f^; 
A 
U A 
A 
9 
3.0. 
sheet 3 
cont'd ® ® 96 ee 90 
dependent changes and thermal response times may be investigated 
both in transient handling and during slow handling and stabilized 
running sequences. The capabilities of the imaging system must 
be matched to the information rate required. 
In a radiographic exercise on a development engine it may not 
be possible ‘a priori' to determine or specify the exact point in 
time when the event may occur within the engine that it is desired 
to record, analyse and measure. 
For this reason, a real time X-ray television or fluoroscopic 
system has been developed (Fig. 5) for use at high energies and 
provides a continual surveillance of events within the engine at 
T.V. framing rates. This permits a qualitative evaluation of the 
engine movements and allows a determination of the point in time 
at which the event occurs and also the rate of change of com- 
ponent movement. An understanding may thus be achieved of the 
mechanism producing the change - for instance thermal or pressure 
changes. The images are recorded on a video tape recorder for 
subsequent evaluation and analysis and may, if desired, be trans- 
ferred to ciné film using teleciné equipment for magnified frame 
by frame analysis. 
Once the time of occurrence of the event has been determined, 
then a film radiograph is exposed to record the event for quan- 
titative analysis. Examples of the films used in our programmes 
are shown in Fig. 6, and vary from the slow high resolution films 
used for steady state radiographs through the medium speed films 
to cover the 0.5 to 5 seconds exposure range to the high speed 
film capable of exposures within durations of 0.05 to 1.0 sec. 
Examples of radiographs produced from our programme are 
shown on the light boxes. 
The images on the radiographs show a fiducial datum bar which 
is arranged adjacent to the cassette and is made of a radio- 
opaque material. This is present on the static and all the dynamic 
radiographs and permits the differential movements to be determined 
in X and Y axes. 
For analysis the radiograph is placed on an X-Y co-ordinate table 
and using a magnifier and graticule eyepiece, the position of the 
component edge is read by the photogrammetrist and punched into the 
computer. The radiographs are individually analysed by a single 
photogrammetrist who may make multiple readings. They are then 
further analysed by other photogrammetrists who punch their 
readings into the computer. These are then automatically analysed 
and a statistical distribution obtained. Errors or 'rogue' readings 
are highlighted and subjected to further examination. 
Using this method, despite the apparent subjectivity it pro- 
vides reliable data with a typical 95$ confidence interval of 
0, 10-0, 15 mm. 
However, many problems require greater accuracy or a less 
subjective approach and these difficult radiographs are analysed 
on a custom made microdensitometer to provide a very accurate 
density profile across the area of interest. Suitable algorithms