€— 
res 
pre re orne oran 
— 
IIS IMAGES 
The work of marking the more than 200 points, 
among which the data points to be used for 
orientation and those to be measured, was very 
expensive. 3 people were needed for 4 days and a 
crane for access to the structure's higher areas. 
The onshore data  point's coordinates were 
determined using the topographic technique of 
forward intersection carried out using a WILD 
T2000 theodolite. 
3 sets of measurements were made, each with double 
straight/reverse readings. The points on the 
Structure were defined during the loading phase. 
In this case an operator collimated the 3 points 
with a prism using a diastimeter in the few 
seconds that the structure was stationary. 
A single reference system was used for all the 
data points' coordinates defined topographically. 
This system will then be used also for 
measurements carried out at the stereoplotter. 
Precision reached in defining the data points was 
t 1 mm. 
4.3 Photogrammetr 
The survey to give an undeformed reference 
geometry was carried out from bases 1 and 2 
(figure 1), shooting respectively the structure's 
bottom and top. 
The bases were distributed following an alignment 
parallel to row 2 of the jacket and 50 meters 
away. One metric camera placed at each of the 
four positions was used. 
8 photograms were taken with a vertical 
inclination of +15°. 
The survey of the loading was carried out from a 
double base with 3 cameras. In fact it was 
necessary to sufficiently cover the  jacket's 
surface with the photograms. And, if this was 
easy to achieve onshore it was not easy for the 
part in the boat. The problem was overcome by 
horizzontally inclining the cameras. 
Three cameras were used, two were connected to 
each other and a third was independent but 
synchronised manually. 
Loading operations were carried out from 5.00pm to 
midnight on 10/5/90. The most important phases of 
the operation therefore took place at night and 
very long exposition times were needed. 
Compatibly with the loading movements, exposition 
times of 15-20 seconds during some of the extended 
stops. 27 shots were taken using a KODAK PAN 200 
film with a 24 DIN sensitivity. Every camera had 
a container for roll-films equipped with an 
aspirator to keep the film flat at the moment of 
shooting. 
5. MEASUREMENT OPERATIONS 
5.1 Photogram orientation 
The topographic calculations were made using a PC 
Olivetti M24 with dedicated software. 
Plotting was made using a WILD AC1 analytical 
plotter conected to a Nova 4X Data General, with a 
PMO package programs. 
The photograms were selected on the basis of 
qualitative criteria and in such a way as to have 
a detailed description of the loading phases. 
Despite the availability of three photograms for 
each phase, having used 3 metric cameras to shoot 
the loading phase, only the photograms of the two 
A and AR cameras of base 4 (sea-side) were used. 
In fact, the third photogram did not offer great 
improvements for the number of points and the 
precision of the measurements. This photogram was 
used only in one case where the orientation of a 
sea side couple was difficult. 
   
   
     
  
  
   
     
    
   
    
    
    
   
    
    
  
    
    
    
    
   
   
    
  
   
    
  
   
   
  
   
   
   
    
  
   
  
    
     
   
  
    
    
  
    
  
   
  
  
    
  
     
     
  
    
   
    
    
   
    
  
   
  
     
  
   
     
     
   
  
   
    
    
    
    
    
      
5.2 Plotting 
14 pairs of photograms were plotted giving a good 
description of the structure's deformative 
phenomenon. 
The point which can be collimated by the operator 
were measured for each stereoscopic model. 
Contrary to expectations, the measuring 
procedure’s limits were not the lack of visibility 
of the points due to lack of light or great 
distance. The biggest problem was the blocked 
view because of different types of obstacles. In 
fact, to be used, each point had to be seen on 4 
photograms, two for the reference pair and two for 
the comparison pair. 
On the other hand it was decided not to measure 
the points without a target to avoid reducing the 
precision. 
The quality of the measurements obtained by the 
plotting may be estimated approximately on the 
basis of the measurements at the data points 
defined topographically. 
These data were included on the plot protocols. 
The average of the standard deviation on the data 
point readings was the following: 
reference pairs 
Py = 0.0014 mm 
Dn = 1.43 mm Dh = 1.69 mm 
pairs during loading 
Py = 0.0023 mm 
Dn = 2.21 mm Dh = 2.83 mm 
where Py is the residual parallax value, Dn is the 
accuracy in the plane normal to the optical axis 
and Dh is the accuracy of the coordinates in the 
direction parallel to the optical axis. 
6. DATA PROCESSING 
The data obtained by the plotting measurements 
were recorded according to the string type: 
ID, X, Y, Z 
where ID is the identifier for each point, and X, 
Y and Z are the coordinates according to the 
absolute reference system. A file was created for 
every position of the platform during the 
load-out. 
These data were then transferred to a personal 
computer where the subsequent processing was 
carried out. 
In the first phase the points were ordered 
according to the identifier's progressive order. 
Then every data file for each loading phase was 
compared with the file obtained during the 
reference phase following this process: 
ID, X(0), Y(0), Z(0) reference phase file 
ID, X(i), Y(i), Z(i) loading phase file 
DX(i) » X(i) - X(0) 
DY(i) s Y(i) - Y(O) 
DZ(i) = Z(i) - Z(0) 
In order to simplify the interpretations, only the 
DZ(i) vertical displacement were considered in the 
following steps. 
Table 1 shows the results obtained by this 
operation. Each column represents the movements 
of all the points for each phase. The points 
which were not measured in any of the loading 
phases or in the reference situation were