spotting-scope which is used first. assessment (Kennell 1998, 2001, and 2010 and Kennell &
In the present paper the author explores the use of Kennell's Templeman 1998). McKesson has built upon this work by
Transport Factor "TF" as a Very Simple Model, using TF as a tool for design synthesis (McKesson 2011).
demonstrating good success in using this method to By application of the relationships revealed in Kennell’s
normalize a family of dissimilar ships, causing their relative analysis, it is possible to rapidly predict the characteristics
merits to become more obvious. that a ship will have, from a very sparse set of early design
requirements.
Using TF to predict ship characteristics has many powerful
applications: It can immediately indicate basic feasibility, This prediction of ship characteristics has many powerful
saving the team from pursuing infeasible missions. But applications: It can immediately indicate basic feasibility,
even more important than this “go/no-go” sort of insight, the saving the team from pursuing infeasible missions. But
TF-driven design prediction can identify how hard the even more important than this “go/no-go” sort of insight, the
design task is likely to be, and where resources will be most TF-driven design prediction can identify how hard the
needed. Conversely, a TF-driven synthesis can be used to design task is likely to be, and where resources will be most
investigate the benefit of technology changes, so that best needed. Conversely, a TF-driven synthesis can be used to
allocation of R&D investments can be made. investigate the benefit of technology changes, so that best
allocation of R&D investments can be made.
And best of all, these benefits can be pursued very quickly –
in a fraction of the time of ‘conventional’ design synthesis. And best of all, these benefits can be pursued very quickly –
in a fraction of the time of ‘conventional’ design synthesis.
The paper takes the reader through the derivation of the
method, and then into a novel example of the virtues and 2.0 TRANSPORT FACTOR (TF) DEFINED
power of the method. The example consists of a design-to- Kennell introduced the Transport Factor in Kennell (1998).
design comparison, wherein a design is evolved to a new set At the same time McKesson was working along similar
of requirements so that differences with competing designs lines, developing a Very Simple Model that was described
can be cast into greater contrast. in McKesson (2006 & 2009) and more recently in
KEY WORDS McKesson (2011). McKesson has also been teaching both
the Kennell and McKesson methods as part of the
Ship Synthesis, Very Simple Models, Transport Factor curriculum of Naval Architecture at the University of New
1.0 INTRODUCTION Orleans.
The author has been a proponent of Very Simple Models Recently these three threads have converged and the author
(VSMs) for ship design. VSMs offer the ability to explore a has been building a VSM around TF, as follows:
design space rapidly and with a minimum of input, First, let us recall the definition of Transport Factor:
frequently saving the design team many weeks of effort
exploring fruitless corners of the space. Of course, VSMs
are limited in fidelity and do not replace detailed design
models. Instead they should be viewed as complementary,
in the same way that a high-powered telescope is often fitted Where “K” is a constant as needed to make TF non-
with a small spotting-scope which is used first. dimensional. In metric units with ship weight in tonnes,
speed in meters per second, and power in kilowatts, K is
9.81.
© 2011 American Society of Naval Engineers 735
3.0 THE DECOMPOSITION OF TF g = gravitational constant (e.g. 9.81 m/sec^2)
Kennell has shown importantly that TF can be decomposed vol = volumetric displacement of ship (e.g.
to parallel the ship weight breakdown. Thus for example, displacement in tonnes / 1.025 (tonnes/cubic meter))
consider a USN SWBS-based weight breakdown as follows:
Lightship weight =
And
Total weight = Lightship + Cargo + Fuel
This weight breakdown then leads to a completely parallel
TF breakdown, as:
TFTOTAL = TF100 + TF200 + TF300 + TF400 + TF500 +
TF600 + TF700 + TFFUEL + TFCARGO
Now let us consider the individual characteristics of these
TF components:
3.1 TF-TOTAL Figure 1- McKesson's 2011 presentation of the curve of
"Observed Best Attainable" TF for a large number of ships
Specific Power, TF, and L/D are concepts that are nearly
interchangeable. Specific Power is von Karman’s metric for
transport effectiveness, and is calculated as Power / (Weight As we move forward from this point, let us first pause to
x Speed) (von Karman 1950.) Transport Factor is Kennell’s reconsider the meaning of the “Best Attainable” curve. This
inversion of Specific Power, and is calculated as Weight x curve is an observed frontier of performance, without stating
Speed / Power. Lift to Drag ratio is the observed ratio of the what efforts must be undertaken to actually attain this level
ship weight (lift) to ship resistance (drag). In most cases of performance. In general it is reasonable to expect that
ship resistance is not known, only ship power is known, so this level of performance is attained by selecting optimal
the drag is estimated to be OPC x (Power / Speed), where values for ship slenderness, prismatic coefficient, and so
“OPC” is an overall propulsive efficiency. An important forth, unfettered by service constraints. Other ships, which
point to take from this derivation is that L/D = TF/OPC. are not free to take optimum L/B, or optimum Cp, will
Note that in the case of powered-lift craft such as SES, the generally have inferior TF performance.
