Derailment at Ealing Broadway 2/3/2016 (and RAIB report)

I thought driver training required to stop short of points, if they were the first train through and check they were correctly scotched and clipped. The RAIB report makes no mention of this.

An example of how not to do things!

Although for this move the points were trailing, they are also facing points for trains leaving Platform 7. Facing points are of course required to be locked but I understand that points which are always trailing do not necessarily have to be locked. As these points were used in both directions presumably some sort of substantial mechanical lock would always be engaged, therefore the blades probably could not be moved by much, if at all, by the train as it trailed the points. It would all depend on the strength of the lock and the weight of the train.

It is true in days gone by that trailing points did not need to be locked but this has now changed in the last 10 to 15 years. As the A end is facing and trailing then a full set of points are provided. The forces going through the mechanism are so great if they are being pushed by the train in the wrong direction then something has to give. I do have pictures of the points before and after and they were in a right mess. The train actually came off at the back of the points and the wheel went straight through the middle of the points and nearly cut everything in half.

I would suggest that a number of factors combine to determine the likelihood of a derailment occurring when a train tries to force its way through a set of points locked in the other position. These factors would include train weight ( the heavier a train is the less likely a derailment is), the train speed (I suspect that a derailment is more likely at very low or very high speeds than it is at medium speeds; at low speeds because of the train's lack of momentum). The third factor, I suggest, would be the geometry of the points involved; points laid out for higher speeds may well be easier to trail through because the initial contact between wheel and switch rail is at a shallower angle.

There was already a train in platform 8. Paragraph 40 of the report says that the signaller was unable to set a route for it to depart. (That is also why points 38 had to be reversed - they were still set normal for the route that previous train had taken into the station)

Also, they didn't know there was a knowledge gap - it was an "unknown unknown". Moreover, even if platform 8 or 9 had been empty, getting there would have involved negotiating three sets of points (38, 40B, and either 40A or 41A), instead of just the two (38, 39A) at the slip. (40/41 are the scissors crossing giving access to platforms 8 and 9) All three would have to be scotched and clipped before the train could proceed.

Can I reiterate a point made some weeks ago on a related thread where the question was asked as to what training any of the staff would have received on a slip set of points. From the RAIB report it seems the answer is none, which is what I originally suspected. One cannot blame the staff for this as having a model railway isn't one of the attributes LU seeks as part of its recruitment process!

Unfortunately I think these sorts of incidents will happen more and more in future years if LU continues with what appears to be a policy of removing knowledgeable old-school staff from the network. Local training on both trains and stations is paramount, and there is a lot of benefit in knowledge passing down through generations of staff. The Station Supervisor should have never been placed in the position where he simply wasn't adequately trained to carry out a task expected of him.

The only criticism I would direct towards any of the staff is the "heated" nature of at least one of the conversations. The controller involved simply should have known better, especially as he evidently didn't know the answer to the question either.

I suspect that no matter how much time you spend playing with 'Hornby' if you were asked at 1am to go and clip a set of points in between rain storms - you probably wouldn't notice that another set of points (that nobody told you to look at) wasn't set correctly. I know I wouldn't - I'd be taking care not to slip on a wet sleeper or trip over the various bits of 'track' because not only are point clips a little heavy, they make brilliant electrical conductors.

It is all very well to state that a walk of the route would have prevented this, but it was dark, it was raining, and a new track arrangement in operation, as well as the joys electrified rails.

While the BR rule book (1950) I was tested on does not say it, but I was told that before authorising a train past a signal the route must be walked to check the position of all the points, in cases of failure of signalling apparatus.

As for the heated exchange... If I called up a signalbox and the signalman informed me that he had no indications as to the positions of his points, I too would be incredulous! But then again I have only worked in mechanical boxes where we have the box levers and also the supplementary electrical detection of the position of the points and signals. I find it bizarre that these power boxes have no indicators, now I know why it takes so long to get things moving on the modern railway when signalling apparatus fails! No holding flags out of 'box windows!

Anyway I presume that there is not much weight on the leading bogie of a D stock... In my experience the point drive rods, locks etc. usually deform when the point is run through!

