#110 — Why Digital Signalling Never Got Cheaper, with Andy Harrison

Twenty years ago the promise was simple. Digital train control would get cheaper, and ETCS and CBTC would become the default for every new scheme. Neither happened. Costs keep rising, and the system knowledge keeps migrating into the supply base.

My guest is Andy Harrison, a long-time friend and consultant to top railway clients across Great Britain, with four decades in the industry. We get into why the tipping point never came, what changes when the railway no longer holds competence in its own systems, why "change management" is far more than a new job description, and why zero disruption should be the default for bringing new signalling into service, not the exception.

Subscribe to the channel "Railway Conversations with Doc Frank" on your podcast platform of choice (YouTube, Apple, Spotify) to get notified first.

If are interested in deeper understanding of modern train control, visit my training portal at www.docfranktraining.com. There you find the most comprehensive suite of online courses in advanced railway signalling anywhere. CBTC, ETCS, high capacity signalling, and more. All in one place, all vendor neutral.

Transcript

0:04

Andy Harrison, welcome to the conversation. Hi Frank, it's good to see you. Thanks very much. We had a bit of a pre-talk a couple of minutes ago and you said something that the signalling industry has changed quite a bit, which is by the way something that I noted as well. What do you mean by that? How do you see the recent changes that we are seeing in the industry? I think it's, in all the years that we've been working in railway signalling, and personally that's almost 40 now, I was expecting to see over that period of time the application of the ERTMS, ETCS, CBTC systems to become much more widespread and to become almost the de facto standards for new signalling schemes across the world, for costs to have reduced, for life cycle support costs to reduce. And I think in my recent work with clients, I'm not really seeing that. I'm seeing the cost of signalling installations continue to rise, both in the UK and in other places. And I'm seeing a real nervousness within railway authorities about taking on life cycle support costs for these high technology systems, which have very different maintenance requirements, very different obsolescence management requirements, than existed previously with the technology that we've previously used. And so I think the worry I have is that we're working very hard to create some extremely interesting technology to support improved signalling, but perhaps we're not doing it in a way which is very cost effective, very deliverable, and very maintainable over the life cycle of the system. And as a result, actually, railway authorities, and more importantly, passengers, the freight companies, are not getting the benefits that the improved signalling systems can provide. And I worry, I think, that I don't see a tipping point where that cost increase changes to become a cost decrease. What are your thoughts? Yeah, well, definitely the costs of projects go up. And I also noticed that I think it has a lot to do with the mindset of the clients still being stuck in the old ways of doing things and the old technology. So for example, one thing which I think has to change significantly with digital systems such as ETCS or CBTC is the relationship between the client organisation and the supplier organisation. You can't just say, ah, I buy three more widgets, like two signals and one point machine, and then I can kick the supplier out. And if anything happens, then I go to another supplier and buy three other widgets from them. So with a system like CBTC, where the single point of responsibility for design, for design changes, for obsolescence management, for all these things sits with the supplier, the client needs to have a long-term, almost like a partnering relationship with the supplier. So basically as some kind of technology partner. But that's an approach that was never used, that was never needed in conventional signalling. And this is a mind shift that client organisations have to do, which seems pretty obvious for you or me, because we have the perspective of many other projects elsewhere. But even if we try our best to convince client organisations as consultants, that this is how it needs to be, they listen, but they can't get their head around it. That's my impression anyway. It's a really valid point I think you're making. The whole issue of partnership is becoming, I think, much more obvious in the industry now. I think people are starting to wake up that instead of having a short-term particular delivery contract for a particular scheme or a particular route, what's required now is a long-term relationship between the railway authority and the supply company, both for the delivery, but also for the management of that system through its life. And that's a fundamental shift. Railway authorities used to, and that's true for many technologies which hold maintenance responsibility, first-line maintenance responsibility, particularly. For a lot of these modern technologies, that's simply not appropriate anymore. So the railway authority is changing the fundamental nature of the way it maintains to the railway because it's actually contracting out to the supplier of the system over perhaps 15, 20, 25, or 30 years. And more than simply first-line