The Underground is Closed

The persistent closing of large parts of the London Underground (Tube) system at weekends must stop. The chaos, confusion and overcrowding it causes on the odd routes that are open has to be experienced to see how bad it is. 

I first wrote about this in 2011 and here we are, 15 years later and nothing has changed. In fact, it's got worse. The Piccadilly Line regularly gets shut down all weekend to do new train testing. One might wonder why they can't do it at night and one might ask why it takes so long. Are they so afraid of their engineering that they really don't trust it without years of closures to try it out?

London Underground 2024 Tube Stock being tested on the Piccadilly Line at Turnham Green one Sunday during the summer of 2026. Although the station is in use for the District Line, the platforms on the Piccadilly side have been taped off to prevent public access. Photo: Fortek67.

Actually, it's not their engineering, it's the suppliers' engineering. Since the trains were designed in Germany. Transport for London's engineers have little idea of what's been done and whether it meets their strict London-centric standards. Gone are the days when the Underground's engineers designed their own trains, using their long experience of operations in London, ensuring that it would work pretty much out of the box and getting it into service within a few weeks of delivery.

I feel sorry for the tourists who will go home saying, "Don't visit London, the Tube never works at weekends." I feel sorry for the staff who have to answer repeated questions on alternative routes. I feel sorry about the loss of income for shops and businesses.

Whatever happened to the "customer led railway" we hear so much about? We seem to have reverted to the old "operator led railway"

There must be a better way of managing the railway's renewal programme.

Railway Technical Website Update

We are modernising the Railway Technical Website. There is a new look and new pages. Many pages are being updated. We are gradually archiving the old pages where they are likely to be useful.

While we will try to get everything in place correctly as we update each page, there are bound to be some minor teething problems. Please let us know if you find any errors we have missed.

Thank you for visiting us and for your understanding during our refurbishment programme.

Are standards for the railway too rigid?

Maybe. A case study on door operation.

Sometimes, it is worthwhile to re-examine your standards. Is your use of standards too rigid? Are you locking yourself into over engineering your railway? Are you sticking to the rules because they are the rules? Are there opportunities to gain time, reduce costs, improve visual impact or upgrade the passenger experience by adjusting standards, re-analysing risk or getting a derogation? This article offers a small but important example of how a standard could be modified to good effect.

This case study, first written in 2018 expressing concern about the rigid use of standards, is on train door operation. If we assume door opening is activated by a member of the train crew, we would want to ensure that the doors remain closed until the train has stopped. Once wheel stop (zero speed confirmed) has been detected, the on-board system can unlock the door opening controller and the crew member can release or open the doors.

Waiting for Doors

Fine? Yes, as far as it goes. It is definitely safe but this arrangement causes a considerable lag between wheel stop and door opening. Time is needed for wheel stop to be detected, for a release indication in the cab to appear, for the train crew to operate the release buttons, for the door controllers to receive the signal and for the doors to begin opening. On many main line trains, it can be 7-8 seconds before the doors are opened for passengers. On a metro or suburban service, it can be 3-4 seconds. An extra four seconds at every station on a line like the London Overground route between West Croydon and Highbury & Islington adds 76 seconds to the train’s journey time. This is half a train path and buying a peak-hour train path is expensive, so both time and money are being wasted. Perhaps there's a better approach? Perhaps we could use human experience to inform automated system protection?

In the days of guard door operation on the London Underground (finally abolished in 2000), there was no interlocking of doors with train movement. Doors could be opened while the train was moving. They weren’t of course (usually) but the facility was used to speed up station operation. The experienced guard knew that there was always going to be a lag between the train wheels stopping and the doors opening fully and he knew also that station stop time had to be kept to a minimum. To keep dwell time short, the guard would press the door open buttons a second or two before the train stopped so that the doors were beginning to open at wheel stop. Valuable seconds were saved and no one got hurt. So, why can't we replicate this in an automated system?

Speed Detection

Could human experience be replicated in automation? Of course it could. It fact, it already has. In January 1980, the first D Stock train went into service on the District line with a new door control system that kept the train door control locked while the train was moving but which allowed the guard to release the doors if train speed was detected below 4 mph. It replicated the human experience that allowed time to be saved at stations. It remained in use for many years but the idea was lost in the safety panic that ensued after the Kings Cross fire of 1987. Perhaps, now, with efficient train throughput at the top of the railway agenda, is the time to review the rigid zero speed restriction on door opening and make better use of the technology available for the digital railway. And, are there other examples where this approach could be beneficial?