After reaching the quasi-steady state the wave gage was positioned 

 at the first station. The motion was observed on the paper tape of the 

 Sanborn recorder and was visually determined to be periodic. With the 

 maximum paper speed, 100 millimeters per second, three to four wave periods 

 were recorded. The gage was moved to the next location, 10 centimeters 

 away, and the procedure was repeated. This was done over a distance of 

 approximately one wavelength of the incident waves so that at least two 

 maxima and minima were recorded. 



Of the three to four wave periods recorded at a given station one 

 was chosen for analysis. This variation of the free surface during 

 one wave period was digitized manually at intervals of 1/20 the wave 

 period. Since the motion is assumed periodic the two end points of the 

 digitized data should be identical and the mean value of the two end 

 points was chosen whenever they were not exactly the same. This way 

 20 equally spaced (in time) values of the surface elevation were 

 obtained and these values were used as input to a simple computer 

 program which performed a Fourier series analysis of the data and gave 

 the amplitude of the first harmonic motion as output. 



When realizing that each experimental run required the measurement 

 of the wave motion at some 30 stations it is quite obvious that the 

 manual procedure of digitizing the wave records means that it is an 

 extremely time-consuming effort to obtain the reflection coefficient. 

 However, the end result (Fig. 20) rewards the effort in producing 

 reflection coefficients which vary only within +_ 0.02 with choice of node 

 and antinode. The experimental data listed in Appendix B (Tables B-2 

 and B-4) were all obtained and analyzed in this manner. 



Modifications of the Hewlett-Packard Computer System at the Ralph M. 

 Parsons Laboratory made it possible to interface this computer with the 

 wave tank experiments. This enabled the tedious manual reduction of the 

 data to be circumvented and to feed the experimental data directly into 

 the computer. A multifunction meter triggered the computer to start 

 taking data by imposing a high voltage. After being triggered the 

 computer starts taking data at the rate of 14.5 readings per second, i.e., 

 a reading per At seconds, where: 



At = 0.069 seconds. (C-2) 



The computer program was designed to take a total of 50 readings, 

 i.e., cover a period of approximately 3.5 seconds, thus ensuring that an 

 entire wave period is recorded. Of these 50 equally spaced values the 

 first 15 are discarded to avoid transient effects and from the 16'*-" 

 reading on, the next Dl + 1 values are adopted for computations where Dl is 

 the integer most closely approximating 



Dl = 1^ . (C-3) 



30 