maintenance, making sure that the system is performing and is available as required. It's a supply chain issue that's component changes, component unavailability over the life cycle of the system. And obsolescence, simply obsolescence management. In 10 or 15 years, the base components of a lot of the systems that we're dealing with, they'll be no longer available. And that obsolescence management approach needs to be fundamentally different. So I think you're right. It's kind of my summary of that. The change from a contract-driven environment to a partnership-driven environment, just clearing the way forward for a lot of these technologies. Yes. And one additional aspect. So I subscribe to everything you just said. Obsolescence management is a huge challenge for these kinds of systems and it does require a long-term support by the contractor. But another thing that I would like to bring up in this regards is configuration management. So if you consider CBTC in London, you've got the first applications by Thales on the Jubilee and the Northern Line, which is a CBTC based on induction loops that some people would call a transmission-based train control. Anyway, now the latest project that they're doing also with Thales is the four-line modernisation, which is also Thales CBTC, but with radio communication. Now, in order to sort that problem, you would probably need to have some kind of configuration management where the supplier comes in and says, well, actually the system on the Jubilee line, which is the older one might need to be upgraded to radio-based communication so that you have basically the same technology, the same system configuration for all the CBTC applications in London. So you have CBTC from the same supplier. You would assume that you get interoperability out of this, but you don't unless you do a revamp of the older systems. But who pays for that? Who is doing that? Nobody has budgeted for that. And this is something where things like obsolescence management, things like midlife upgrades of technology, they need to be planned for, they need to be budgeted for, they need to be arranged for in collaboration with the supplier. And none of these things were business as usual in the conventional signalling world, which I think is a big problem. And it comes back to the mindset of the client organisation. Everybody's talking about change management. Everybody's talking about, ah, it's not just technology. We need to focus on the people, on the process. But when it comes to the crunch, and this is one of the examples, then things just fall short. I think you're absolutely right. I think it's a broader issue. Well, let's call it configuration management. Traditionally, by the way, engineering disciplines have certainly had interface points, but they've been reasonably standalone. The civil engineering teams do what the civil engineering teams do, the signalling teams do what the signalling teams do, rolling stock do what rolling stock do. And their points of interface have been well-known and well-recognised and well-established over many years. What a lot of the new signalling technologies are doing is it's fundamentally changing the way that those interface points exist. So a traffic management system sitting on top of either an ETCS or a CBTC system requires an intimate knowledge of the performance of rolling stock and the performance in the the specification, if you like, that the civil engineering structures and track work the permanent way that's out there on the railway today. And that management of the configuration of trains, of infrastructure needs to be maintained through the life of the system, the life of the signalling system, whether it's TMS or anything else. Now, those sorts of interfaces, those sorts of relationships are not something that traditional railway authorities have really got their head around yet. And so the idea of the signal department operating as a change control board, if you like, say for the civil engineers, well, of course, you can relay that piece of track. But if you change the gradient or you change the cant, then I need to know about that because that affects the way the signalling system operates. Or to the rolling stock engineers, of course, we can have battery-powered trains. But if they change the performance of the train set, then I need to know about that because the way in which my TMS algorithms do train planning or do train contingency planning. So there's a new set of relationships in this new world that the railway needs to get its head around and really start planning for because the old ones don't work anymore. And the particular difficulty I think that we face in the sort of work we do is that we have a railway authority that needs to maintain a railway which has been around for a long time using its more traditional mind. And it's trying to change to deliver and maintain a railway using these new technologies, which requires a fundamentally different way of working. So you have these two things running together in parallel. And that's why I think there's a lot of talk in our world about change management. But it's really, it's how to manage two parallel ways of working because existing railways are going to exist for a long time, many, many years. And it's going to take a long time to transition to a railway based on new technology. So we're going to be faced with having these two operating paradigms running together for a