Water hammer: the pressure spike your logger never sees
TL;DR. Water hammer is the pressure surge that runs through a pipe when flow is stopped or started suddenly. It is over in a fraction of a second, which is why a logger taking a reading every 15 minutes records 96 numbers a day and misses every single one. The damage is cumulative: each surge works joints, fatigues the pipe wall and widens existing leak paths, so the failure arrives months later and gets written up as an old pipe.
Key takeaways
- Water hammer is governed by the Joukowsky equation, where the pressure rise equals fluid density multiplied by wave speed multiplied by the change in flow velocity.
- A pressure logger on a 15-minute interval takes 96 readings a day, so a surge lasting under a second is invisible to it.
- England lost 2,690 megalitres a day to leakage in 2023 to 2024, which the Environment Agency puts at 19% of the water put into supply.
- CIWEM states that lowering or stabilising pressure, including reducing transient effects, helps to reduce bursts.
- Small reductions in maximum pressure over large areas cut burst rates more than large reductions over small areas.
- Catching water hammer needs high-frequency pressure logging at the sources of surge: pumps, actuated valves and fast-closing fittings.
Table of Contents
On this page
Last updated: 21 August 2026
What is water hammer?
Water hammer is the pressure wave created when moving water is forced to change velocity quickly. Water is nearly incompressible and a column of it in a pipe carries real momentum, so when a valve slams or a pump trips, that momentum has nowhere to go. It converts into pressure, which travels along the pipe as a wave, reflecting off every boundary until friction kills it.
The banging noise most people associate with water hammer is the least important part. The noise is the pipe moving against its brackets. The engineering problem is the pressure itself, which can briefly reach several times the normal operating pressure of the system, and then swing the other way into a negative phase that can collapse a thin-walled pipe or pull contaminated water in through a joint.
How big is the pressure spike from water hammer?
The size of a water hammer surge is set by the Joukowsky equation, which states that the magnitude of the pressure wave equals fluid density multiplied by the speed of sound in the fluid multiplied by the change in flow velocity. Written out it is ΔP = ρa₀Δv. The important thing about that relationship is what is missing from it: the operating pressure of your system does not appear at all.
So water hammer does not scale with how hard you are running the network. It scales with how fast the water was moving and how abruptly you stopped it. A modest velocity change stopped instantly produces a large spike on top of whatever pressure the pipe was already holding. The wave speed term is not a fixed number either; it depends on the compressibility of the water and the elasticity of the pipe, so a rigid metal main and a plastic service pipe behave differently under the same event.
That is the counter-intuitive part for anyone managing an estate. A site running at a comfortable pressure with no complaints can still be generating punishing transients every time a booster set cycles or a solenoid valve on a washroom or a cooling system snaps shut. Nothing on the standard reporting pack will show it.
Why does a normal pressure logger miss water hammer?
Because of the sampling interval. A typical network logger records pressure every 15 minutes, which is 96 samples in 24 hours. A transient event begins and ends inside a fraction of a second. The chance of a 15-minute sample landing on the peak of a sub-second event is effectively zero, and even if one did, a single high reading in a day of flat data reads as a glitch and gets filtered out.
This is the central measurement problem. The logger is not broken and the data is not wrong. It is simply answering a different question: what is the average pressure in this zone. It was never designed to answer what is the maximum pressure this pipe experienced, and maximum pressure is the number that predicts failure.
Catching water hammer needs logging at a rate fast enough to resolve the wave, which means sampling many times a second rather than a few times an hour, and it needs the logger sited where surges are generated rather than at the notional zone centre. Those two changes are what separate transient monitoring from ordinary pressure monitoring.
What damage does water hammer actually do?
Water hammer almost never causes a single dramatic burst. The realistic failure mode is cumulative. Each surge flexes the pipe wall, works mechanical joints, loosens fittings and enlarges the paths water already escapes through. Enough cycles and a pipe that was serviceable fails at a point of stress concentration, typically months or years after the surges that caused it.
This is why the industry treats pressure as a lever on asset life rather than only on flow. CIWEM’s policy position statement on distribution network leakage sets out that lowering or stabilising pressure, including reducing transient effects, can also help to reduce bursts. It adds a finding that reframes most pressure work: small reductions in average and maximum pressures over large areas are likely to be more beneficial in reducing burst rates, on both mains and services, than large pressure reductions over small areas.
There is a second effect on leakage volume as well as on failure. The FAVAD concept, proposed in 1994, describes how the flow from a leak varies with the square root of pressure while the area of some leak paths also changes with pressure, so a leak in a flexible material opens wider as pressure rises. Cutting the peaks therefore reduces what is escaping through leaks that already exist, before any repair is made.
How much water is at stake in the UK?
England lost 2,690 megalitres a day to leakage in the year to March 2024, which the Environment Agency puts at 19% of the water put into supply. That is the network-side number and it is the one that gets quoted, but the same physics applies to the pipework inside a boundary, where nobody is measuring at all.
