Abstract
The hydraulic design of centrifugal pumps has in the past been almost entirely empirical, since there is a large discrepancy between conventional theory and actual tests. Several papers† have already considered this problem of providing a reconciliation of theory and practice. The author suggests that a possible solution may be obtained by considering the total head as being made up of (a) head due to impact between impeller blade and water; and (b) head due to centrifugal force. It is assumed that the flow conditions may be such that the impeller blades shear the water without accelerating it. This is suggested as representing the maximum quantity-zero head state where the increase in velocity from branch to casing throat is obtained at the expense of suction pressure. Neglecting losses there will be no torque on the shaft and therefore no head generated. At less quantities the velocity differences between blade and water will cause impact and therefore generate a head. It is also assumed that at less quantities the “pressure” existing between blade and water will transmit to the water the additional energy of the centrifugal head. This latter assumption is supported in some measure by the fact that in practice the head is affected by variation of the ratio of impeller diameter to eye diameter. Head and quantity values are calculated for zero head quantity QF, maximum efficiency quantity Qf, half maximum efficiency quantity Qf/2, and zero quantity. These results compare very reasonably with actual test curves.