In the last week, I’ve spent four days roaming Scotland on a Spirit of Scotland rover ticket gathering data for step offsets and validating theories I’ve had for a little while.
To start, Robroyston is a new station on the Stepps line which opened in December - it’s served by Class 385s between Glasgow Queen Street and Edinburgh Waverley via Cumbernauld and Falkirk Grahamston. Over the last few weeks we’ve been acquiring timings for the station, and I completed the set with 4 visits on Friday and Saturday, so we’ll soon be able to provide reporting times for that station. The signalling layout in that area also means we had to remeasure Stepps in the northbound direction as trains now pass the appropriate signal at a slower speed meaning they now arrive at Stepps around half a minute later.
The other area of interest was the Highland Mainline. The Highland Mainline is one of the most interesting routes to handle from my perspective. It has a range of manually controlled lever frame signal boxes through to modern remotely controlled panels. This means there’s a diverse mix of manual and automatic train reports. Between Perth and Inverness, we currently have automatic coverage at Pitlochry, Aviemore and Carrbridge stations along with the loops at Kincraig, Slochd, Tomatin and Moy from the Stanley Junction and Inverness Highland signalling panels. South of Perth, the trains reappear on the train describer feed at Hilton Junction for services towards Edinburgh, and Bridge of Allan for those towards Glasgow and LNER’s Highland Chieftain.
Last June, I spent a week shuttling back and forth on the Highland Mainline collecting data for use in our step offset tables and I wrote a post about Aviemore in some detail. It’s now around six months since that trip and there are now more ScotRail 2+4 short set HSTs (both classic and Inter7City) and the Class 800 ‘Azuma’ trains have taken over operation of the Highland Chieftain service so it was time to make another visit.
Over the last few years, I’ve been developing a new model to determine the likely offsets at a given station and can also adjust this for different lengths of trains. Offsets at present only work using a ‘typical train’ with no variation for the various different types that exist on the network. This will soon be about to change as part of the next stage of the upgrade to RTT.
Following the upgrade of the processor last month, we’re now moving on to stage 3 which adds support for different offsets based on additional data over and above that normally available from the train describer. The Highland Chieftain and Caledonian Sleeper run trains with a longer formation than ScotRail on the HML which means their characteristics are different from the norm.
I’ve taken several trips on the Chieftain this week to collect data to validate against the model, and also provide a basis for offsets if the model proves to be invalid, and will be able to provide a ‘long train’ offset soon for those stations by applying this to just LNER and Caledonian Sleeper services. Unfortunately, there’s no public data available about what’s operating a service so the ability to apply this in other areas will be challenging - but it’s there to prove the functionality works.
At Pitlochry, long trains show up a fault with our offset processing. Longer trains have to go beyond the departure signal, stepping accordingly on the train describer, and therefore have a computed departure time before the arrival time resulting in the processor determining it passed the location. This is also a commonly reported fault at Inverkeithing. The new ‘long train’ offset approach has an approach to mitigate this which will be making its appearance in due course.
Updates to the Highland Mainline and the addition of Robroyston will be available in the next few weeks once we’ve had the time to process and review the data collected through our normal auditing process.