long, long time. And that's a challenge. Oh, it is. And many people hear change management and they think we just have to change the job description. So you get a new piece of paper, which outlines a couple of your functions of your job functions have changed. And that's about that. But it does really require a quite substantial change of mindset, a fundamental mindset that had been moulded for decades and decades. The interface between signalling and rolling stock is such a classic. I've just done a free training where I'm trying to help people to identify whether their jobs, their careers might be impacted by changes in the signalling technology. Because when you introduce these new technologies, and it doesn't matter whether it's ETCS or CBTC, anything to do with automatic train protection, one big change that's going to happen is that for the first time, maybe ever, you've got signalling equipment on board of the train with an interface to the train. And this is significant because as you said, all of a sudden the signalling engineer comes along and asks the rolling stock engineer for a guaranteed emergency brake rate in order to design the signalling system in a safe way that the trains can be stopped where they need to be stopped to avoid accidents. And there is one word in this guaranteed emergency brake rate where a rolling stock engineer has massive problems with and that's the word guaranteed, right? So they don't want to give you any guarantees. They say, our trains will brake as good as they can. And the signalling engineer will say, well, sorry, that's not good enough. I need an absolute deterministic value so that I can do my safety approval for my system and I can make sure that your trains are not crashing into each other. So I think it's a really good point. This is a very, very difficult subject and I'm sitting here in London, it's the middle of December and today is a Friday and there's a railway strike. And it's the second railway strike of the week. And part of the reason for that industrial action is very much around the idea of modernisation and the changes in working practices and the implications that that's going to have on staff. And I'm not close enough, I wouldn't profess to be an expert in industrial relations, I'm not close enough to understand what the real issues that they're facing are. But the reality as we see it, sitting here today is that those existing working practices are going to change and they're going to be driven, not just by managers feeling that they want to do things in a different way. They're going to be driven by the fact that technology requires things to be done in a different way. And that's particularly the case for the sorts of systems that we're dealing with, the ETCS, the CBTC, the TMS and other things. And it gets even more complicated when you start thinking about the role of the supplier, the technology supplier, against the role of the railway authority, where competence in the system, the knowledge of how these systems work and operate, invested in the supplier organisation rather than invested in the railway authority. So the railway authority requirements for competence are fundamentally different in the brave new world than they are at the moment. And that requires a shift in thinking and a shift in process, as you say, whilst still maintaining enough of the knowledge and competence of the existing railway to make sure that that continues to work during this very long transition period when technology starts arriving. So I think it's a hugely challenging time to be a railway authority, to be part of a railway authority. If this conversation makes you wanting to go deeper, I run what I believe is the most comprehensive portfolio of online training courses in advanced railway signalling. CBTC, ETCS, high capacity signalling and more. You find them all at my own training platform docfranktraining.com. That's docfranktraining.com. And everything is right there waiting for you. But now back to the show. Yes, yes. So what you're touching on is something that I would categorise under stakeholder management, with the union obviously being a very important and very influential stakeholder. I can recall the first meeting I had here in Perth when I did the planning for the CBTC rollout here, what we call high capacity signalling by now. First meeting with the union representative here, who was in charge of the union where all the train drivers were organised. The first question he came in with, are you going to replace train drivers with automated systems? And so my response must have been convincing enough and compelling enough that I said, no, there's no way that we can ever replace drivers in Perth because, you know, it's an open air railway where you've got external impacts from kangaroos and stuff like that, whatever you have in Australia. And so he was happy with that response. And the second question he had just after that was, if the system misbehaves, the automatic system doesn't do what the driver's comfortable with doing, does the driver have any means of overriding the automatic system? And when I told him the driver just puts the hand on the lever and then he's back in control, the entire conversation shifted. We were best friends from that moment onwards. We had a very constructive discussion. I explained to him what the system is going to do. He felt well informed so he could go back to his members appearing knowledgeable, which