Flixborough disaster

1. Flixborough disaster
By:
Paradigma Carlo
Giovanni

2. Introduction
•Largest peacetime explosion ever to occur in the UK
•Date: Saturday, 1 June 1974
•Location: Flixborough chemical plant owned by Nypro
(UK) Ltd
•Deaths of 28 workers on the site
•Widespread damage to property within a 6 mile radius
around the plant

3. •1967: 20,000 TPA caprolactam plant built by DSM at the
Flixborough site using process involving the hydrogenation
of phenol
•1972: 70,000 TPA ; new process used
•New process based on oxidation of cyclohexane
•Posed much greater hazard than phenol process
C6H5OH + 2 H2 → (CH2)5CO

4. Description of the cyclohexane
process
•Process operates by injecting compressed air into
liquid cyclohexane at a working pressure of about 9
bar and temperature of 155°C
•Cyclohexanone and cyclohexanol produced
•Conversion is low and it is necessary to recirculate
the cyclohexane continuously through a train of six
large SS-lined reactors

5. C6H12 + O2 → (CH2)5CO + H2O
Cyclohexane Cyclohexanone
Caprolactum Nylon6

6. Simplified flow diagram of cyclohexane oxidation plant
before March 1974 (Whittingham, 2005)

7. Events leading to the accident
2.Miners overtime ban of Nov, 1973
•Resulted in the government passing legislation to restrict
the use of electricity by industry to 3 days a week
•Was not possible to operate the process on this basis
•Was decided to utilize existing emergency power
generation on-site

8. •Major electricity user: 6 stirrers in the cyclohexane
reactors
•Primary purpose: disperse compressed air that was
injected into each reactor via a sparger
•Also ensured that droplets of water formed within
the reactor system were dispersed into the
cyclohexane

9. 1.The No. 5 reactor problem
•Jan, 1974:normal electricity supply resumed
•Was found that the drive mechanism for the stirrer
in the No. 4 reactor had been subject to severe
mechanical damage
•No reason was found for this. It was therefore
decided to continue to operate the plant with the
No. 4 reactor stirrer shutdown.

10. •Cyclohexane reactors were MS vessels fitted with an
inner SS lining to resist corrosion
•March,1974: Cyclohexane found leaking from 6 feet
long vertical crack in the MS shell of the No. 5 reactor
•Due to technical problems experienced earlier and the
effects of the 3-day week, the plant owners were keen
to make up lost production
•Therefore decided to remove No. 5 reactor for
inspection and continue operation with the remaining
five reactors

11. 1.Installation of 20” bypass pipe
•This pipe connected together the existing 28 inch
bellows on the outlet of reactor No. 4 and the inlet of
reactor No. 6
•Dog-leg shape of pipe
•Company did not have qualified mechanical engineer
on site to oversee design and construction

•No hydraulic pressure testing of pipe carried out,
except for a leakage test using compressed air.

12. Simplified flow diagram of cyclohexane oxidation plant
after March 1974 (Whittingham, 2005)

13. 1.Resumption of production
•Plant restarted and operated normally, with occasional
stoppages, up until the afternoon of Saturday, 1 June 1974
•Previous day: plant had been shut down for minor repairs
•Early hours of 1 June: plant in process of being restarted
•Start-up involved charging system with liquid cyclohexane to
normal level and then recirculating this liquid through a heat
exchanger to raise the temperature.

14. •The pressure in the system was maintained with nitrogen
at about 4 bar until the heating process began to raise the
pressure due to evaporation of cyclohexane.
•The pressure was then allowed to rise to about 8 or 9 bar,
venting off nitrogen to relieve any excess pressure. The
temperature in the reactors by then was about 150°C.

15. •On 1 June this procedure was followed except it was noted
by the morning shift that by 06.00 hours the pressure had
reached 8.5 bar even though the temperature in the No. 1
reactor had only reached 110°C
•Was not realized at the time that this discrepancy might
have indicated the presence of water in the system
•The start-up continued until, at about 16.50 hours, a shift
chemist working in the laboratory close to the reactors
heard the sound of escaping gas and saw a haze typically
associated with a hydrocarbon vapour cloud.

