Advanced Design for Safety

CME3034
Tutorial
Questions
Solutions
John Dalton
Room C527

Engineering Design Centre

CEAM, Merz Court
www.edc.ncl.ac.uk
John.Dalton@ncl.ac.uk
0191 222 8556

CME3034: Advanced Design for Safety

Assessment
● Assignment on:
Thermal runaway 40% of the marks
● Exam:
90 minute exam 60% of the marks
- Exam is 3 from 5 questions
so that’s approx. 30 minute per question

1
CME3034: Exam Scope:
● (JD) Pressure Vessel Standards
● (JD) Pressure Relief Sizing
● (JG) Bio. Safety
● (MH) Maintenance and Safe Systems
● (MH) Management of Change
● (SB) Plant layout
● (SB) Instrumentation and Protective Systems

Scope Does NOT include

Lecture 1: Introduction (JD), and
the Thermal runaway lectures (with Prof. Wright)

You can of course contact me

at any time with questions:

I am in Room C527, or

email: John.Dalton@ncl.ac.uk

2
Tutorial:
Q&A

Q 1

3
Question 1) Target Safety Integrity Level (SIL) Assessment

a) In terms of Safety Instrumented Functions, what is meant by revealed

and unrevealed failures?

b) A key outcome of SIL assessment of Safety Instrumented Functions is

identification of the correct functional test interval. Describe why is
this important in relation to Process Safety?

c) What are the disadvantages of functionally testing with:

- a test interval that is too long
- a test interval that is too short

d) There are several methods for carrying out Target SIL Assessment:
- Describe the relative benefits and limitations of each method
- Describe under what circumstances each method would typically be
used

Q1) Answer

a) “In terms of Safety Instrumented Functions, what is meant by

revealed and unrevealed failures?”

a) Revealed and unrevealed failures

Revealed (Spurious) failures due to component failure – the function

operates and ‘trips’ the plant to a safe condition unnecessarily

The failure is self revealing and ‘fail safe’

You may incur a business loss due to an upset in production due the
process safety hazard is not present

Unrevealed (Dangerous) failures where the component failure is not

revealed and there is nothing to indicate to the operator that the
function will not operate correctly when a demand is placed on it

Should a demand occur the function would fail and this may lead to
safety or environmental consequences

4
Q1) Answer

b) “A key outcome of SIL assessment of Safety Instrumented Functions

is identification of the correct functional test interval. Describe why
is this important in relation to Process Safety?”

b) Routine functional testing

The only way you can identify an unrevealed or dangerous failure is by

testing the function (but under safe conditions)

If you do not functionally test then you will miss an unrevealed failure
and then you could have a process safety incident

i.e. a fire, explosion, release of toxic gas,

when a subsequent real demand is placed on a function that then fails

to operate as per design

Q1) Answer

c) “What are the disadvantages of functionally testing with:

- a test interval that is too long
- a test interval that is too short”

c) Test interval

Too Long:
If the test interval is too long then a real demand on the function
may occur in between functional tests.

For example,
if your demand rate is 1 per year, and your
test interval is every 2 years,

… meaning demands are more frequent that your tests. You are at
increased risk of a process safety incident.

5
Q1) Answer

c) “What are the disadvantages of functionally testing with:

- a test interval that is too long
- a test interval that is too short”

c) Test interval

Too Short:
A shorter test interval often means improved reliability of a function (a
lower PFDavg), but this is not always the case if the test interval
becomes too short. If the test interval is too short then you may not
get the reliability you require out of your function due to:
● More tests mean more chance of human error during a test as the
technician makes a mistake and leaves the function in a failed state
● To do a test you may need to defeat the function (this is to prevent
the function operating fully and tripping out the plant under test
conditions which is undesirable). Testing too often may mean the
time the function is in a failed (defeated) state during testing
becomes significant
● Testing too often also means you are wasting money in terms of
technician time and effort

Q1) Answer

d) “There are several methods for carrying out Target SIL Assessment:
- Describe the relative benefits and limitations of each method
- Describe under what circumstances each method would typically be
used”

d) Target SIL methods

i) Risk graph

Advantage – Quick and easy to use (10 minutes per assessment)

Disadvantage – Difficult to include common cause failures

Disadvantage – Difficult to include human error
Disadvantage – Difficult to include multiple initiating events
Disadvantage – Tends to be conservative, i.e. gives high SIL ratings
Disadvantage – Does not calculate a PFDavg

Use: – Screening assessment of many functions

6
Q1) Answer

d) Target SIL methods

ii) Layer of Protection Analysis (LOPA)

Advantage – Quick and easy to use (1 hour per assessment)

Advantage – Quantitative – calculates a PFDavg
Advantage – Easy inclusion of multiple initiating causes

Disadvantage – Difficult to include common cause failures

Use: – More detailed assessment of high SIL rating functions, …

For example, if a Risk Graph predicts a SIL2 system, then a LOPA

assessment is less conservative and would generally predict SIL1

It is more cost effective to go to the trouble of repeating the SIL

assessment using LOPA than spending the money designing a SIL2
system based on the results of the Risk Graph

Q1) Answer

d) Target SIL methods

iii) Fault tree analysis

Advantage – Fully quantitative

Advantage – Accurate, not overly conservative
Advantage – Allows for full inclusion of human error, common
causes of failure

Disadvantage – Takes a long time,

e.g. 1 week per function by an expert

7
Q1) Answer

d) Target SIL methods

iii) Fault tree analysis

Use: – Only when other simpler methods such as LOPA are

predicting SIL2 or above

The cost of installing hardware for a SIL2 or above system is very

high. It is generally worth the time and effort of a Fault Tree analysis
which takes out some of the conservatism of a LOPA assessment

The Fault Tree may then predict a lower Target SIL.

So only use Fault Tree analysis:
● In exceptional cases where other methods are
predicting SIL2 or above
● Where the consequences of an event are very high,
e.g. multiple on-site fatalities or off-site fatalities,
i.e. need a more robust method in front of the Competent Authority

Q 2

8
Question 2) Target Safety Integrity Level (SIL) Assessment

An atmospheric storage tank is used to store gasoline. The gasoline is

delivered via pipeline. A Safety Instrumented Function is installed on the
tank that automatically closes a valve in the delivery pipeline in the event of
high liquid level in the storage tank. The function consists of a single level
instrument that acts on a single trip valve via an electronic switch.

a) Describe what is meant by ‘unrevealed’ and ‘revealed’ failures in terms of

the high tank level Safety Instrumented Function.

b) The Target SIL Assessment for the tank high level Safety Instrumented
Function has identified a required SIL1 target with a PFDavg of 0.025.

