Trains all at a stand in the Manchester area... (Thursday afternoon 06AUG2026)

[EYFanatic]

So there hasn’t been a single train in or out of Manchester Victoria in about 10 hours? Wow.

The light engine move 0M68 seems to have gone through a little while ago

1J43 2142 Leeds to Manchester Victoria is on the move and heading towards MCV and might be the first passenger arrival

 

[Parham Wood]

What worries me most is what happens if there is a major fire or terrorist activity and the centre is put out of action for good or a very long time? Are any contingency arrangement really suitable to take over until a replacement centre is built? I am glad I don't have to make decisions on this as they all will be costly in one way or the other!

 

[dosxuk]

Like with Heathrow, power failures of this scale are often not a simple "nobody specced a backup supply" problem - instead multiple small failures - some planned, others not - resulting in a loss of redundancy at exactly the wrong time.

 

[Acathater]

Northern now saying expect disruption until end of tomorrow

Service updates

Real-time updates on any disruption or issues affecting your Northern route, station or train.

www.northernrailway.co.uk

this is showing against multiple routes

Description:​

Failure of the electricity supply

Impact:​

Train services running to and from this station may be cancelled or delayed. Disruption is expected until the end of the day on 07/08/26.

 

[JamesT]

Something as major as this should have two separate supplies from different substations, hopefully sufficiently isolated from each other so an electricity network problem affecting both would be rare. Surely this should have been thought of and identified in a risk assessment. It will be interesting to see who signed this off and whether cost saving had anything to do with the decision. Will the Government be bothered though? Certainly getting a second supply from a separate independent substation would be a costly business depending on geography.

Post #147 of this thread states that there are two independent power supplies to the ROC. The reporting that there was only a 90 second outage does suggest it’s not an ongoing loss of power that is the problem, but the response of the systems within the ROC to losing power.

 

[Belperpete]

FWIW the BBC 10 o/c news said the public supply outage lasted all of 90 secs, this might correspond to the changeover @Elecman talks about. So if the UPS can't cover a gap that short it probably wasn't working in the intended fashion. And wouldn't have been able to bridge until a genny started up either.

You are assuming that the system would automatically switch back to the DNO supply as soon as it comes back on. On systems that I am familiar with, you don't switch back until you are sure that the supply is stable and can be relied on. If so, the system would switch to the generator when the DNO supply failed, and would remain switched to the generator even though the DNO supply came back on, until a technician manually triggered switching back. So how long the DNO supply was off for would be irrelevant.

The issue would be how long it took to get the generator up and running, and how long the UPS maintained the supply for. Did it take the generator too long to get on line, and/or did the UPS not maintain the supply for as long as it should have?

 

[lowe1409]

We took the X2 from Lime Street to Preston, Preston was like a ghost station, as was Lancaster when we finally made it. A kind train manager said we could get on the Northern to Windermere to make it to lancaster despite there not being ticket acceptance for that route between Northern and TPE

Northern and TPE along with LNER have had long-standing automatic acceptance in the event of any cancellation. I believe this also applies now to all DFTO TOCs. Plus during disruption like this, youll rarely see conductors checking tickets anyway.

It's the School Holidays.
There surely isn't that much charter work to cover all of that, for the right price buses would be available, definitely not a quantity enough to run on every line, but they could have at least tried (TfW did and got some buses running Manchester-Warrington, everyone else gave up).

TFW only had the one route affected and hence will have had controllers spare to arrange it. Imagine northern control, bearing in mind they had no phones for a while, aranging busses for most services around the North west. It would be chaos. And assuming they could source the busses and they turn up. You'd have to communicate to station staff about the busses and running that many services without coordinators would be a nightmare at large stations such as Piccadilly. Plus the staff abuse when passengers inevitably get left behind because a 50 seater coach can't take as many passengers as a 6 car 331.

By arranging acceptance with bee network, although slower, provided passengers with alternatives quickly and easily. A reliable but slow journey beats a chaotic, unpredictable one for the sake of a few minutes.

 

[MichaelTrains]

Three others and I are at Preston where we have been told the last train to Bradford and Leeds is definitely running.

The monitors are saying delayed and not cancelled, so hopefully said train turns up.

