was due to the fact that the wave generation system did not allow absorption of reflected 

 and re-reflected waves within the flume; so wave energy built up in the flume during 

 testing. It should be noted that systems that absorb reflected energy are only partially 

 effective, and so it is virtually impossible to prevent wave energy increase due to re- 

 reflected waves in stability tests. 



The wave energy in this experiment reached an equilibrium state quickly 

 during each run due to the relatively long wave period. A simple analysis was done to 

 determine the rate of increase of wave energy. For this analysis, the wave time series 

 were divided into eight segments, each of 4,096 points or 3.41 min, except the first 

 segment which was 2,048 points. Both the first and second segments started at the 

 400th data point or r = 20 s and the final segment ended at the 16,784th data point or f = 

 14.0 min. The segments overlapped by 2,048 data points, and each segment represented 

 approximately 120 mean wave periods, except the first segment, which was about 60 

 wave periods long. For each segment, the incident H„g was calculated using the 

 technique described above. Figures 4.6 and 4.7 show the wave height variation for h, = 

 1 1.9 cm and h, - 15.8 cm, respectively. In these figures, the ratio of the incident 

 segment wave height H^^tsd) to the overall wave height H„Jit^ is shown as a fiinction of 

 normalized duration t^^/t^ where t^^ is the duration to the center of the segment and t^ = 

 15 min is the total run duration. As can be seen from the figures, the wave height 

 reached an equilibrium very early during each run. Based on these observations, the 



70 



