For this validation three different test design cases are applied to the network and the
output data are
compared to six existing yachts of an equal type that are scaled to the required length using the formulae by M. Barkla
performance of the The validation is
[1]. To get a fair judgement of the network these yachts are not
included in the list of basis yachts that were used to produce the training data.
only performed for the In-Out
approach, because the performance targets of the basis yachts are unknown.
The results are displayed in the Figures 6 to 8. Design Case 1
Loa Boa
SA Figure 6:
[m] [m]
T [m] [m²]
Mast Height [m] Design Case 2
Loa Boa
SA Figure 7:
[m] [m]
T [m] [m²]
Mast Height [m] Design Case 3
Loa Boa
SA Figure 8:
[m] [m]
T [m] [m²]
Mast Height [m]
Arithmetic Mean Variation -
-0.08 -0.11 -6.54 2.45
Validation of Test Design Case 3
These three examples show that the values derived by the network, even for input values that are not included in the training data, meet dimensions of existing yachts with satisfying accuracy.
The aspect ratio of the sails, the Dellenbaugh angle and vertical centre of gravity of the basis yachts are unknown and therefore the values derived for the test design cases cannot be validated.
Arithmetic Mean Variation -
-0.28 0.03
-10.33 -0.69
Validation of Test Design Case 2
Arithmetic Mean Variation -
-0.12 0.01
-17.63 -0.75
Validation of Test Design Case 1
The sensitivity analysis is a procedure to determine the sensitivity of each output parameter to changes in the input parameters. If a small change in an input parameter results in relatively large changes in the output parameters, these parameters are said to be sensitive to that input parameter. [8]
The values derived by the sensitivity analysis are
basically the rates of change in the value of one parameter achieved by changing another parameter in a defined step. This
step is often chosen to be the
standard deviation, however in the case at hand the ranges of the different parameter are just divided into 10 parts of the same length.
The sensitivity analysis is a very useful approach to give a better understanding of complex systems. It is a tool
to map the functional relations between the
different parameters. With this knowledge a more aimed variation of parameters can be performed to achieve a certain result.
The sensitivity analysis is performed both for the In- Out approach and the Out-In approach. The results can be seen in the Figures 9 and 10.
Sensitivity About the Mean
1,2 1,4 1,6
1
0,2 0,4 0,6 0,8
0
Displ. Dellenbaugh Angle
Aspect Rat io Sails
Input Name Figure 9:
10 12 14 16
0 2 4 6 8
Loa Boa
T Mast Heigth above Deck
Input Name Figure 10: Sensitivity Out-In
The sensitivity analysis maps successfully the relations between the
different ©2007: Royal Institution of Naval Architects parameters. Most of B-37 these
VCG above Water-line
Sensitivity In-Out Sensitivity About the Mean Sail Area
Loa Boa T
Mast Heigth above Deck VCG above Water-line
8. SENSITIVITY ANALYSIS
Displ.
Dellenbaugh Angle Aspect Ratio Sails
Sail Area
Sensitivity
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54