 

[EYFanatic]

Trains seem to be leaving and arriving at Piccadilly now according to RTT

Will this cause trouble with trains LDS-BPN tomorrow do you think as some trains there have been cancelled

 

[MichaelTrains]

Trains seem to be leaving and arriving at Piccadilly now according to RTT

Will this cause trouble with trains LDS-BPN tomorrow do you think as some trains there have been cancelled

I doubt it.

We are waiting on platform two and have been told everything is fine now by the Avanti member of staff here.

 

[Andyh82]

Northern and TPE along with LNER have had long-standing automatic acceptance in the event of any cancellation. I believe this also applies now to all DFTO TOCs. Plus during disruption like this, youll rarely see conductors checking tickets anyway.

TFW only had the one route affected and hence will have had controllers spare to arrange it. Imagine northern control, bearing in mind they had no phones for a while, aranging busses for most services around the North west. It would be chaos. And assuming they could source the busses and they turn up. You'd have to communicate to station staff about the busses and running that many services without coordinators would be a nightmare at large stations such as Piccadilly. Plus the staff abuse when passengers inevitably get left behind because a 50 seater coach can't take as many passengers as a 6 car 331.

By arranging acceptance with bee network, although slower, provided passengers with alternatives quickly and easily. A reliable but slow journey beats a chaotic, unpredictable one for the sake of a few minutes.

Unfortunately acceptance with Bee Network only gets you so far, there didn’t seem to be acceptance with other buses or implicit instruction that they’d get their fare back if they paid

 

[QueensCurve]

Oh we do...

But this has been happening since the 60s with PSBs etc.

No secondary site for Saltley that I know of, not sure where you would put it.

This makes me think of the Warspite submarine fire in the 1960s. The fire destroyed all the control an instrumentation cables for the reactor and for someting like 48 hours they didn't know anyting about what was going on in the reactor. As a result they designed an independently acting control and instrument panel for all the submarines. Something similar happened at the Shippingport PWR in the US.

 

[EYFanatic]

1J43 2142 Leeds to Manchester Victoria is on the move and heading towards MCV and might be the first passenger arrival

Never mind... Caped at Rochdale

Going to be a nightmare for not just passengers but also the train crews

 

[RiponHall]

I quite vividly remember sitting in a meeting at Waterloo about a revised new signalling scheme in the extended saga during which the Dorset Coast Resignalling became 'Bournemouth Beach', and when the remaining intermediate signalboxes between Bournemouth and Dorchester were going to be recontrolled to Basingstoke as part of the overall ROC strategy.

When questioned about the resiliance of having the entire route between Waterloo and Exeter controlled from a single location, the Territory staff sat quietly whilst those from HQ responded that there would be duplicate installations - but that they would not be needed anyway as resiliance would be built into the installation, and as that would be impregnable, have backup systems, retain local interlockings and duplicate controls, there would be no issues.

When they had left, the territory staff said that no amount of operational knowledge or wisdom would be able to stop the business case for saving money by centralisation and simplification - but that inevitably major events would, in future, become catastrophic ones that were simply too big to manage.

 

[EYFanatic]

This makes me think of the Warspite submarine fire in the 1960s. The fire destroyed all the control an instrumentation cables for the reactor and for someting like 48 hours they didn't know anyting about what was going on in the reactor. As a result they designed an independently acting control and instrument panel for all the submarines. Something similar happened at the Shippingport PWR in the US.

HMS Warspite had a fire when alongside in Liverpool in 1976 - not the 60s

 

[The Mercian]

You are assuming that the system would automatically switch back to the DNO supply as soon as it comes back on. On systems that I am familiar with, you don't switch back until you are sure that the supply is stable and can be relied on. If so, the system would switch to the generator when the DNO supply failed, and would remain switched to the generator even though the DNO supply came back on, until a technician manually triggered switching back. So how long the DNO supply was off for would be irrelevant.

The issue would be how long it took to get the generator up and running, and how long the UPS maintained the supply for. Did it take the generator too long to get on line, and/or did the UPS not maintain the supply for as long as it should have?

a very good summation this, the second paragraph in particular where attention I’d imagine will be focused as they investigate what went wrong.

 

[Prime586]

Worth pointing out a complete loss of all the phones initially, in the ROC as well. Backup mobile phones in use for the RCM and controllers.

Ah, the curse of VOIP systems. Their (lack of) performance during power failures leaves a lot to be desired over the conventional TDM PABXes with Central Battery powered phone lines that they replaced.

