Take a Test    Article Library    CEEJ Home    Submit an Article     Contact CEEJ

Article # 0063

Determination of the Cause of Water Hammer in Two Steam Pressure Letdown Stations

By Richard L. Jones PE

Steam Pressure Letdown Stations are common in large industrial facilities such as refineries and chemical plants and especially in cogeneration plants that are providing them with steam.  Often the steam being reduced in pressure and temperature starts out as very high pressure superheated steam.  The stations are comprised of a control valve for pressure reduction and a separate desuperheater for temperature reduction but many have more recently evolved to steam conditioning valves where both the pressure reduction and temperature reduction are accomplished within a single valve body.  In either case the application has the potential for experiencing significant water hammer and requires proper design and maintenance to ensure that water accumulation and hammer is avoided.   

In 2013 a Houston area cogeneration plant experienced a major water hammer event in one of two identical and parallel steam letdown stations.  The main line was moved several feet from the shock destroying concrete pipe supports and breaking the smaller drain piping.    Each station includes a steam conditioning valve (SCV), spray water control valve (SWV), safety valve, and drip leg with strainer and steam trap. Process conditions are 850 psig and 750 degrees F inlet and 175 psig and 400 degrees F outlet.  Installation was in 2004. 

The system layout has two parallel and interconnected stations each with a 10” inlet vertically down into the angle-bodied SCV with a 20” outlet horizontal in a straight run of 20 ft.  The outlet line then makes a 90 degree horizontal turn and extends another 20 ft. at which point the line turns vertically up and ultimately ties into the plant’s 175 psig steam header.  An isolation valve is located in the vertical up section of the discharge line.  The SWV is installed approximately 10 ft. upstream of its connection to the SCV.  The SCV has a body drain and steam trap at the low point of the valve which is also the low point of the high pressure steam inlet line.  The low pressure drip leg is located immediately prior to the elbow where the line turns vertical upward. The line from the SCV to the drip leg is sloped toward the drip leg to ensure any condensate or spray water fallout drains into the drip leg.

There are several ways water could become present in the line in sufficient quantity that could result in water hammer. The source could be either upstream or downstream of the SCV.  The upstream option is that condensate can form in the dead-leg section of the high pressure superheated line upstream of the SCV during periods when the SCV is closed and there is a failed-closed, plugged, valved-out or missing steam trap in the body drain of the valve.  When water hammer occurs due to water present at the inlet of the valve at start-up there is usually valve damage that includes the diffuser (when one is installed) being blown out of the valve outlet into the downstream piping.  The downstream option can occur from one or more of multiple possibilities: a) Condensate can form in the dead-leg section of the low pressure  line downstream of the SCV during periods when it is closed and there is a failed-closed, plugged, valved-out or missing steam trap b) A poorly performing SCV !
 or poorly designed installation can result in excess spray water which falls out of the steam stream without evaporating and in sufficient volume that the steam trap is unable to handle the load c) an incorrectly controlled or a leaking or open SWV can allow water to be injected in the line when there is inadequate or no steam.

The preliminary site inspection began with the SCV which was in the shop with the outlet accessible to view.  There was no apparent damage to the valve internals including the outlet diffuser.   Had the water hammer originated upstream of the SCV there would probably have been visual damage.  The diffuser is generally blown out of the valve into the downstream piping.  It was therefore assumed that the water build-up was after the SCV.

There was concern that the other letdown station could experience the same type of damage and become unavailable.  While still in operation the insulation was removed and a temperature profile was made 360 degrees around the 20” pipe just upstream of the drip leg.  The results were that there was a 100-110 degree difference between the upper and lower areas of the pipe indicating significant accumulation of water in the pipe.  Further the trap was found to be in service (not valved-out) but cold and not operating and either failed closed or plugged.  The trap bypass was opened and significant water drained from the line.  It was clear that the second station was in danger of experiencing a serious water hammer.  The station was removed from service until the cause of the excess water was determined and the system modified to resolve the problem.

The following actions were taken and details assimilated for analysis and the best possible prediction of the cause of the water hammer:
a) Inspection of the physical system
b) Review of isometric drawings
c) Review of operating trends including past start-ups
d) Calculations using the process conditions to confirm suitability of installed equipment
e) Discussions with operators and witnesses to the event
The analysis yielded the following results:
a) The system appears to have been designed properly and consistent with the manufacturer’s general recommendations and good engineering practice except that the location of the temperature sensor/transmitter may be a bit close to the SCV. 
b) The application is not an easy one as high flow rangeability is required.  This high rangeability is easily handled in the pressure control but is most difficult for the temperature control.  This is even more challenging considering the small difference between the set point and the saturation temperature.  Any other issue such as wear on equipment, low water temperature, etc. only makes the temperature control more difficult.
c) Based on the original conditions the percentage of water addition is fairly high at approximately 15%.
d) A study of trends does not reflect an abnormality in the SWV stroke and water flow relative to the SCV stroke and steam flow.  Thus there does not appear to be an excessive amount of spray water being added to the steam.
e) Operators had frequently heard loud hammering noises consistent with small volumes of water in a steam line being picked up by the fast moving steam and impacting the pipe at changes in direction.
f) Significant water was seen discharging from the safety valves during the event.  The safety valves are located a few feet upstream of the elbow where the 20” line turns vertical up.   This indicates significant water in the line being moved at high velocity and resulting in an effective high pressure at the safety valves.

Based on the analysis, it appeared that the source of the water was not a serious failure of the steam conditioning equipment with large water fallout overwhelming the drainage system but was probably due to a combination of condensate from a period where the steam conditioning valve was closed for an extended period and/or some spray water fall-out from the desuperheating process at low flow conditions over an extended period.  However at the reasonably small rates of water accumulation the installed steam traps should have readily handled the load.  It was therefore reasoned that the entire issue was the failure of the steam trap to drain the low pressure side of the SCV.

The steam traps and upstream strainers were cut from the lines and disassembled in the shop.  The strainers located immediately upstream of the traps were totally plugged thereby eliminating the only downstream condensate drainage possible in both lines.  With the trap removed from service, even a small amount of spraywater fallout would eventually create adequate water for such a hammer to occur.

Recommendations were provided to minimize the possibility of such water hammer events from occurring in the future:
a) Install a level indicator with switch on each drip leg to allow visual observation of the water level and send an alarm to the control room. 
b) Install suitable trap bypass valves in case an extended blowdown is required in the future.
c) Install the trap discharge piping to drains such that the discharge can be visually observed and this used in the determination of trap performance and estimation of load.
d) Interlock SCV and SWV to ensure the SWV never opens without SCV being open.
e) Regularly inspect the system and review the trends for any sign of problems or change in operation and performance.
f) Regularly blow down the Y-strainers upstream of the traps.
g) Begin a regular program of steam trap testing.  Prioritize the most critical traps and test them more often. 

About the Author

Richard Jones, PE is the President of Richard L. Jones, Inc. and a 1975 Nuclear Engineering graduate of Texas A&M University.  P E Registration Number: 51266


Article # 0063        


Take a Test    Article Library    CEEJ Home    Submit an Article     Contact CEEJ