In Part 1 the point was made that the only constant in life is change! Variations in duty point can come from changes in static heads (dynamic levels in ground water systems, tanks full/tanks empty etc), increased friction losses (deposition of solids on the pipe wall and throttling of flow to name a few) and the inevitable fluctuations in production processes. What is the nature of these variations? How do they affect the energy consumed, Mean Time Between Failures (MTBF), plant reliability and finally, the total cost of owning the whole system?
A bad selection process that saw the installation of a significantly oversized pump.
This is a constant variance from the design operating range with the installed duty points close together but far off the BEP. Usually this is found out after commissioning. Correcting the mistakes becomes a strategy of trying to modify the pump, and/or the system and/or the controls to get an acceptable match between the pump and the system. The most popular solution is the throttling valve. Invariably the isolating gate valve on the discharge is used for this purpose. It's energy intensive, inefficient, and does the valve no favours. It also creates noise and vibration in the system. For this an impeller trim could be a far better solution although it is permanent. A Variable Speed Drive (VSD) is also a candidate but look carefully at the make up of the Total Dynamic Head (TDH). Systems that have a high proportion of friction loss in the TDH respond well to the application of VSDs. If the curve tends to be less steep, then look to applying other strategies.
To summarise, an oversized pump will tend to operate on the right hand side of the performance curve at lower efficiencies and with a higher NPSH required. Cost per cubic meter of liquid pumped (specific energy) and, in all likelihood, pump maintenance costs will increase significantly and plant reliability will decrease.
Large changes in flow rates due to changes in processes or system demand
Inflows into waste water plants and outflows from potable water pump stations are classic cases of this scenario. How to handle 400l/s during a particular time of day and 700l/s at another.
The installation represented in the diagram is an example of a system that has to be carefully managed and operated. If only one pump is running, there should be no problem. When demand increases and the second pump is started, the duty ranges from point B to C. What each pump will "see" however is a TDH of +/-34m which is left of the Best Efficiency Point (BEP) A, but this could be the maximum left of BEP that the pump could be safely operated. If this was a two operating, one standby installation, operating all three pumps together (blue) moves the duty point from marginal (D) to disastrous (F)! This could happens because expansion in the plant, suburb, mine etc. requires the 700l/s already mentioned. The three pumps have to be run for extended periods to keep up. Operation on the left hand side of the curve? Suction recirculation, cavitation, shaft deflection, seal, bearing, shaft and bearing failures!
A flat(ish) curve will tend to favour pumps in parallel while a steeper curve could indicate a combination of pumps in parallel and VSDs. For parallel pump installations, operator training is essential. Most of these installations have at least one pump as a standby unit while the others handle variations in demand. Due to a lack of understanding of the interaction of pumps in parallel, many of these installations I have seen, have ALL the pumps running. Our diagram warns of: low Mean Times Between Failures (MTBF) i.e. poor plant reliability, vibration and noise. It get's worse! As the pumps wear, so the individual curves start to sag, droop or just head for the intersection of the two axis of the published curves. Just to add even more to this negativity, the individual pumps don't do this to the same degree. This could mean one or more of the pumps operating against a closed valve.
Another "solution" to the problem of variability in TDH is the use of a bypass line fitted between the discharge and the suction. The bypass approach is a last resort as it comes with far more disadvantages than advantages. The valve used to control the amount of flow in the bypass will be a weak point. This is not a straight forward isolator but a correctly sized control valve that has been selected only after all the accepted procedures for selecting, sizing and installing a control valve have been followed. This is a process that should be left to suitably qualified staff. The day to day operation of the valve should be checked on a regular basis as the first hint of trouble will be failed pumps and/or control valve. Essentially energy is added to the liquid which is then converted to heat and other forms of energy before the liquid is returned to the sump for energy to be added again once it passes through the pump. This approach to handling variable duty points, is, in most cases, the most energy intensive and, because energy represents, by far the largest proportion of the pumping system's Life Cycle Cost, a bypass should only be used once all the other options have been throughly investigated and discarded.
Designing and operating pumping systems is an exercise in managing change. It is seldom that a particular system will operate at a fixed duty point for it's entire life cycle. The most important event of any successful design occurs when there is recognition that changes can happen in the duty range and these variations are defined and catered for through good design techniques and role player empowerment.
Operating a pump "all over" it's performance curve will mean a significant proportion of the life cycle will be spent operating at low efficiency duties. This means high specific energy costs (cost per cubic metre pumped) and be the subject of regular attention from the maintenance department. Overall plant reliability will also have to be factored in due to the low MTBF of the pumping system.