A discharge pipe jumps when a pump stops. A check valve bangs after a power interruption. Mechanical seals keep leaking, even though the pump runs normally once it reaches speed. These symptoms justify investigating pressure transients, but they don’t prove water hammer by themselves.
Pump starts and stops can create damaging water hammer when liquid velocity changes faster than the piping system can accommodate. The event may begin as a pressure rise or a pressure drop. Finding the cause means recording pressure during the transition and comparing it with pump speed, valve movement, and system conditions—not just reading a gauge during steady operation.
What Happens During a Pump Pressure Transient?
Moving liquid carries momentum. When a pump or valve changes its velocity, a pressure wave travels through the liquid and piping. That wave reflects at tanks, closed valves, branches, and other boundaries.
Water hammer can occur with liquids other than water. Fluid density, pipe material, wall thickness, restraint, and entrained gas all influence the response.
The familiar bang is only one possible symptom. A transient can also produce:
Brief pressure spikes that ordinary gauges never display.
Low-pressure conditions that approach the liquid’s vapor pressure.
Check-valve impact, pipe movement, or support damage.
Repeated loading of seals, gaskets, instruments, and valve components.
Damage depends on both the highest and lowest pressures, component ratings, and repeated loading. A piping system that tolerates steady discharge pressure may not tolerate recurring surges or vacuum conditions.
Starts and Stops Create Different Problems
During startup
A pump starting into a full line accelerates an existing liquid column. A rapid speed increase, abrupt discharge-valve opening, or sudden check-valve opening can generate a pressure wave. Starting another pump on a shared header can also disturb pumps already operating.
A partially drained line presents a different problem. Liquid can accelerate through an empty section, compress trapped air, and strike a closed boundary or another liquid column. Simply lengthening the motor acceleration ramp may not address that filling event.
For centrifugal pumps, startup valve position must match the application and manufacturer’s operating limits. Don’t transfer that sequence to a positive displacement pump: it needs an available discharge path and properly designed overpressure protection.
During shutdown or a power trip
When pump head falls, forward flow decelerates. In an uphill discharge line, static head can drive flow backward. If a check valve remains open until appreciable reverse velocity develops, its eventual closure can stop that reverse flow abruptly and create a surge.
This is why “install a slower-closing check valve” isn’t a reliable blanket recommendation. The valve’s dynamic response must suit the system’s flow deceleration.
A pump trip can also create an initial low-pressure wave. If local absolute pressure reaches vapor pressure, vapor cavities may form and separate liquid columns. Their later collapse can create a severe pressure rise. The damaging event may arrive after the motor has stopped.
Use Event Timing to Narrow the Investigation
The most useful first question is: Exactly when does the disturbance happen?
Immediately at startup: Investigate acceleration, valve opening, line fill condition, and trapped air.
Shortly after shutdown: Investigate flow reversal, check-valve closure, and reflected waves.
Only during power failures: Compare uncontrolled coastdown with the normal shutdown sequence.
When another pump starts or stops: Review shared-header interactions and pump sequencing.
Throughout steady operation: Consider cavitation, reciprocating-flow pulsation, valve instability, or mechanical vibration rather than assuming transition-related water hammer.
These are starting points, not diagnoses. A sound can travel through structural steel, and the loudest location may not be where the transient originates.
Consider a hypothetical North Mississippi facility pumping water to an elevated storage tank. Scheduled VFD stops are quiet, but a thunderstorm-related outage produces a discharge-line bang. The normal ramp-down isn’t controlling the outage event. The investigation should address pump coastdown, static elevation, reverse flow, and check-valve response—not just extend the programmed stopping time.
Capture the Event Before Changing Equipment
A steady pressure gauge or slow historian trend can miss a short surge entirely. Use appropriately rated, fast-response pressure transducers with synchronized recording of relevant operating signals.
Useful measurements include pump suction and discharge pressure, pressure at vulnerable high points or remote boundaries, pump speed, motor status, valve position, and flow where instrumentation can resolve the event. Recording only at the pump may miss the lowest pressure elsewhere in the line.
