3/30 Monju Safety Analysis 'Inadequate & Unconvincing'

[the.greenbase@green2.greenpeace.org]

----------
Original-TO:      World Press (Green2:Green2:Gnl:INET)
Original-Cc:      The Greenbase (Green2:Green2:Gnl:Main)
----------
                   ************************
                   GREENPEACE PRESS RELEASE
                   ************************
 
MONJU SAFETY ANALYSIS 'INADEQUATE AND UNCONVINCING', EXPERT
SAYS AFTER PLANT TOUR 
 
Tsuruga, Japan, 30 March 1996 (GP) -- Nuclear expert Professor 
Jochen Benecke said after meetings with PNC officials Friday 
that he is not convinced by the company's safety analysis to 
protect against a disastrous nuclear explosion at their Monju 
fast breeder reactor.  He said that in addition he is worried 
that the December 8th sodium leak may be an indication of 
incomplete analysis in other components of the plant. 
 
Benecke is a nuclear physicist with an extensive background 
analysing the safety of fast breeder reactors.  He is touring 
at the invitation of Greenpeace Japan.  They met with
officials  of PNC's fast reactor and safety divisions at the
plant on  Friday. 
 
Benecke told a public meeting in Tsuruga today that PNC 
admitted to him that a severe 'core-disruptive' accident is 
possible at Monju, as at any other fast breeder reactor, and 
they are studying the topic.  However, they insisted that 
quality control in the design and analysis of the reactor
would  prevent such an accident from happening.
 
"But the sodium leak in December seems to be a failure of 
quality in both design and analysis," Benecke told the 
meeting.  "If it is possible to fail in such a simple and 
well-known matter as the thermo-couple which caused the leak, 
how much confidence can we have that PNC has done everything 
absolutely perfectly with respect to a major accident?"
 
There is no reliable scientific method for estimating the size 
of an explosion which could occur during a worst-case, or 
so-called 'Bethe Tait', accident in the Monju reactor.  The 
safety calculations(1) used by PNC were originally not meant
to  provide the type of firm numbers needed to design a
reactor  vessel so that it could withstand a severe core-
disruptive  accident.  
 
But since there are no alternative methods available to 
engineers, they use the existing calculations to design 
safety-critical components -- such as reactor vessels -- of 
FBRs anyway.  Since the computer models are not adequate to 
this task, technicians are forced to make assumptions based on 
'best guesses'.
 
"The resulting safety analysis is unconvincing and 
inadequate", Benecke reported to the public meeting today.  
"Ultimate protection for the public against a nuclear disaster 
is not based on real scientific analysis, but has a rather 
large component of guesswork."
 
PNC officials declined to provide documents to support their 
view that their safety analysis is adequate.
 
"I had hoped to discover that Japan had safety analysis which 
is much better than what I have encountered in Germany and the 
USA, but I'm disappointed to find the Japanese experts using 
standard procedures which have already been shown to be 
inadequate to protect against accidents with very severe 
consequences," said Benecke.
 
Greenpeace Japan campaigner Hideaki Takemura declared himself 
unsatisfied with the meeting:  "PNC engaged in a public 
relations exercise, and that meant they arrived at the meeting 

with no documentary evidence to back up their statements.  If 
they cannot provide even the safety analysis of the plant --
we  find it hard to take their new 'openness' very seriously".
 
ENDS
 
NOTES TO EDITORS 
 
It has been understood since the 1950s that fast reactors (eg. 
Monju) can experience very severe accidents which may result
in  a nuclear explosion.  Such an event could 
completely destroy the reactor vessel and containment 
structure, spreading radioactive material across a wide area, 
as happened at Chernobyl.  This type of accident is known as a 
'Bethe Tait' accident, after the two physicists who provided 
the first estimate of the damage which would be caused by this 
type of accident.  
 
However, Bethe Tait accidents occur very infrequently so there 
is almost no actual experience from which to analyse them in 
detail.  It is especially difficult to determine how large the 
energy released ('explosion') from a Bethe Tait accident might 
be.   
 
Although nuclear reactors are designed with multiple 'safety 
systems', specific methods to prevent and mitigate major 
accidents of this type are difficult to design because little 
is known about their evolution from initiation to termination. 
 Because the potential energy release is unknown, it is 
therefore not possible to be certain that the reactor vessel, 
which contains the nuclear core, will survive an explosion 

during a Bethe Tait accident.
 
The Power Reactor & Nuclear Fuel Development Corporation (PNC) 
follows the standard procedure for analysing severe accidents. 
 Since there is almost no actual data nor coherent theory to 
work with, PNC use a series of computer models of event- 
sequences, known as 'codes', to make judgments about the 
evolution of a severe accident.
 
Each of these 'codes' or models -- SAS-3D, VENUS PM, 
PISCES-2DELK, PPP-M, and MIMIR-N2 -- applies to a specific 
event or series of events.  For example, the PPP-M code is
used  to analyse the effect of pressure waves on the pipe
nozzles  that penetrate the reactor vessel at various points. 
 
Unfortunately, there is no model to analyse the whole sequence 
of events of a major core-disruptive (Bethe Tait) accident.  
The available codes can only provide partial modelling of an 
accident.  There is no proven means available today to 
synthesise the sequences into a realistic picture of a Bethe 
Tait accident from beginning to end.
 
Therefore, much of the synthetic safety analysis for FBRs is 
dependent on assumptions and 'best guesses' made by
technicians  to move from one calculational model to another
(eg. moving  from a code which models the initiating phase,
such as SAS-3D,  to one modelling events later in an event
sequence such as  VENUS-PM).  
 
For this reason, the standard computer codes which are used by 
PNC and other fast breeder operators are only useful for 
judging relative values, not absolute values.  In other words, 
one can use such models to help make a distinction between a 
component that needs to be 'stronger' versus 'weaker', but not 
to judge exactly how strong or weak a component should be.    
The use of these codes to claim that PNC are 'certain' that
the  reactor vessel at Monju will withstand a Bethe Tait
accident  because they 'know' the energy release of such an
accident,  is not scientifically justified.
 
Newer computer models which seek to overcome these
difficulties  (eg. SIMMER) are still not verified, so
designers continue to  fall back on partial and inadequate
codes.
 
For More Information:
Jochen Benecke, John Willis
Greenpeace Japan
(03) 5351 5400
 

----------