Examination of Witnesses (Questions 19840
- 19859)
19840. When you say it is "alternative",
is it not going to happen?
(Mr Thornely-Taylor) If
that is what is in the agreement, it will happen.
19841. Why can you not put 15 metres down straightaway?
(Mr Thornely-Taylor) Because
the tunnel boring machine is constantly advancing and you need
to be able to get the tunnel lining segments immediately behind
the newly excavated bare tunnel—which is in danger of causing
more settlement than we want to achieve if you do not get the
segments in as quickly as you can—so you need to get the
train continually advancing up behind the tunnel boring machine.
The way you do that is to put in new track. Six metres is quite
long but I am calling it short increments.
19842. Is there a difference in performance
between 16 and six?
(Mr Thornely-Taylor) Only
that we need to avoid joints underneath the Grand Central Studios
and the longer the rail the less of a problem dealing with joints
is.
19843. What is the improvement then if you reduce
the number of joints?
(Mr Thornely-Taylor) We
did see two slides yesterday—perhaps we only actually saw
one—but there is another one that shows what happens with
a joint and that took us to NC30.
19844. The presence of a joint takes you to
NC30?
(Mr Thornely-Taylor) Yes.
19845. How does that match up with NC 25?
(Mr Thornely-Taylor) It
does not. We are going to have to undertake to have no joints
underneath Grand Central Studios after the tunnel boring machine
passes through and that includes the period when there are these
temporary six-metre short lengths.
19846. Thank you for that. Can you help me with
this, Mr Thornely-Taylor? What do you claim is the uncertainty
of your individual third octave band level predictions?
(Mr Thornely-Taylor) We
saw on yesterday's slides little "t" marks which showed
what happened with band five to each band. I explained that actually
the consequences of the sort of uncertainties we were discussing
yesterday are that peaks in the predictions occurred in slightly
different frequencies than the ones you think they were and it
is not so much that they go up, but they appear in a different
part of the screen where the NC curve might be more demanding.
Again, that is covered by the kind of uncertainty that I talked
about yesterday.
19847. There is a constant 5dB on each and every
individual prediction?
(Mr Thornely-Taylor) That
is how it is presented.
19848. That is not quite what I had asked you.
(Mr Thornely-Taylor) Yes,
it is the way it is addressed in this process.
19849. Forgive me for just wanting to be clear
on this, you are saying that for each individual third band octave
level prediction that you have encountered, every single one of
them is plus and minus 5dB?
(Mr Thornely-Taylor) My
evidence yesterday was that it is plus 5dB—might be more
than five because sometimes we do over-predict how railways will
turn out.
19850. But on the plus factor it is five?
(Mr Thornely-Taylor) Yes.
19851. On each and every occasion?
(Mr Thornely-Taylor) As
we saw in yesterday's evidence.
19852. Thank you. I think you know Dr Hunt,
do you not?
(Mr Thornely-Taylor) I do
very well. I regard him as a friend, I hope he will continue to
be such and he was kind enough to present my papers for me at
a conference in Lisbon because I had to come home to proceedings
like these.
19853. Yes, he told me about that. It was the
roughest ride he ever had!
(Mr Thornely-Taylor) It
was a very interesting piece of new work and we had a long briefing
session, and I understand he did extremely well.
19854. You know him well and, indeed, you are
friends. I have misunderstood the tenor of your evidence, you
are not seeking to rubbish what he saying, are you?
(Mr Thornely-Taylor) I gave
a lot of evidence in chief yesterday about the interpretation
of his work and his findings in the context of the Crossrail case.
He was clear that his model is not suitable for Crossrail. His
particular interest is in comparing models and he talked a lot
about weather forecasting models. I am very pleased to say we
have not got violent, swirling clouds underneath the Grand Central
Studios. We have not got the forecast where clouds will be at
any particular time, it is actually a bit easier than that. He
was quite right in saying that if you do not know important perimetres
in modelling, the consequence of that is a difference between
what you predict and what happens, but it is very important to
interpret that in a way which is directly relevant to the case
in point, which is advising the Secretary of State whether he
can commit to levels which will be delivered in the event with
a minimum risk of failure.
19855. Certainly he did talk about weather forecasting,
but his evidence was not limited to that. He was talking about
mathematical models.
(Mr Thornely-Taylor) Exactly,
his particular interest is modelling and particularly comparing
different models.
19856. He also applied his evidence to mathematical
models dealing with sound and his analysis of many mathematical
sounds, if I can put it that way, was they do not purport to be
accurate to any greater degree of accuracy than 10dB. That was
the tenor of his evidence. That is on a wide basis of academic,
rigorous analysis of models.
(Mr Thornely-Taylor) It
was but I explained in chief that he is quite right. When you
get this phenomenon that I mentioned a moment ago, that a peak
comes out in a different place in the spectrum, the consequence
of that is that in any one frequency there is a big difference
because what was a peak was moved to the left or the right and
left in its place a trough but always when it comes to designing
a railway, we are addressing commitments given either in terms
of the familiar LAmaxS that we have been talking about for the
last year, or the special case we now have of NC curves. In both
cases once you turn the predictions into assessments against the
NC system or the LAmaxS system, those differences caused by peaks
drifting to the left or the right virtually disappear. We saw
a slide yesterday about validations at Greenwich which Dr Hunt
has expressed concerns about. Once those are expressed in terms
of NC the difference between measured and predicted drop right
down to 1.3dB, much less than the five that we are currently using.
19857. What I understood you to say yesterday,
Mr Thornely-Taylor, was that even if you take Dr Hunt's 10dB as
opposed to your 5dB, you can cope with that, is that right?
(Mr Thornely-Taylor) The
slide we saw yesterday for the noise level in NC terms from the
temporary construction railway with no rail joints in the studio
was NC20. With the 5dB uncertainty, it was NC25 which was the
basis of the draft undertaking but if you were to say there is
an uncertainty of 10 it only takes you up to a prediction of NC25.
19858. Would you be prepared to adjust your
standards to accommodate Dr Hunt's 10 dB since you appear to be
saying it does?
(Mr Thornely-Taylor) No
further work is necessary. We could show it again if necessary
but our prediction for the temporary construction railway is NC20,
De facto 20 plus 10 is 30. Sorry, it is NC20 plus the 5dB
uncertainty, so you increase the 5dB to 10dB De facto,
it is NC25.
19859. You are saying, if I understood you correctly,
in effect in practical terms, you can accommodate Dr Hunt's observations,
can you?
(Mr Thornely-Taylor) The
probable outturn is that it will be better than NC25.
| 
