Water Hammer

Water hammer effect is an impact load that is the most misunderstood force known to pressure transducers today. A water hammer is created by stopping and/or starting a liquid flow suddenly. The results of a water hammering or pulse load are devastating to a pressure sensor. The impulse load occurs suddenly, in the millisecond time frame, but the effects of it last a life time. Water hammers occur in almost all pressure systems and usually can not be stopped without extensive time, energy and studies.

A common example of a water hammer effect occurs in most homes everyday. Simply turning off a shower quickly sends a loud thud through the house; this is a perfect example of a water hammer. Dishwashers and washing machines make these same sounds, because inside them small solenoid valves are being opened and closed quickly, producing this pulse noise. The key phrase in the examples above was turning on or off the water "quickly" verses turning it off slowly. In the shower example, if you turn the water off slowly, the water hammer will not occur. Common industrial hardware like relief valves, solenoid valves, valves in general, centrifugal pumps, positive displacement pumps, and regulators can and will cause heavy hammer effects. A simple solution to this devastating effect is to protect each sensor with a pressure snubber. Snubbers are low ticket items that will insure that this liquid hammer effect will not render your costly sensor useless. All pressure sensors should utilize snubbers for all installations.

The hammer occurs because an entire train of water is being stopped so fast that the end of the train hits up against the front end and sends shock waves through the pipe. This is similar to a real train, instead of slowing to a stop, it hits into a mountain side. The back of the train continues forward even though the front can not go anywhere. Since the water flow is restricted inside the pipe, a shock wave of incompressible water travels back down the pipe deflecting everything in its path. An unprotected transducer in the path of this monstrous wake is without question, going to sustain heavy damage.

To understand the damage caused by the water hammer forces, it is necessary to understand the principles behind the sensor. Most pressure sensors utilize a rigid diaphragm as the primary sensing element. The diaphragm deflects due to the pressure, and its deflection is transformed to an electrical output via various methods. The key component is the rigid diaphragm. The rigid diaphragm deflects only on the order of a thousandth of an inch. With a large wake of fluid hitting the sensor, it is no wonder the diaphragm is bent beyond its elastic limit and permanent damage is done. Remember that a snubber eliminates this effect and therefore should always be installed on every pressure system.

Snubbers are chosen by the media that they will be used on such as liquids, gases or dense liquids like motor oils, and their physical mounting fittings. Snubbers only let so much fluid pass through per unit time, eliminating the surge from hitting the diaphragm. Liquids possess a large hammer effect because they are incompressible, but gases can also possess a hammer effect large enough to render a sensor useless. A practical analogue to a snubber is a sponge in the drain of a sink. The sponge ensures that the sink empties slowly, instead of all at once. A lot of common questions are asked about hammer effects; the following are just a few.

Will a snubber affect the response time of my pressure transducer?

In most cases, the transducer is connected to a meter of a recorder that updates at 2 to 3 times a second; therefore a snubber will not affect it at all.

What are the symptoms that my sensor has been damaged by a fluid hammer?

Most sensors will exhibit a higher than normal output at zero pressure (a zero shift). This occurs because the diaphragm can not return to zero. In severe cases no output occurs or the output does not change with an increase in pressure.

If my sensor has a large zero offset caused by this hammer effect can it be repaired?

Most sensors are non-repairable. The diaphragm is the main building block of the sensor. When building a sensor the diaphragm is first built and the all the other components are chosen to achieve the rated specification. When a diaphragm bends beyond its elastic limit, it can not be bent back to original shape or replaced because of the unique components associated with the original diaphragm. If a diaphragm does have a slight zero shift, less than 10%, it probably is still linear and can be used. Before reinstalling it in the system, please acquire a snubber or the hammer effect will occur again and possibly damage the unit further.

Will a snubber stop an overpressure?

Snubbers stop spikes only, they do not perform miracles. An overpressure will not be stopped a snubber. A spike lasts only on the order of milliseconds; any overpressure for more than that time will damage the sensor.

How is a snubber installed in a pressure system?

The snubber would screw on to the front end of the transducer and then thread into the piping system. The snubber is located between the piping under pressure and the pressure transducer. The following brief equations summarize the hammer effect and is followed by an example of the destructive force of water hammer. The following equation determines the maximum pressure change that occurs during a fluid hammer. The equation assumes that the piping is inelastic.