he was after that kind of mini training or so. But these are things where just a couple of answers can steer the entire conversation into completely different directions. And if you get this wrong once, once only, if you get this wrong and you start on the wrong foot, then things can go pear-shaped very, very quickly and very, very stubbornly. It will be very hard to bring things back afterwards. So maybe I was just lucky that I was off on a good start, but I can imagine what's going to happen if the start of such a conversation is more controversial or if there is already a history. I didn't have any baggage with the union. It was the first time of me talking to them. But maybe if somebody does have baggage with them, where both parties basically think that the other party keeps lying at them or something like that, or speaking through teeth or something like that, as you would say, then it will be much more difficult to really get those things right and really come up with a convincing argument that no, we are not going to get rid of the train drivers. I think stakeholder management is absolutely fundamental to what we're doing. It's curious, isn't it? I think when we started our careers, I never really imagined sitting here having this conversation with you. We were sitting there over scheme plans, looking at headways, breaking calculations, and all sorts of things. And the idea of active stakeholder management was a world away. It was what other people did in other industries. But yeah, I seem to spend a lot of my professional life now sitting between people, doing exactly what you're talking about, doing stakeholder management and about that mediation, and trying to help people, I suppose, to understand the positive benefits that being a person in a railway environment brings, because the people are creative, and they're knowledgeable, they're committed, and they're competent. And, you know, the aim of technology is not to, in my opinion, anyway, it's not to lose that. It's to make the best of it and to find a way of helping technology work with the best of people. So that whatever, you know, potential safety pitfalls and things like that we still have on the railway, technology can help to reduce, and people can help to reduce technology or sometimes break. So it's, you know, to find yourself in the middle of a stakeholder mediation process, it's not something, as a signal engineer, I ever thought I was going to do. So it's been quite an interesting learning curve for me. I think the other point I wanted to make, though, is that, you know, railways don't exist in isolation. And so, you know, a lot of talk about electric vehicles, railways are responding to that. A lot of talk about artificial intelligence, driverless cars, et cetera, et cetera, et cetera. So those things are becoming part of the common vernacular of people, you know, people in the pub talk about driverless cars. And, you know, the railway can't ignore the fact that technology is moving on in other areas. And so, you know, over time, whilst I'm not advocating that we start getting, you know, start thinking about reducing our needs for drivers, over time, the change in competence, the base in competence will move. We will end up in a situation where, you know, technology is taking a different role, an extended role. And the role of the human is much more about monitoring and supervision, rather than necessarily active. And I think that's a reality in the world outside railways. I suspect railways will not be able to hold that torrent of water at bay for a long time. I think we will end up moving too. And I think from a signal engineering perspective, again, that's an area that we never really imagined we would have to get, have to start thinking about, you know, how do we take artificial intelligence and the work that's being done in AI, explainable AI and trustworthy AI, and make the most of that within the railway environment? Yeah. So it's a completely different skill set, but it's one that likely falls within the roles of signal engineering. The change of roles and responsibilities is quite an important aspect of change management. And I come back to my previous example where people think, oh, well, I've got a couple of new paragraphs in my job description, and then we're done with that change management thing. And that's just not the case. So one example in case is that I quite like mentioning in my training courses, but which upsets many people with a signalling design background is where I'm saying the task of signalling design will change fundamentally under a system like CBTC. And not only will it change, it will move. It will move from one kind of organisation to another kind of organisation. So if you wanted to do detailed data design for CBTC, there's only one company you can go and work for, and that's the CBTC supply organisation. Or maybe they have a low cost design subcontractor sitting somewhere, most cases in India. But then if you come from a high salary country, you wouldn't really want to work there either because you don't want to give up, I don't know, 90% of your salary just to keep working in signalling design. And many signalling designers struggle with that notion enormously. They think they're going to get out of a job. And what I'm doing actually, I'm showing them opportunities where they can move with their knowledge and experience and where they can assume slightly different roles