What Pressure Relief Really Means

Chemical Engineering Progress

September 1, 2010 | Kelly, Brian D
Poor design, inadequate maintenance, and improper operation can all lead to
mechanical failure. The basic concepts of pressure relief, when understood and
applied properly, can help prevent this failure.
A pressure-relief system protects pressure-coded equipment from
overpressure. Such equipment includes pressure vessels, heat exchangers,
and piping systems that are designed and constructed to meet American
Society of Mechanical Engineers codes (I, 2) with a maximum allowable
working pressure (MAWP) exceeding 15 psig. Even though regulations around
the world have strict requirements that the MAWP of coded equipment never be
exceeded, many major process incidents have been caused by relief system
failures. Therefore, engineers need to focus on pressure relief as a matter of
process safety.
This article will familiarize the novice engineer with some of the challenges in
pressure-relief systems - by first reviewing some important concepts that must
be understood, then exploring relief system design.
Relief protection strategy
Internal pressure in any closed system is a function of the quantity of material
within that system, as well as other internal forces. Pascal's Principle states that
the static pressure will be uniformly distributed across the system over all
surfaces. The only practical way to alleviate high system pressure is to remove

or release energy from the system. Pressure-relief methods accomplish this,

and when a pressure-relief device is connected to a closed blowdown system,
material can be transferred to a location where it is unlikely to harm people or
damage the environment.
In practice, relief protection is sometimes misapplied as a substitute or
afterthought for poor design. Good design requires a thorough understanding of
the process under both normal and upset conditions. Equipment selection
should consider the full range of operating conditions likely to be encountered
during a typical run cycle. This will determine the equipment's metallurgy and
the material wall thickness.
Once a pressure rating has been established for a piece of equipment, the
facility in which that equipment is installed must ensure that actual operating
conditions never exceed that level. The MAWP for an operating system is
based on the lowest MAWP of all the components in the system. This is
typically the pressure vessel. Instrumentation systems that use control valves or
pressure regulators moderate process pressures to tight tolerances. Although a
process circuit may safely demonstrate an ability to withstand operating
pressures higher than its MAWP, it should not be subjected to those conditions.
A relief protection strategy addresses higher pressures than a process control
system can deal with, as well as emergency situations. Human errors and
equipment malfunctions can disrupt a process and cause vapor and/or liquid to
accumulate within a closed system. A material balance illustrates how this
causes system pressure to increase - and ultimately results in loss of
containment. If material is removed at the same rate as it is generated, pressure
will remain in balance until safe operating conditions can be restored.
The principal challenge in designing relief systems is to recognize and define
upset conditions that threaten pressurecoded equipment. These conditions may
include flow, pressure, temperature, phase and composition.
Pressure relief valves

The heart of a pressure-relief system is a spring-loaded pressure relief valve

(PRV) that is actuated by the upstream static pressure. Several different types
of relief devices are available.
The simplest PRV is the conventional relief valve (Figure 1) used to protect
systems filled with incompressible fluids, such as liquids and slurries, which are
sensitive to thermal expansion. As the fluid expands, the pressure against the
disc or plug increases; when it reaches a predetermined setpoint, the spring
compresses and the valve opens in proportion to the pressure increase over the
set pressure. To minimize pressure drop across the valve, the spring
mechanism is enclosed within the cavity of the bonnet, which is outside the
normal flow path. The valve element is a fiat disc or plug that normally rests on
an annular seat. The valve opening (trim size) is based on its ability to relieve
the excess volume associated with an overpressure scenario. The inlet line and
bottom flange on the valve must be sized so that frictional losses will not exceed
3% of the set pressure. Otherwise, the valve will experience variable pressure
and will chatter, which causes damage to the seat. The outlet flange on a relief
valve is typically one flange size larger than the inlet.
Selecting a conventional relief valve involves determining the appropriate set
pressure and the associated relief load needed to maintain that pressure. The
orifice size required to pass the liquid flow volume within a specified period of
time is determined by:
where Q^sub L^ is the liquid flowrate, C^sub v^ is the orifice discharge
coefficient and equal to 0.62 for a sharp-edge orifice, A is the orifice area, g is
acceleration due to gravity, and h is the pressure differential across the valve
orifice.
When the valve is fully open, the pressure on the plug will normally exceed the
set pressure by 10%; this is allowed under emergency conditions. The set
pressure must never exceed the maximum allowable working pressure of the
weakest component in the system being protected.

