Troubleshooting pumps

SUBJECT : Pump selection practices that cause high seal and bearing maintenance problems. 6-9

Purchasing well designed hardware does not bring automatic trouble free performance with it. The very best equipment will cause problems if it was not designed for your particular application. Here are a few of the more common selection problems we find with centrifugal pumps:

  • Buying the same size pump as the one that came out of the application. That’s O.K. If the old pump was the correct size, but the odds are that it was too big because of the safety factors that were added at the time of purchase. This will cause the pump to run off of its best efficiency point (bep) and you will spend a lot of production money for the additional power that is needed to run against a throttled discharge valve or orifice installed in the discharge piping.
  • Buying to a standard, or making a decision based on efficiency, and believing that these two some how relate to quality. Standards were written for packed pumps. When a mechanical seal is being used the shaft L3/D4 number is almost always too large. Efficiency is always gained at the expense of maintenance. Efficiency means tight tolerances and smooth passages that will eliminate reliable double volute designs and keep the maintenance department busy adjusting tight tolerances to maintain the efficiency you paid for.
  • Series and parallel installation problems. We often find pumps installed in parallel, but no one knows it because the second pump was installed at a much later date and no one has bothered to trace the piping. Pumps in parallel require that they have the same diameter impeller and that they run at the same speed, or the larger pump will throttle the smaller one causing it to run off the best efficiency point, deflecting the shaft. The capacity should be looked at if the higher capacity pump might exceed the N.P.S.H. available.
  • When pumps are installed in series the impellers must be the same width and they must run at the same speed or the higher capacity pump will either cavitate because the smaller capacity pump can not feed liquid at the proper capacity, or it will run throttled if it is feeding the smaller pump. In either case the larger of the two pumps will be adversely affected.
  • Purchasing a larger pump because it will be needed in the future. Will raise the operating cost to unacceptable levels (Power = head x capacity) as the pump is run against a throttled discharge valve. This inefficient use of power will translate to a higher heat environment for the seal along with all of the problems associated with shaft deflection.
  • Using a variable speed motor to compensate for a pump curve that is not flat enough. Many boiler feed pumps require a flat curve so that the pump can put out varying capacities at a constant boiler pressure (head). We see this same need if we are pumping a varying amount of liquid to a very high constant height.
  • Varying the speed of a pump is similar to changing the diameter of the impeller. If you look at a typical pump curve you will observe that the best efficiency point (bep.) comes down with impeller size to form an angle with the base line (capacity line) of the graph. This means that if you vary the speed of the impeller, the pump always runs off the bep. except in the case where the system curve intersects the pump curve, or in the case of an exponential system curve such as we find in a typical hot or cold water circulating system.
  • Double ended pumps installed in a vertical position to save floor space makes seal replacement a nightmare unless you are using split or cartridge designs.
  • Specifying a desired capacity without knowing the true system head. You can’t guess with this one. Some one has to make the calculations and “walk the system”. The present pump is not a reliable guide because we seldom know where it is pumping on its’ curve. Chart recorders installed on both the suction and discharge side of the pump will give a more accurate reading of the present head if they are left on long enough to record the differences in flow. The trouble with this method is that it will also record a false head caused by a throttled valve, an orifice, or any other restriction that might be present in the piping.
  • Requesting too low a required N.P.S.H. will cause you to end up with a different kind of cavitation problem. See another paper in this series for information about “Internal recirculation”.
  • Failure to request a “center line design” when pumping temperature exceeds 200°F (100°C) it will cause pipe strain that will translate to wear ring damage and excessive mechanical seal movement.
  • The use of “inline” pumps to save floor space. Many of these designs are “close coupled” with the motor bearings carrying the radial and thrust loads. Because of typical L3/D4 numbers being very high, the wear rings act as “steady bearings” after the pump is converted to a mechanical seal. The pump should have been designed with a separate bearing case and a “C” or “D” frame adapter installed to connect a motor to the bearing case.
  • Thrust bearings being retained by a simple snap ring. Beyond 65% of its rated efficiency most centrifugal pumps thrust towards the pump volute. The thin snap ring has to absorb all of this axial thrust and most of them can not do it very well .
  • The mechanical seal has been installed in a packing stuffing box that is too narrow to allow free seal movement. If a mechanical seal was specified, the pump back plate should have been manufactured with a large diameter seal chamber. In most cases the stuffing box recirculation line should be installed from the bottom of this large seal chamber to the suction side of the pump or a low pressure point in the system. There are some exceptions to this, however:
    • If you are pumping at or close to vapor point.
    • If the entrained solids have a low specific gravity.
    • If you are using a Flowserve (Duriron pump) that adjusts to the back plate.
    • If you are using a double suction pump where the stuffing boxes are at suction pressure.
  • High temperature applications have several special needs:
    • A jacketed stuffing box that isolates the pumpage from the stuffing box contents by a carbon bushing to retard heat transfer.
    • A centerline design to compensate for thermal expansion.
    • A cartridge seal design that allows open impeller adjustment after the pump has come up to operating temperature.
    • A stainless steel shaft to retard heat transfer to the bearings.
    • A method of cooling the bearing oil, but never the bearings.
    • A coupling that will compensate for axial expansion.