Dry running happens when a pump operates without enough fluid reaching the pump inlet or internal wetted parts. In some cases, the pump has lost prime. In others, the source tank is empty, a suction valve is closed, air has entered the suction line, or the pump is running against a condition that gives little or no useful flow.
Dry running can damage mechanical seals, bearings, impellers, stators, bushes and other internal parts. Booster pump guidance commonly identifies shaft seal and bearing damage as key risks when pumps run dry, while low-flow or no-flow operation can also increase fluid temperature inside the casing.
This article explains the main ways to detect dry running, including power monitors, paddle flow switches, level switches, pressure instruments and temperature monitoring. The right method depends on pump type, fluid, suction arrangement and how fast the system must trip.
Start with what you need to prove
Dry-run protection works best when the control logic answers one clear question: “Does the pump have enough fluid to run safely?”
That can be proven in several ways:
| Detection method | What it proves | Main strength | Main limitation |
|---|---|---|---|
| Source level switch | Fluid is available before the pump starts | Prevents dry start | Does not prove flow through the pump |
| Flow switch or flow meter | Fluid is moving through the line | Direct no-flow detection | Needs correct installation and minimum velocity |
| Power monitor | Pump load has changed | No wetted sensor required | Needs correct baseline and trip setting |
| Pressure switch or transmitter | Pump is building expected pressure | Simple on fixed-duty systems | Pressure does not always prove flow |
| Temperature sensor | Casing, seal or motor temperature is rising | Good backup protection | Usually detects later than flow or power |
| Vibration or acoustic monitoring | Air, cavitation or abnormal running may be present | Useful for condition monitoring | Usually not a primary dry-run trip |
A reliable system often uses two signals. For example, a low-level switch can prevent a dry start, while a flow switch or power monitor can detect loss of prime after the pump has started.
Power monitoring detects dry running from motor load
A power monitor measures the electrical load on the motor. For many centrifugal pumps with radial impellers, active power generally changes with flow rate. When a centrifugal pump loses liquid, draws air, or runs at very low hydraulic load, the motor often consumes less active power. A power monitor can trip the pump when active power drops below a learned or calculated threshold. Technical guidance on active power monitoring describes it as an indirect dry-running protection method that can switch off the motor when active power falls below a defined minimum value. It also notes that the method suits continuously operated centrifugal pumps with radial impellers and avoids inline wetted sensors.
Power monitoring is useful where installing a flow switch is difficult. It can also work well on corrosive or fouling fluids because the sensor sits in the electrical panel, not in the pipe. This reduces exposure to corrosion, coating and blockage.
Power monitoring needs careful setup. The trip point should be based on normal operation at the expected duty point, low-flow condition and minimum safe flow. A motor that is much larger than the pump can make underload detection less sensitive because the motor’s no-load current remains a significant part of the measured current. For this reason, active power in kW is usually more useful than current alone.
Power monitoring also has limits. Viscosity changes, speed changes from a Variable Frequency Drive (VFD), worn hydraulics, blocked strainers and process changes can all affect load. On some positive displacement pumps, dry running may not create the same clean underload pattern as a centrifugal pump. For these applications, power monitoring should be checked against real site data before it becomes the only trip.
Flow switches detect the actual no-flow condition
A flow switch detects whether fluid is moving through the pipe. A paddle flow switch uses a vane or paddle in the flow stream. When flow pushes the paddle far enough, the switch changes state. Flow switch guidance describes paddle devices as mechanical switches activated by the pressure of the flowing medium, and notes that flow switches can trigger alarms, pump shutdowns or other control actions when flow falls outside set limits.
A discharge flow switch is one of the most direct ways to detect loss of flow. It can identify several dry-running causes:
- empty source tank
- closed suction valve
- blocked suction strainer
- air lock
- failed prime
- broken coupling or impeller issue
- closed discharge path, depending on installation point
Paddle switches are common because they are simple and easy to understand. They suit clean fluids and stable flow. They need the right pipe size, minimum velocity and installation position. Flow switch guidance commonly calls for straight pipe upstream and downstream to reduce disturbed flow at the sensor.
The main weakness is false tripping. Turbulence, pulsating flow, low velocity, slurry, stringy solids or coating on the paddle can cause unreliable switching. A paddle switch also needs a start delay. The pump may need a few seconds to prime and establish flow, so the logic should allow a timed proving period before tripping.
A typical control sequence is:
- Start the pump.
- Ignore the flow switch during a short proving delay.
- Confirm flow after the delay.
- Stop the pump if flow is not proven.
- Alarm and require operator reset after repeated failures.
Thermal, ultrasonic, magnetic or inline flow devices may suit applications where a paddle is likely to foul or wear. Non-intrusive ultrasonic options avoid wetted moving parts, while magnetic devices need a conductive fluid.
