Pressure-Relief Valve Selection and Transient Pressure Control
Pressure-Relief Valve Selection and Transient Pressure Control
The valve, which is normally closed, is designed to open rapidly once its pressure setting is
exceeded. A PRVs effectiveness depends on the properties of the system, the characteristics of
the surge experienced, and the way in which the valves attributes and settings are configured.
This article illustrates the challenges inherent in PRV design and shows that an appropriately
designed PRV can protect some systems from excessively high or low pressures and that
inappropriate use can actually worsen a systems transient response. The general principles of
PRV use and selection are presented along with a sensitivity analysis of PRV parameters. Although
this understanding is essential to effective system design, a PRV is selected by evaluating PRV
T
ous problems in water distribution systems. High water hammer pres-
BRYAN W. KARNEY, sure can permanently deform or rupture a pipeline and its components;
AND DAVID L. MCPHERSON low pressures can collapse a pipeline, causing leaks, disrupting service,
and contaminating pipelines.
Numerous transient control strategies have been developed, including changes
within the distribution system (pipeline diameter, thickness, alignment, profile, and
other hydraulic components), wave speed reduction, optimal operational pro-
cedures, and installation of dedicated devices such as automatic control valves,
surge tanks, and air chambers (Karney & Simpson, 2007; Wylie et al, 1993). Auto-
matic control valves, including pressure relief valves (PRVs), flow- or pressure-
regulating valves, air valves, and check valves, are common and often cost-effec-
tive. Depending on the type, a valve is used to control transient conditions either
by reducing the rate of net change in flow velocity in the pipeline or by dis-
charging or admitting air into the pipeline. When triggered by pressure that is val-
ued beyond a preset limit, a PRV opens to allow flow. The resultant outflow
causes a pressure drop and thus has the potential to reduce the maximum pres-
sure; inflow compensates for reduced water flow and can limit low pressures
and even cavitation. A PRV must have a low physical inertia so that it can
respond rapidly to the sensed pressure and open before the set point is greatly
exceeded (Chaudhry, 1987). As the case studies presented in this article show,
delays in valve opening can compromise protection of the distribution system.
Pipeline Maximum pressure head Steady-state pressure head Minimum pressure head
190 190
Pressure Headm
Pressure Headm
80 80
30 30
140 140
250 250
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.9 m, TV1 = 5 s, TV2 = 80 s
C Simulation with an undersized PRV D Simulation with a too-low-pressure set point PRV
300 300
190 190
Pressure Headm
Pressure Headm
80
80
30 30
140 140
250 250
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.2 m, TV1 = 5 s, TV2 = 80 s d = 0.9 m, SET2 = 2 m, TV1 = 5 s, TV2 = 80 s
E Simulation with a too-slow-opening PRV F Simulation with a too-fast-closure PRV
300 300
190 190
Pressure Headm
Pressure Headm
80 80
30 30
140 140
250 250
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.9 m, TV1 = 20 s, TV2 = 80 s d = 0.9 m, TV1 = 5 s, TV2 = 40 s
dvalve and port size, PRVpressure-relief valve, SET2low-pressure set point, TV1opening time period, TV2closing time period
Default SET2 is equivalent to SET2 15 m, i.e., the PRV is activated instantly when main valve 1 closes.
A B
80 40
Water Hammer Pressurem
C D
30 50
20 40
30
10
20
0
10
10
0
20 10
30 20
0 5 10 15 20 25 0 20 40 60 80 100 120 140
PRV Opening Time Period (TV1)s PRV Closure Time Period (TV2)s
d = 0.9 m, TV2 = 80 s d = 0.9 m, TV1 = 5 s
dvalve and port size, PRVpressure-relief valve, TV1opening time period, TV2closing time period
collapse. Cavitation will lead to vapor cavity formation and how this sensed pressure is itself changed by both
and subsequent cavity collapse (the so-called column rejoin- the ongoing transient in the system and action of the con-
der effect), which can cause extremely high pressures. The trol valve. Sensitivity analyses would reveal how the sys-
large negative pressures indicate that the response is unac- tems transient performance changes with variation of
ceptable when the PRV is not operated, and thus a protec- each PRV parameter, and this could aid in identifying
tion from transient pressure is necessary for this system. rules for PRV design. Therefore, the sensitivities of each
As shown in Figure 3, part B, operation of an appropri- PRV1 parameter (d, SET2, TV1, TV2) were analyzed and
ately designed PRV will limit maximum and minimum tran- are shown in Figure 4.
