2014 Aurecon recruitment banner


Rail signalling: Q&A

Rail Signalling

In this roundtable, four of our leading rail signalling experts in both passenger and heavy haul rail networks, talk about the issues facing the industry.

The project management 'pain point'

John Cranley: I don’t know any signalling organisation that is particularly strong in project management. Throughout their career, signallers are trained to be risk averse and have one way of doing things which tends to inhibit the development of good project management and of course innovation. Proven safe designs tend to be the focus. Because there are not many pure signalling projects, the majority have to integrate with other disciplines; we really need project managers that can see beyond pure signalling to effective systems integration.

David Ness: However, let’s remember rail signalling is a safety environment, you can’t apply project management theory in an uneducated way to the discipline. If you have multidisciplinary business, you have the opportunity to bring project management skills from other industries such as defence, and temper with strong business skills.

David Radcliffe: Like software projects, signalling projects often run late and have similar challenges - logic based projects that you are trying to apply in a wide variety of environments and satisfy a wide variety of stakeholders. There is an opportunity to bring the project management techniques of the software industry to what we do.

Malcolm Lauder: You tend to get two types of project managers that look at signalling - technical professionals who get bogged down in technical detail or project managers who have no understanding of the technical. There is a risk in both, as rail signalling is a system, each part affects the other, sometimes if something happens you have to go right back to the beginning. An expensive and time consuming process.

John: Professionals on a project can be scared of signalling and leave it to technical guys, and it is often at the tail end of a job when everyone else is already late in their delivery, so signalling gets squeezed. Problems need to be identified earlier on in the job, but if the senior managers on the job don’t understand it then this won’t happen.

How will technology affect the delivery of rail signalling projects?

David R.: The technology is changing and the way signalling projects will be delivered will change. The tail end delivery will be affected by new technology - software based systems such as ERTMS or positive train control or the next generation beyond that, which is likely to be more communications based, involving less way side infrastructure, more GPS based and more computer based. This will make it more of a ‘black box’ in some senses. The good side is that you don’t get interfered with by other infrastructure, you don’t need to worry about putting cables down or putting signals next to the track, however you do need a very accurate understanding of what the track looks like to create the virtual models to do the signalling design and implementation work. Signalling will become more independent in some ways from the rest of the project and more technically based, moving further away from civil engineering.

Malcolm: I agree to a certain extent, but you have to understand the track lay out and the interlocking will always have to be designed with civil engineers, the track engineers and everyone involved in the track lay-out. Where those requirements are not known upfront, that is what causes the delays. Signalling will always have to respond to the operators’ vision for the railway. With new technology, the signalling will be in-cab, but how the railway operates is defined by the operator - the number of passenger trains or amount of tonnes they want to transport are typical drivers and the interlocking has to service those railway operation requirements.

David R.: As we discussed at a mining automation conference recently - driverless metros and driverless haulage systems are where we are going. Some of the operations are the same as with drivers, but some things do change. The systems will become more and more optimised as they become more common, so the amount of dead time and recovery time will become less and less and there will be a requirement for the systems to be more reliable. Driverless systems are increasingly common - all new metros around the world are driverless and in next 10-20 years the same thing will happen on the iron ore lines.

Malcolm: The communications emerging technologies will push us down the software development type project delivery. This systems approach will give us an opportunity to try and use those tool sets and approaches.

John: Our modelling software Open track, Open Power-net and Train, can model client requirements and give them options. There is a really intelligent client approach that lets us use these. The worry is if the client has gone straight to tender without defining the requirements. There are some clients who have no idea on the specifics of how they want to operate which is a challenge.

Malcolm: The role of the consultant has disappeared to the detriment of the whole approach. If a tender is lacking information it causes real problems down the line.

Performance specifications - it's all in the definition

David N.: Performance specifications in a signalling context are high risk, but clients think that gives them maximum flexibility for less amount of work. However, all it does is expose them to risk, relying on the vendor interpreting it but they can’t be effective in that way. History is showing us that the current approach doesn’t work. There is a need to spend money upfront on definition work to minimise delivery risk. We need a shift in the way we go to the market to get the balance right.

