A district heating network has to deliver enough pushing force to every consumer on the loop, including the one furthest from the pumps, without over-pressurising the ones close by. The way this is managed is differential pressure control, which regulates the network circulation pumps to hold a target pressure difference between the flow and return pipes at the point where it matters most. This guide explains what that setpoint is, why it is referenced to the hydraulically most-remote consumer, how a floating setpoint saves pump energy, and how a SCADA system trends zone differential pressure to catch branches that are being starved of flow.
Differential Pressure Control in one line: Differential pressure control in district heating throttles the network distribution pumps, usually through variable-frequency drives, to maintain a target pressure difference between the supply and return pipes so that every consumer has enough available head to draw its design flow. The setpoint is chosen to satisfy the critical consumer, the substation that is hydraulically most remote and therefore hardest to supply, because if that one has enough differential pressure then every closer consumer does too. Holding just enough dP rather than a fixed high value lets the pumps slow down when demand drops, which is where most of the energy saving comes from.
The available driving force at any point on a heating network is the difference in pressure between the flow pipe and the return pipe, and that difference is what pushes water through each consumer's substation heat exchanger. As water travels out along the distribution mains it loses pressure to friction, so the flow-to-return difference is largest near the pumps and smallest at the far end of the loop. The consumer that sees the smallest differential pressure is the one that is hardest to supply, and network operators call it the critical consumer. Because it is the weakest point, if the control system guarantees enough differential pressure there, every consumer closer to the plant is automatically satisfied with room to spare.
Differential pressure control works by measuring the flow-to-return pressure difference and adjusting the distribution pumps to keep that measured value on a chosen setpoint. When demand rises and consumers open their control valves, the extra flow drags the differential pressure down, and the controller responds by speeding the pumps up to restore it. When demand falls and valves close, the differential pressure tends to climb, and the controller slows the pumps to bring it back to setpoint. The reference point for the measurement can be a sensor installed near the critical consumer, or a sensor at the plant whose setpoint has been calculated to leave the required differential pressure at the far end after friction losses.
The reason for controlling to a differential pressure rather than simply running the pumps flat out is that the network must serve a constantly changing pattern of demand. If every valve on the system opened at once the pumps would need to work hard, but for most of the year only a fraction of the load is active. Referencing the setpoint to the critical consumer means the control target is defined by the genuine worst case on the network, so the operator can be confident that meeting it protects everyone else while still leaving the pumps free to ease off whenever the full load is not present.
The energy a circulation pump consumes climbs steeply with speed, so the biggest lever for saving pump power is to run the pumps no faster than the network actually needs at any moment. A variable-frequency drive on each distribution pump lets the control system trim speed continuously instead of leaving pumps at a fixed output, and pairing that drive with differential pressure control is what turns the network into a demand-following system. At low load the pumps drop back and draw a small fraction of their full power; at peak load they ramp up to hold the same differential pressure at the critical consumer.
A fixed differential pressure setpoint already saves energy compared with constant-speed pumping, but a floating, or reset, setpoint saves more. The idea is that the setpoint needed to satisfy the critical consumer is not the same in every condition. When network flow is low, the friction losses between the plant and the far end are small, so a lower plant differential pressure still leaves enough head at the critical consumer. A floating scheme lowers the pump differential pressure setpoint as measured flow or valve demand falls, and raises it as they climb, so the pumps hold only the head that the current conditions require rather than a worst-case constant.
Implementing a floating setpoint needs a signal that reflects how hard the network is working. Some schemes reset the setpoint from total network flow, some from the position of the most-open consumer control valves, and some from a differential pressure transmitter located out at the critical consumer that feeds its reading directly back to the pump controller. Whatever the input, the control logic keeps a margin so that no consumer is ever pulled below the differential pressure it needs to draw its flow. The result is a pump that is genuinely following the load, spending energy only in proportion to the heat being delivered.
On a real network the theory of a single critical consumer meets the messier reality of many branches, some long and some short, whose balance can drift as loads change or as valves and strainers foul. This is where a SCADA layer earns its place: by bringing the differential pressure at the plant, at key network nodes, and at instrumented substations into one place, it lets an operator see the pressure profile across the whole system rather than trusting that the design assumptions still hold. A cloud SCADA platform such as Merobix can trend each zone's differential pressure over time, so a branch whose available head is quietly falling shows up as a downward drift long before consumers complain of being cold.
A starved branch has a recognisable signature in the trends. If a particular zone's differential pressure sits persistently near or below the value its substations need while the plant is holding its setpoint, that branch is not getting its share of the driving head, which points to under-sized balancing, a throttled or fouled valve, or a demand pattern the network was not balanced for. Comparing that zone's differential pressure against network flow reveals whether the shortfall appears only at peak, when friction losses are highest, or all the time. Seeing several zones together also tells the operator whether the problem is local to one branch or a sign that the pumps or their setpoint are set too low for the whole system.
Beyond fault-finding, logged zone differential pressure is what lets an operator tune the control scheme with confidence. Trends show whether the floating setpoint is tracking demand smoothly or hunting, whether the margin above the critical consumer's requirement is comfortable or too tight, and how much the pumps back off during mild weather. Alarms on low zone differential pressure give early warning that a branch is approaching starvation, at a point where a balancing adjustment or a pump setpoint change can fix it. Treated this way, the same measurement that drives the pump control also becomes the operator's window into the hydraulic health of the network.
The critical consumer is the substation that is hydraulically most remote from the network pumps, meaning it sees the smallest flow-to-return pressure difference and is therefore the hardest to supply. Differential pressure control references its setpoint to this consumer because guaranteeing enough differential pressure there guarantees it for every consumer closer to the plant. It is effectively the weakest point that defines what the pumps must achieve.
A floating setpoint lowers the target differential pressure when the network is running at low flow, because friction losses between the plant and the far end are small then, so less pump head is needed to satisfy the critical consumer. Since pump power rises steeply with speed, letting the pumps ease off whenever the load allows saves considerably more energy than holding a fixed worst-case setpoint. The control logic keeps a safety margin so no consumer is ever starved as the setpoint floats.
Flow tells you how much water is moving but not whether each consumer has enough driving force to draw its share, and a single pressure reading does not capture the difference between the supply and return pipes that actually pushes water through a substation. Differential pressure is that flow-to-return difference, which is exactly the available head a consumer works with. Controlling to it directly ensures every consumer, especially the remote critical one, has the driving force it needs regardless of how the load is distributed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.