The same thing happens with BT's Digital Voice lines - you are expected to use your mobile during a power cut, but it usually turns out the mobile basestations don't have sufficient (or sometimes any) UPS capacity either.

 

[QueensCurve]

HMS Warspite had a fire when alongside in Liverpool in 1976 - not the 60s

I stand corrected. My mind had rewrote the date as 1967 for some reason.

 

[Belperpete]

Post #147 of this thread states that there are two independent power supplies to the ROC. The reporting that there was only a 90 second outage does suggest it’s not an ongoing loss of power that is the problem, but the response of the systems within the ROC to losing power.

I would question whether two feeders from the same substation are truly independent. If the substation goes down, you lose both feeds. As appears to have happened here.

My experience was that you have two parallel feeders from your main supply, AND a feeder from a totally independent supply (such as a backup generator). So on the Southern, for example, there would be two parallel feeders from the traction substation, and a backup feeder from the DNO supply.

 

[lowe1409]

I would question whether two feeders from the same substation are truly independent. If the substation goes down, you lose both feeds. As appears to have happened here.

My experience was that you have two parallel feeders from your main supply, AND a feeder from a totally independent supply (such as a backup generator). So on the Southern, for example, there would be two parallel feeders from the traction substation, and a backup feeder from the DNO supply.

The ROC had a back up generator. However for whatever reason, it didn't function as intended and hence no power

 

[EYFanatic]

I stand corrected. My mind had rewrote the date as 1967 for some reason.

No worries... Not to go off topic, but her Captain actually was going to scuttle the sub to put the fire out

#

Is there a reason trains do not seem to be going in and out of Vic apart from a light engine move but trains are running from Picc? I can see this stretching into Saturday too

 

[cin88]

No worries... Not to go off topic, but her Captain actually was going to scuttle the sub to put the fire out

#

Is there a reason trains do not seem to be going in and out of Vic apart from a light engine move but trains are running from Picc? I can see this stretching into Saturday too

Picc wasn't affected as much. Manchester North and Central Workstations have had a total loss of correspondence in every set of points so nothing can move until the position of the points is established and they're secured in the correct position for any movements. There's a unit trapped next to Newton Heath depot and 3 units* trapped in Vitriol Works' patch all on the Up Rochdale line, that also happens to be why nothing ran Leeds to Manchester Victoria.

* Edit: 3 trains.

 

[Fokx]

Picc wasn't affected as much. Manchester North and Central Workstations have had a total loss of correspondence in every set of points so nothing can move until the position of the points is established and they're secured in the correct position for any movements. There's a unit trapped next to Newton Heath depot and 3 units trapped in Vitriol Works' patch all on the Up Rochdale line, that also happens to be why nothing ran Leeds to Manchester Victoria.

Liverpool workstation was also badly affected, no services passed Huyton Jn for a number of hours. (Hence two TPE class 802 units being sent back to Earlestown and Eccles loops after detraining passengers).

Signalling was only restored at 2138.

 

[Belperpete]

What worries me most is what happens if there is a major fire or terrorist activity and the centre is put out of action for good or a very long time? Are any contingency arrangement really suitable to take over until a replacement centre is built? I am glad I don't have to make decisions on this as they all will be costly in one way or the other!

When Cannon Street signalbox burnt down, the Southern had a temporary replacement up and running very quickly (within a week IIRC). But it only had limited facilities. And the Southern had a standby frame available.

The equipment used in the ROCs is largely off the shelf equipment. It shouldn't take too long to buy replacements, and instal the necessary programs and data. The main problem would probably be arranging the accomodation and telecoms links.

 

[Bow Fell]

Well, they're saying on the News at 10 that the control centre lost power for 90 seconds, but I guess there's much more to it than what they're actually reporting.

No, I don’t think you’re getting it.

There’s nothing been hidden or covered up, regardless of the power being lost for 90 seconds, it’s the fact that the signalling didn’t recover after the initial power outage, multiple track circuit occupations across the North West, too many to even imagine, signals showing blank on the ground, signallers unable to exactly ascertain a trains exact location for a period of time as well as loss of phones to add into the mix.

I know there will be the usual criticism for not getting things back up and running but I think one thing that did come out well today was the priority of the trapped trains between stations, this was the priority for everyone, running a train service was on the back burner for any period of time until this was dealt with and secondly the reliability and integrity of what had restored after the initial outage.

