1986 RBMK disaster

["Cuttler, Jerry" <CuttlerJ@spkb.candu.aecl.ca>]

The detailed explanation is more complex.  It is necessary to model both the 
neutronics and thermal hydraulics of the RBMK reactor.  This was done by 
AECL physicists in 1986.  The flux distribution just before the transient 
was double humped, not convex.  The bottom of the core was decoupled from 
the top and behaved independently.  With the combination of axial burnup and 
flux distributions, a destructive power pulse on scram is obtained only when 
positive scram reactivity is inserted.  Without the latter, the power rise 
is slow enough for a modest shutdown system to terminate the pulse and 
prevent severe core damage.
Jerry Cuttler
 ----------
From: radsafe
To: Multiple recipients of list
Subject: Re: Design of Nuclear Power Plants
Date: Thursday, May 02, 1996 12:19AM

It seems to me that the two things (control rod design and positive void
coefficient) worked together to cause the power excursion.  The rods
inserted positive reactivity because of the positive void coefficient.  The
"displacer" section of the rods was graphite and was relatively weak
absorber compared to the water.  When the rods moved, the displacer pushed
out the water, causing the equivalent of a void (even though it wasn't a
steam void).

So, given a reactor with a positive void coefficient (a "less than adequate
design" to start with), yes the control rod design was flawed.

But that is not the same as saying that the rod design was the root cause of
the accident.  The report I have (NUREG 1250, is that outdated now?) and
other articles I have come across seem to indicate that "root cause" was a
total breach of all reasonable operational safety parameters brought on by
overconfidence.  The operators went way beyond the limit of safe operation
of the plant during the test.  Even though there were weaknesses in the
design, there's no reason to beleive the RBMKs couldn't have continued to
operate safely given properly trained personnel and the proper attitude
toward safe operation.  When you look at the sequence of events leading to
the accident, it's hard to imagine any better way to blow up a reactor.

I do not agree with the comment that the continued operation of the reactor
was not forbidden by operating procedures.  The operators had data showing
that they had lost adequate shutdown margin *before* the initiation of the
test.  No matter what the design of your reactor, when you're handed that
information, and you proceed with the test, it's not realistic to say the
accident was caused by poor design.

Standard disclaimers.
Keith Welch
welch@cebaf.gov
 ----------------------------------------------------------------------------  
 ---
 --


>The "root cause" of the Chernobyl power pulse was the "less than adequate"
>design of the control rods.  If the rods had not inserted a large amount of 

>reactivity into the bottom part of the core, when the reactor was tripped,
>the power pulse would not have occurred.  This was first identified by
>reactor physicists at AECL (Sheridan Park) immediately after the event. 
 The
>analysis is described in report AECL-9426 (1987).  Adi Dastur can supply
>more details.
>
>Positive void coefficient aggravated the effect of the power pulse, but the 

>analysis shows clearly that the power rise would have been quite slow, if
>the rods had not inserted reactivity during the reactor trip (scram).
>
>I was told that this design weakness had been identified at the Ignalina
>plant, long before the Chernobyl event, but no corrective action was taken
>at that time.  I understand that the control rods in all RBMK reactors were 

>modified, after the disaster, by lengthening the tie rod (by 1 metre)
>between the boron absorber section and the graphite "displacer" section to
>eliminate the water volume below the graphite section, when the rod is 
fully
>withdrawn.

>Continuing to operate the reactor at 50% full power for an additional 9
>hours put the reactor in a very (axially) unstable state.  The top part
>became "neutronically decoupled" from the bottom part of the core.  So a
>"contributing cause" was the continued operation that put the reactor in an 

>unstable state.  The operating procedures did not forbid such operation.
> Had this mode of operation been analyzed in advanced, this weakness would
>have been identified.  Hindsight is 20/20.
>