Using the data supplied:

i) Explain the difference between total failure rate and dangerous failure
rate
ii) Calculate the required functional test interval to achieve the target
PFDavg
iii) If the functional test lasts 1 hour, during which time the function is
deliberately placed in a defeated state to avoid a spurious trip that
causes an interruption to the transfer of gasoline, comment on its
significance

Q2) Target Safety Integrity Level (SIL) Assessment

Total failure rates for items:

Level instrument 0.3 per year

Electronic switch 0.2 per year
Trip valve 0.1 per year

c) How could the reliability of the high level Safety Instrumented Function
be improved?

d) The Hardware Fault Tolerance of the high level Safety Instrumented

Function is zero.

What do you understand by this and how would you change the design of
the function to achieve a Hardware Fault Tolerance of 1?

9
Q2) Answer

a) “Describe what is meant by ‘unrevealed’ and ‘revealed’ failures in

terms of the high tank level Safety Instrumented Function”

a) Revealed and unrevealed failures

Revealed (Spurious) failures due to component failure

● The function operates and ‘trips’ the plant to a safe condition
unnecessarily
● The failure is self revealing and ‘fail safe’
● In this case the trip valve in the feed pipeline to the tank would
close even if the liquid level is not at a high level in the tank
● This is a safe condition but would cause a business interruption
● The process hazard is avoided as the tank cannot overflow

Unrevealed (Dangerous) failures where the component failure is not

revealed and there is nothing to indicate to the operator that the
function will not operate correctly when a demand is placed on it

Should high level occur in the tank the function would not operate,
therefore if the operator does not detect the failure of the function
(which is unlikely) and intervene manually then the tank would overflow

Q2) Answer

b) i) “Explain the difference between total failure rate and dangerous

failure rate”

b) PFDavg calculation

i) Dangerous failure rate is typically 1/3 (to 1/2)

of the total failure rate,

i.e. 1/3 of all failures are dangerous (and unrevealed) and

2/3 are safe (and revealed)

For Safety Instrumented Functions it is only dangerous failures that

are important because this is where the function has failed and
provides no protection.

Dangerous failure rate = (0.3 + 0.2 + 0.1) / 3 = 0.2 /year

10
Q2) Answer

b) ii) “Calculate the required functional test interval to achieve the

target PFDavg”

b) PFDavg calculation

ii) PFDavg = 1/2 x dangerous failure rate x test interval (T/years)

PFDavg = 1/2 x 0.2 x (T/12) = 0.025

(T/12) = 0.25

so, T = 3 months

Overall this function achieves SIL 1 (PFDavg between 0.1 and 0.01)

Q2) Answer

b) iii) “If the functional test lasts 1 hour, during which time the
function is deliberately placed in a defeated state to avoid a
spurious trip that causes an interruption to the transfer of
gasoline, comment on its significance”

b) PFDavg calculation

iii) The PFDavg should include the PFDavg for the instrumentation
hardware {as calculated in b) ii)} and also the PFDavg for the time
when the function is defeated during testing

But this is 1 hour every 1 months, i.e. a probability of 0.0005,

which is not that significant compared with

the 0.025 hardware PFDavg

11
Q2) Answer

c) “How could the reliability of the high level Safety Instrumented

Function be improved?”

c) Reliability improvement

● Functionally test more often,

e.g. 1 test per month would give PFDavg becomes 0.008
If you test more often than this then the time during which the
function is defeated during testing becomes significant
The test defeat PFDavg is 0.001 for a 3 month test interval, so it is
starting to become significant, therefore testing even more often
than 1 per month would not be recommended
● Use more reliable instrumentation hardware,
i.e. a smaller dangerous failure rate

● Use multiple valves or level instruments in a voting system

● Use self-diagnostics on the instrumentation so they self-detect and

warn of a fault

Q2) Answer

d) “The Hardware Fault Tolerance of the high level Safety Instrumented

Function is zero. What do you understand by this and how would you change
the design of the function to achieve a Hardware Fault Tolerance of 1?”

d) Hardware Fault Tolerance, HFT

This is how many faults a function can have and still operate correctly

The HFT of zero means one failure (either the valve, switch or level
instrument) and the function would fail to operate on demand, this is
typical of a SIL1 function

A HFT of 1 would mean that a single failure could occur and the
function would still operate (typical of a SIL2 function)

To do this you would need 2 sets of valve, switch, level instrument in

parallel each acting as a 1 out of 2 system

So that if one switch failed then there is still another switch left to
activate the function

12
 Q 3

Question 3) Plant Layout

A chemical plant complex is to be located on a green-field site. The site is
square and has a plot plan area of 1 km2. A short distance from the eastern
boundary of the site is a small town. All other sides of the site face onto
open farmland.

The complex is made up of 5 process units that have been qualitatively risk
ranked as follows based on material properties and inventory:

Unit 1 Feedstock storage unit (Medium hazard – environmental toxin)

Unit 2 Reaction unit (Very high hazard - flammable)

Unit 3 Side-stream unit (Medium hazard – toxic)

Unit 4 Product storage tank farm (Low hazard)

Unit 5 Services (power, steam) (Low hazard)

13
Q3) Plant Layout
A chemical plant complex is to be located on a green-field site. The site is
square and has an plot plan area of 1 km2. A short distance from the
eastern boundary of the site is a small town. All other sides of the site
face onto open farmland.

a) In terms of plant layout, what would be the benefit of conducting a

hazard study?

b) Using the principles of inherent safety describe how you would arrange
the process units on the site?

c) Describe how you could reduce the domino hazard due to vapour cloud
explosion events on Unit 2.

d) What should you consider when locating the Reaction Unit Control Room?

e) What should you consider when locating the office building for the Site
Sales and Administration Team?