A relief valve is not custom-designed for a specific scenario. PRVs of various

body and orifice sizes are available as standard stock items from equipment
suppliers (Table 1). Relief valves are selected based on body and orifice sizes
and identified by orifice letter designations. The spring is specified for the
design set pressure, and final shop calibration takes place before the valve is
installed.
Safety valves
A safety valve has a different spring characteristic than a relief valve. Externally,
both valves look similar. A safety valve will pop, whereas a relief valve will
gradually lift off the seat. When handling gases or compressible fluids, a safety
valve is used. The mechanics of a safety valve are similar to those of a relief
valve. A safety valve's movement is characterized by a sudden or popping
action. If the flow through a safety valve is less than 25% of the valve's capacity,
the valve will close. This places stringent requirements on the sizing of a safety
valve. If the valve is too large, it will quickly relieve the system of overpressure
and slam shut. If the condition causing the overpressure has not been
corrected, the valve will again pop open, and the cycle will repeat. This
repetitive chattering can damage the valve and/or seat and result in leakage.
Emergencies
The designer must understand how system pressure, flow and composition may
vary with time during an emergency release. The orifice equation for gas or
vapor service includes a density correction factor:
where M^sub v^ is the vapor mass flowrate, Y is the dimensionless vapor
expansion factor, P1 is the vapor density upstream of the orifice, P1 is the
system pressure upstream of the orifice, and P2 is the pressure downstream of
the orifice. As the system depressurizes, the density of the vapor decreases,
thereby contributing to a continual change in the rate of release.

The designer must also understand the various overpressure scenarios that
might occur and be able to identify the limiting release scenarios for the system.
For most processes, this may be one or some combination of the following:
1. system upset resulting from instrument air failure, loss of power, or loss of
cooling
2. circuit blockage, such as line plugging or accidental closure of a valve, that
exposes the system to pump or compressor shutoff conditions
3. fire that adds heat to the system, exposing equipment to extremely high
temperatures.
Scenario I often establishes the limiting relief load for a system. However, all
three scenarios might occur at the same time, or multiple scenarios might occur
at different times, which makes relief system design very difficult. Each release
contingency must be independently evaluated and a systematic comparative
analysis conducted.
Relief calculations begin by defining the normal operating envelope for the
process. The maximum operating pressure must be sufficiently below the
MAWP to avoid occasional disturbances or encroachments on the design limits
of the equipment. Otherwise, the operating envelope will need to be tightened.
Using a blanket or buffer system, which introduces an external medium such as
nitrogen to control system pressure and reduce swings, may be needed to
prevent such excursions. A nominal set pressure between the maximum
operating pressure and the system MAWP must be selected to enable the relief
valve to handle the initial upset. If the setpoint is too low, the relief capacity may
need to be increased, especially in vapor or gas applications. The potential for
premature release may also increase at low set pressures.
To address Scenario 2, a nominal set pressure must be established to protect
the equipment that generates the pressure and the associated piping. A relief
system in liquid service must be able to accommodate the flowrate represented
by the point on the pump curve that corresponds to the set pressure of the relief

valve. Similarly, a relief system in gas or vapor service must be able to

accommodate the corresponding flowrate on the compressor curve. If the circuit
contains a reciprocating compressor, the volumetric flowrate can be determined
from the speed of the machine and the cylinder capacity.
Evaluating Scenario 3 is more difficult, because more variabies must be
considered. Normally, the analysis assumes that a pool fire is located under the
largest vessel in the process circuit. API Recommended Practice 2218 (3)
provides guidance on determining the size and magnitude of a pool fire for a
particular plant layout. The amount of liquid in the vessel, the exposed area on
the bottom plane of the vessel, and the heat flux will determine the excess
vapor generated per unit of time. The presence of passive fire protection (e.g.,
fireproofing) on the vessel will also affect vapor generation. It may be necessary
to carry out several iterations, including analysis of jet fire from a neighboring
circuit, to determine the maximum heat load on the system. Alternatively, API
Standard 520 (4) provides a simplified correlation for estimating fire load.
Although fires are rare, the associated relief load is often higher than that of a
blocked-in circuit (i.e., a piping system where closure of a valve or blockage of
the line suddenly impedes flow throughout the line, causing a dramatic increase
in the static pressure).
Scenario 3 is triggered by a failure in a major process system, such as power,
instrument air, or cooling water. It does not include a pressure pulse generated
by an internal explosion or runaway chemical reaction. A heat and material
balance for the upset condition will indicate the incremental liquid and vapor
load on the system. This defines the relief capacity required to address this
scenario.
These three scenarios will likely have different relief capacity requirements.
Because oversized valves tend to be problematic, a stepped or staggered
strategy with multiple PRVs should be considered. The set of conditions that
produces the smallest load is dealt with first, and the set pressure ofthat valve
must not exceed the system MAWP. A PRV is selected based on the standard
orifice and valve body sizes (Table 1); if a precise match cannot be found, it

