51 



creating substantial differences between the pressure at the transducer, and the pres- 

 sure that would exist if the transducer were not there to disturb the flow. Another 

 source of error is the frequency response of the transducer (Raichlen et al. 1990; 

 Ippen and Raichlen 1957). The frequency response of pressure transducers is not flat, 

 and as a result, the recorded signal could be quite different from the actual pressure. 



In the experimental setup described above, the pressure transducers were mounted 

 on a large plywood panel oriented in line with the flume. The measurement face of 

 the instrument was flush with the surface of the plywood to minimize dynamic effects 

 near the gauge. The panel was large enough for the boundary layer to be fully 

 developed near the surface of the plywood, to prevent edge effects from the edge 

 of the plywood affecting the measurements. This arrangement is expected to have 

 resulted in minimal dynamic effects on the measured pressure. 



There is no information available about the frequency response of the pressure 

 transducers used for this experiment. However, a spectral analysis of the records can 

 help identify any potential problems. The top plot of figure 3.17 gives the measured 

 water surface and subsurface pressure for a short segment of a record. There are 

 some clear higher frequency fluctuations in the water surface that do not appear in 

 the pressure record. In this particular segment, there is a sharp secondary crest in 

 the first trough (near the 2s point) in the water surface record. 



This loss of high frequency information in the pressure records is confirmed by 

 an examination of the discrete Fourier transform of the records. The second plot in 

 figure 3.17 is the smoothed variance spectra of the two records from which the above 

 segments were taken. The water surface record has a great deal more variance at 

 the higher frequencies. Note that the spectrum of the pressure record has decayed to 

 almost zero at a; = As~^ point, while there is still considerable variance in the water 

 surface at that frequency. 



Linear wave theory predicts that the motion of high frequency waves decays with 

 depth much faster than low frequencies. In this example, the non-dimensional water 

 depth at the peak frequency is co'^h/g « 1. This is generally considered intermediate 

 depth water, and the decay with depth is expected to be moderate. At the frequency 

 where there is essentially no energy in the pressure record, the non-dimensional water 