Q3) Answer

a) “In terms of plant layout, what would be the benefit of

conducting a hazard study?”

a) The hazard study would identify:

● What type of process hazards are present,

i.e. fire, explosion, toxic gas

● The location of the process hazard relative to other units

● The hazard study would also identify measures that would

reduce the risk of a loss of containment, fire, explosion by
recommendation of additional controls where necessary

14
Q3) Answer

b) “Using the principles of inherent safety describe how you

would arrange the process units on the site?”
b)
● Do not locate Unit 2 (flammable) next to Unit 3 (toxic).
- must avoid an explosion having a domino effect and causing a leak
of toxic material as this may hinder the emergency response

● Install a low hazard unit between Unit 2 and Unit 3 to avoid the
domino effect

● Do not locate Unit 2 or Unit 1 near the off-site residential area.

Minimise off-site risk to public.

● Locate low hazard units near off-site residential area

● Do not locate Unit 2 (flammable) next to the Services (Unit 5).

- must avoid an explosion having a domino effect that damages the
services, e.g. causing a power failure co-incident with a major
accident on Unit 2

● Reduce the inventory of Unit 1, e.g. supply by pipeline

Q3) Answer

c) “Describe how you could reduce the domino hazard due to

vapour cloud explosion events on Unit 2”

c)
● Ensure plant structure is open, well-ventilated and not congested
– this allows leaks to disperse and should an explosion occur then
the lack of congestion would means the force of the explosion
would be minimised

● Separation distance from Unit 2 to other units

– larger separation reduces the potential for the blast wave and
missiles to damage adjacent units

● Gas detection and remote isolation

– detect a leak of flammable vapour then automatically isolate the
unit and inventory

• Blast wall around high hazard equipment on Unit 2

– limit potential for missiles

15
Q3) Answer

d) “What should you consider when locating the Reaction Unit

Control Room? ”

d)

● Do not locate close to Unit 2,

i.e. locate outside of hazard range

● If cannot locate outside of hazard range may require a

strengthened building

● Because of Unit 3 the control room should be a toxic gas

safe haven so that it can be occupied in a toxic emergency

Q3) Answer

e) “What should you consider when locating the office building

for the Sales and Administration Team? ”

e)

● Locate outside of hazard range of all hazards where

practicable

● All personnel in this building are non-essential and could be

located off-site as they are not needed for the day-to-
day running of the plant

16
 Q 4

Question 4) Maintenance & Safe Systems

You are designing an export pump station for use on a high pressure liquid
propylene supply grid within a petrochemicals complex. The design calls
for the provision of a typical ‘running pump plus spare’ arrangement, so
that each pump can be taken off line for maintenance whilst its partner
remains in operation to ensure continuity of supply

To
Grid
From
Supply

Develop a more detailed version of the above diagram to indicate the

basic layout of ‘isolations for maintenance’ for the pump station and
describe how, when maintenance is required, each pump can be:
● safely isolated in all necessary respects
● de-inventoried and prepared for safe maintenance
● prepared for service and de-isolated following maintenance
● re-inventoried for return to service

17
Q4) Answer
“Develop a more detailed version of the above diagram to indicate the
basic layout of ‘isolations for maintenance’ for the pump station …”

Double block and bleed installations around each pump, combined with
slip-plate positions allow each pump to be safely isolated.

To
Grid
From
Supply

Slip Plates

Spectacle
Pieces

18
Q4) Answer
“ … and describe how, when maintenance is required, each pump can be:
● safely isolated in all necessary respects
● de-inventoried and prepared for safe maintenance
● prepared for service and de-isolated following maintenance
● re-inventoried for return to service

When a pump is taken out for required maintenance, the sequence is:

● Isolate via ‘outer’ isolation valves

● Connect one of the two ‘bleed’ valves and the pumps drain to flare
disposal and the other ‘bleed’ to a pressurised nitrogen supply
● Displace the propylene inventory to flare, maintaining pressure to
avoid pipework chilling from adiabatic flash of propylene
● Once free of bulk propylene, purge pump to flare until hydrocarbon
free
● Isolate electrical supply to pump
● De-pressured nitrogen and fit slip-plates to either side of pump
installation
– pump can then be removed for maintenance

Q4) Answer
“ … and describe how, when maintenance is required, each pump can be:
● safely isolated in all necessary respects
● de-inventoried and prepared for safe maintenance
● prepared for service and de-isolated following maintenance
● re-inventoried for return to service

When returning a pump from maintenance, the sequence is:

● Remove slip-plates
● Connect pressurised nitrogen to one of two bleed valves and purge
until pump and pipework is proven oxygen free
● Pressurise pump and pipework and then disconnect nitrogen
● Introduce propylene via gradual opening of one of the two isolation
valves
● Maintaining pressure, purge nitrogen from pump and pipework to flare
● Remove connections to flare and open pump isolations
● Reinstate electrical supply to pump

19
 Q 5

Question 5) Management of Change

In response to production losses resulting from repeated failure of a
reactor feed pump, it is proposed to install a second feed pump in
parallel with the existing unit.

The new installation will be fed from existing power supplies and will
include an auto-start / auto change-over facility, added to the
existing process control computer system, to ensure continuous
operation in the event of a single pump shutdown.

As a result of the expected improvement in production reliability, a

significant reduction in the amount of operator attention during normal
operation is anticipated, and it is therefore also proposed to reduce
the level of staffing on the production unit.

Describe four factors you would include in the EHS assessment within
the management of change process for the proposed modifications to
the process unit and in each case why the factor needs to be
considered?

20
Q5) Answer

● Introduction of new pump

– same design, operability and ability to deliver duty?

● Introduction of new pump and isolation/changeover valves

– additional leak and failure points?

● Existing electrical supply able to support new duty including changeover

valves?

● Risk of both pumps online together owing to failure of changeover –

impact on downstream process and potential overpressure risks?

● Changes/addition to process control computer system

– introduction of error/additional failure mode(s)?

● Impact of reduced manpower on operational and emergency capability –

enough resource to handle potential scenarios?

Q 6

21
 Question 6) PRV Sizing
Determine the flow area of a safety valve used on a vessel containing
nitrogen gas, where a set pressure of 55 barg is required. Coefficient
of discharge Kd = 0.87.

