A large single stage overhang process pump has had multiple mechanical seal failures due to excessive shaft deflection presumably due to high hydraulic radial thrust attributable to its single volute lip design.
To reduce the high hydraulic radial load and minimize the shaft deflection it was decided to modify the casing to simulate a double volute lip design by welding a piece of metal splitter opposite the single volute lip at the periphery of the volute cutwater.
Initial calculations were made to estimate the axial thrust values using empirical thrust factors for both single volute and double volute configurations based on using the same model impeller. The equipment was then tested at normal operating conditions at 1185 RPM. The objectives were:
To verify that the modification will result in the expected reduction in hydraulic radial thrust.
To compare the calculated (or theoretical) radial thrusts with the actual (or tested) radial thrusts to validate the widely used empirical thrust factors that were used in the thrust calculations.
The pump was run at different flow rates ranging from shut-off (zero flow), to its best efficiency point (BEP), and to its normal end-of-curve flow. Both the radial thrusts and shaft deflections were measured at these different flow rates. The equipment was similarly tested before, and after, the casing modification.
Based on the test results, the following observations were made:
The volute modification resulted in big reduction in shaft deflection.
There was no significant change in its Q/H/E performance resulting from the volute modification.
As expected, the shaft deflections (both before and after the modification) showed a decreasing trend as the flow rates were being increased, and the deflections were at minimum values at the BEP.
Additionally, a comparison was made between the calculated shaft deflection and the measured shaft deflection prior to the volute modification. The comparison showed that up to about 50% of BEP the empirical thrust factor used in calculating the hydraulic radial loads of single volute pumps was conservative and resulted in higher calculated thrust value and shaft deflection. At above 50% of BEP, the empirical thrust factor was slightly at the low side, but close enough to be considered in line with the test results.
The casing modification to simulate a double volute lip design, resulted in very significant improvement to the equipment's mean-time-between-failures (MTBF).