Grid Turbulence in Dilute High-Polynner Solutions 



The towing speed was chosen from considerations of sensor frequency re- 

 sponse, towing-tank length and test run duration, grid towing power, and pres- 

 sure drop across the grid. Sensor frequency response and test run duration 

 have been discussed already. The grid towing power is about 1/3 horsepower. 

 The hydrostatic head corresponding to the pressure drop through the grid is 

 about 1/4 in. There was no need to restrain the free surface in this work to re- 

 duce wave generation, since spectral measurements showed no wave components. 

 Waves moving relative to the grid were observed only ahead of the grid. 



A common problem with water velocity measurements by hot-element sen- 

 sors is contamination. In this work the problem was unimportant, as illustrated 

 by the fact that one particular quartz-coated sensor was used at intervals over 

 periods of weeks without cleaning and with only minor calibration shift. The 

 most important step in obtaining this reliability was tank water conditioning by 

 filtering and deaeration. It was also found to be important to stir the tank by 

 several passes of the grid before beginning calibration tests, to eliminate any 

 temperature gradients. 



The tank solution preparation problem was sizable, since the 6000 gallons 

 of solution had to be homogeneous but not appreciably degraded by the method of 

 preparation. Two methods were used. In the first method the polymer particles 

 were dispersed directly into the tank filled with conditioned tap water, using an 

 aspirator type of disperser. The towing-tank fluid was then mixed by towing the 

 grid back and forth, sometimes with an off-center board on the grid to produce 

 larger-scale mixing. The grid turbulence helped to dissolve the gelatinous 

 blobs which the polymer particles become after being wetted. The results with 

 the first Polyox solutions prepared by the "direct-dispersal" technique, led to 

 work with the polyacrylamides. Prolonged grid stirring or stirring by air bub- 

 bling did not dissolve the gelatinous blobs of Separan AP-30 and JIOO. Thus it 

 was necessary to prepare a master solution at a relatively high concentration 

 (e.g., 0,8%) and mix it for a long period of time to encourage homogeneity. This 

 solution was then diluted to the desired concentration with conditioned water and 

 fvirther mixing. 



Sensor calibration was done by towing the sensor carriage along the tank 

 and measuring bridge output voltage as a function of speed for a fixed bridge 

 operating condition. The slope of the bridge voltage curve at the mean speed 

 behind the grid was taken as the sensor sensitivity. Each sensor was calibrated 

 before and after each series of grid-turbulence tests. Sensor deterioration was 

 detected as a significant loss of sensitivity and as an upward drift of cold re- 

 sistance. When the drift of sensitivity in 2 hours was as much as 4%, the sensor 

 was replaced. 



Turbulence samples were obtained using the equipment indicated in Fig. 4. 

 Each set of grid-turbulence tests for various x/m and amplifier and filter set- 

 tings was paralleled by a corresponding set of noise tests. For these tests the 

 tank was allowed to become quiet, as for the regular tests, and only the sensor 

 carriage was run along the tank. All signal generation, conditioning, and re- 

 cording conditions were kept the same as in the grid-turbulence tests. Usually 

 for each test condition, one or more tests were made to record the unfiltered 

 signal, and one or more additional tests were made to record the filtered, 



45 