Isentropic coefficient of the gas [k]: 1.40

Gas temperature at relieving pressure: 20 °C (= 293 K)
Required gas flow rate: 18000 kg/h
Set pressure: 55 barg
Back pressure: Atmospheric

Q 6) Answer

First determine the maximum relieving pressure Po

max relieving pressure
= (1.1  55) barg
= 60.5 barg

Pressure
(barg) relieving
10% pressure

set pressure
= 55 barg

time (s)
 Po = (1.1  55) + 1 = 61.5 bara

22
 Q 6) Answer

Next, check if flow is critical or not

Po = 61.5 bara  k 
 2   k 1 
 
Pb = 1 bara
Pc  Po   
k  1
 1.4 
2
 
   1.4 1 
 61.5   
 1.4  1 
 32.5
Pb  1  32.5

Since backpressure (Pb =1 bara) is much less than

the critical pressure (Pc = 32.5 bara)
Pb < Pc,  flow is critical

PRV Criticality

Check for criticality

PR = 61.5 bar PR=61.5 bara Pb=1 bara
Pb = 1 bar
k = 1.40
 k 
PC  2   k 1 
 

  
PR k  1 P
 1.4 
2
 
   1.4  1  PR
PC  61.5   
 1.4  1 
 32.5 bara
Pb>PC

Pb=PC PC=32.5 bara

Pb < PC  Critical Pb<PC Pb=1 bara

distance

23
Q 6) Answer
m
So using the critical formula: A  1
 PO 2
0.2883 C Kd  
vg 
 
1
 k 1  2
 
2   k 1  

 
where: C  3.948 k   
  k  1  

for k = 1.40, C = 2.703

Q 6) Answer

Also determine the specific volume vg

Po = 61.5 bara RT
R = 287 J/kg K vg 
T = 293 K P
287  293

61.5  10 5
 0.01367 m 3 kg 

24
Q 6) Answer

Finally:
m
A  1
P 2
0.2883 C Kd  o 

v g 
18000
A  1
 61.5  2
0.2883  2.703  0.87   
 0.01387 
 395 .8 mm 2

Therefore a diameter do of 22.45 mm,

or next size up, i.e. 23 mm

Q 7

25
 Question 7) PRV Sizing
Repeat the previous example, but this time with a back pressure
of: 36 barg.

reminder of Q6)

Determine the flow area of a safety valve used on a vessel containing

nitrogen gas, where a set pressure of 55 barg is required. Coefficient
of discharge Kd = 0.87.

Isentropic coefficient of the gas [k]: 1.40

Gas temperature at relieving pressure: 20 °C (= 293 K)
Required gas flow rate: 18000 kg/h
Set pressure: 55 barg
Back pressure: 36 barg

Q 7) Answer

First determine the maximum relieving pressure Po

max relieving pressure
= (1.1  55) barg
= 60.5 barg

Pressure
(barg) relieving
10% pressure

set pressure
= 55 barg

time (s)
 Po = (1.1  55) + 1 = 61.5 bara

26
 Q 7) Answer

Next, check if flow is critical or not

Po = 61.5 bara  k 
 2   k 1 
 
Pb = 36+1 bara
Pc  Po   
k  1
 1.4 
2
 
   1.4 1 
 61.5   
 1.4  1 
 32.5
Pb  37  32.5

Since backpressure (Pb =37 bara) is greater than

the critical pressure (Pc = 32.5 bara)
Pb > Pc,  flow is sub-critical

PRV Criticality

Check for criticality

PR = 61.5 bar PR=61.5 bara Pb=37 bara
Pb = 37 bar
k = 1.40
 k 
PC  2   k 1 
 

  
PR k  1 P
 1.4 
2
 
   1.4  1  PR
PC  61.5   
 1.4  1 
Pb=37 bara
 32.5 bara
Pb>PC

Pb=PC PC=32.5 bara

Pb > PC Sub-Critical Pb<PC

distance

27
Q 7) Answer
m
So using the sub-critical formula: A  1
P 2
0.2883 C Kd Kb  O 

vg 
1
 k 1  2
 
2   k 1  

 
where: C  3.948 k   
  k  1  

for k = 1.40, C = 2.703

Q 7) Answer

For kb:

1
 2k  P  2 k  P  k  1  k   2
  b    b  
 k  1  o 
P P
 o  
Kb   k  1  k  1  
  2  
k 
 k1 
 

Can either use the formula, or …

28
Q 7) Answer

Finding Kb, the back pressure correction:

Pb/Po = 37/61.5 = 0.602

From the table, we see for k=1.40 & Pb/Po=0.602

note:
Kb = 0.989

Q 7) Answer

Also determine the specific volume vg

Po = 61.5 bara RT
R = 287 J/kg K vg 
T = 293 K P
287  293

61.5  10 5
 0.01367 m 3 kg 

29
Q 7) Answer

Finally:
m
A  1
P 2
0.2883 C Kd Kb  o 

v g 
18000
A  1
 61.5  2
0.2883  2.703  0.87  0.988   
 0.01387 
 400 .6 mm 2

Therefore a diameter do of 22.6 mm,

or next size up, i.e. 23 mm

So not that much different from the other example !

Q 8

30
 Question 8) PRV
A safety valve is to be sized for the superheater outlet of a large power
station boiler. The valve will be expected to handle 27000 kg/hour of
steam when it is relieving pressure, which will discharge to the
atmosphere.

The normal operating pressure at the superheater outlet is 57 barg and

the temperature is 450 °C.

Based on this information, make an estimate of the relieving pressure and

thus determine the size of a suitable safety valve for this duty. Assume
the isentropic coefficient for superheated steam is 1.30 and the discharge
coefficient for a typical valve will be 0.71.

Q 8) Answer

First determine the maximum relieving pressure Po

max relieving pressure
= (1.1  62.7) barg
= 68.97 barg

Pressure
(barg) relieving
10% pressure

set pressure
= 1.1  57 10%
= 62.7 barg
Pop
57 barg Op Pressure
time (s)
 Po = (68.97) + 1 = 69.97 bara

