[SystemSafety] Ladbroke Grove account

[Peter Bernard Ladkin]

On 2019-10-17 16:42 , Andrew Banks wrote:
> The sad saga of both Labrooke Grove and Southall (where drivers passed signals at "danger") are sadly not the last time human behaviours are seemingly disregarded.  While the incidents were complex, the simple fact is that both crashes would have been prevented had drivers obeyed the signals... 

Ah yes, reminds me of some years back when Carl Sandom's IEC New Project Proposal for a report on
Human Factors and System Safety was made. It came to the German National Committee and one gentleman
from the process industries said "why do we need this? The kit is there; how to use it is described
in the handbook; if somebody doesn't do what it says, they should be disciplined or fired. What else
is there to say?" My answer: quite a lot, actually. The handbook can be incomplete or misleading.
The data on which to make a decision may not be manifest to the decision maker. The decision
algorithms might be wrong. The human capabilities needed, including biorythmic assumptions, may not
have been appropriately assessed by the equipment designers, installers or operator organisations.
Those are all system properties (since the operator is part of the system) and can be assessed. But
it is of course much easier to close your eyes and fire the operator. Until no one wants to work for
you any more......

It is manifestly not so that "human behaviours [were] seemingly disregarded" in the Ladbroke Grove
reports. Cullen goes into considerable detail on the human behaviour, including the planning, the
assessment, the sightlines, other drivers' opinions of the signalling and track arrangement, and the
incident handling and "organisational learning" (SN 109 was known as a "black spot" - Cullen Part 1
5.22).

As far as "drivers obey[ing] the signals", that suggests a line of communication between driver and
signal that Cullen was convinced was not there: "I consider that I can also put aside any idea that,
for whatever reason, driver Hodder passed SN109 at red in the knowledge that this aspect was being
displayed. On the contrary I have no difficulty in accepting that he believed that he was driving in
reliance on having received, at some stage, a proceed aspect, that is to say, an aspect other than
red. (Cullen Part 1, Para 5.5)"

Cullen Part 1 goes into considerable detail on the siting and the sight lines of SN109. Even the
organisational assessment and learning part: "Dr Ian Murphy, a lecturer in mathematics at Glasgow
University, who has made a study of various aspects of risk in the railway system over the last nine
years, provided a report for the Inquiry in which, on the basis of past experience in the years
1996, 1997 and 1998 and the evidence of an underlying annual frequency of two SPADs at SN109, he
estimated that there was an 86% chance in a given year of at least one SPAD at that signal. Having
regard to the collision opportunities, there was also a 7.2 % chance of a collision in any given
year, or in other words one collision in about 14 years. (Cullen Part 1, Para 5.22)"

1993 was when the sight lines around Gantry 8 were considerably altered (and then adapted).
According to Murphy's calculations, the chance of a collision up to and including 1999, when the
accident happened, were about 43%. That is a system property (including the humans). The pressing
question for me is not why poor driver Hodder passed SN109 at danger on that particular day, but how
Railtrack could have learnt about that system property earlier, before the chance of an accident
became high.

PBL

Prof. Peter Bernard Ladkin, Bielefeld, Germany
MoreInCommon
Je suis Charlie
Tel+msg +49 (0)521 880 7319  www.rvs-bi.de





-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20191017/94416a43/attachment.sig>