Flixborough Disaster

At about 1653 hours on Saturday, June 1, 1974, the Nypro (UK) site at Flixborough was severely damaged by a large scale explosion. Twenty eight employees were killed and 36 injured. Outside the works injuries and damage were widespread, with 53 people being reported injured. Varying degrees of damage was caused to 1,821 houses and 167 business premises. The loss of life could have been far higher if the accident had occurred on a weekday.

The Flixborough plant was operated by Nypro Chemicals, owned jointly by Dutch State Mines and the National Coal Board (NCB) at the time of the disaster. The plant was built between 1964 and 1967 for the production of Caprolactum, which is the basic raw material for the production of Nylon.

The part of the plant on which the explosion occurred, involved the oxidation of cyclohexane to cyclohexanone using air injection at 8.8 bar g (130 psig) and 155°C (31 1°F). The oxidation reaction was carried out in the liquid phase in six reactor vessels arranged in series, each set 14 inches (360 mm) below its predecessor to allow the flow to progress through the reaction train by gravity.

The reactors were made of 13 mm steel and a 3 mm thick stainless-steel liner, and each was of 5 m height and 3.5 m in diameter, built with an interior overflow weir, baffle plates, and an agitator.

**Flixborough Disaster Events Sequence**

Flixborough Accident Description

Two months prior to the explosion on 27 March 1974, a 2 m long vertical crack leaking cyclohexane was discovered on Reactor 5. Subsequently, the plant was shut down and the reactor removed for repair.
In order to permit continued operation of the plant, the site management team decided to remove Reactor 5 and install a 0.5 m (20 Inches) diameter temporary bypass pipe between the bellows to take its place. Due to elevation changes, it was necessary to incorporate a ‘dogleg’ shape into this bypass.
The pipe was constructed and supported by scaffolding; however, no account was taken for the turning moment that would act on the pipe due to fluid flow. Consequently, the scaffolding support was not adequate to resist the shear forces.
The temporary pipe performed satisfactorily for two months until May 29, a leak appeared from the bellow that forced the plant to shut down to repair a leak. During the late afternoon on 01 June 1974, whilst the plant was being restarted, a slight rise in pressure occurred, well below the relief valve set point, causing the temporary pipe to twist.
The installed bypass ruptured, and 30 tons of cyclohexane at 10 bar (150 psi) and 150°C escaped in 30 seconds, forming a cloud more than200 meters across and 100 meters high (650-330 ft), which was ignited at 4:53 PM by a nearby furnace at Hydrogen plant, resulting a massive vapor cloud explosion.

Key Findings from Flixborough Explosion

The officially accepted cause of the disaster is the mechanical failure of the “dog leg” 20 inch by-pass. The “dog leg” by-pass consisted of three lengths of 20 inch pipes with flanges at each.
The by-pass pipe was of smaller diameter (20 inch) than the reactor flanges (24 inch) and in order to align the flanges short sections of pipes with stainless steel bellows were added at each end of the by-pass. Under pressure these bellows tend to squirm or twist. The assembly as constructed was of unknown strength and did not comply with the British Standard.
No design calculations were carried out for the modification, with no account being taken of the appropriate Standards or the bellows unit manufacturer’s recommendations. The modification was fabricated onsite without any engineering drawings, calculations or hydraulic testing.
The welding on the temporary pipework connections had not been tested or checked. Although the assembly had been pneumatically tested to 9 bar g (130 psig), it had not been tested up to relief valve pressure of 1 1 bar g (160 psig).
The pipe was constructed and supported by scaffolding; however, no account was taken for the turning moment that would act on the pipe due to fluid flow. Consequently, the scaffolding support was not adequate to resist the shear forces.
The team tasked with making and installing the temporary pipe were no professionally qualified to do so.
Furthermore, the cause of the crack in Reactor 5 can also be linked back to a process modification. Prior to the crack there was a leak of cyclohexane from the stirrer gland and to condense the leaking vapor, water was poured on the top of the reactor. However, the water contained nitrates which caused stress corrosion cracking of the mild steel pressure vessel
Prior to the incident, the works engineer had left at the start of the year, and by June 1974 the company had yet to find a replacement. At the time the bypass line was being planned and installed, there was no engineer on site with the qualifications to perform a proper mechanical design, or to provide critical technical review on related issues. There were chemical and electrical engineers on 80 staff, but no other mechanical engineers.
At the time of the disaster, large scale explosions were not considered in the design and location of occupied buildings. The control room was located close to the plant for operational reasons.
As explained earlier, the plant site contained excessively large inventories of cyclohexane, naphtha, toluene, benzene, and gasoline. These inventories likely contributed to the fire after the initial blast, which burned for ten days following the explosion.