There is a common perception in the fluid movement (pump) industry that a selection is done based on "a duty point". Admittedly, some systems do have very small changes in the flow...
Possibly one of the best or worst examples of applications that have large variations in TDH are groundwater or borehole installations. Not only do water levels in the hole vary from season to season but porosity and transmissivity within the aquifer always cause significant changes in water levels. Recently, we had the opportunity to share in the process of optimizing a system that had a static water level of 20 m while the level fell a further 65 m when the pump was switched on. Figure 1 graphically represents the problem. Curve A shows the system curve for a 32 mm class 12 HDPE rising main at a static level of 20m. Curve B is the system curve for the same rising main at the new (dynamic) level of 85 m (20 m + 65 m drawdown). The blue curve shows the pump performance while the dotted red extensions show duty points that are not recommended by the manufacturer. At a static head of 20 m (start up) the pump will operate at 31 l/m against a head of 22 m. This is well outside the manufacturer's recommended range. As the water level falls to the dynamic level (inflow now equals outflow), the duty point becomes more acceptable at 8 l/m against a head of 88 m. If for any reason the borehole could deliver an increased amount of water (higher rainfall, reduction in inflow losses) then the time the pump will spend on the extreme right hand side of the curve will increase. The usual failure symptoms for operation on the right-hand side of curve are motor thrust bearing and/or coupling and/or winding failure. The pump end will also have a sharply reduced operating life due to failure of the built in (up) thrust mechanisms. While this case study covers a relatively small pump set, it still represents a substantial investment for the end-user. The application could just as well be a multi megawatt shaft dewatering pump. A recent example of this required the water in an abandoned mine shaft, in central Africa, to be lowered from 120 m to 280 m. Failure in a system such as this is not an option. Stopping the pump for a matter of hours, allows not only the recovery of the water level but very costly delays in the reopening of the mine. Removal, repair and reinstallation of the pump set and rising main is a hazardous, time consuming and expensive exercise. Having defined the problem, what can be done about selecting pumps that are expected to handle some level of variation in the system Total Dynamic Head (TDH). In both cases we had a borehole that had maximum and minimum TDHs that effectively pushed the pump to points that were too far left and right of the Best Efficiency Point (BEP). Some of the usual solutions to this problem are:
1- Installing a throttling valve at the discharge. Some of the drawbacks? Unauthorized tampering, use of cheap non-control pattern valves (yes this is a control valve application). If globe, characterized butterfly and ball type valves are used they become expensive and require another level of skill to set and maintain. They are also prone to the "fiddle factor".
2- If the static head is the main culprit (as was the case with our borehole) a smaller rising main can be installed to prevent the pump "falling off" the r.h.s. of the curve. The disadvantage here is that the losses incurred in the pipe remain for both the start up flow and for the minimum flow. Admittedly the losses do decrease as the flow decreases but what is needed here is something with a bit of intelligence!
In searching for an elegant solution to this vexing problem, I came across a valve seems to have some good potential for making a positive contribution to improving overall system efficiency and reliability. Figure 2 graphically shows the same installation fitted with a commercially available control valve.
Curves A & B are the system curves for the 32 mm class 12 HDPE rising main. Curves C & D are the curves for a Control valve with a set point of 26l/m. Notice how the curves become very steep as flows approach the rated figure. This is where the work of supporting the pump needs to be done. At lower flows the gradient is flat which is good as this represents lower losses. As the flow decreases because of increasing static heads, so the losses need to be decreased. From the diagram, the head loss at maximum flow is about 22 m (the gap between curve A and the intersection of curve C and the blue pump curve). As the water level in the borehole falls, so the control valve's curve rises. At the minimum flow (8l/m) the gap between pipe curve B and the intersection of control valve curve D and the pump curve is now only +/-3 m. While this approach might not solve all high drawdown borehole problems, it could make a very positive contribution to solving some vexing groundwater system related problems.
Web: www.jtatraining.co.za