may be necessary to select two valves of different sizes with set pressures
staggered by a few psi to avoid chatter.
Staggering set pressures
The next relief situation will require a larger valve with a set pressure a few psi
above that of the previous valve. The orifice areas of all the valves that release
at lower set pressures must be factored into the system design to prevent
oversizing the valve at the higher set pressure. This process is repeated until all
major contingencies have been addressed.
The staggering of set pressures when multiple PRVs are used to protect a
single process is governed by several key principles. At least one PRV must
open at or below the system MAWP. The full relieving capacity of this valve will
occur at 110% of the MAWP. This pressure differential is referred to as
accumulation. Subsequent PRVs may be set at slightly higher pressures, but
the relieving pressure must not exceed 116% of the MAWP. In the case of a
fire, the accumulation may reach 121% of the MAWP. These values are
highlighted in API Standard 520 (4).
Table 2 presents the specifications of a relief system for a chemical
fractionation process that operates at 135 psig and has a MAWP of 160 psig.
The discussion to this point has dealt only with liquid and gas applications.
During an upset, flashing liquids and two-phase mixtures are often released.
As liquid flashes across an orifice, it expands. As the operating pressure drops,
the compositions of the liquid and vapor streams change. This can affect the
capacity of the relief device. Iterative calculations will be needed to determine
the pressure and flow along the entire circuit.
Two-phase flow presents a significant challenge and is beyond the scope of this
article. There are currently no pressure-relief devices with certified capacities for
two-phase flow, since there are no approved test methods. More information on
two-phase flow can be found in API Standard 520, Appendix D (4).

The blowdown header

When material passes through a relief device, it is only partly through its route
to an atmospheric vent or elevated flare stack. It enters a blowdown header i.e., is a piping network that collects releases from multiple relief valves across a
plant site. During a plant upset, several relief valves may discharge
simultaneously. This imposes a backpressure on each valve. If the
backpressure is high enough, it will interfere with the operation of valves with
low set pressures.
To avoid this problem, the designer should size the blowdown header to
accommodate all possible releases and reduce the backpressure on the PRVs.
Pilot-operated valves
Alternatively, balanced-bellows or pilot-operated valves, which are less
sensitive to backpressure, may be selected. In a bellows valve (Figure 2), the
spring is exposed to ambient pressure through a vent on the bonnet. The spring
is isolated from the internal system pressure by a protective bellows.
A pilot-operated safety valve (Figure 3) contains a piston that is controlled by a
servo mechanism, which is enclosed in a dome on top of the main valve
housing. A pilot supply line transmits the system pressure to a pilot valve within
the dome. When the set pressure is reached, the pilot valve opens and releases
the pressure from the dome. The piston is then free to open so the main valve
can release the system fluid. The control pilot opens either to the main exhaust
pipe or to atmosphere. A pilot-operated valve is less sensitive to pressure
fluctuations than a conventional valve. It may be used in situations where the
system operating pressure is approaching the MAWP.
Although these specialty valves offer benefits in certain applications, they are
mechanically more complex and can introduce additional modes of failure. The
mechanical bellows in a balanced-bellows valve can be prone to leakage or
failure. The pilot-operated relief valve's supply line is shock-sensitive and may
plug or stick in some applications, and the rubber O-ring seals in the dome are