Sensor response, sampling rate, and instrument connection details matter. Long impulse lines and heavy signal filtering can hide the waveform. A transient specialist should select recording settings based on expected wave travel times and event duration.
Before testing, assemble:
Pipe lengths, inside diameters, materials, elevations, and major branches.
Liquid temperature, density, vapor pressure, viscosity, and gas or solids content.
Pump curves, operating flow, speed settings, and available rotating-inertia data.
Check-valve type, size, orientation, condition, and dynamic performance information.
Control sequences, actuator timing, tank levels, and existing surge-protection details.
Plant personnel can document event timing, visible pipe movement, leaks, alarms, and operating conditions from safe locations. Don’t deliberately reproduce a suspected damaging trip without an approved engineering test plan. Instrument installation and internal inspections require applicable isolation, depressurization, lockout/tagout, and facility procedures.
What a Surge Calculation Can—and Cannot—Tell You
For a rapid velocity change, the Joukowsky relationship estimates pressure change as ΔP = ρaΔV. Here, ρ is liquid density, a is pressure-wave speed, and ΔV is the change in liquid velocity.
The practical message: a modest velocity change can produce substantial pressure change in a stiff, liquid-filled system. Wave speed isn’t simply the liquid’s sound speed; pipe flexibility and other system conditions affect it.
Engineers also compare the transition time with the approximate round-trip wave travel time, 2L/a, where L represents the relevant pipe length to a reflecting boundary. This helps assess whether the change is hydraulically rapid.
Neither relationship replaces transient modeling for a branched network, multiple pumps, trapped gas, column separation, or complex valve behavior. Those applications need a time-based analysis that checks pressure throughout the system, not just at the pump flange.
Match the Remedy to the Recorded Mechanism
Replacing a pump or adding a generic surge device before identifying the event can leave the original problem in place.
Startup acceleration: Review VFD ramps and valve sequencing together. Motor speed and liquid flow don’t necessarily change at the same rate.
Check-valve slam: Select for the expected deceleration and reverse-flow behavior. “Non-slam” labeling alone isn’t enough; request applicable dynamic data.
Low pressure after a trip: Evaluate engineered surge storage, hydropneumatic vessels, or other protection suited to the line profile and fluid.
Air accumulation or line draining: Review filling procedures, high points, and application-appropriate air-management devices. Air admission and subsequent expulsion can themselves affect surges.
High-pressure excursions: Evaluate surge-relief response, capacity, location, and safe discharge destination. An ordinary relief valve may not react quickly enough.
Shared-header disturbances: Review pump staging, minimum operating speeds, and interactions between control loops.
Variable frequency drives can help with controlled transitions, but a programmed ramp cannot be assumed to remain available during loss of power. Likewise, stronger pipe supports may address restraint deficiencies without correcting the hydraulic load.
What Buyers Should Request
For replacement valves, pumps, or surge-protection equipment, provide the event records and system layout—not just pipe size and normal pressure.
Ask the supplier or engineer to address minimum and maximum transient pressure, fluid compatibility, solids handling, operating temperature, maintenance access, and behavior during power loss. Pressure ratings and allowable transient limits should be checked against current manufacturer data for all affected components.
After modifications, repeat approved measurements under representative conditions. A quieter shutdown is encouraging, but pressure records provide better evidence that the problem has been addressed.
Bottom Line
Investigate pump water hammer during starts and stops as a system event. Establish the sequence, capture both pressure peaks and dips, and match the correction to the mechanism. Don’t assume a new pump, longer ramp, or different check valve will solve every surge.
Process & Power can help facilities throughout Tennessee, Arkansas, and Mississippi review pumping conditions and discuss equipment or service support for recurring transition problems.
Call Process & Power, Inc. at 901-362-5500 or visit us at 1721 Corporate Avenue, Memphis, TN 38132 for help with industrial air compressors, pumps, blowers, vacuum systems, and compressed air equipment and service throughout Memphis and the surrounding area.