31
 Q 8) Answer

Next, check if flow is critical or not

Po = 69.97 bara  k 
 2   k 1 
 
Pb = 1 bara
Pc  Po   
k  1
 1.3 
2
 
   1.3 1 
 69.97   
 1.3  1 
 38.2
Pb  1  38.2

Since backpressure (Pb =1 bara) is much less than

the critical pressure (Pc = 38.2 bara)
Pb < Pc,  flow is critical

PRV Criticality

Check for criticality

PR = 69.97 bar PR≈70 bara Pb=1 bara
Pb = 1 bar
k = 1.30
 k 
PC  2   k 1 
 

  
PR k  1 P
 1.3 
2
 
   1.3  1  PR
PC  69.97   
 1.3  1 
 38.2 bara
Pb>PC

Pb=PC PC=38.2 bara

Pb < PC  Critical Pb<PC Pb=1 bara

distance

32
Q 8) Answer
m
So using the critical formula: A  1
 PO 2
0.2883 C Kd  
vg 
 
1
 k 1  2
 
2   k 1  

 
where: C  3.948 k   
  k  1  

for k = 1.30, C = 2.634

Q 8) Answer

Also determine the specific volume vg

Po = 69.97 bara
T = 450 °C = 743 K

from Steam Tables:

v g  0.0441 m 3
kg 

33
Q 8) Answer

Finally:
m
A  1
P 2
0.2883 C Kd  o 

v g 
27000
A  1
 69.97  2
0.2883  2.634  0.71   
 0.0441 
 1257 mm 2

Therefore a diameter do of ≈ 40 mm,

Q 9

34
Question 9) PRV
A pressure vessel with a volume of 3.99 m3 contains 73.5 kg of a gas
at 130 °C. A pressure relief valve has been sized for this vessel with a
throughput of 3700 kg/hour.

A HAZOP has identified a risk that if an outlet valve is accidentally

closed, 49 kg of gas could be added to the vessel over a time period of
38 seconds.

Assuming this accumulation will be linear, determine if the pressure

relief valve will be fit for purpose.

Q 9) ans

First let’s look at what could happen … max press

without a prv

122.5 kg

+ add
49 kg
Po

Pset

Pnorm
73.5 kg
38 seconds

Volume = 3.99 m3, Temp = 130 °C = 403 K

35
Q 9) ans

First let’s look at what could happen … max press

without a prv

122.5 kg

max press
when prv
operates
Po

Pset
prv opens
here
Pnorm
73.5 kg
38 seconds

Q 9) ans

Work out some corresponding pressures:

mRT
Pnorm 
V
73.5  287  403

3.99
 21.3  10 5 Pa
 21.3bara
mass corresponding
Pset  1.1  21.3
to 23.43 bara is
 23.43 bara
80.83 kg
PO  1.1  23.43
 25.77 bara

36
Q 9) ans

Putting some values in gives:

122.5 kg

tinc=?

PO = 25.77 bara

Pset = 23.43 bara

80.83 kg
prv opens
here
Pnorm = 21.3 bara
73.5 kg
38 seconds 122.5  80.83
tinc  38 
122.5  73.5 
 32.32 s

Q 9) ans

Consider the rate of accumulation and

the rate of discharge through the PRV
(i.e. above set pressure)

During accumulation
we add:
49 kg
 1.29 kg s  32.32 s  41.69 kg
38 s
During PRV operation
we discharge:

3700 kg hr
 1.028 kg s  32.32 s  33.23 kg
3600 s

So the net addition:

41.69 kg  33.23 kg  8.47 kg

37
Q 9) ans

Viewing this:

122.5 kg

PRV disch’s
33.23 kg we add
41.69 kg

What’s left
8.47 kg
80.83 kg
32.23 seconds
73.5 kg
38 seconds

Q 9) ans

The question is:

Does this correspond to the MAWP, (or the maximum relieving pressure) ???

PV  mRT
mRT
P 
V
80.83  8.47   287  403

3.99
 25.88  10 Pa,
5
or 25.88 bara

Since 25.88 bara ≈ 25.77 bara (PO)

Then we may assume the PRV is
correctly sized and Fit-for-Purpose
However it is borderline and we might want
to increase the flow-rate a bit.

38
 Q 10

Question 10) PSV and PV Sizing

During the HAZOP stage of a plant design it is identified that a pressure
vessel, used for reactions, may be subject to thermal runaway if a
particular fault occurs. The data for the exothermic reaction as a
characteristic between temperature and pressure is shown in the table.

Using the data supplied redesign the vessel to determine a nominal wall
thickness and size a safety valve suitable for releasing the excess pressure.
The safety relief system operates under back pressure.
Assume:
i) a Kd value of 0.71 for the safety valve, and
ii) an estimated life of 10 years for the pressure vessel, and
iii) a mass flow rate, ṁ, for the safety valve of 3500 kg/h

Use the BS EN 13445 standard for

pressure vessels and EN ISO 4126
standard for safety valves.

data for the exothermic reaction:

39
 Pdes = 20 bar
Pnorm = 18 bar
Pb = 8.5 bar
De = 1.1 m
1.1 m
L = 4.2 m
Corrrate = 0.08 mm/year
Life span = 10 years
Mat. tol. = ± 0.2 mm
RT/M = 530 MPa
D 2
V L Kd = 0.71
4 Solvent = 1200 kg, 37.5 kMol
1.12 Reactant = 675 kg 5.11 kMol
  4.2
4 ΔH = -13.2104 kJ/kMol
mass rate = 3500 kg/h
 4 m3
Tmix = 40 °C
Cp = 4 kJ/kg K

Reaction 1200 kg Solvent 37.5 kMol

675 kg Reactant 5.11 kMol
Cp = 4 kJ/kg K

40
Reaction 1200 kg Solvent 37.5 kMol
675 kg Reactant 5.11 kMol
Cp = 4 kJ/kg K

H  13.2  10 4 kJ / kMol of reactant

Q  moles  H
 5.11  13.2  10 4
 675  10 3 kJ

Reaction 1200 kg Solvent 37.5 kMol

675 kg Reactant 5.11 kMol
Cp = 4 kJ/kg K

H  13.2  10 4 kJ / kMol of reactant

Q  moles  H
 5.11  13.2  10 4
 675  10 3 kJ

Q  m  C p  T
Q
T 
m  Cp
675  10 3

1200  675   4
 90  C

41
Reaction 1200 kg Solvent 37.5 kMol
675 kg Reactant 5.11 kMol
Cp = 4 kJ/kg K

H  13.2  10 4 kJ / kMol of reactant

Q  moles  H
 5.11  13.2  10 4
 675  10 3 kJ

Q  m  C p  T
Q 22 bar
T 
m  Cp
675  10 3

1200  675   4
40 °C 130 °C
 90  C
ΔT = 90 °C

PRV Criticality

Check for criticality

PR = 22 bar
Pb = 8.5 bar
k = 1.56

42
PRV Criticality

Check for criticality

PR = 22 bar
Pb = 8.5 bar
k = 1.56

 k 
PC  2   k 1 
 
PR  k  1 
 1.56 
 
 2   1.561 
PC  22   
 1.56  1 
 11 bar

PRV Criticality

Check for criticality

PR = 22 bar PR=22 bar Pb=8.5 bar
Pb = 8.5 bar
k = 1.56

 k 
PC  2   k 1 
  P
PR  k  1 
PR
 1.56 
 
 2   1.56 1 
PC  22    Pb>PC
 1.56  1 
 11 bar
Pb=PC PC=11 bar

Pb < PC  Critical Pb<PC Pb=8.5 bar

distance

43
PRV Sizing
Po = 22 bar
k = 1.56
Kd = 0.71
T = 130 °C
m = 3500 kg/h

PRV Sizing
  3500  
kg
m
 h
Po = 22 bar 1
k = 1.56  
 1.56 1  2

  2   1.561  
Kd = 0.71 C  3.948 1.56     2.8
  1.56  1  
T = 130 °C
m = 3500 kg/h
RT 287  130  273 
vg    0.0526 m3 / kg
P 22  10 5

