Motto: “I am not the Messiah!”, Brian
For quite some time I wanted to write a post about two articles published in the Permanent Way Institution Journal, co-authored by David Marriott and me, articles referring to the “inherent standard deviations”. I’ll try to do that but first I will need to tell you, dear reader, how my point of view on this concept of “inherent standard deviations” evolved over time.
I want to be sincere and open with this story so I will share here my personal uncensored thoughts. Please excuse me, dear reader, if they sound disrespectful, judgemental, offensive or odd in any way. I feel the need to write them like this for you to see the story through my eyes. Here it starts:
Call me Constantin.
I am originally from Romania and there the track is measured using a French-inspired versine measurement system, and the way the track (riding) quality is assessed differs from the one used in the UK. Although, before moving to the UK, I was involved in a research project where I learned a bit more about the inertial measurement system, about track measurement Standard Deviations and other similar concepts, my mindset was still ‘en Francais’, not fully embracing this smart system of measuring track quality. In fact, as you will see, these concepts were still not clear to me.
Since my first projects designed in the UK, I got to use a small software called the TGSD Calculator, developed by a certain David Marriott. This software – I was told – could tell how good my design was. It was calculating a set of standard deviation values and I had to make sure they were below what the standard limits for these were at the time.
I was puzzled by how small this software was – a bit more than 1 Mb and it could tell how good my design was!
That’s quite amazing, isn’t it?
I developed a software for calculating cubic parabola coordinates and it was double this size…
As I don’t like to work with things I don’t understand, I asked around to find out more about this amazing software. The explanations I received initially didn’t satisfy me enough, they even frustrated me, because, nobody knew exactly what the software does – I heard only superstitions, myths, stories and not well justified engineering science.
I sent an email to the software developer asking some ‘smart’ questions based on the superstitions I almost started to believe. No answer received. Later I realised my questions were quite silly and did not worth an answer. Probably David Marriott received hundreds of emails like that over the years.
The software help file mentioned something about a Butterworth Filter but at the time this section was cryptic to me, although it tells what this software actually does.
I started to dig for information in a database of standards and I discovered a few old documents that provided a bit more light into the issue of what the TGSD Calculator magically does. These documents tell stories about track recording vehicles recording vehicle accelerations and that any irregularity is an unnecessary acceleration. The advice given to the designer was to adjust his design to eliminate the unnecessary vehicle acceleration caused by his/hers design. And the story tells that the TGSD Calculator can actually show you where those unnecessary accelerations are.
My notes on the documents I found record some early doubts, mainly on the cant, curvature and centrifugal forces. But my critical eye closed to these doubts and, by using this TGSD Calculator quite often, I slowly became a believer, because these documents were making sense to me. I’m sure whoever wrote them did it in good faith but, as I will find out four years after, these documents created a TGSD Calculator magical mythology because … Let’s not spoil this now!
I started to believe this mythology – that this magical piece of software can calculate these unnecessary vehicle accelerations that appear only where there are alignment changes. Where the elements are constant – straights,curves, cants, even continuous transitions – no unnecessary accelerations appear. As soon as the alignment element changes, these demoniacal accelerations suddenly mess up the design, even though it complied to all the design rules we work with. My belief was that this fine tuning of the Calculator was due to some dynamic modelling and that on constant elements the accelerations are whatever they are – you cannot change them – and that this dynamic model was calibrated with what the track recording coach produces. I believed that through this magical identification of these accelerations, the TGSD Calculator can tell how good a design is.
What enforced more my trust and faith in the software was what I got from the raw data export function of the software. By reading this output I could see some waves that the software was modelling at every point where the design changes.
Only there and nowhere else! (Only there, my friend! Nowhere else! Remember this!)
Ooooo! This software is so smart!
These waves are, I thought, magic dynamic responses to those unnecessary accelerations of the vehicle when running over my design. I started to call these “inertial waves” and “damped oscillations” and I kept referring to them like that. (If it looks like a wave it must be a wave, isn’t it?)
This concept, of “inertial waves”, came to me naturally because the system of measurement is called inertial. And that graph looks like a “damped oscillation” – so it must be one!
