How to Verify Parallel Pump Load Sharing
When two pumps run in parallel into a common header, they are supposed to split the flow, but a mismatched pair can leave one pump doing most of the work while the other barely contributes or even runs dead-headed. This procedure verifies that parallel pumps are actually sharing the load, using the fact that pumps on a common header must make the same head, and reading each pump's own flow and amps to see who is really pushing. It catches the lazy pump before it becomes an overloaded survivor and a starved partner.
Verify Parallel Pump Load Sharing in one line: To verify parallel pump load sharing, use the rule that pumps discharging into a common header must all make the same head, then read each pump's individual flow and motor amps at that shared head. Equal, healthy pumps should carry similar flow and draw similar current; a pump that reads much lower flow and amps than its partner at the same head is weaker, and one near shutoff head with almost no flow is being pushed off its curve by the stronger pump.
Understand How Parallel Pumps Must Share
Two pumps piped into a common discharge header are forced to the same discharge head, because they are connected to the same pressure. That single fact is the key to the whole check: at any instant both pumps make the same head, so the way they share the total flow is set by where that common head intersects each pump's individual curve. If the two pumps have identical curves, they split the flow evenly; if one curve is stronger, that pump carries more flow at the shared head.
The combined curve of parallel pumps is built by adding their flows at each head, so the pair delivers more flow than one pump but not double, because pushing more total flow into the system raises the head on the system curve and both pumps slide back up their curves to a lower individual flow. This is why adding a second pump in parallel gives less extra flow than people expect, especially on a steep system curve. Understanding this is the difference between judging the pumps fairly and blaming a pump for physics.
The trap that this check exists to catch is a mismatched pair. If one pump is weaker, worn, running slower, or has a lower shutoff head, then at the shared header head it may deliver little flow, and if its shutoff head is below the header head it delivers nothing at all while still spinning, dead-headed by its stronger partner. That starved pump can overheat on its own recirculation while the strong pump is overloaded carrying the whole duty, so verifying the split protects both machines.
Read Each Pump's Head, Flow, and Amps
Take the measurements that reveal the actual split. Read the discharge and suction pressure at each pump and confirm both pumps are indeed making the same head at the common header, which they must be if they are truly in parallel on one header; a difference means they are not sharing a common point or a check valve is partly closed. Then read each pump's individual flow, if a per-pump meter exists, and each pump's motor amps, because amps are a good proxy for how much work each machine is doing when a per-pump flow meter is missing.
Compare the two pumps at that shared head. Two matched, healthy pumps running in parallel should draw similar motor current and, where measured, carry similar flow. A pump that reads noticeably lower flow and lower amps than its partner at the same head is the weaker machine, running higher on its curve toward shutoff while the partner carries the load lower and to the right. The verification method for placing a single pump on its curve, in the note on how to verify a pump curve against its operating point, applies to each pump individually at the shared head.
Watch for the dead-headed pump, the worst case. A pump whose amps sit near its shutoff-condition value with almost no flow, or that reads low amps and is clearly not contributing, is being held off its curve by the stronger pump and may be running below its minimum continuous flow, heating on internal recirculation. That is the condition the note on how to set a pump minimum-flow recirculation line guards against, and in a parallel set it can happen even when the header flow looks fine, because the total is being carried by one machine.
Pin Down Why the Sharing Is Unequal
When one pump is carrying less than its share, work through why its curve sits lower than its partner's. A worn pump makes less head at any flow, so its curve has sagged and it loses the sharing contest at the common header, which the note on what a pump wear ring is explains as widening internal clearance. If the pumps are supposed to be identical and one consistently reads weaker, that pump is the wear candidate, and comparing each against its published curve confirms it.
Check for a speed or setup difference. On variable-speed pumps, a small difference in commanded speed makes a large difference in head, because head scales with the square of speed per the note on what pump affinity laws are, so two pumps at slightly different speeds will not share evenly. A partly closed isolation valve, a fouled strainer on one suction, or a stuck check valve on one discharge also handicaps a pump and shifts the load to its partner, and those are cheap to find by walking each pump's piping.
Confirm the two pumps were ever a matched set to begin with. Pumps with genuinely different curves, different impeller sizes or different models, cannot share evenly on a common header no matter how healthy each is, because their curves cross the shared head at very different flows. If that is the situation, the unequal sharing is by design and the question becomes whether the weaker pump is being pushed below its minimum flow, which is a real risk. Continuous trending of each pump's amps and flow on a platform such as Merobix makes a drift toward unequal sharing visible over time, catching a pump that is slowly wearing out of the match.
When to Escalate
Escalate to maintenance when the unequal sharing is traced to a worn pump whose curve has dropped below its partner, because restoring the match means rebuilding the weak pump, which field checks cannot do. Running a worn pump in parallel with a healthy one loads the healthy pump harder and starves the worn one, so the pair keeps drifting further out of balance until the strong pump is overworked. A rebuild returns the pair to a matched set.
Escalate to engineering when the pumps were never a matched set or when a pump is being pushed below its minimum continuous flow by its partner, because those are design questions about whether the parallel arrangement is safe. A pump running dead-headed or near shutoff in a parallel set is at real risk of overheating, and deciding whether to add recirculation protection, resize a pump, or change the operating philosophy is a design decision for qualified personnel, not a field adjustment.
Frequently Asked Questions
Why do two pumps in parallel not simply double the flow?
Because pushing more total flow into the system raises the head the system demands, and both pumps slide back up their curves to a lower individual flow at that higher head. Pumps on a common header must make the same head, so the combined curve is found by adding their flows at each head, and where that combined curve meets the steeper system curve gives less than double the single-pump flow. On a steep system curve the extra flow from a second parallel pump can be quite modest, which is normal physics rather than a fault.
How do I tell which parallel pump is doing more work?
Read each pump's motor amps, and its individual flow if a per-pump meter exists, while confirming both are making the same header head. Two matched healthy pumps draw similar current and carry similar flow. The pump reading noticeably lower amps and lower flow at the shared head is the weaker one, running higher on its curve toward shutoff while its partner carries the load. If one pump reads near its shutoff condition with almost no flow, it is effectively dead-headed by the stronger pump and at risk of overheating.
Can one parallel pump run dead-headed while the header flow looks normal?
Yes, and that is the dangerous case this check exists to catch. If one pump is weaker or its shutoff head is below the shared header head, the stronger pump holds the header at a head the weak pump cannot exceed, so the weak pump delivers little or no flow while still spinning. The total header flow can look fine because the strong pump is carrying it all, but the starved pump is running below its minimum continuous flow and heating on internal recirculation, which will damage it if left unaddressed.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.