Total Flow for 2 parallel pump operation

Hi Leman

As the flow rate increases with the second pump in operation the friction also increases. The duty point therefore moves to the left of the pump curve to offset the higher head. that is less water per pump.

You can get 2000 us gpm if you add another pipe or one equivalent bigger pipe.

If a single pump is run with the bigger pipe you will get more water than 1000 gpm and may overload. the pump then will have to be throttled.

Tq very much to all for the info given.

You may not get 2000 gpm because of slippage in the pumps. Even new pumps do not necessarily flow their full rated capacity.

-Keith

95% sure that both pumps do not have identical head capacity curves. They must be within 2-3% or one pump will fight with the other. Even if both pumps are identical in model, identical in impeller trim, they may be worn out differently if they are old. If they are new, and they are not "engineered" pumps, the two curves will likely have a sloppy head capacity curve tolerance. If you had a performance test done when new, compare the two curves. If they are not identical models in every way, and with identical impellers with identical impeller trim (and identical under filing if so performed) then they will ALWAYS be fighting each other. I'm sure of this.

Good answer.

As well as the pumps having a pressure/flowrate curve, the system has a pressure/flowrate curve. The operating point is where the pump pressure/flowrate curve intersects with the system pressure/flowrate curve.

An analogy for the original poster would be to connect two batteries in parallel in an electrical circuit. Why is there not double the current compared to running with a single battery (rhetorical question)?

I would appreciate some clarification regarding static head. It has been my understanding the static head is the difference in liquid levels between the suction and discharge side of the pumps and would be the head the pump see's when not in operation. As such the static head does not change when the pumps are in operation. When the pumps are started, the head they see will be a function of the losses in the piping and other components in the system like valves and heat exchangers. If I am incorrect her, please let me know and explain.

combined flow requires a larger pipe then the single pump. in this case 1+1 does not equal 2. please give more info.

<please give more info.>

As others pointed out:

If you take electric analog--- the discharge line --Or 2 different Outlet pipes to final sink?

You have to consider resistance of outlet pipe/s.

If you have 2 different Outlet pipes-- you WILL get Total flow out of 2 Pumps=2xQ

Hello Mr. Leman!

How is your pump connected? Are they supplied from a common pipe and also discharge
to a single pipe? Try running each pump one at a time and measure the capacity. If each pump gives you the desired 1000 USG capacity, then there is no problem with your pumps. You have to increase the size of the pipes or if not, pipe them separately.

Using the PUMP AFFINITY LAWS you can establish a piping system curve, and a pump curve. Look at Reply #5 which shows typical system and pump curve (s). The following formulas are invaluable in piping system design.........particularly where multiple pumps are used.

Q1/Q2 = N1/N2 = D1/D2

H1/H2 = (N1/N2)^2 = (D1/D2)^2

BHP1/BHP2 = (N1/N2)^3 = (D1/D2)^3

Q = Quanity, GPM

N = RPM

D = Diameter of impeller

H = Head-ft., dynamic pump head, differential pressure (ft) across the pump.

Substute as required, just keep the "power" straight. Ex: H1/H2 = (Q1/Q2)^2, etc.

Note in the example in Reply #5, the system curve commenced at 0 ft and 0 gpm, whereas the pump curve started a some head,ft, above 0 ft. This represents system static head, required to keep the system full of water.

_____________________________________________________________________

When I ponder, I sits and thinks, sometimes I just sits.

OH ! SHOOT !!! I got it bass ackwards..............The system curve shows a static pressure, not the pump curve. The pump curve is the dynamic head of the pump, which is the head added by the pump. The system curve is the curve through which the system operates. Reading the differential pressure across the pump is the dynamic head, the discharge pressure is the total head of the system, dynamic head plus static. AND the intersection of the pump curve and system curve is the actual operating point of the system and pump, for that particular condition, etc.

This not only occurs in water pumps.

It occurs when paralleling hydraulic pumps, boilers, feed pumps etc., etc., you never get double the output. When operating in restricted waters Navy ships always paralleled boilers, steering pumps, feed pumps etc., what this actually did was decrease the time in which manoeuvring was carried out. e.g. a steam ship may be able to reach a speed of 27-28 knots on one boiler, parallel a second boiler (exactly the same) you would only get may be 35-40 knots out of the vessel. However, the vessel manoeuvrability increases say three fold plus.

So far I can see all relevant arguments are presented in previous articles - especially #5 did provide a good visualisation of the general results.

In message #2 already the friction was addressed, to be responsible for a lower flow rate of 2x for operating two identical pumps in parallel. I would like to add the rule behind this:

A "Flow Systems" flow resistance usually is related to the flow velocity in square. The flow resistance also is dependend on flow patterns like laminar or turbulent flow. Flow resistance creates a pressure drop. The pressure drop can be expressed:

delta-p = ZETA x v² . These effects are described in many literature for fundamental flow principles. The most important is the BERNOULLI equation.

If you are increasing the flow rate you are increasing the flow velocity in same proportion because Q = A x v (law of continuation); A = pipes cross section area.
Consequently the pressure drop is increasing as well. Therefore you never will get a double flow when connecting a second pump in parallel. A higher pressure drop in the system changes the pumps operating point - and the result will be a smaller total flow than 2x the flow from the single pump. This is shown very well in the diagram of reply #5.

Other aspects on this matter were mentioned as well: one pump fighting the other (two pumps characteristics will differ more or less typically - not to talk from wear conditions). The pressure drop (flow resistance) may change suddenly in case the flow pattern would change from laminar to turbulent characteristic (vice versa; check the Reynolds Number for determination of the flow pattern).

With centrifugal pumps, the flow is determined by where the system curve crosses the pump curve and only at this point. With two pumps operating in parallel. The flow may be doubled while the head stays the same, assuming that the head loss in the pipe with the combined flows is not so great as to move the point the system curve crosses the pump curve/s to the left. This is the same effect as if we install valve in the line to control flow. Another potential problem is the possible effect one pump may have on the other when operating two pumps in parallel. That is, it is impossible to manufacturer two pumps exactly the same. As a result one pump is most likely more efficient than the other resulting in some loss by the other pump to perform as designed. A solution here is the use of a "combination check and minimum flow by-pass valve" see hbe-engineering.com for details. In conclusion, when you bring on the second pump, the total head the the two pumps see moves the flow to the left on the curve mainly due to increased head loss in the pipe due to increased friction.

Flow rates are vary dependent on the size and length of a pipe. The effect of pipe size is not linear with flow. The restrictive effect of the pipe size increase at a rate greater than the flow rate. If you 2X the pumping capacity and leave the pipe size the same, the resulting flow will not be 2X. Double the pipe size (crosssectional area) at the same time and you double the pumping capacity and you will get twice the flow rate.
By putting two pumps in parrallel, you undoubtably added elbows which are very restrictive as well. Paul Eckerson

With a good design and proper selection of pipe sizing of discharge manifold, you can attain the required flow rate 2000 USGPM.

No way can that happen in a practical (usefull ) situation. The pumps will wear and not stay in identical specifications for very long apart from everything else aformentioned. I have seen this many times and these instances were all promoted by the designers as being capable of meeting the system curve requirements. If they worked from the start which was rare, they did not meet the performance criteria for very long.