Corrections were made to the mean values of the measured loading 
components to account for centrifugal loads and the influence of the dy- 
namometer boat, and to the first harmonic of the measured loading compo- 
nents to account for gravitational loads. The derivation of these correc- 
tions is discussed later. 
From the known values of the three measured force components and 
three measured moment components, the values of the bending moments about 
the shaft centerline and bending moments normal to the nose-tail line at 
the 0.3 and 0.4 radii were calculated. These bending moments were calcu- 
lated at every 4 degrees of blade angular position, and harmonically ana- 
lyzed. The wave form was reconstructed by using the first 10 harmonics of 
blade angular position, in exactly the same manner as was used for the 
other components of blade loading. 
Plots of the data were generated by the CDC computer system using a 
-Calcomp Plotter. 
ACCURACY 
During the experiments for steady operation, V=0, and dynamic 
pitching, W#0, where many revolutions of data were averaged during a sin- 
gle run, the standard deviations of speed V, rotational speed n, forces, 
and moments were computed, assuming the distribution in these variables at 
a given condition follows the normal probability distribution. For forces 
and moments, the standard deviation was calculated at every increment of 
blade angular position at which forces and moments were recorded. An 
error band around the data mean was then represented using the standard 
deviation multiplied by a factor dependent on the confidence level chosen. 
For the present analysis, the factor of 1.96 was selected which corre- 
sponds to a confidendence level of 95 percent. A confidence level of 95 
percent indicates a confidence (or probability) that 95 percent of the 
data considered falls within the error band. For a given run the average 
error (95 percent confidence band) in model speed V was approximately 
+0.005 foot per second (1.6 mm/s), while the error in rotational speed n 
was less than 0.001 revolution per second. The very low error in n re- 
sulted from the use of a precision solid-state motor controller as dis- 
cussed in the section on facility and dynamometry. 
Za 