John: I think that is 100% right, but there are two further drivers that should stop that process happening - for example, in the mining boom 5-6 years ago the drivers for getting the project going were “get this done yesterday” so railways didn’t have opportunity to go into the detailed definition phase. For passenger railways, the drivers tend to come from politicians, so they want things done before next election, which also forces the definition phase to be bypassed. There really needs to be longer term planning.

David R.: I wouldn’t throw away performance based spec as a mechanism for going to market, particularly in green field sites. What is necessary is a better understanding on the client side of what they’re actually asking for and being more precise about what they actually want. I think Greenfield performance specs work reasonably well, but brown field performance specs usually fail dismally.

The other interesting thing is whether new technologies will overcome these issues by creating greater flexibility much later in the project. New technologies will in part help the process by creating greater flexibility.

David N.: I think the opposite, just like rolling stock, where we have a global market ... where you take what’s made or you pay an incredible premium. I think the globalisation of the signalling market will mean railways have less choice and will take the standard product or pay a huge premium to customise it. May be 20-30 years away but that is where we are heading.

David R.: I agree. And I would argue that customisation doesn’t often help them. Signalling is a safety product and is very expensive to change.

Malcolm: There is a real risk around this. People may be seduced into thinking tweaks possible, as it is a software product and can pick up later and won’t do the things we need to up front - the specifications and the requirements capture and defining the scope - there is a real risk that relying on flexibility of software will create even greater log-jam at the back end, worse than now.

John: Comes back to needing the right person making these decisions up front and specs, who understand the implications. As David [N] said, the defence industry has rigorous high safety integrity levels in their projects so people from that industry understand those issues as well as railway systems engineers.

Is the signalling engineer becoming a systems engineers?

John: The signalling engineer will turn into systems engineer as opposed to your traditional relay interlocking type engineer. Systems engineers make better project managers than your traditional engineers. I am confident that PM expertise in the signalling profession will get better because the type of people coming in to the industry now are systems people.

David N.: There is a recognition that project management in this industry is not great, there are changes afoot in the market with regard to technology and globalisation of the market which demand a change in attitude from clients to customisation and perhaps open the opportunity for systems engineering approaches as used in oil and gas, defence, software to be applied more rigorously to the signalling market.

Innovation in rail signalling and the safety imperative

David R.: There is a tendency to keep doing what we’ve always done because it’s proven and because we use past practise as a touchstone of safety. It tends to become an immoveable benchmark ... try to do anything new and it gets knocked back. The safety imperative drives this tendency to doing what has always worked in the past. How does an industry get around that? In other safety sensitive industries, like the nuclear industry usually a safety board would be making decisions, often the buck stops with one guy in signalling, as well as the CEO, and it is onerous responsibility.

Malcolm: There is also blame culture of investigations in rail industry ... whereas in aviation industry there is usually an attempt to learn and make sure it doesn’t happen again. The Australian Transport Safety Bureau (ATSB) is changing things for the better, but there is still a way to go in developing more creative responses to problems/disasters.

John: In many Asian countries, railway operators want the best of everything, and have strong safety system assurance culture. There is a willingness to pay for quality. There is high patronage on public transport, so operators can afford to invest in technology and systems assurance processes required. Australia is still very much a 'car country'. Although this is changing, but it is still some way behind Asia.

David R.: Europe is highly train dependant. Must increase patronage, must build better and bigger and better networks to encourage more people onto the railways and with that will come the investment.

David N.: Traditional thinking has been that in order to fix problems will just build more tracks/extra lines. A lot of money has gone on optimising the infrastructure that’s already there. The question is can we optimise what’s there without increasing the number of tracks? New signalling technologies will be a lot of help ... will improve capacity without spending so much money. The world is 'moving closer together' though, everywhere corridors are constrained, we do need to get better at using infrastructure that's there.

John: Even in Europe, lots of the systems are very old ... how are we going to move from technology that is 50 years old? Not so long ago internet banking was a foreign concept, that sort of transition will happen in rail signalling. Those that come up with the ideas that can meet safety assurance standards will create a quantum leap in terms of the overall cost and effectiveness of the rail systems.


To top