3500
A 1
 22  2
0.2883  2.8  0.71   
 0.0526 
A  298.6 mm2

44
PRV Sizing
  3500  
kg
m
 h
Po = 22 bar 1
k = 1.56  
 1.56 1  2

  2   1.561  
Kd = 0.71 C  3.948 1.56     2.8
  1.56  1  
T = 130 °C
m = 3500 kg/h
RT 287  130  273 
vg    0.0526 m3 / kg
P 22  10 5

3500
A 1
 22  2
0.2883  2.8  0.71   
 0.0526 
A  298.6 mm2

1 1
 4 A  2  4  298.6  2
D   
     

 19.5 mm  20 mm

PV
RM / T 530 safety
De = 1.1 m fd   factor
Po = 22 bar 2.4 2.4
= 2.2 MPa  220.8 MPa
Z = 0.9

UTS
Rm/T = 530 MPa

45
PV
RM / T 530 safety
De = 1.1 m fd   factor
Po = 22 bar 2.4 2.4
= 2.2 MPa  220.8 MPa
Z = 0.9

UTS
Rm/T = 530 MPa

remember

Some typical properties for PV steel

R e 355 RM / T 510 RM / T 650

fd   fd   fd  
1.5 1.5 2.4 2.4 2.4 2.4
 237 MPa  212 MPa  271 MPa

Using the Using the

Yield Strength U. Tensile Strength

http://www.masteel.co.uk/p355gh.htm

46
PV
RM / T 530
De = 1.1 m fd  
Po = 22 bar 2.4 2.4
= 2.2 MPa  220.8 MPa
Z = 0.9

UTS Po  D e
Rm/T = 530 MPa e
2f  Z  P
2.2  1100

2  220.8  0.9  2.2
 6 mm

PV
RM / T 530
De = 1.1 m fd  
Po = 22 bar 2.4 2.4
= 2.2 MPa  220.8 MPa
Z = 0.9

UTS Po  D e
Rm/T = 530 MPa e
2f  Z  P
2.2  1100

2  220.8  0.9  2.2
 6 mm
tol = ± 0.2 mm
e = 0.2 mm
en  6  10  0.08   0.2
Crate = 0.08 mm/year
 7 mm
Life = 10 years

47
 Q 11

Question 11) PV Sizing

There is a lot of debate about optimum pressure vessel proportions.
This is normally given as a ratio of Length/Diameter or L/D. Many values
are suggested and these range between 3 up to 10. Often the L/D ratio
is dependent on operating conditions, i.e. fluid throughput, surges.
Usually what is required is the maximum volume for the least surface area
and thus weight. This leads to spherical vessels, which while desirable
are usually uneconomical.

Using simple calculus, for a cylindrical vessel with hemispherical ends,

prove that the minimum surface area for a fixed volume gives a length
L of zero and that the vessel is thus a sphere.

48
Q 11) ans

Volumes and areas of a sphere:

vol:
4  D3
V   r3 V
3 6
area:

A  4 r 2 A   D2

Q 11) ans

so for a sphere, we wish to

minimise the Surface area SA
for a given ‘fixed’ volume V

1
SA: S A   DL   D 2
;i.e. 2hemispheres = 1 sphere

2
 D2  D3
Vol: V  L
4 6

49
Q 11) ans
Using 2 make L the subject

 D2  D3
V L
4 6
4   D3 
L V  
 D2  6 
4V 4  D3
  
 D2  D2 6
4V 2 3
L  D
 D2 3

Q 11) ans

4V 2
Using 3 L  D
 D2 3 put 3 into 1
and 1 S A   DL   D2
 4V 2 
S A   D  D   D2
 D 3 
2

4 V D 2
   D2   D2
D 2
3
4V  D2 4
SA  
D 3

50
Q 11) ans
To minimise SA, we simply differentiate 4
w.r.t. D 4V  D2
SA  
D 3
dS A 4V 2
 2   D
dD D 3
For max or min this = 0
4V 2
   D0
D2 3
2 4V or
 D3
 D 2 V
3 D 6
6V Volume of a sphere
D3 


Q 12

51
 Question 12) PV Sizing
A pressure vessel operating at 7.5 bar, 30 °C is to be designed for a
brewery to the BS EN13445 standard. The cylindrical vessel will be
vertical, requires a volume of 16m3 and will be made from:
304 Stainless Steel, to which the following properties apply:

Chemical Analysis (%) Mechanical Properties

C Cr Ni 1.0% Proof Tensile Elongation
Stress Strength %A
MPa MPa
0.08 18.5 9 235 530-730 40
The vessel will have a L/D ratio of 4. Based on operating conditions,

i) Select the best type of ends for this vessel.

ii) Determine the dimensions and the required wall thicknesses

for all parts of the vessel.

iii) Comment on the final design. Is this practical?

End closure selection

Hemispherical
Cost £
Ellipsoidal
Flat Ends
Low pressure
small diameter
vessels
Torispherical

1 .. 1.7 10 .. 15 35 .. 40
Pressure barg

52
Q 12) ans

For a dished end vessel we have the general

formula:
Vtotal  Vcyl  2  Vend
where:
Vend   D3

i.e. for ASME6%:  = 0.081

Kloepper:  = 0.09897
Korbbogen:  = 0.13116

Q 12) ans

Vtotal  Vcyl  2  Vend L

n4
D
 D2 L  nD
 L  2 D 3
4
 D2
 nD  2 D 3
4
n D 3
  2 D 3
4
 n 
Vtotal    2  D 3
 4  1
1

Vtotal  4 Vtotal  3
D3 D 
n 4  2  n   8 

53
Q 12) ans

Once we have D, using 1

1
 4 Vtotal  3
D 
 n   8  
We can calculate L, using

L
4
Vtotal  2   D3 
D 2

Q 12) ans

Assume a Korbbogen end, so:  = 0.13116

and: L D  n  4
1
 4  16  3
D 
 4   8  0.13116 
 1.675 m

L
4
16  2  0.13116  1.6753 
 1.675 2

 6.7 m

54
Q 12) ans

For wall thicknesses, we need

Dimensions :

Design stress : ???