temperaturesensitive, so the valve may need to be housed in a protective

shelter. Generally, both of these valves require more maintenance than the
standard relief or safety valves.
Preventing overpressure
Rupture discs are also commonly used to relieve pressure. A rupture disc
(Figure 4) consists of a thin metal membrane secured between two flange faces
that ruptures when a specified pressure is reached. Although they are not as
precisely calibrated as PRVs, they may be more suitable in corrosive or fouling
services. A rupture disc is normally specified with a burst pressure tolerance of
5%. Rupture discs are fast-acting. They do not reseat and must be replaced
after they discharge. They are sometimes installed upstream of relief valves as
protection against corrosive substances.
Blowout panels, or explosion vents, are sometimes used to protect large gasfilled systems from sudden overpressure. They are typically installed with a
welded frangible joint that fails on sudden impact. Because these devices are
activated in an emergency situation and can release a large volume, it is not
practical to connect them to a blowdown header. Instead, these devices release
directly to atmosphere.
Relief devices handling nonhazardous materials are commonly discharged
directly to atmosphere or to grade. Such services may include steam, air, water,
nitrogen and carbon dioxide. If these materials are at higher temperatures or
could affect visibility or air quality on a plant site, they are generally discharged
to an elevated vent pipe.
A complex network of collection drums, piping, and instrumentation is required
to ensure that all pressure releases are discharged to a safe location. A closed
blowdown header is commonly used to transport hazardous releases from the
outlet flanges of all PRVs to a vapor/liquid knockout drum and subsequently to
an elevated flare stack. Blowdown headers in large process plants are
sometimes 3 to 5 ft in diameter.

The design of the blowdown system is a difficult task that requires considerable
judgment. The blowdown header must be sized to accommodate the
simultaneous release from several relief valves during a major plant upset. Any
unforeseen pressure drop will result in a buildup of backpressure, which will
restrict the capacity of the relief valves upstream. Detailed pressure-drop
calculations are required along the entire blowdown header, including at all
points of entry. A slight positive pressure should be maintained to prevent air
ingress through any system leaks, since this could create an explosive
environment. The design of the blowdown header must address potential solids
(sludge), liquid slugs, water condensate, freezing, incompatible fluids, corrosion,
and heat loss. The blowdown header should slope 1 deg. toward a common
liquid knockout drum. If there are concerns about freezing or liquid setup (i.e., a
phenomemon where fluid viscosity reaches a point where flow ceases, which
typically occurs with viscous or heavy hydrocarbons at low temperature), the
main blowdown header and branch connections should be insulated and/or heat
traced.
A main blowdown drum separates liquid from vapor. The liquid stream is cooled
and circulated back to slop tankage, where it can subsequently be reprocessed.
A water seal drum is commonly employed to ensure that air cannot be drawn
back into the system when the header cools down. This drum must also be
sized based on the maximum relief load as well as the total volume of the
blowdown system.
Gas exiting the seal drum enters the base of an elevated flare stack. A flame
arrester is commonly installed at this point to prevent flame from receding down
the stack during low flow conditions. The stack is sized to pass the maximum
gas load for a major plant upset. A gas burner installed at the upper tip of the
flare provides effective fuel-air mixing. The tip is equipped with an electric igniter
and a pilot gas line to protect against flame-out and subsequent explosion. Lowpressure steam is commonly introduced to a flare stack to reduce smoke. The
height of the flare and its location relative to other process equipment and
roadways is based on the anticipated radiant heat flux. The maximum heat flux
at adjacent structures must not exceed 6.3 kW/m^sup 2^.

The concepts and principles discussed so far are well established, with
additional detailed information available in API Standards 520 (4) and 521 (S).
Even if a process does not handle oil or hydrocarbons, these practices are
standard throughout most of the chemical process industries (CPI).
One of the most difficult challenges for relief systems is dealing with runaway
chemical reactions. The Design Institute for Emergency Relief Systems
(DIERS) is a consortium of companies that develops, shares and maintains
technology in this area. Information on the DIERS Users Group can be obtained
at www.aiche.org/technicalsocieties/DIERS/index.asp.
Pressure-relief valves are mechanical devices with moving parts, so they are
not suitable for situations that that could experience a sudden pressure impulse.
Other protective measures, such as a rupture disc or a diverter drum (which is
similar to a pulsation bottle), should be considered for such installations.
Pressure relief is a recognized safeguarding strategy in the CPI. When
designed and installed properly, it may be considered an independent layer of
protection. It is not a passive strategy - it requires considerable care and
attention once installed.
Testing
Unless PRVs are periodically calibrated and benchtested, they may not provide
the necessary protection when it is required. Many jurisdictions require relief
valves and safety valves to undergo annual bench testing at a certified facility.
Sometimes, these tests are not possible because a valve has extended run
cycles. Therefore, provisions must be made to remove such valves during
normal operation. This is typically accomplished by blocking in the valve and
replacing it with a similarly calibrated backup relief device. A bypass circuit must
be routed to the blowdown header to address an emergency situation that might
occur during the swingover. Block valves installed upstream and downstream of
PRVs for isolation purposes must be chain-locked or car-sealed open when the
PRV is in service. (A car-seal uses a tagged aluminum strap to prevent a valve

