harmonics. In addition, the wake data show no significant amplitudes 

 for harmonics greater than the tenth. Therefore, all data and analyses 

 except Figures 8 and 9 are based on reconstructed signals using the 

 first 10 harmonics. The symbols shown in Figure 8 indicate unfiltered 

 values determined from the experiment; each represents the average value 

 at the indicated blade angular position for over 200 propeller revolu- 

 tions. The variation in measured values at a given angular position is 

 discussed in the section on accuracy. The lines on Figure 8 indicate 

 that the variations of the signals with blade angular position are 

 adequately represented by the number of harmonics retained. 



The variations of all measured hydrodynamic loading components 

 with blade angular position for simulated propulsion in calm water with- 

 out hull pitching are shown in Figure 10. The amplitudes and phases of 

 the harmonics of these loading components are presented in Figure 9. 



These data show that for hydrodynamic loading the variation of all 

 loading components was predominantly a once-per-revolution variation. 

 The extreme values for all loading components, except Fz and Mz, occur- 

 red near the angular position of the spindle axis, 6 = 114 and 270 

 degrees; i.e., within 24 degrees of the horizontal. The propeller eval- 

 uated has a projected skew angle at the tip of approximately 11 degrees; 

 therefore at the positions of extreme loading the blade tip is within 

 approximately 13 degrees of the horizontal. This suggests that the 

 tangential component of the wake is the primary driving force; see 

 Figure 6. The extreme values of F^ and M^ occur within 20 degrees of 

 the extreme values of the other components. The reason for this varia- 

 tion in location of extreme values is not clear; however, it may be 

 partially due to experimental inaccuracy with the T^-^z flexure as dis- 

 cussed by Boswell et al. (1976a, 1976b, 1978). Further the net stresses 

 in the blades are generally less sensitive to the F^ and Mz components 

 than they are to the other force and moment components. 



The results presented here for circumferential variation of hydro- 

 dynamic loads follow trends similar to results presented by Boswell et 

 al. (1976a, 1976b) for a single-screw transom-stem configuration and 

 results presented by Boswell et al. (1978) and Jessup et al. (1977) for 

 a twin-screw transom-stern configuration. 



The circumferential variations and first harmonics of all loading 

 components except F^ and Mz were substantially larger fractions of their 

 time-average values for the condition evaluated here than they were for 

 the conditions evaluated previously on the models r^eported by Boswell 

 et al. (1976a, 1976b, 1978). For example, (FxH^l/^xtj was 0.66 for the 

 present case, 0.40 from Boswell et al. (1976) , and 0.52 from Boswell et 

 al. (1978). The differences in the ratios of the circumferential vari- 

 ations of loads to the time-average loads for these three cases arise 

 from many factors including the propeller time-average loading coef- 

 ficients which are essentially independent of the unsteady loading, the 

 magnitude of the circumferential variation of the wake (primarily the 

 amount of shaft inclination for the three cases under consideration 

 here) , and propeller geometry especially the blade width and pitch- 

 diameter ratio. The ratio of the unsteady loading to the time-average 

 loading is useful for evaluating the unsteady loading of a given 



12 



