Kaplan, Sargent and Goodman 
fication for vertical plane motions of SES craft. 
Although there has only been a somewhat limited degree of 
success in identifying (i.e., finding correct numerical values) the 
various important stability derivatives for heave and pitch motions 
using the trajectory data obtained from simulation studies with a 
complete mathematical model, important insight into the require- 
ments for achieving improved results was obtained. The importance 
of seal dynamics and forces ; the lack of significant influence of 
certain coupling coefficients ; the effect of measurement of more 
state variables ; and the influence of sampling time and integration 
step size on stability and convergence of the solutions, are examples 
of the various conclusions obtained from these particular studies. In 
order to obtain more understanding of methods for treating full scale 
data, which is the ultimate objective, the methods were applied to 
some full scale data. 
This data was recorded on magnetic tape, in digital form, 
with the 'important'' information necessary for dynamic analysis 
provided by samples at the rate of 100 cps. The data useful for the 
present analysis was CG heave acceleration (z), bubble pressure 
(p), and pitch angle (6). Pitch rate data (@) was recorded ata 
much slower rate (10 cps) and hence was not compatible with the 
pitch angle data, as well as possibly losing important information due 
to improper sampling at too low a frequency (the Nyquist sampling 
rate for the present SES craft corresponds to 20 cps, or sampling 
every 0.05 sec.). No information on the heave motion except heave 
acceleration was available and hence z and z would have to be 
found by integration operations, which have more stringent sampling 
and time increment requirements. In addition, the actual data mea- 
sured during the full scale tests suffered from telemetry errors that 
result in sudden ''drop-out'' of data. This caused ''spikes'' in the re- 
sulting data, above and beyond the generally spike-like appearance 
of much of the data, especially for bubble pressure and accelerations, 
and this data was specially treated by filtering and interpolation pro- 
cedures to assure a relatively continuous and ''smooth"' data output. 
Thus the measured data itself may have certain inherent defects from 
the point of view of its applicability as a source of trajectory data for 
system identification, while it is still perfectly good for its original 
purpose of providing information on vehicle response characteristics 
(e.g. level of accelerations, maximum and average pitch angles- 
bubble pressure variations, etc), 
The experience in applying this method of system identi- 
fication to full scale data obtained from tests of the SES craft inir- 
regular waves provided valuable guidelines for future applications. 
1658 