to still have their importance and their value in the new world of CBTC just without doing exactly the same what they used to be doing during conventional signalling. And there's another mind shift change which is significant and where people really, really struggle to get their head around. Like I want to do my conventional signalling design. And they said, yeah, well, you can keep working on conventional signalling design as long as your railway still has conventional signalling, which may be a few decades, depending on how long the migration process will take from the old system to the new system. But eventually at some point in time, it will start dying out. And I think the earlier you can get your head around and say, hey, I need to adjust my career for these future opportunities so that I'm falling forward and not just waiting for the tsunami to wipe me out. That's a big thing. I would agree with you. And I think what you're alluding to is kind of where we started this conversation, which is that signalling and control has traditionally been about boxes and lights on sticks, lights in cabs and bits of cable and all that kind of stuff. And the skill set, the competency in the railway authorities and the supply side and the technology suppliers has grown to support that environment. The moment that you change that environment to become much more computer-based in the future, much more AI-based, then the skill sets need to develop. And the skill sets will develop, but they will change. But it's not just the railway authority staff that will need to change. It's the railway authority. It's the way that we regulate. It's the way that we carry out safety accident investigations, for example, that will need to change. Skill sets there will need to be different. Skill sets on the supplier side, skill sets in the consultancy environment, they will all need to develop to reflect the new reality we find ourselves in, because as sure as X is X, technology will drive change in railways, and it will continue to do so. So I think the point that you're making is hugely valid. A signalling designer who is experienced in designing conventional signalling schemes will have a job as long as there are conventional signalling schemes. The moment that there's a tool set, and if I look at the way tools have developed over my career, it's incredible, really. The moment that there's a tool set where I feed in the answers that I need in terms of the performance of the train service I'm looking for, and the infrastructure base that I'm using, and the train performance that I've got, the moment that I feed that into a tool, out at the end of that tool pops a set of control tables and a scheme plan and all the other bits of books we need. Well, my role as a designer has disappeared. I have no job anymore. And I think we're not far away from the point where that tool set will exist. I'm absolutely sure it will happen. Yeah. And that reflects a need for people to really wake up, smell the roses, and decide, smell the coffee rather, and decide which side of the game they're going to play. Do we play the old side and eventually we're all out of the job, or do we shift our competency and become better at dealing with the technology changes that are arriving? Yeah. And it's not just the client organisation and the third-party signalling designers which have to change. We haven't talked much about suppliers. The role of supplier organisations will change as well. They will have to adapt as well to a new model of partnering with the client organisation, and they have to stop behaving as if they were still delivering widgets, which they did in conventional signalling. So I had supplier telling me straight into the face, Frank, we are not a consulting organisation. Well, maybe that has to change in a new environment where the technology that they are delivering, such as CBTC, for example, has such a massive impact on the operation and the performance of the entire railway, not just that bit of signalling. And helping a client organisation applying this technology in a sensible manner helps everyone. One of the concepts that I keep hammering on is what I call zero disruption delivery. So if you have a new signalling system, such as CBTC, and you have a system design that basically builds the new system completely in parallel to the old system, you can have a setup where the old system keeps running trains during the day, and the new system can be tested in shadow mode and installed in shadow mode. And then there can be some tests done during night windows. So at the end of the day, it should be possible, and some railways have done it, to introduce a new technology without losing a single timetabled train service. And people say this is impossible, but what they mean is it's impossible if we keep applying the old methods that we had to apply for 100 years in conventional signalling. They can't get their head around that suddenly because of the system architecture of two systems which are completely in parallel and almost independent, you've got these new opportunities that could be extremely beneficial. They struggle so much with it. It's amazing. And this is where a supply organisation could help by just explaining what their technology enables, not just in terms of trains per hour and stuff like that, but in terms of process improvements that were never possible or even comprehensible under conventional signalling. It's