ΔP=rcΔv/g
where
c for liquids =(Eg/r)^½
and
c for gases =(KgRT)^½

Where

P is the change in pressure resulting from the fluid hammer (pounds per square foot)
r is the fluid density (pound mass per cubic foot)
cis the speed of sound in the fluid (feet per second) v is the change in velocity of the fluid (feet per second)
g is the gravitational constant (32.2 feet per second per second)
E is the bulk modulus of the fluid media (listed in PSI but must be converted to PSF)
k is the ratio of specific heats (k = 1.4 for air)
R is the specific gas constant (foot pounds per pound mass per degree Rankine)
T is the absolute temperature in Rankine


Example of water hammer occurring in typical house piping. Assuming you have one inch water piping, how much of a change in pressure will be created from a water hammer?

Assume that the water is flowing in 10 gallons per minute and the temperature is about room temperature (70°F). A 1 inch schedule 40 pipe has an internal area equal to 0.00600 ft2.

Fluid velocity V = Q/A = 10 gpm (1/448.83 gpm/cfs)/.006ft2 = 3.71 ft/sec.

Where Q is the flow rate, and A is the internal area in the pipe.

In this example, a 1 inch pipe with a flow rate of 10 gpm had a hammer effect resulting in an increase in pressure of 243 PSI above normal operating conditions. Considering normal city water pressure of 50 PSI, most end users would select a sensor of approximately 100 PSI full scale to be on the safe side. A 100 PSI sensor usually has an over pressure of 200% associated with it, meaning it will be able to withstand 200 PSI. Now the hammer increases the system from 50 PSI to 293 PSI (50 + 243), which is overpressurizing the transducer and causing damage to it. Most end users are puzzled as to how a system that is supplied with only 60 PSI is capable of producing over 200 PSI. After reading this article it should be evident that fluid hammers are a complex phenomena with a simple solution: installing snubbers on all pressure transducers.

c=(Eg/r)^½=[(320x10³lbs/in²)(144in²/ft²)(32.2ft/sec²)/62.3lb/ft³] ^½=4880ft/sec

ΔP=rcΔv/g=(62.3lb/ft³)(4880ft/sec)(3.71ft/sec)/32.2ft/sec²=35,029lbs/ft² or 243lbs/in²

Properties of water at atmospheric pressure

Temp.	Density	Density	Kinematic Viscosity	Viscosity	Surface Tension	Vapor Pressure	Bulk Modulus
°F	lbm/ft3	slug/ft3	lbf-sec/ft2	ft2/sec	lbf/ft	Head ft	lbf/in2
32	62.42	1.940	3.746 EE-5	1.931 EE-5	0.518 EE-2	0.20	293 EE3
40	62.43	1.940	3.229 EE-5	1.664 EE-5	0.514 EE-2	0.28	294 EE3
50	62.41	1.940	2.735 EE-5	1.410 EE-5	0.509 EE-2	0.41	305 EE3
60	62.37	1.938	2.359 EE-5	1.217 EE-5	0.504 EE-2	0.59	311 EE3
70	62.30	1.936	2.050 EE-5	1.059 EE-5	0.500 EE-2	0.84	320 EE3
80	62.22	1.934	1.799 EE-5	0.930 EE-5	0.492 EE-2	1.17	322 EE3
90	62.11	1.931	1.595 EE-5	0.826 EE-5	0.486 EE-2	1.61	323 EE3
100	62.00	1.927	1.424 EE-5	0.739 EE-5	0.480 EE-2	2.19	327 EE3
110	61.86	1.923	1.284 EE-5	0.667 EE-5	0.473 EE-2	2.95	331 EE3
120	61.71	1.918	1.168 EE-5	0.609 EE-5	0.465 EE-2	3.91	333 EE3
130	61.55	1.913	1.069 EE-5	0.558 EE-5	0.460 EE-2	5.13	334 EE3
140	61.38	1.908	0.981 EE-5	0.514 EE-5	0.454 EE-2	6.67	330 EE3
150	61.20	1.902	0.905 EE-5	0.476 EE-5	0.447 EE-2	8.58	328 EE3
160	61.00	1.896	0.838 EE-5	0.442 EE-5	0.441 EE-2	10.95	326 EE3
170	60.80	1.890	0.780 EE-5	0.413 EE-5	0.433 EE-2	13.83	322 EE3
180	60.58	1.883	0.726 EE-5	0.385 EE-5	0.426 EE-2	17.33	313 EE3
190	60.36	1.876	0.678 EE-5	0.362 EE-5	0.419 EE-2	21.55	313 EE3
200	60.12	1.868	0.637 EE-5	0.341 EE-5	0.412 EE-2	26.59	308 EE3
212	59.83	1.860	0.593 EE-5	0.319 EE-5	0.404 EE-2	33.90	300 EE3