Power box diagrams do show the normal position of points and they show the actual lay of the points at any given time, with knobs available to manually set points.

Earl's Court control room's overview diagram does not show the normal position of points nor does it show the actual lay of the points at any given time. Points cannot be manually set individually from Earl's Court control room, which it must be remembered, is of a regulating room design, rather than a PSB design.

If my interpretation of the report is accurate, it seems that the signaller only had an out-of-date diagram to refer to, not realising an up-to-date diagram was present. The signaller knew his diagram was out-of-date and wasn't sure about the exact current layout out on the ground. I understood that that was partly why he didn't recognise 39A points as being part of the route and why he became confused and why discussions with the controller went the way that they did. Had he had point indications, I myself don't see any reason to suppose the signaller would have been unable to identify 39A points as being part of the route. But then, of course, we can't say, nobody can tell. So best not go there, but I certainly thought the report suggested that the signaller wasn't aware of the up-to-date paper plans present in the control room and I'm not certain that the report supports the view that the signaller was unable to interpret the diagrams, although you are right, of course, that he doesn't seem to have been trained on slip points.

I saw this quote from the report... "understand that the signaller had no means of knowing the positions of any of the points under his control," but then I also thought that this was true "Power box diagrams do show the normal position of points and they show the actual lay of the points at any given time, with knobs available to manually set points." So WHY did the signalman state he had no indications?

I have known a few signalling inspectors and they would have exploded if they found a signalman unable to understand N/R indications and not being able to give instructions as to the positions the points need to be in.

O.K. I accept that two versions of the track plans being available in different places, can cause a lot of confusion... and I also accept that I would not assume that all railwayman would understand those diagrams and understand what Normal and Reverse means! In conversations with staff I have resorted to "are the points set for x platform / road"

This is getting a bit off-topic.

It is true that the "service operator" (referred to as the "signaller" in the report) has no means of knowing the positions of any of the points under his control. (Actually those more knowledgeable than me will able to let me know whether or not Trackernet shows this information. I have a feeling it does, but Trackernet is not to be used for safety critical decision making.)

This is how Earl's Court Control Room looked in 2005. You can see in the report (for example Figure 7 on page 16) that the lay of the points is not displayed on the overview diagrams. Indeed, track circuit indications flood at Earl's Court. The diagrams are also not designed to show the normal position of points. For example, you can clearly see in Figure 7 that 39B points are shown with the road from platform 7 merging smoothly with the eastbound main. 39 points would actually have to be reverse for this. So it's obvious that the diagram is not designed to show all points in the normal position, allowing one to infer what's normal and what's reverse.

This (by way of example) is Derby PSB. You can see that all points are drawn in their normal position. This is not the case at Earl's Court. Along the top you can see a row of knobs allowing manual control of individual points. No such control is available at Earl's Court. This knobs have lights confirming the position of all points at any one time as you can see. No such point indications are shown at Earl's Court. You can see that route and track circuit lights do not flood, but illuminate along the route that is set. In this picture, you can see the red lights of 1V50 are illuminated along its precise course. By contrast, this is Earl's Court. Notice the Piccadilly line train on the westbound at Down Street just passing signal PB3. Observe the flooding of track circuit AG. Is this train heading for Hyde Park Corner or Down Street Siding? We cannot tell by the diagram. What is the lay of number 6 points? We do not know.

Earl's Court is not a Power Signal Box. It opened as a regulating room. Day to day signalling is done by program machines. The primary role of the service operator (regulator) was intended to be supervising the program machines, stepping them, pausing them and so on. The routesetting buttons were designed for emergency use and were initially housed in draws. They select full routes from signals not on an NX basis. Observe this picture. If I wish to set the route from Boston Manor to Northfields platform 4 for via signals WR1^A, WR1^B and WR1^C (main aspect only), I will press the WR 1⁽¹⁾ button. If I wish to set the route to Northfields platform 3 via signals WR1^A, WR1^B and WR1^C (with harbour lights) I will press the WR 1⁽²⁾ button. I have no control over points.