Sec. 70.5 



SCREW-PROPELLER DESIGN 



591 



7/0 of Fig. 70. A are related to each other by a 

 vertical ordinate intersecting all three of them 

 at the particular advance coefficient J (or 8) 

 at which the propeller is operating. Such an 

 ordinate is drawn in broken lines on the figure 

 through the points C, E, and F. Corresponding 

 values on the three double and one single inclined 

 scales are determined by dropping perpendiculars 

 on them from the three intersecting points C, E, 

 and F, indicated in the diagram. A single perpen- 

 dicular CM is dropped from the r/o-curve inter- 

 section to the scale of TD/Q. Two perpendiculars 

 are dropped from the point E on the Kr curve, 

 one to the double scale of first basic coefficient 

 and the other, EG, to the double scale of thrust- 

 load coefficient. A single perpendicular is dropped 

 from the point F at the Kg-curve intersection to 

 the point H on the double scale for the second 

 basic coefficient. By having the two scales of 

 each pair opposite each other at the feet of the 

 perpendiculars EG, FH, and the second short 

 perpendicular from E, it is convenient to pick 

 off either dimensional or non-dimensional values, 

 or to enter the diagram with these values. 



Assume that a symmetrical-section propeller 

 such as that depicted on Fig. 70. A is to be used 

 and that the 0-diml thrust-load factor Ctl is 

 L065. The scale of Ctl is entered at the point 

 G and a line GE is drawn perpendicular to that 

 scale until it intersects the Kr-curve at E. The 

 ordinate CEF is erected through the point E 

 and the 0-diml J value is read from either the 

 top or the bottom scale as 0.596. If the speed of 

 advance Va and the propeller diameter D are 

 known, the rate of rotation is obtained directly 

 from the relationship n = Va/(.JD). The actual 

 working efficiency at the point C on the ijo curve 

 is read off from the side scale as 0.6L 



The 0-diml value of the second basic coefficient 

 bg is determined by dropping a perpendicular 

 from F to the inclined scale at H, whereupon the 

 0-diml value of bg is read off as 0.313. With 

 the values of mass density p, speed of advance 

 Va , and rate of rotation n all known, the torque 

 Q is determined from the bg formula given on 

 the diagram. The power Pp which will be absorbed 

 by the propeller is calculated from the derived 

 values of n and Q. 



If the rate of rotation n is given and the power 

 Pp is known, corresponding to the situation in 

 Sec. 59.15, the torque Q is derived by direct 

 calculation. The thrust T is found eitheiifrom the 

 value of TD/Q at the intersection M, or from the 



first basic coefficient br at the intersection near E. 



By and large, the use of any particular series 

 chart for the preliminary design of a screw pro- 

 peller gives essentially the same kind of answer. 

 This is on the basis that the test data for the 

 model propellers from which the charts were 

 constructed did not suffer from scale or surface 

 effects, that the observed data are accurate and 

 carefully plotted, and that the use of dimensional 

 expressions does not omit important factors or 

 introduce unknown errors. For example, since 

 model propellers are almost invariably tested in 

 fresh water, the derived data are also for fresh 

 water. A mass-density factor p(rho) which is 

 omitted for convenience or simplification, as 

 was done by D. W. Taylor, may make a 2 or 3 

 per cent difference in ship-propeller data calcu- 

 lated for salt water [Kane, J. R., SNAME, 1951, 

 p. 626; Schoenherr, K. E., SNAME, 1951, p. 628]. 



A comparison of five kinds of propeller-series 

 charts then in use was made some two decades 

 ago by H. F. D. Davis [ASNE, Feb 1932, pp. 

 8-24]. The discrepancies between the five sets of 

 preliminary-design characteristics worked out 

 from the charts was rather more than would now 

 be acceptable. It must be remembered, however, 

 that the parent propellers all had rather different 

 characteristics, especially with regard to mean- 

 width ratio and blade-thickness fraction. Unfor- 

 tunately, a more modern comparison is not 

 available, worked out in the same detail. Some 

 rather general comments are to be found in the 

 discussion of a recent paper by F. M. Lewis 

 [SNAME, 1951, pp. 621-641]. 



For a beginner in the field, it is bewildering to 

 find so many kinds of charts, all ostensibly for 

 the same purpose, but actually varied to suit the 

 type of initial data and the nature of the answer 

 desired by several groups of people, experienced 

 in these procedures. It is likewise most confusing 

 to find different symbols on each kind of chart, 

 with some expressions dimensional and others 

 non-dimensional. After trying them all, or all 

 that are available, he is in better position to 

 decide which meets his own particular needs, 

 either for analysis or for practical design. 



It is characteristic of any and all propeller- 

 series chart-design procedures that the numerical 

 values required for the full-scale ship are obtained 

 only by estimating the wake fraction w and the 

 thrust-deduction fraction t. The Vq of the open- 

 water test corresponds only to Va on the ship, 

 whence V = Fx/(1 — w), and the thrust T 