Pressure : 7.5bar  0.75MPa

Weld coefficient : assume z = 1

Errata: Question should read

1.0% Proof Stress = 235 MPa

Nominal design stress (Maximum allowed values)

Note: Steel Normal operating load cases
Steels other than  Rp 0.2 T Rm 20 
fd  min  ; 
2.4 
6.2 austenitic
A < 30%  1.5
Steels other than  Rp 0.2 T Rm 20 
6.3 austenitic fd  min  ; 
A < 30%  1.5 1.875 

Austenitic steel  Rp 1.0 T 

6.4 fd  min  
30%  A < 35%  1.5 

Austenitic steel  Rp 1.0 T  R R 

6.5 fd  max   ; min  p 1.0 T ; m T  
A  35%
 1.5   1.2 3 

 Rp 0.2 T Rm 20 
fd  min  ; 
3 
6.6 Cast steel
 1.9
Yield strength ReH may be used instead of Rp0.2 if this value is not available

BS EN 13445-3:2009, Clause 6: Table 6-1

55
Q 12) ans

Using Austenitic, A  35%

 Rp 1.0 T   Rp 1.0 T Rm T  

fd  max  
 ; min  ;  
 1.5   1.2 3 
 235   235 530  
 max   ; min  ; 
  1.5   1.2 3 
 max156.7 ; min 195.8 ; 176.7 
 176.7 MPa

Q 12) ans

For wall thicknesses, we need

Dimensions :

Design stress : 176.7 MPa

Pressure : 0.75MPa

Weld coefficient : z = 1

56
Torispherical wall thickness
The maximum of:
PR
es 
2 fz  0.5P

For the sake of simplicity we will stick

to the formula for membrane stress for
these types of ends

Q 12) ans

Wall thicknesses
PDi
cylinder: e 
2 fz  P
0.75  1675
e
2  176.7  1  0.75
e  3.563 mm  4 mm

PR R=0.8D for
end: e Korbbogen
2 fz  0.5P
0.75  0.8  1675
e
2  176.7  1  0.5  0.75
e  2.848 mm  3 mm

57
Minimum wall thickness
Pressure Vessel Min. thickness
Diameter (m) (mm)

1 5
1 to 2 6
2 to 2.5 9
2.5 to 3.0 10
3.0 to 3.5 12

Values include a corrosion

allowance of 2 mm
Coulson & Richardson: Chapter 13. Mechanical Design of Process Equipment

Q 12) ans

3mm
Comment:
4mm This is a very large vessel
and the wall thicknesses
seem rather small.
6.7m Would need to check the
structural integrity
further to see if this would
‘stand up’!

1.675m

58
 Q 13

Question 13) PV Sizing

You are designing a pressurised reactor vessel to the BS EN13445
standard. The reactor, which operates at 15 bar is shown below. It is in
the form of a vertical cylinder. The bottom end is hemispherical for
operating purposes.
The reactor operates at a maximum temperature of 350°C.

59
 Question 13) PV Sizing
The selected material for construction is CrMo4-5, which is suitable for
operation at high temperature. The material has a density of 7850kg/m3
and the following material properties.
Rp0.2/20 Rm/20 Elongation
MPa MPa A%
e16: 300
e>16: 290 450-600 19

The yield strength reduction factor at temperature is characterised by the

formula: -0.0011T+1.0623, for temperatures between 100°C and 400°C

You have not yet decided which type of end will go at the top.
Determine, from the following ends, which will give the least overall mass
for the vessel.
(i) ASME 6%,
(ii) Kloepper,
(iii) Korbbogen,
(iv) Ellipsoidal,
(v) Flat end with hub (uniform thickness)
Assume the plate is supplied from the manufacturer to the nearest mm.

End closure selection

Hemispherical
Cost £
Ellipsoidal
Flat Ends
Low pressure
small diameter
vessels
Torispherical

1 .. 1.7 10 .. 15 35 .. 40
Pressure barg

60
Q 13) ans

First job is to determine the design stress.

Material (e16mm) has: A% = 19
Rp0.2/20 = 300 MPa, and
Rm/20 = 450 MPa

Nominal design stress (Maximum allowed values)

Note: Steel Normal operating load cases
Steels other than  Rp 0.2 T Rm 20 
6.2 fd  min  ; 
2.4 
austenitic
A < 30%  1.5
Steels other than  Rp 0.2 T Rm 20 
6.3 austenitic fd  min  ; 
A < 30%  1.5 1.875 

Austenitic steel  Rp 1.0 T 

6.4 fd  min  
30%  A < 35%  1.5 

Austenitic steel  Rp 1.0 T  R R 

6.5 fd  max   ; min  p 1.0 T ; m T  
A  35%
 1.5   1.2 3 

 Rp 0.2 T Rm 20 
fd  min  ; 
3 
6.6 Cast steel
 1.9
Yield strength ReH may be used instead of Rp0.2 if this value is not available

BS EN 13445-3:2009, Clause 6: Table 6-1

61
Q 13) ans
First job is to determine the design stress.
Material (e16mm) has: A% = 19
Rp0.2/20 = 300 MPa, and
Rm/20 = 450 MPa

 Rp 0.2 T Rm 20 
fd  min  ; 
 1.5 2.4 
Using: -0.0011T+1.0623 Rp 0.2 350   0.001  350  1.0623  300
 214 MPa
 214 450 
fd  min  ; 
 1.5 2.4 
 min 142 .5 ; 187.5 
fd  142 .5 MPa

So our Design Stress fd = 142.5 MPa

Q 13) ans

Operating parameters:
D: 1.5 m (1500 mm)
L: 3 m (3000 mm)
P: 15 bar (1.5 MPa)
T: 350 °C
fd: 142.5 MPa
z: 1 (assumption)
density : 7850 kg.m-3

