ICAN Scottish Partner

Latest Events

Donate to SCND

Amount to donate:
£  GBP  




ScrapTrident


Risk of nuclear disaster at Faslane shiplift

ImageFolllowing a Freedom of Information request from Scottish CND the Ministry of Defence have released risk assessments for a major nuclear accident on the shiplift at Faslane.   The heavily redacted internal reports say that the risk is "close to the tolerability criterion level".

In November 2007 Scottish CND requested copies of two risk assessments carried out into the dangers from a major accident involving a Trident submarine in the Faslane Shiplift which had been published in 2000.  After a long delay the reports were recently released, however all the probability figures have been redacted from the documents.

This was reported by Rob Edwards in the Sunday Herald on 30 August 2009.

The documents were a Radiological Probabilistic Risk Assessement and an Accident Probability Assessment with a Fault Tree Annex.  A 2003 review of these documents by Atkins had been released to Scottish CND in 2007 under Freedom of Information legislation.  The Atkins review revealed a number of shortcomings.

Below are some comments on the content of the assessments:

 

A Radiological Probabilistic Risk Assessment of the Faslane shiplift for Vanguard class submarines with Strategic Weapon System embarked.  (Atomic Weapons Establishment, November 2000)

This is an analysis of the effect of all the missiles detonating and dispersing the plutonium from all the warheads on a submarine.  The report has a list of assumption which include:

“The quantity of Pu and explosive used in the assessment was based on the assumption that an SSBN has xx RBSs (A or C) fitted and that there are xx missiles with xx tonnes TNT equivalent explosive”

“all xx missiles detonate resulting in the dispersion of xx RBAs worth of plutonium”

“If an RBA is insulted the high explosive (HE) will detonate with a probability of unity”

[RBA is Reentry Body Assembly, ie warhead.]

Three criteria are considered – Individual Fatality, Societal Fatality and Societal Contamination.  Of these it would appear that the one where the risk comes closes to the criterion is Societal Contamination.

There is a table with three columns, for the three criteria, and four rows – single point estimate, optimistic estimate, best estimate and pessimistic estimate.

For Societal Contamination (SC) the optimistic estimate is “nearly two orders of magnitude below” the SC criterion, the  best estimate is ”approximately one order of magnitude below the SC criterion”, and the pessimistic estimate is above unity but “close to” the tolerability criterion for SC.

“The low values of the Compliance Factors (Societal Contamination) for the CESO(N) pessimistic case indicate that the risks are close to the tolerability criterion level; there is, therefore, a strong argument for ensuring that the risks are As Low As Reasonably Practical (ALARP)”

 [CESO(N) is the Chief Environment Safety Officer (Navy)]

This report refers to two types of RBAs - A and C.  This may be two different types of UK Trident warhead.

“The RBA-C is comparable to the RBA-A in reaction probabilities and consequences”

The reference could be to the full-yield and lower-yield warheads.  This study was looking at Pu release and the two types of warhead will have the same amount of Pu.  The naming suggests a third option, B, was also considered but not deployed.

 

Accident Probability Assessment of Faslane shiplift for Vanguard class submarines with Strategic Weapon System embarked (August 2000)

This has a series of diagrams which illustrate the sequence of events which could lead to missile detonation or missile cookoff.  For example if the submarine rolls and has a side impact - the RBAs could detach and impact against the 3rd stage of the missile - leading to a missile/RBA detonation.

Flaws in these reports

There are two obvious omissions from the assessments.  The consequence modelling only looked at the dispersal of plutonium from the warheads.  It did not consider the potential for nuclear yield, or the effects of dispersing the reactor's radioactive contents at the same time as warhead plutonium.  This latter failure is an obvious failing, because at several points the reports refer to scenarios in which the reactor containment would be breached.

In describing the scope of their work the AWE assessment (page 5) it says:

“The RPRA assessed the risks to the public of a release of plutonium (pu) from the SWS.  The radiological hazard from an NRSP [reactor] accident was not considered”.

The Accident Probability Assessment says:

“all large aircraft impacts are assumed to directly lead to a primary Loss of Coolant Accident (LOCA), but in this assessment they are assumed to lead to an explosive reaction from the SWS and a subsequent release of Pu”  (p xii)

A large aircrash would lead to the dispersal of a proportion of the reactor contents, from the LOCA, plus Pu from warheads, but this assessment is only looking at Pu dispersal.

A torpedo accident could damage the reactor (p 8).  The missiles are much nearer to the reactor than the torpedo compartment so the risk of damage to the reactor from the detonation of all the missiles would be greater.  There is a serious risk that a missile explosion would breach the primary containment.  In a Primary Containment Failure accident the entire contents of the reactor can be dispersed in a short period of time.

The diagrams in Annex A include a number of sequences in which there is a Primary Containment Failure accident which subsequently leads to a missile/warhead explosion.  If the submarine rolls by more than 90%, following an aircrash, then there can be prompt criticality leading to a failure of the primary coolant circuit (prompt crit/fail of prim cct/frags generated).  Fragments from the reactor could then cause a missile explosion (Annex A p 7).  Dropping a load or the collapse of a crane can also result in prompt criticality (Annex A p 13 &14).  In these cases the reactor contents would be dispersed as well as the Pu.

If the consequences are close to the tolerable criterion with just Pu dispersal, what would they show if the reactor contents were added ?

Furthermore, in the complex scenario of the detonation of 14 Trident missiles, AWE are wrong to ignore completely the possibility of a nuclear yield.  Even a very low nuclear yield would further complicate the sequence of events and would increase the nuclear hazard.

Commenting on the report nuclear engineer John Large said:

"If the containment of a nuclear weapon was breached, the consequences would be dire indeed, particularly if the plutonium was lofted high into the air by a fire.  No civil contingency plans could cope with it.  The risks are not minimised to an appropriate level."