For an EX pump handling acids or solvents, choose dry-run protection only after confirming both hazardous-area suitability and wetted-material compatibility; no single sensing method fits every fluid or reliably separates very low flow from no liquid without commissioning tests.
Commission the protection in a gated sequence
Use this sequence to select and prove the protection. Do not set a trip threshold from a nameplate or a single normal operating point: the installation must show how the chosen measurement changes between normal pumping, low flow, and dry running.
- Record the operating envelope. Before selecting hardware, collect pump type, normal and minimum operating conditions, fluid identity and concentration, temperature, expected flow range, and the location of each device relative to the hazardous area. Confirm the written operating envelope and area classification with the site engineering process; proceed only when the installation requirements are known.
- Check instrument suitability. For every wetted sensor, compare the actual fluid and temperature with the manufacturer’s compatibility data for the sensor body, probe, seals, and coating. For every device in or connected to the classified area, verify its approval and installation conditions against the applicable equipment documentation. Proceed only when both checks pass.
- Choose the measured variable. Decide whether the application needs direct flow detection, a local indication of liquid at the probe, or pump-load monitoring from the electrical room. A thermal flow switch or calorimetric probe measures a flow-related condition; a power relay infers pump load. Choose the method that can distinguish the relevant operating states, then verify its signal at normal operation.
- Capture reference readings. Record the selected instrument output or pump electrical power at the normal operating points and at the lowest acceptable flow. If safe and part of the approved test plan, capture readings for loss of liquid and other expected abnormal states. Do not set the trip until the recorded states are distinguishable.
- Set and test the trip. Configure the low-flow or underload response, delay, and reset behavior using the device documentation and the measured envelope. Test the protective action using an approved method. Proceed only after the measured trip condition causes the intended pump shutdown and the normal operating condition does not cause a nuisance trip.
- Document the limits. Record the measurement, reference readings, trip setting, test result, device location, and conditions under which the protection cannot distinguish low flow from dry running. Confirm the record against the approved operating procedure before returning the pump to service.
Distinguish very low flow from a dry pump
The critical decision is whether the process requires detection of zero flow, a minimum acceptable flow, or a broader pump fault. These are not interchangeable. A centrifugal pump can operate at very low flow while still containing liquid, and its electrical load can change with operating condition. A dry-run function based on a low-load threshold may therefore trip at an unacceptable low-flow point—or fail to distinguish that point from true loss of liquid—unless testing demonstrates a usable separation.
Use the following readings to direct the next check:
| Observed reading or condition | Likely interpretation | Next check |
|---|---|---|
| Flow indication falls below the required minimum | Low flow or no flow; a direct flow device can identify the threshold it senses. | Compare its response at the minimum acceptable flow and at the no-liquid condition. |
| Thermal probe output changes unexpectedly when air passes | Air bubbles can disturb calorimetric measurements, particularly in large quantities. | Repeat the test under the expected entrained-air conditions; assess whether another measurement is needed. |
| Pump power falls below its normal load band | Possible underload, including dry running, but the reading is an inference rather than direct proof of liquid flow. | Compare power readings at normal flow, low flow, and dry-run test conditions. |
| Current alone appears to indicate a dry pump | Current-only measurement was reported as ineffective for centrifugal-pump dry-run protection. | Do not adopt current alone as the trip variable; evaluate power measurement or direct flow sensing. |
| Power rises or falls outside normal load | Power relays may detect overload or underload; blockage and other abnormal conditions can also affect load. | Determine which fault states the device detects and test each relevant state before assigning the trip a dry-run meaning. |
Select a direct flow device for the wetted path
A flow switch or probe provides a measurement in the process line, so it is the most direct candidate when the protective requirement is a flow threshold. The cited options include FCI flow switches described as available for Ex areas, thermal flow instruments from IFM Electronic and Pepperl+Fuchs, calorimetric probes described as EExia, vibrating blades with PTFE or other coatings and Ex versions, and the KROHNE DW181 target flow controller. Treat these as candidates, not as blanket approvals for an installation.
For aggressive liquids, check the complete wetted assembly against the actual acid or solvent. One described calorimetric probe construction uses 316 Ti stainless steel, but that material’s compatibility must be verified for the specific product. A vibrating blade may be offered with PTFE or another coating; verify the coating, joints, and temperature limits in the device documentation. For the DW181, confirm that the required flow range and fluid compatibility match the application. The flow-switch recommendation for Ex service does not establish compatibility with aggressive liquids.