62
Q 13) ans
Calculate wall thicknesses for cylinder and
hemispherical end
cylinder: e  PDi
2 fz  P
1.5  1500
e
2  142.5  1  1.5
e  7.939 mm  8 mm

hemi-end: e  PDi
4 fz  P
1.5  1500
e
4  142.5  1  1.5
e  3.959 mm  4 mm

Q 13) ans

Calculate volume and mass of cylinder material

cylinder: Di = 1.5 m
L = 3.0 m
e = 8 mm
Do = 1.5+20.008 = 1.516 m

 D2o  D2i 
Vcyl  L
4
 1.5162  1.52  mcyl   Vcyl
 3
4  7850  0.1137
 0.1137 m3  892.55 kg

63
Q 13) ans

Calculate volume and mass of hemi-end material

hemi: Di = 1.5 m
e = 4 mm
Do = 1.5+20.004 = 1.508 m

 D3o  D3i 
Vhemi 
12
 1.5083  1.53  mhemi   Vhemi

12  7850  0.0142
 0.0142 m3  111.57 kg

Q 13) ans

So total mass, so far (without top end)

cyl: mcyl = 892.55 kg

hemi: mhemi = 111.57 kg

Without top end

mass = 1004.12 kg

64
Q 13) ans

We now need to consider the size, volume and

weight of:
(i) ASME 6%,
(ii) Kloepper,
(iii) Korbbogen,
(iv) Ellipsoidal,
(v) Flat end with hub (uniform thickness)

Remember Operating parameters:

D: 1.5 m (1500 mm)
P: 15 bar (1.5 MPa)
fd: 142.5 MPa
z: 1 (assumption)
density : 7850 kg.m-3

Torispherical wall thickness

The maximum of:
PR
es 
2 fz  0.5P

For the sake of simplicity we will stick

to the formula for membrane stress for
these types of ends

65
Ellipsoidal wall thickness
According to BS EN 13445-3: 2009 we treat ellipsoidal
ends the same as torispherical ends
Di
hi 
4

Di

Conditions: Di Di
K  2
2hi 2D i 4 
K is the shape
factor 1.7  K  2.2

BS EN 13445-3, Section 7.5.4

Ellipsoidal ends

Ellipsoidal wall thickness

According to BS EN 13445-3: 2009 we treat
ellipsoidal ends the same as torispherical ends
Di
hi 
4

Di
Designed as nominally equivalent torispherical
ends with: Di
K … so K = ???
2hi
 0.5  
r  D i    0.08 
 K  
R  Di 0.44K  0.02 

66
 Q 13) ans Wall thicknesses
ASME 6% 1.5  1500
e  7.918  8 mm
2  142.5  1.5

Kloepper
1.5  1500
e  7.918  8 mm
2  142.5  1.5

Korbbogen
1.5  0.8  1500
e  6.334  7 mm
2  142.5  1.5
Ellipsodal 1.5  0.9  1500
e  7.126  8 mm
2  142.5  1.5
Flat end with hub (uni thickness) …

Circular flat end with a hub: Uniform thickness shell

eaf

Deq
Lcyl
es Di

67
Flat end with hub
Required min thickness e, (eaf>e)
P
e  C1D eq
f
 D  es  e 
C1  max 0.40825  A 1 i ; 0.299   1  1.7 s  
 Di  Di 
 es 
A 1  B1 1  B1 
 2D i  es  

3 P  D i  3 e2s 2D i  es 
2 4
3 f  D  es 
B1  1   i       
P  D i  16 f  D i  es  4 D i  es 

Uniform D  D  r Tapered 1
eq i D eq  Di  DF 
Thickness Shell 2

Flat end with hub: Method

r = es

Uniform ? Tapered
Thickness Shell
Deq=Di - r Deq=½(Di + DF)

C1={P,Di,f,es} inc(r)
e ={C1,Deq,P,f}
r < 1.3e
?
r 1.3e use e, r and lcyl  0.5 es Di  es 

68
Flat end, +hub Start with:
P 1.5MPa f 142.5MPa
es = 7.939 mm, and
Di 1500mm ecyl 7.939mm
r = 2 mm
D eq  Di  r
r>1.3e?
es r Deq A1 B1 C1 a C1 b max(C1) e
7.939 7.939 1492.1 0.9914 0.9940 0.4069 0.3017 0.4069 62.295 
62.295 9.939 1490.1 0.5404 0.5463 0.2298 0.3201 0.3201 48.944 
48.944 11.939 1488.1 0.7077 0.7158 0.2983 0.3156 0.3156 48.188 

46.651 57.939 1442.1 0.7329 0.7412 0.3085 0.3148 0.3148 46.583 

46.583 59.939 1440.1 0.7336 0.7419 0.3088 0.3148 0.3148 46.515

46.515 61.939 1438.1 0.7344 0.7427 0.3091 0.3148 0.3148 46.447 

Q 13) ans Wall thicknesses

ASME 6% 1.5  1500
e  7.918  8 mm
2  142.5  1.5

Kloepper 1.5  1500

e  7.918  8 mm
2  142.5  1.5

Korbbogen
1.5  0.8  1500
e  6.334  7 mm
2  142.5  1.5
Ellipsodal 1.5  0.9  1500
e  7.126  8 mm
2  142.5  1.5
Flat end with hub (uni thickness) e  46.447  47 mm

69
Q 13) ans Volumes & masses
ASME 6%
 =7850 kg.m-3
V  0.081 1.5163  1.53   0.00844 m3 69.4 kg
Kloepper
V  0.09897 1.5163  1.53   0.01080 m3 84.8 kg
Korbbogen
V  0.13116 1.5143  1.53   0.01251 m3 98.2 kg

Ellipsodal

V 1.516 3
 1.53   0.01429 m3 112.2 kg
24
Flat end with hub (uni thickness) - Approx
 1.5162
V  0.047  0.08484 m3 666 kg
4

Q 13) ans Summarising total masses

ASME 6%: 1004.12 + 69.4 = 1073.5 kg
Kloepper: 1004.12 + 84.8 = 1088.9 kg
Korbbogen: 1004.12 + 98.2 = 1102.3 kg
Ellipsodal: 1004.12 + 112.2 = 1116.3 kg
Flat end: 1004.12 + 666 = 1670.1 kg

total
vessel
mass
kg

70