Level switches stop the pump before it runs dry
A level switch detects whether there is enough fluid in the source tank, sump, vessel or well. Common options include floats, conductive probes, pressure-based level measurement, ultrasonic level sensors and radar level sensors.
For open tanks and sumps, level protection is often the first safeguard. A low-low level switch can stop the pump before air enters the suction line. In a sump, it can also provide automatic restart once the level recovers, provided the process allows automatic restart.
Level switches work well when the source vessel is the main dry-run risk. They are less effective when the fault occurs between the tank and the pump. A tank may contain fluid, but the pump can still run dry if a suction valve is shut, a strainer is blocked, a foot valve leaks, or the suction line pulls air through a joint.
Use level detection as prevention, not as proof of pump performance. For higher-risk duties, pair it with flow, power or pressure monitoring.
Pressure switches can help, but pressure is not flow
A pressure switch or pressure transmitter can detect whether the pump is developing the expected suction or discharge pressure. This can work well in fixed-duty systems where the pump should always build a predictable pressure shortly after startup.
Discharge pressure can help detect loss of prime. If the pump starts and discharge pressure does not rise within the proving time, the control system can stop the pump. This is common on booster sets and transfer systems.
Pressure has a major weakness: it does not always prove flow. A centrifugal pump can generate pressure with a closed discharge valve while the casing has little or no through-flow. In that condition, the fluid inside the pump can heat rapidly. Dead-headed operation is described as no-flow operation where the pump churns the same volume of fluid, increasing temperature and risking wear or damage.
Suction pressure can also mislead the control system. A flooded suction may show positive pressure when liquid is present. Once the pump runs, suction pressure depends on static head, friction loss, vapour pressure and pump demand. In low-head systems, the difference between a safe and unsafe condition may be too small for a basic switch to detect reliably.
Pressure monitoring is useful, but it is stronger when combined with flow or power monitoring.
Temperature monitoring detects the heat caused by dry running
Dry running often causes heat. The heat may appear at the mechanical seal, casing, stator, bearings or motor winding. Temperature sensors can detect that rise and shut the pump down before damage progresses too far.
Temperature monitoring suits:
- magnetic drive and canned motor pumps, where internal bearings need liquid for cooling and lubrication
- progressive cavity pumps, where stator temperature can rise during dry running
- pumps handling hot fluids where casing temperature trends matter
- systems where low-flow heating is a major risk
The drawback is response time. Temperature usually rises after the fault has already started. Heat must transfer from the damaged or heating component to the sensor. For fast-damage applications, temperature should support faster methods such as flow or power monitoring, rather than replace them.
Vibration and acoustic monitoring can detect related faults
Dry running often brings other symptoms: cavitation, air entrainment, unstable hydraulic forces, bearing noise and vibration changes. Condition monitoring can detect these patterns.
Vibration monitoring is useful where pumps already have condition monitoring in place. It can help maintenance teams identify recurring suction problems, air ingress or unstable operation. Acoustic monitoring can also detect cavitation-like noise in some systems.
These methods are better for diagnosis than primary dry-run shutdown. The signal can vary with speed, pipework stiffness, fluid properties and nearby equipment. Use them to support reliability work, not as the only dry-run protection on a critical pump.
Choose the method by failure mode
The best dry-run detection method depends on how the pump is likely to lose liquid.
For a tank transfer pump, start with low-low level protection in the source tank. Add a discharge flow switch if a closed valve, blocked suction line or loss of prime could still occur.
For a booster pump set, combine source pressure or level protection with discharge pressure and flow proving. If the pump runs on a VFD, include speed-aware logic so the trip limits change with operating speed.
For a centrifugal process pump with clean liquid, a power monitor can provide fast dry-run detection without cutting into the pipe. Validate the underload threshold during commissioning.
For slurry or fouling service, avoid small wetted parts where possible. Consider non-intrusive flow measurement, pressure trend logic, power monitoring and planned inspection of sensors.
For positive displacement pumps, dry running behaviour depends strongly on pump type and materials. Flow proving and temperature monitoring are often more reliable than assuming underload will always occur.
Where flammable fluids or hazardous areas are involved, dry-running protection becomes part of the safety concept. Guidance on centrifugal pumps in hazardous service states that effective monitoring should initiate motor shutdown when appropriate limits are violated.
Key takeaways
- No single dry-run detection method suits every pump system.
- Flow switches prove actual movement of fluid, but paddle types need correct installation and clean, stable flow.
- Power monitors can detect underload quickly on many centrifugal pumps and avoid wetted sensors.
- Level switches prevent dry starts, but they do not prove flow through the pump.
- Pressure switches are useful, but pressure alone does not always confirm flow.
- Critical systems often need two independent signals and clear shutdown logic.
