So realistic have the Simulators become they are now even used as a method to convert pilots from one aircraft type to another, often with zero airborne time involved in the conversion. This paper discusses the technical aspects and application of similar technology for training and regularly assessing Signallers in operating complex electronic Signalling Centers. Over 150 replica simulations of each Signaller Workstation for the controlled areas have been supplied in the UK and incorporate accurate models of the interlocking, the exact Workstation screen layout and controls, current timetable, track topology and train performance.
Full-fidelity Simulators are now a regular part of the training curriculum for Signallers as they allow them to practice the routine tasks in controlling trains and crucially, emergency and perturbation scenarios in a safe, off-line and totally realistic environment. The signaller is constantly observed and monitored during each session by an experienced Trainer / Assessor who has the ability to change the session from a normal situation to one involving a failed train, faulty points, signals and track circuits, SPAD, adverse weather conditions etc.
Background and History
Training of Signallers traditionally has been achieved in many countries by having a new and inexperienced trainee shadow a Signaller with a number of years of experience. This was deemed to be “on-the-job” training and was sometimes combined with a model railway and miniature model of the equipment. In the days of lever frame signal boxes and simpler NX style panels this sufficed, as the geographic area covered and therefore workload was fairly limited. However, direct training of emergency situations and handling perturbations in the most efficient manner was still difficult due to not being able to practice these safely on the “real” railway. Training therefore by default was limited to learning the theory and being tested on the rulebook and receiving practical on-the-job experience based on control of normal traffic operation.
Therefore a trainee (and even a Signaller with extensive experience) could go many months or even years without experiencing an abnormal situation or emergency. Many case studies from the railway and similar safety critical industries such as aviation, nuclear power etc has shown that thorough knowledge of the rules and theory provided no guarantee of how a real emergency will be handled and only regular exposure to and practice of these events can achieve a high level of safe operation, predictable emergency handling and service recovery. Therefore the rules and theory classroom based approach is now heavily supplemented by an activity based / hazard recognition method using realistic simulation.
In the late 1980’s the UK installed the first VDU based Signalling Control Centre (SCC) at London’s Liverpool Street and this technology provided a further step change in both complexity and geographic area of control. Coupled with the introduction of VDU based SCC’s came Automated Route Setting where the fundamental tasks of route setting to the timetable and overcoming simple perturbations and delays became increasingly automated. Therefore the Signaller’s environment has changed considerably from the lever frame / panel environment where every movement was directly controlled, to one where the geographic area has enlarged and the normal minute-by-minute tasks increasingly automated.
Increased automation means a Signaller can frequently be in an environment with too little to do for quite long periods of time – but suddenly too much to do when an emergency or significant perturbation or any other special circumstances arises. Therefore fundamental and up-to-date knowledge of the geographic area of control must be retained at all times to ensure a minimal time lag when intervention becomes necessary. With decreasing day-to-day exposure of abnormal operation, a replica simulation of the control method of the real railway is the only method where all emergencies and any other abnormal events can be practiced in a safe off-line manner.
High Fidelity Simulators
High-fidelity Simulators across a number of industries are classed as those that fully replicate the true environment in order for the trainee to be unaware of whether they are in the real or simulated environment. As previously mentioned, such is the realism achieved in the aviation industry that simulators are used as the sole method for pilots of conversion between aircraft types.
With regards to simulators for railway signalling, full-fidelity versions for VDU based control centres are generally agreed to require the following capability:
a) A model of the actual interlocking embedded with identical system response times. This ensures when failures are introduced, the display and response the trainee experiences is the exact same way as the real control system.
b) The display screens and primary control devices (mouse or trackerball and keyboard) the trainee uses are exact replicas of the real ones.
c) The communication methods the Signaller would use are duplicated, ie Cab Secure Radio and Signal Post Telephone
d) The actual timetable for the area covered is run and Automatic Route Setting (if implemented).
e) All track topology (gradients, line speeds, signal sighting) and train performance characteristics (acceleration / braking) are used.
f) The environment is as close to the real control centre as possible, ie the furniture, lighting conditions etc.
g) Ability to introduce any infrastructure failures and perturbations such as failures of track circuits, points, signal lamps, interlocking (whole or modules), train failures, SPAD, non-timetabled freight train etc.
h) Recording and playback of the Trainees performance.