lift-power is included in the L/D or TF calculation, without Because this is a curve of TF, it also includes the propulsive
distinguishing it from the propulsion-power, thus creating a efficiency. Thus the ‘best attainable’ frontier suggests that
‘virtual drag’ that accounts for the lift power. the designer was free to select the best performing and
McKesson (2006 & 2009) and Kennell (2001) (and earlier, optimal propulsion suite.
von Karman) both provide contours for the observed state- The crux of this portion of the present model is to
of-the-art of “best attainable” TF. (In McKesson (2006) this understand that the overall TF performance, including
is erroneously reported as best attainable L/D. In fact, this performance predicted by the ‘best attainable’ curve,
is only L/D if OPC = 1.0.) McKesson shows that a large represents all of the hydrodynamics of the ship: It contains
body of data collapses into a single frontier when analyzed all of the hull resistance performance and all of the
using a non-dimensional speed, and he offers a simplistic propulsion efficiency.
equation for predicting the “best attainable” whole-ship TF
When using the ‘best attainable’ curve there are several
for any combination of size and speed. McKesson’s 2009
interesting realizations. The most important of these is the
data is presented in Figure 1. The solid black line is
realization that the curve does not tell the user how to get
McKesson’s curve of “observed best attainable” for TF
the ‘best attainable’, merely that it should be possible. That
performance. As may be seen the curve is not rigorously
is to say, the curve provides no information as to what hull
the absolute best performance ever observed, but is at least
parameters to use for any given point, what Prismatic
on the upper edge of the attainable performance space. It is
Coefficient to pick, what the Length-to-Beam ratio is, etc.
also computationally simple, being:
Indeed, it does not even tell whether the ship is monohull or
multihull, waterjet or propeller driven.
Where: 3.2 TF-FUEL
FN = volumetric Froude Number: Kennell (2010) has shown that TF-fuel depends linearly
upon range and specific fuel consumption (SFC), but not
FN = upon speed nor any ship hydrodynamic feature. (Note that
this has been derived for a Navy-style “DDS 200” fuel
calculation, and not a Breguet range calculation.) This is a
very surprising insight: In other words a ship’s TF-fuel will
be the same for the same range and machinery type
(specifically SFC), regardless of whether the ship is a
736 © 2011 American Society of Naval Engineers
monohull, trimaran, sea train, etc. So if two ships have the He found that W-100 was proportional to CN0.90, and other
same propulsion technology – e.g. diesel-with-waterjets – lightship weight groups were proportional to CN0.825. For
and the same range, then they will have the same TF-fuel. purposes of a Very Simple Model, however, we can safely
More generally, if we change the ship’s propulsion ignore these exponents and state that the weight is
technology we can capture the effect of this simply by approximately linearly proportional to the pseudo-volume of
changing TF-fuel in proportion to the change in SFC. “Cubic Number.”
The formula for this is: Cubic Number is the volume of a rectangular prism that
encloses the entire ship. For ships of similar form, we may
TFFUEL = K•(SFC•Range)• assume that this prismatic volume is some constant times
(1.0+fuel rate correction factor)• the “real” total volume of the ship, meaning the sum of the
underwater and above-water volumes.
(1.0+plant deterioration factor) ÷
However, we may further choose to assume that the
(1.0-tailpipe allowance factor) abovewater volume is some constant multiple (or fraction)
Or approximately: of the underwater volume for a fixed mission (e.g. tanker,
ferry, etc.) This then would collapse this volume
TFFUEL = .003622 •SFC •Range formulation to state that total ship volume is directly
proportional to underwater ship volume, or in other words to
ship weight. Thus in this simplification of Benford’s
(with SFC and Range in English units of lbs/hp-hr and formula, we can replace the structural weight density with
nautical miles respectively, and standard values of the three simply a structural weight fraction: W100/WTOTAL. It is
“factors”) simple to see that this is equivalent to TF100/TFTOTAL.
