Differential-Pressure Sensor Placement and Reset Control for Variable-Flow Chilled-Water Pumps
A variable-speed pump needs a pressure target. Excessive differential pressure wastes pumping effort; too little can leave coils short of water. Sensor location establishes what the controller sees, while reset logic adjusts the target as loads change.
Measure the pressure the load needs
Identify the hydraulically critical circuits from the pressure-loss calculation. The most distant coil is a candidate, but physical distance alone does not establish the critical branch. Different branches can become limiting as valves modulate, so larger networks may need several remote sensors or demand requests.
Connect the high-pressure port to the supply and the low-pressure port to the return at the selected branch. Show exactly which coil, valve, strainer and pipe losses lie between the taps. Taps at comparable elevation simplify interpretation. A sensor across the pump measures total pump differential, which includes distribution losses outside the remote branch.
Calculate an initial branch setpoint
Consider an illustrative secondary chilled-water loop. At design branch flow, the coil loses 28 kPa, its control valve 12 kPa, the strainer 5 kPa and the branch piping 5 kPa. All four lie between the remote sensor taps.
Δpbranch = 28 + 12 + 5 + 5 = 50 kPa
If the remaining supply-and-return path loses 80 kPa at design conditions, the pump must develop 130 kPa. For water with ρ = 1000 kg/m³:
Hpump = Δppump / (ρg) = 130000 / (1000 × 9.81) = 13.3 m
The remote control target is initially 50 kPa, not 130 kPa. As system flow falls, distribution losses change. Confirm the required branch pressure by testing and balancing; the example values are not commissioning settings. In a filled closed circuit, building height is not added as circulating pump head.
Reset slowly from valid demand
Use a fast, stable pump-speed loop to maintain measured DP at its setpoint. A slower supervisory loop can trim that setpoint downward while loads remain satisfied, then raise it when valid requests persist. For conventional valves, near-full opening with unmet demand can indicate insufficient pressure. Verify the feedback point: a flow command to an electronic pressure-independent valve is not necessarily its physical opening.
Define adjustable limits, deadbands, request persistence and response delays in the sequence. Diagnose a stuck valve, dirty strainer or undersized coil before allowing it to keep the plant at maximum pressure. Any exclusion needs an alarm and documented engineering review. Coordinate DP reset with chilled-water temperature reset so both do not react independently to the same valve demand.
Example secondary-loop pressure boundaries: 50 kPa across the branch and 130 kPa at the pump.
Keep minimum-flow protection active
In variable-primary systems, the minimum pressure must still permit the required evaporator flow through the bypass. Verify flow for every enabled chiller and stage; a remote DP reading alone does not prove it. Also preserve pump operating limits and the manufacturer’s minimum differential across pressure-independent valves.
Commission the complete control chain
Check sensor range, calibration, port orientation and air-free sensing lines. Establish the upper setpoint under the specified balancing procedure. Trend DP, target, pump speed, critical-valve feedback, flow and load temperatures during low load and staging. Test signal-loss fallback and recovery. Stable load delivery is the acceptance test; lower pump speed alone is insufficient.
Build the pressure-loss schedule before writing the reset sequence. Nexora’s Pump Head Calculator Design Suite (Excel) helps organise pipe, fitting and equipment losses for engineering review. Carry the reviewed branch requirements into the controls specification and field tests.
Comments