The same Environment Agency report notes that pipes are particularly prone to bursts during both freeze-thaw events and hot dry conditions due to ground movements. Ground movement and transient loading stack. A pipe already stressed by a cold snap is the one that fails on the next surge, which is why winter and the end of a long dry spell produce clusters of failures rather than an even spread.
Where does water hammer come from on a typical site?
Water hammer is generated wherever something changes flow velocity quickly. On a commercial or public-sector estate that is a short and predictable list, which makes it a practical monitoring plan rather than a survey of everything.
- Booster pump sets, especially fixed-speed units that start and stop rather than ramp.
- Solenoid and actuated valves that close in well under a second, common on washrooms, irrigation and process lines.
- Pump trips on power interruption, which is the uncontrolled version of the same event and usually the worst case.
- Fast-acting float and ball valves on tanks and cisterns, which slam as they seat.
- Rapid closure of fire system test valves and any manual quarter-turn valve on a large main.
Monitor for water hammer at those points, not at the point of entry. A transient attenuates as it travels, so a logger at the incoming main will under-read an event generated in a plant room three floors up.
The full AQUAIOT water hammer capability stack
Flow: establish what the velocity actually is
Before you can reason about surge you need the velocity term. The Clamp-on Ultrasonic Flow Meter reads bi-directional flow and total volume on live pipework from DN8 to DN100 without cutting the pipe or taking a shutdown, so a plant room can be characterised while it stays in service. Knowing the normal running velocity and how sharply it changes at each pump start is the input the Joukowsky relationship needs.
Consequence: find the leaks the surges are widening
The NOAH Multifunction Leak Sensor covers the downstream half of the problem. Its Leak Sensing Membrane triggers the moment water makes contact, it runs on LoRaWAN Class A with battery life up to 10 years, and it goes into the plant rooms, risers and voids where a fatigued joint finally lets go. Water hammer causes the failure; contact sensing tells you which room it happened in.
Level: the tanks at the end of the run
Tank fill valves are both a cause of water hammer and a casualty of it. AQUAIOT Radar gives non-contact level to within ±2 mm through foam and condensation, so a float valve that has started hunting or failing to seat shows up as a changed fill signature rather than as a callout.
Telemetry and alarms
All of it reports over encrypted cellular and LoRaWAN into the AQUAIOT Cloud, with threshold alarms, trendlines and exports, plus RS485 and Modbus on site and open APIs into SCADA, BMS and CAFM. That matters here because transient work produces a before and after that has to be defensible to whoever funds the surge vessel or the soft starter.
How do you actually reduce water hammer?
The engineering answers are old and well proven. Slow the closure: fit soft starters or variable speed drives so pumps ramp instead of slamming, and specify slow-closing valves where a fast one is not needed. Absorb the wave: surge vessels, accumulators and air valves give the energy somewhere to go. Reduce the velocity: an oversized pump on an undersized main creates the conditions for surge every time it runs.
None of those is difficult. Knowing which one you need, and proving it worked, is the hard part, and that is a measurement problem rather than a hydraulics problem. Fit the monitoring first, characterise the site for a few weeks, then intervene where the data points.
Frequently asked questions
Is water hammer the same as high pressure?
No. High pressure is a steady state you can read off any gauge. Water hammer is a transient event caused by a rapid change in flow velocity, and the Joukowsky equation shows the surge depends on density, wave speed and velocity change rather than on the system’s operating pressure. A site at a perfectly normal pressure can still be producing severe transients.
Can you hear water hammer?
Sometimes, but silence proves nothing. The bang is the pipe moving against its fixings, so a well-clipped or buried pipe can carry a severe surge without making any noise at all. Audible hammer tells you there is a problem; the absence of noise tells you nothing about the pressure the pipe is seeing.
How fast does a pressure logger need to sample to catch water hammer?
Fast enough to resolve an event that begins and ends within a fraction of a second, which means many samples per second rather than the 15-minute interval used for zone pressure. A 15-minute logger takes 96 readings in a day and will not record a transient at all. High-frequency logging is a different instrument configuration, not a setting change on the same device.
Does reducing pressure fix water hammer?
It helps but it is not a cure, because the surge sits on top of whatever pressure is already there rather than being caused by it. CIWEM notes that stabilising pressure and reducing transient effects both contribute to fewer bursts, and that small reductions in maximum pressure across large areas do more for burst rates than deep reductions in small ones. Slowing the closure that creates the wave attacks the cause directly.
Where this leaves an estates team
Water hammer is the clearest example of a failure mode that is invisible to the instrumentation most sites already have. The pressure data looks calm because it is averaged, the pipe looks fine because the damage is cumulative, and the eventual burst gets recorded as age. None of that is wrong. It is just not the whole record.
If you have pumps that start and stop, valves that snap shut, and a burst history nobody can explain, the gap is measurement rather than maintenance. Speak to an AQUAIOT engineer about characterising flow and pressure behaviour on your plant rooms and incoming mains.
By GP, Digital and IoT lead at AQUAIOT.