The ship of course has other consumables, but it seems
3.5 TF200
reasonable to assume that their weight may also be a linear
function of range. So for purposes of a VSM we may The TF of propulsion machinery is, of course,
combine the consumables with the fuel into a single TF, TF200 = K •W200 •V / SHP
altering the constant “0.003622” as required, perhaps to
Note that this contains a very common naval architect’s
simply 0.004.
“VSM” of machinery: W200/SHP. In McKesson 2006 and
3.3 TFCARGO 2011 this was the parameter McKesson labeled “weight of
Up to this point we have developed prediction formulas for power.” And it is commonly understood that this “weight of
TF-fuel (based on Range and SFC) and TF-TOTAL power” or “power density factor” is mainly a function of
(assuming ‘best attainable’ performance.) In terms of the machinery technology choice: Low speed diesels are heavy
total ship, this leaves only TFLIGHTSHIP and TFCARGO to be per kW, gas turbines are light per kilowatt. But within a
found. In the paragraphs that follow we will look at the technology class the “weight of power” may be treated as
governing relationships for these items as well. constant: Twice as much power weighs twice as much. (A
more complex, but still simple, model of the “weight of
TFCARGO is in most cases either (a) given, or (b) the ultimate power” could easily be constructed, and would not
result – rarely is it an intermediate variable. materially change the benefits of this use of TF as a VSM of
In the case of an analysis the TFCARGO is given, because the ship synthesis.)
cargo weight is given as a part of the highest level This suggests that we can have a Very Simple Model of
description of the ship. machinery technology, based on a ‘menu’ of technologies. I
Alternatively, the TFCARGO may be the output from a TF- offer the following as starting points in such a menu:
based VSM: The sum of all the TF components must equal High speed diesel / Waterjet (Fast Ferry propulsion):
the TFTOTAL. This means that one predicts the best
attainable TFTOTAL, and then calculates the component TF W200/SHP = 10 lbs / hp
values for TF100 through TF700, TF-fuel, etc. Simple Diesel – Electric / Azipod:
subtraction will yield the amount of TF that remains
W200/Ps = 35 kg / kW
available for cargo carriage…if any.
TF200 must be calculated on a ship-specific basis by taking
3.4 TF100
the appropriate machinery power density from the menu and
McKesson (2006) proposes the use of a Cargo Carriage multiplying it by ship speed. Thus the TF200 for Azipods on
Multiplier which is approximately the same as saying that a 10 knot ship will be precisely half the TF200 for Azipods
W100 is proportional to WCARGO. This reflects the fact that on a 20 knot ship – but that should be the ONLY difference
one part of the ‘job’ of the ship structure is to enclose the between these two cases.
cargo.
3.6 TF300
Benford (1976) earlier proposed to model structural weight
The electrical plant weight on a ship will depend
using ship Cubic Number “CN”, where:
approximately-linearly on the electrical load of the ship.
CN = L •B •D / 100 The electrical load of the ship (Redmond 1984) varies
New Ship
Assumed new displacement 5280 6151 7172 8356 t
Change of speed will change TF attainable
Fn‐old 0.99 0.99 0.99 0.99
TF‐predicted‐old 46.47 46.47 46.47 46.47
TF‐attained‐old 39.18 39.18 39.18 39.18
k 0.84 0.84 0.84 0.84 k = TF‐attained / TF‐predicted
Speed‐new 30 30 30 30 knots
Fn‐new 1.19 1.16 1.13 1.10
TF‐predicted‐new 29.00 30.90 32.97 35.19
TF‐"attained"‐new 24.45 26.05 27.80 29.67 << This is the new TF‐TOTAL "budget"
for the new ship
Range changed, so the TF expended on range will change
Old range 4000 4000 4000 4000 n.mi.
New range 4000 4000 4000 4000 n.mi.
TF‐consumable‐new 13.73 13.73 13.73 13.73 << 0.00343 x Range
We assume machinery technology will not be changed
TF‐200 1.78 1.78 1.78 1.78 << 0.05936 x SPEED
We assume the “13456” technology will not be changed
TF‐13456 5.70 6.07 6.48 6.91 << 0.233 x TF‐TOTAL
Subtraction identifies the amount of total TF available for Cargo
3.25 4.48 5.81 7.25
We have to re‐dimensionalize the TF to find the total power of the new ship
Predicted Power 32698 35746 39066 42642 kW
We know what cargo we WANT to carry, so we can calculate the NEEDED TF‐CARGO
Cargo 1500 1500 1500 1500 t
Needed TF Cargo 6.95 6.35 5.81 5.33
‐53% ‐30% 0% 36%
^^ This is the ship that meets the new requirements
Table 3 - A range of reconverged designs derived from the Baseline by the VSM
Where: FN =