Skip to main content
Tag

railway

Hornby BIG DATA Project

Overall Purpose of the Role:

There is a big push to make railways more effective and increase capacity. The main opportunity is to run trains closer together and autonomously. We are building a model railway that can be used to explore possibilities for this. The main theme will be to use an existing model railway set-up that already has digital control. See the set up on https://www.youtube.com/watch?v=AjSHDU2B474

Hornby BIG DATA Project Layout Under DCC Control

Hornby BIG DATA Project Layout Under DCC Control

The idea will be to fit proximity sensors to the track and model trains and tachometer on the trains so that the system knows exactly the whereabouts on the train-set of the trains. The movement of points will be automated based upon the routes chosen for the trains. The trains will be kept apart by knowing their own speeds, lengths and braking distances and those of the other trains on the train-set. There should be some facility to establish that a train formation is complete (i.e. it has not left a coach behind due to an inadvertent decoupling.)

It may be necessary to fit a vision system of the office ceiling to ascertain where the trains are on the network and also potentially on the front of the train to detect objects fouling the line or other trains. We are using Arduino technologies and commercially available sensors, cameras and radio, see the attached appendix.. The finished system will be used as a training tool and a means to experiment with different traffic management strategies and autonomous technologies. We will also look to experiment with detecting the health and resilience of the various components such as the electric train drives using available data such as current usage, noise etc.

Before the project starts we will have preliminary system design worked out however this will need to be detailed, adapted and implemented.

The system is currently controlled by Hornby OO train set and E-Link/Train master Digital Control system. See pictures below.

Main Task to do:

  • To develop technology demonstrator of a moving block automated signalling system based upon a Hornby OO train set and E-Link/Train master Digital Control system.
  • To write any code necessary.
  • To configure electronic and sensors and implement them with the software.
  • The job will involve certain practical tasks such as soldering, electrical testing, using basic modelling hand tools etc.

Bill of Material So Far:

Bill of Material so far

 

Will a data driven approach replace RAMS and conventional safety engineering?

If one looks through the railway safety and reliability standards, EN50126 / EN51028/9 https://www.linkedin.com/groups/4985242 and the recommended activity at each life-cycle stage it will be seen that nearly all the items are data driven. Using Big Data and Machine Learning we could eliminate a lot of the work done now by safety engineers. Of course you would have to follow a more formal system engineering approach to ensure that the data was created and logged properly but requirements management databases like DOORS and system modelling using for example SysML (Systems Modeling Language) would enable that to happen couple with BIM (Business Information Modelling). BIM means that railway design and development is properly documented through the use of CAD.

factors-influencing-railway-rams

The diagram above shows factors affecting RAMS and they are all based upon some form of data either structure/unstructured and real time/historical. They all provide management information that must be acted upon by using some form of decision logic. All of which is amenable to replication by an intelligent machine.

I conclude that RAMS is a now a past-its-sell by date of a commodity. The Fault Tree below might soon be redundant replaced by real data and modelling.

fault-tree