You see, I contributed also tho the TGSD mythos – on top of the “unnecessary accelerations” invented by someone else, I added my own “inertial waves” and “damped oscillations”.
I keep on using here the words magic and mythos because they are the right words – as long as you don’t fully understand something, as long it’s logic doesn’t fully makes sense but you still believe it, that’s magic, mythology, superstition, faith, not science. Pseudo-science the most. But its nicer to call this “magic” and “faith” than “pseudo-science” or “pseudo-engineering” … it sounds almost honourable!
For cant it was something kind-of wrong as it did not correlate at all with the horizontal alignment. If you read these lines, dear reader, you for sure know that there is a strict correlation between the cant and the horizontal geometry. Well, in this TGSD Calculator the same cant alignment was producing exactly the same SD no matter where it was applied on a horizontal alignment – on straight, normal curve, reverse, or on any kind of silly geometrical shapes.
This is a thing that puzzled me but my magical explanation was that the software models the compliant change from one element to another and was not designed for the silly tests I was subjecting it.
These dynamic-modelling-like graphs gave me the Faith!
Especially because they looked similar to other dynamic modelling results I have seen – damped oscillations, inertial waves … you know!
So, I started to fully believe in this software!
As many others, I joined the TGSD Faith!
I really believed that the TGSD Calculator is the Saviour of the design track quality!
The railway track Holy Grail has been found!
I started to sacrifice my design judgement for better and better SDs!
But I have never, oh, never, achieved perfection!
I never got the SD to be ZERO – the Nirvana of the Design Track Quality.
All my design changes were still imperfect, still generating those “inertial waves” and non-null SDs, no matter how hard I tried.
As soon as I became a real believer in this magical software, I called my Railway Track Professor and dear friend to tell him about this tool I now have. I explained him what I believed the software does and I sent him some of the old documents I found.
“Well, Constantin, this doesn’t make sense! I don’t think this thing does what you say it does!”
I remember him reminding me that cant deficiency, although measured in mm, is in fact an acceleration. Above a certain value this non-compensated acceleration is unnecessary, even dangerous – but the TGSD Calculator was ignoring it completely.
Anyway, no matter what I told him about this TGSD Calculator, I could not convince him.
Eh! Either I didn’t explain it very well or he actually doesn’t understand this … magic … most likely the later…
At some point someone mentioned that this David Marriott disagrees with how his software is used and reacts a bit odd when discussing it. He might have lost his Faith?! He might have been jealous of not making enough money from developing this great magical software?! Who knows?
The time passed, I used this software more and more and my early doubts were long forgotten.
My faith grew that much that I became a prophet of this Faith – I invented clever explanations for what the software does and how smartly it calculates these magic “inertial waves“, I taught my younger colleagues how not to sin against the SD, I gave a few “very smart and well justified” presentations about what I believed the TGSD Calculator does to my colleagues and even publicly to a section of the Permanent Way Institution.
And then, four years after my first contact with this software, I was invited to give my clever presentation at a national PWI conference.
But, you see, the engineer who invited me said that also at this conference was supposed to present Mr David Marriott, the engineer who developed this magic software. And he didn’t agree with a few of the things I had in my own presentation.
How could he?!
I was a believer! I was spreading the Faith in the great tool he created!
I was advised that it would be better for us to have a chat about this subject to avoid a public confrontation on the subject of the design (inherent) standard deviations.
I got David’s email address and phone number. I thought before contacting him that he might be upset that I revealed some of his magic that makes this software work. Nothing was further from the truth …
We exchanged a few emails and then I called him and we had a 90 minutes discussion that ruined everything.
This guy, David, demolished all the well-constructed magical justifications I had for what his software does.
Because, dear reader, that magic does not exist!
I’ll spoil a bit of the next post on this subject with one of the very first questions David Marriott asked me when discussing about the TGSD Calculator:
How can this software be about accelerations, dynamic waves and oscillations when it doesn’t have any speed input?
I thank him again today for opening my eyes then.
Thank you, David!
To be continued …