16. The accident
•16.53 hours on 1 June 1974: massive aerial explosion
occurred with a force later estimated to be about 15 to
45 tonnes of TNT equivalent
•Explosion heard up to 30 miles away and damage
sustained to property over a radius of about 6 miles
around the plant
•28 plant workers killed with no survivors from the
control room
•All records and charts for the start-up destroyed
•Following the explosion, 20 inch bypass assembly was
found in a ruptured condition

17. The Public Inquiry
•Following the disaster, public inquiry
conducted under the chairmanship of Roger
Parker QC
•To establish the causes and circumstances of
the disaster
•To identify lessons to be learnt from the
disaster

18. Conclusions of the inquiry
•The immediate cause of the main explosion
was the rupture of the 20 inch bypass
assembly between the No. 4 and No. 6 reactor
•Two main theories to explain

19. The 20 inch pipe theory
•The 20 inch bypass assembly failed due to its
unsatisfactory design features
• However, the assembly had survived 2 months of
normal operation.
• A number of independent pressure tests were
commissioned to determine unusual conditions

20. •The normal working pressure = 8 bar
•practice during start-up to allow the pressure to
build up to about 9 bar.
•The safety valves for the system, were set to
discharge at a pressure of 11 bar
•At pressure above 11 bar, squirming motion
which distorted the bellows.

21. •Even when the assembly squirmed, no rupture
until pressure crossed 14.5 bar, a pressure not
achievable in reactors.
•Inquiry concluded that a rupture of the 20 inch
bypass due to pressure, temperature conditions
•Report conceded ambiguity in the hypothesis
•Simulation tests could not replicate failure at
similar conditions

22. The 8 inch hypothesis
•Alternative theory
•50 inch split in an 8 inch line connected to
separator below bypass
•this failure led to a smaller explosion causing
failure of the main 20 inch bypass
•Zinc embrittlement had caused the split
•Small lagging fire at a leaking flange causing zinc
to drip onto the 8 inch pipe
•Brittle failure – Vapour release – Explosion – 20
inch failure

23. •Inquiry Report had devoted discussion of this
two-stage theory
•Finally dismissed as being too improbable
• No other theories considered by them to explain
failure of 20 inch bypass pipe

24. The water theory
•Another alternative theory
•Not considered by the Inquiry
•Much of scientific work after Inquiry closed
•Examined the effects of not operating the No. 4
reactor stirrer during the start-up at a time when
water may be present
•More probable explanation

25. •Cyclohexane and water are normally immiscible
•Azeotrope forms due to the limited solubility of
water in cyclohexane.
•This azeotrope has lower boiling point than
either water or cyclohexane
•Unstable interfacial layer may form
•Under certain conditions can boil and erupt
violently ejecting cyclohexane and superheated
water from the reactor.

26. •Normally impossible for water layer to form due to
dispersion of water by air distribution
•During start-up, the air to the reactors shut off
•If stirrers running during start-up, no water layer
•If stirrer stops, a layer of water could form, together
with the unstable azeotrope.

27. •As temperature of reactor increases, boiling
point of azeotrope is reached
•Possibility of a sudden violent eruption from the
reactor and ejection of slugs of liquid reactant
•Slugs exert high mechanical forces on the bypass
assembly, loosely supported by scaffolding
•Causes bypass assembly to fail without the high
static pressure in the reactors

28. Alternative event sequence
•Most credible explanation
•Explains failure of 20 inch bypass
•Also provides an explanation for the whole
sequence of events

29. •Unexplained failure of drive mechanism of No. 4
reactor
•Crack developing in the lining and shell of No.
5 reactor
•Failure of the 20 inch bypass assembly.
•Any/all failures caused by violent eruption of
reactor contents due to presence of water
•Committee failed to see common thread

30. •Drive mechanism failure for No. 4 reactor
unexplained
•Thought to be irrelevant
•Crack in the shell of the No. 5 reactor due to
stress corrosion
•Proposed by plant owners but not credible

31. •The failure of the bypass concluded by Inquiry
due to reactors being over-pressurized
•Implies human error, not verifiable
•Greatest failing of Inquiry was not taking
account of all the events 6 months prior to the
disaster
•Issue of non-operation of the reactor stirrers
ignored

32. Conclusions
• Human error analysis
Table gives causes against the different types of
error that occurred.
•Direct cause
Failure of the 20 inch bypass pipe led to huge
release of inflammable cyclohexane vapour
which ignited
•Root causes
A badly designed 20 inch bypass pipe installed
rather than finding reasons for the crack in the
No. 5 reactor
Why bypass failed ?

33. (Whittingham, 2005)

34. Safety considerations
•Learnings
•Low inventory especially of flashing fluids
•Before modifying process, carry out systematic
search for possible cause of problem
•Carry out HAZOP analysis
•Construct modifications to same standard as
original plant
•Use blast-resistant control rooms and buildings

35. THANK YOU