For thermal or calorimetric devices, establish the manufacturer’s response direction and switching point in the actual process. Air bubbles in large quantities can disturb calorimetric readings. If the probe’s output changes with entrained air, treat that as a real operating limitation and test whether the resulting trip behavior remains acceptable.
Use pump power when electrical-room mounting is preferred
Power-based monitoring can avoid installing the sensing module in the hazardous area when the selected equipment is mounted in the electrical room and its installation design permits that arrangement. The named candidates include EL-FI DLM20 modules and EMI 491 or 492 wattmetric relays. The DLM20 was described as using power relative to normal installation operation, with overload and underload thresholds for stopping the installation. EMI 491/492 relays were described as measuring power rather than current and detecting both dry running and blockage.
These devices infer process state from pump load. Their usefulness depends on the measured power bands for normal operation and fault conditions being distinct. The reported adjustment was delicate, so use measured operating data and follow the specific device’s setup instructions; do not transfer a threshold from one pump, fluid, or operating point to another. Confirm the unit’s area suitability and electrical connection arrangement from its documentation rather than assuming that mounting it outside the hazardous area resolves every protection requirement.
A detector based on COS PHI, voltage, and current was also proposed in an Ex d enclosure for underload detection. Treat it as a separate architecture: verify its certification, location, measured variables, and response against the installation requirement. Do not substitute current-only monitoring for power measurement on a centrifugal pump.
Use level feedback only when the process supports it
A separate control arrangement uses a variable-speed pump with a 4–20 mA input representing the level in a vessel above the pump. Pump speed is regulated as the vessel fills, with minimum and maximum speed limits. This arrangement was reported to work for level regulation, but it is not the same as a direct dry-run detector. It depends on the relationship between pump flow and vessel level changing quickly and predictably enough for the intended control and protection function.
If considering this approach, verify the level transmitter signal, pump speed limits, and vessel response across the expected operating range. Then test the specific loss-of-liquid condition. If the level loop cannot identify that condition before the pump is at risk, add a separate protective measurement rather than treating successful level regulation as proof of dry-run protection.
Set thresholds from measured operating states
For either flow sensing or power monitoring, collect repeatable readings at representative operating points. Include the lowest acceptable flow because a trip that works only at normal flow may interrupt valid operation. Where the protection must discriminate dry running from low flow, compare both conditions directly. If the measured bands overlap, adjust the measurement location or method, or define a different operating limit; do not disguise overlapping states with a guessed threshold.
For power relays, characterize both underload and overload if the selected relay uses both thresholds. A blockage may produce a different load response from dry running, and the device may detect either or both. Verify each intended fault separately. Record how speed changes, valve positions, fluid properties, and process transients affect the readings, then test the thresholds at the edges of the approved operating envelope.
Prove the trip and its hazardous-area installation
Before functional testing, confirm the installed sensor or monitoring system matches the selected configuration and the site’s area requirements. Use approved test methods to simulate the relevant fault; do not create an uncontrolled dry-running event simply to obtain a reading. Observe the sensor output, relay state, control-system response, and final pump shutdown. If any stage fails to follow the intended sequence, correct the wiring or configuration and repeat the test.
Verify both sides of the trip: normal pumping must remain stable without nuisance shutdown, and the defined low-flow or dry-run condition must produce the protective action. Test blockage separately where overload detection is required. Capture the actual settings and readings in the commissioning record, along with known limitations such as air-bubble sensitivity or inability to distinguish low flow from no product. Release the pump only after the final fault test confirms that the configured trip shuts it down as intended.
Frequently asked questions
How do I choose a dry-run detector for an EX pump?
First confirm hazardous-area approval and wetted-material compatibility for the actual fluid. Then choose direct flow sensing or pump-power monitoring based on which method separates normal flow, minimum acceptable flow, and dry running in tests.
How do I tell very low flow from dry running on a centrifugal pump?
Compare measured flow or pump power at the minimum acceptable flow and at the no-liquid condition. If the readings overlap, that measurement cannot reliably discriminate the states at the selected trip point.
Can I use motor current alone to detect pump dry running?
Do not rely on current alone for a centrifugal pump. The described application reported current-only protection as ineffective; evaluate power measurement or a direct flow sensor instead.
Will a 316 Ti calorimetric probe work with acid or solvent?
Check compatibility for the exact fluid, concentration, and operating temperature against the probe manufacturer’s data. Large quantities of air bubbles can also disturb the calorimetric reading.
How do I verify a dry-run trip without damaging the pump?
Use an approved test method to simulate the relevant fault, then observe the instrument output, relay state, and pump shutdown. The final verification is a confirmed trip under the defined fault condition with stable operation at the lowest acceptable flow.