handwheel from being turned.) Under controlled conditions, the chain may be
unlocked or the car-seal broken. Either contingency is required to ensure high
availability for a critical safety function. Inspection and maintenance work must
be carefully scheduled so that plant safety is not compromised.
Closing thoughts
Relief valve maintenance must be near the top of a facility's mechanical integrity
program. The condition of each valve at the time of its inspection must be
recorded in the equipment database. Improved metallurgy should be considered
when replacing any components showing signs of accelerated wear or
corrosion.
Although pressure relief does not contribute directly to corporate profits or
product quality, system failures may lead to serious accidents and jeopardize an
entire plant operation. Engineers must resist the temptation to treat the design
of pressure-relief systems as a computational exercise. This discipline should
be assigned to the most experienced and talented process engineers. The
simplified discussion presented here is not a guide to PRV design, but rather is
intended to help engineers entering the workforce understand its basic
fundamentals.
[Sidebar]
Nomenclature
A = orifice area, ft^sup 2^
C^sub v^ = orifice discharge coefficient, dimensionless, 0.62 for a sharp-edge
orifice
G = gravitational acceleration, 32 ft/s^sup 2^
H = pressure differential across valve orifice, ft of liquid
M^sub V^ = vapor mass flowrate, Ib/s

P^sub 1^= system pressure upstream of orifice, lb/ft^sup 2^

P^sub 2^ = pressure downstream of orifice, lb/ft^sup 2^
Q^sub L^ = liquid flowrate, ft^sup 3^/s
Y = vapor expansion factor, dimensionless
P^sub 1^ = vapor density upstream of orifice, lb/ft^sup 3^
[Reference]
LITERATURE CITED
1. American Society of Mechanical Engineers, "ASME Boiler and Pressure
Vessel Code, Section Vili," ASME, New York, NY (2007).
2. American Society of Mechanical Engineers, "ASME B3 1 .3 - Process Piping
Code," ASME, New York, NY (2002).
3. American Petroleum Institute, "API Recommended Practice 22 1 8 Fireproofing Practices in Petroleum and Petrochemical Processing Plants," API,
Washington, DC (1999).
4. American Petroleum Institute, "API Standard 520 - Sizing, Selection, and
Installation of Pressure-Relieving Devices in Refineries." API, Washington, DC
(2000).
5. American Petroleum Institute, "API Standard 52 1 - Pressure-Relieving and
Depressuring Systems," API, Washington, DC (2007).
FURTHER READING
1. Center for Chemical Process Safety (CCPS), "Guidelines for Engineering
Design for Process Safety," Chapter 14 - Pressure Relief Systems, AIChE, New
York, NY ( 1 993).
[Author Affiliation]

BRIAN D. KELLY, P.ENG.

BRIRISK CONSULTING LTD.
[Author Affiliation]
BRIAN KELLY, P.Eng., is a process safety consultant in Calgary, Alberta
(Address: 121 Royal Bay NW. Calgary AB T3G 5J6). He has conducted
numerous process safety audits, incident investigations, and process safety
workshops. He has 35 years of engineering and operations experience with
Imperial Oil and Syncrude Canada Ltd. He is also a staff consultant with
AlChE's Center for Chemical Process Safety. He holds BASc and MASc
degrees in chemical engineering from the Univ. of Ottawa. He is a registered
professional engineer in the province of Alberta.
Kelly, Brian D
Copyright American Institute of Chemical Engineers Feb 2009
http://www.highbeam.com/doc/1P3-2143893111.html

HighBeam Research is operated by Cengage Learning. Copyright 2016. All

rights reserved.
www.highbeam.com