a really interesting point, because technology is ubiquitous. It's everywhere. And people's understanding of technology, especially younger people's understanding of technology, is immense. They get it. They understand how apps work. They understand the internet. They understand how to use technology in a way that works for them. And I think what goes with that is that people's tolerance of technology failure is going down. If you look at the reaction in your house when your internet router fails, or you can't get a mobile phone signal, or something like that. And so I think the point I'm making really is that people's acceptance of technology failure. Railways are not, we're not a separate world. People are not tolerant of railways coming along and saying, our system doesn't work for this reason, our system doesn't work for that reason. And they're certainly not tolerant of, we're going to close your railway for a number of days or a number of weeks while we carry out an upgrade project. That is a world which existed before. It's not a world which exists now. And if we fail as an industry, if we fail to take advantage of the fact that new tools and new techniques and new technology gives us a different way of bringing systems into service, which has less impact on passengers and freight customers, then we're not doing our job. Frankly, we're not doing our job. We absolutely ought to be aiming for zero disruption. That's the default. Default is zero disruption. Default is not, we're going to close you for three or four weeks. Default is zero disruption. We will make this change and you as a customer will not see it. That has to be where we're at. It's interesting that the organisations that should benefit most from it because they just stay out of the headlines in the newspapers for shutting down the railways repeatedly and for too long, they have the biggest problem of taking these new concepts on board. It's amazing. And they don't listen, most of the cases, they don't listen to the people who show them the way, who give them the opportunity to explain things to them. I had this before, I was in consulting roles explaining to client organisations how CBTC would work and what all the benefits were. And in hindsight, it turned out they just didn't believe me. They didn't believe me because they thought I was reading from a science fiction book. And then they went on a field trip to other continents where they spoke to metro railways that had already applied this new technology. And they came back and basically recited all the things where the executive said, wait a minute, isn't that the stuff that Frank's been telling us for the last three years? Maybe that wasn't science fiction after all. But as long as people who have this kind of perspective, don't get the credibility and are not being listened to, it's very, very hard to turn people around to really, I mean, you need to keep trying. I'm definitely not giving up on that, but it's an uphill battle, that's for sure. And it's a lot harder than what you would wish it would be. Yeah, I understand where you're coming from. I think that what I'm seeing in the work that I do is that as younger people join railway organisations and start to progress their careers and they become more professional in what they're doing, and eventually they become more senior in an organisation, their expectations are different to the people that were in charge when I started working. Their expectations of technology, their expectations of change, their expectations of the pace of change are fundamentally different. And I think the hope, if you like, at the end of this conversation, the hope is that those young people will continue to develop in a way which actually supports a better use of technology, a more rounded set of competences for more people to enjoy better careers, all for the benefit of passengers and and at a price which, in most countries, the taxpayer can actually afford. And so I think I'm relying on that generational change and generational involvement to start to really get behind the changes that we're talking about here. It's a very nice, optimistic statement to wrap this up. I've had some of these interviews with younger chaps who have been in the rail industry, in the signalling industry, somewhere between six months and six years. And I was quite impressed by the level of perspective that those guys have. And I am carefully optimistic that things may actually develop in the way that you just described, which would be hugely beneficial for the entire industry. So with that, I'm conscious of your time. I'd like to thank you. That was an excellent conversation, which I enjoyed a lot. And I hope we can do something like this again in the future. So thanks, Andy. Hi, it's Doc Frank again. I hope you enjoyed this conversation. If you like the content, then the best way you make sure that you don't miss out on any of the future episodes, just subscribe to the channel. By doing so, you also get access to all the previous episodes, which I'm pretty sure there will be many that are of interest to you. Another thing that you can do to help not just your own environment, but the entire industry is to share this episode with friends and coworkers, because it helps more people in our industry to get educated and an educated rail signalling industry is a better industry. With that said, thank you very much. Hope to see you again soon and bye for now.