[SystemSafety] Two Process-Industry Accidents

[Peter Bernard Ladkin]

Charles Perrow just circulated the following news article about the NRC and the behavior of its
information office just after the Fukushima accident from Bill Dedham of NBC news on 2014-03-10:

http://www.nbcnews.com/storyline/fukushima-anniversary/u-s-nuclear-agency-hid-concerns-hailed-safety-record-fukushima-n48561

I find it odd that information officers, people whose job it is to keep themselves and then us
informed, were hindered from obtaining information by having access to Twitter and YouTube
restricted. Not that that would have worked - they could have tuned in to the BBC Live Feed as I
did, which summarised relevant Twitter traffic. And they probably could just go home and tune in
anyway. But just the fact that somebody decided to keep stuff from their own information officers is
very peculiar.

but I find appalling the deliberate attempt to disown their own risk study.

The NRC is supposed to be an oversight organisation, one which assesses safety amongst other things,
one of whose tasks is to inform the public about scientific and engineering facts, including their
own studies. This shows their information department behaving like a lobby group, trying to hide
info. I find that inappropriate. Surely other NRC personnel do too?

Perrow also pointed privately to a recent article on the Chevron Richmond accident in 2012:
http://www.truth-out.org/news/item/24421-to-serve-the-public-or-to-serve-themselves-the-us-chemical-safety-boards-turmoil-is-playing-with-public-safety


I know the area well, as a former Berkeley resident. The US Chemical Safety and Hazard Investigation
Board (CSB) suggested introducing a safety-case regime in order to improve the safety record of US
process industries. That ignited a controversy. We had a lively discussion of this in January 2014,
starting with a link to a newspaper article in http://www.systemsafetylist.org/0772.htm  Martyn
suggested that indeed US industry might benefit from the UK's work and experience
http://www.systemsafetylist.org/0773.htm and Mike Holloway posted a series of links to the extensive
public consultation over the safety-case proposal http://www.systemsafetylist.org/0774.htm .

Nancy registered her arguments against safety-case regimes for US industry
http://www.systemsafetylist.org/0780.htm as well as noting she was in contact about it with
interested parties.

Apparently the whole affair has got so fussy it has attracted the attention of a Congressional
committee chaired by Darrell Issa, who is the Representative for the 49th District, northern oceanic
San Diego county, a ways away from Richmond.

Prompted by reading the article above, I've just been through the collected public comments on the
CSB draft report, referenced by Mike Holloway.

The engineering situation at Chevron Richmond seems pretty clear. The physical causality was
corroded pipes bursting, along with the presence of an ignition source (just one of the commentators
addressed the safety issues connected with the presumed ignition source). It seems the physical
circumstances - the condition of the pipes - was known. But it seems the company hadn't done
anything about it. So the issue at stake, as one might think at Texas City, seems to be
* how to get known hazards officially recorded in some way, and
* how to get companies to implement elimination/mitigation measures to these known hazards.

There are also related issues, such as
* how to get all hazards identified effectively (so that they are all "known")

Nancy's paper in the SSS in-house journal is directly referenced in the comments by the American
Chemical Council, an industry group; in a joint letter from the Americal Petroleum Institute and the
Western States Petroleum Association; and by John Bresland, a consultant who was former chair of the
CSB and who includes a copy of the paper, along with another report by Sam Mannan of Texas A&M, as
an appendix to his comments. The paper appeared end 2011, well before the accident. How it came to
these organisations' attention is not apparent from their comments. When I was in the SSS a decade
ago there didn't seem to be many, if any, process-industry members. They were mostly US military
contractors (which is where the term "system safety" was popularised).

It is not clear to me what the issue, of getting known hazards officially recorded and getting
organisations responsible for such plants to handle (mitigate or eliminate) those hazards, has to do
with engineering per se. It seems obvious to me that this is a topic for organisational
sociologists. I don't know how engineers who are inexpert in organisational behavior would have
anything worthwhile to say. The engineering issues per se seem to me to be clear: the pipes are
known to be suspect; so inspect them more intensely, and replace when necessary.

It struck me that few of the people who commented on the CSB report (on their own behalf or on
behalf of organisations) demonstrate apparent expertise in what it takes to get large organisations
to rectify known hazards before accidents occur. I do sympathise with the residents who comment that
they really don't want this kind of thing going on on their doorsteps and support any effort to
ensure it doesn't.

Many commentators appear to be communicating opinions about safety cases without really knowing what
one is (cf. http://www.systemsafetylist.org/0822.htm ) but nevertheless being pretty certain they
don't want them.

Many point out that there is no clear statistical evidence that safety-case regimes "work" better
than, say, the California status quo. Yes, UK, Norwegian and Australian operations have demonstrably
far better safety records, but there are all sorts of confounding factors (safety cases weren't the
only things introduced to improve safety performance). The response to that point seems to be
equally clear: introduce *all* the measures that have led to better performance, not just safety
cases alone. As Martyn said, there is a lot of specific UK expertise. And Norwegian and Australian
also, we may presume.

There are some useful articles which we didn't reference during the discussion in January/February.
An article by Rena Steinzor, a legal scholar at the Uni Maryland:
http://lawdigitalcommons.bc.edu/cgi/viewcontent.cgi?article=1695&context=ealr  and a review by Peter
Wilkinson of Nancy's paper on safety cases, on the CSB WWW site:
http://www.csb.gov/assets/1/7/Wilkinson_Review_of_Leveson_Paper.pdf . Dewi Daniels also suggested
that he had reviewed Nancy's paper http://www.systemsafetylist.org/0799.htm , but he hasn't shared
his review with us (yet). Maybe he could?

I don't think Peter Wilkinson is on the list. Could someone here who knows him invite him to join?

PBL

Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de