Provision of the above capability a) to f) ensures the Simulator will provide a replica environment for training purposes with the ability to run normal operations experiencing the timetable of any day and specific time. The capability described in g) and h) extend the Simulator capability beyond normal operations into a tool for experiencing and practicing responses to emergencies and any other abnormal situation and having an auditable record of the actions taken.
The first full-fidelity Simulators for VDU based control centres were installed in the UK in the early 2000’s following the publication of a key report noting the lack of an off-line environment for the safe training and regular assessment of Signallers. These early Simulators had the Trainer / Assessor and Trainee located in the same room but it was soon realised this was not an ideal environment – particularly when the need for communications arose between the Trainee and a third party such as a train driver.
With the role of the train driver played by the Trainer / Assessor it was found necessary to provide a physical separation between the two, mainly for realism of voice communication. However, maintaining visual observation of the Trainee by the Trainer / Assessor is vital for proper assessment of situation handling therefore separation of the two by a glass wall is now the norm. This audio separation also ensures the Trainee is unable to detect in advance when a change in situation is about to be introduced by the Trainer / Assessor.
Failure and Perturbations
In addition to the Trainee being able to practice normal routine operations of route setting and despatching, a vital part of the Simulator is the ability of the Trainer / Assessor to introduce abnormal conditions. Therefore whilst the Trainee is presented with identical screens to the real Control Centre, the Trainer / Assessor has a separate user interface of the overall geographic area of control and an extensive set of controls to introduce these abnormalities.
Normally to start a training session the Trainer / Assessor simply selects the timetable to run and a start time within that day. The Trainees screens will then populate with the trains correctly in position for the selected start time and the interlocking states and trackside signals set. Typically, at some time into the training session the Trainer / Assessor will introduce a single or multiple abnormality and this will be immediately reflected in the correct way on the Trainees screens, including any audible alarms that would occur. The interlocking and infrastructure will also respond as per the real environment ensuring the Trainee will experience the exact same control system sequences and visual and audible conditions. Should the response to the introduced situation require voice communications, either the Trainee or Trainer / Assessor will initiate contact with each other with the Trainer / Assessor playing the role of the third party. All voice communication is automatically recorded and time stamped for subsequent replay during the review process at the end of the training session.
In addition to introducing any number of failures during a session, the Trainer / Assessor can create a script of a session in advance. This allows a preset number of infrastructure failures or perturbations to be automatically introduced at times during the session, allowing the Trainer / Assessor to focus on observing the Trainee. It also allows a number of Trainee’s to be all assessed against the same scenarios.
Recording of Session
Any training session is recorded within the Simulator for subsequent analysis, playback and review with the Trainee. This includes all communications that have occurred. A simple automatic report of the Trainee performance is carried out but this is limited to an analysis of how closely the timetable was followed during the session, aggregate minutes lost or gained and number of routes cancelled with trains approaching, Further automatic analysis has been considered to automatically assess the Trainee’s actions during emergencies but it is currently felt this is better served by a combination of observation and playback with extensive review and feedback by the Trainer / Assessor.
Whilst the original safety and business case for full-fidelity Simulators was the training and ongoing assessment of Signaller’s, their accuracy has opened up a number of other uses. This includes the testing of timetable alterations prior to implementation, planning capacity improvements and reduced services during engineering works.
Additional uses starting to emerge include human factors studies around wholly new and expanded geographic area user interfaces – for example the workload effect of allowing a larger geographic area to be controlled on one workstation by a single Signaller during periods of reduced traffic. A further use is to build the Simulator at a much earlier stage of a new signalling scheme. This will allow the projected layout to be tested by Signallers and the capacity tested with the proposed timetable and any pinch-points and high-risk SPAD signals to be identified before the design is committed to.