sient pressure heads along the pipeline to be in the range of Figure 4, part A, illustrates that an increase in valve
3.6 to 22 m. However, inappropriate selection of PRV para- size would improve transient performance but only to a cer-
meters, such as an undersized PRV, a pressure set point for tain limit; an increase in valve size beyond 0.5 m would not
SET2 that is too low, opening of a PRV that is too slow, and efficiently improve transient performance in this case study.
closure of a PRV that is too fast, would result in poor per- The authors design and field experience indicate that PRV
formance and unacceptable transient pressures in the diameters usually range from one twentieth to one third of
pipeline (Figure 3, parts CF). Yet, Figure 3 shows that the main pipe diameter, with most values near the middle
even a poorly designed PRV provides some protection and of this range. The logic for achieving this type of range is
is certainly preferable to avoiding the PRV completely. evident in Figure 4, part A, although not exactly followed.
Sensitivity analysis of upstream PRV parameters. The slight discrepancy that appears can be explained as
Proper selection of a PRVs control parameters is impor- follows: in this case study, only 3 m of pressure head existed
tant to transient control, particularly because these para- along a 500-m length of pipe; thus, the velocity in this sys-
meters are inevitably uncertain. For example, activation tem is very slow. This implies that in order to discharge a
of a PRV depends on the pressure it senses, which in turn certain amount of water, a larger PRV is needed. The ulti-
depends on exactly how and where the pressure is sensed mate choice of a PRV size is usually a compromise between
Pipeline Maximum pressure head Steady-state pressure head Minimum pressure head
Pressure Headm
100 100
50 50
0 0
50 50
100 100
150 150
200 200
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.5 m, TV1 = 10 s, TV2 = 70 s
C Simulation with an under-sized PRV D Simulation with a too-high-pressure set point PRV
250 250
200 200
150 150
Pressure Headm
Pressure Headm
100 100
50 50
0 0
50 50
100 100
150 150
200 200
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.1 m, TV1 = 10 s, TV2 = 70 s d = 0.5 m, SET1 = 50 m, TV1 = 10 s, TV2 = 70 s
Pressure Headm
100 100
50 50
0 0
50 50
100 100
150 150
200 200
0 100 200 300 400 500 0 100 200 300 400 500
Distancem Distancem
d = 0.5 m, TV1 = 60 s, TV2 = 70 s d = 0.5 m, TV1 = 10 s, TV2 = 40 s
PRVpressure-relief valve, SET1high-pressure set point, TV1opening time period, TV2closing time period
Default SET1 is equivalent to SET1 15 m, i.e., the PRV is activated instantly when main valve 2 closes.
A B
180 100
Water Hammer Pressurem
140 80
Maximum
100
60
60 Maximum
40
20
20
20 Minimum
Minimum
60 0
100 20
0 0.2 0.4 0.6 0.8 0 10 20 30 40 50 60
PRV Diameterm PRV High-Pressure Set Point (SET 1)m
TV1 = 10 s, TV2 = 70 s d = 0.5 m, TV1 = 10 s, TV2 = 70 s
C D
80 120
70 100
Water Hammer Pressurem
80
60
Maximum 60
50
40 Maximum
40 20
30 0
Minimum
20
20
Minimum 40
10
60
0 80
0 10 20 30 40 50 60 0 20 40 60 80 100
PRV Opening Time Period (TV1)s PRV Closure Time Period (TV2)s