 

[Annetts key]

Work in healthcare. We have generator support for the whole site and fuel for 7 days. We have had instances in the past with the generator not starting when required. There’s a monthly test now, running for 3 hours. When the monthly test was first started we had a lot of issues which we had to log and report. Gradually these have been reduced to the point where there are very few problems now.

You need to test generator backup. If it causes problems then don’t stop testing, fix the issues.

I'd say at this point, having basically shut down the network over much of the North of England in the run up to the evening peak, that ought to be enough scale. And frankly given that it's been shown to have been such an issue, someone, somewhere should have asked the question "what if the backup fails?"

As I said earlier, someone has a very difficult day coming up tomorrow.

As said by another poster:

Manchester ROC has 2 independent 6.6kV DNO intake feeders fed from a primary substation either of which can feed the entire building by closing the relevant bus bar couplers on the HV panels ( there are some reports of a large part of Ardwick also bring without power) do sounds like it’s a SPENWL issue.

I can't speak about the Manchester ROC, but I can tell you about the arrangements for the PSB (Power Signal Box) where I was based.
Keep in mind that this PSB where I was based did not have any electrified lines.
There was a "guaranteed" feed from the main sub-station for the city (the DNO supply). Here "guaranteed" just means that there were no other customers fed from the same cable. This was an older installation. In newer installation, there are often two independent feeder cables.

An online UPS was provided to maintain power for several minutes. This should be enough time to allow the standby diesel generator to start. The control system for diesel generator was set up to automatically start it if power was lost for more than about 20 to 30 seconds. The diesel generator was designed to be ready to take the load in around 1 minute.

I can't remember the frequency (I think it was three monthly), but "crash" testing of the standby diesel generator (where the incoming mains is switched off and you check that the UPS can supply the load, that the generator automatically starts and that it takes the load) was part of the maintenance schedule. Although, because permission was needed from the signallers, this did not always happen. In which case the generator had to be run off-load. The test run was typically around one hour IIRC.

The fuel tank that feeds the generator was suppose to hold enough fuel for at least 24 hours. The fuel tank may of course be bigger than this. If the tank only holds 24 hours worth, there should also be a second fuel tank with more fuel.

In addition to the automatic starting system, standby diesel generators also have a manual push button start system and some may also have an emergency starting system.

Everyone seems to be "blaming the railway".
Has anyone considered that this surprising combination of issues could actually be due to external factors e.g. hackers / foreign interference?
I seem to remember one incident last year was quietly tentatively blamed on Russian interference - could this be a scaled up attack with the aim of causing disruption?

The railway uses closed control systems for the signalling system and the power system for the signalling system. In order to hack it, you have to be on railway property.

Oh we do...

But this has been happening since the 60s with PSBs etc.

This is true, however, some 1960s/1970s PSBs did have emergency through routes for the remote interlockings, some of which would still work during a total loss of mains power to the PSB. How? Because the through route control circuit was fed from a 50V battery at the PSB. Further, some remote interlockings had emergency panels that could be switched in.

None of this helped for the area of the railway that had lost power, but it could have kept some of the area controlled by the PSB operational.

But of course, this cost money, so, because it was hardly ever used, some of the emergency panels were taken out of use.

No secondary site for Saltley that I know of, not sure where you would put it.

Alas, with computerised signalling, because it's not a design requirement, it's much harder to have a secondary site, and because the interlockings are now normally at the ROC/signalling centre, there are no remote interlockings to control...

Is Network Rail a basket case?

By which I mean guilty of putting more and more eggs in the basket until it breaks.

Case 1) the current disgrace - as commented by others earlier, full testing with switchover onload and prolonged running in that state is mandatory.
If you decide to replace staff in signal boxes with levers, with digital stuff, you need to be more resilient than they were over the whole area in question.

My background in broadcasting involved very regular tests involving killing the mains and testing that the diesels and battery UPS covered the outage and that nobody noticed it on the output!

Case 2) bigger basket, more eggs, more people affected and more frequent - OHL!
There seems to be precious little understanding of 'Single Point of Failure' (SPF), whilst there is much excitement and whooping about new areas going under the wires, OHL has an enormous disaster disruption potential that steam or diesel locos just don't have.
Reducing the number of Thunderbirds as opposed to increasing them seems to indicate the basic lack of understanding of risk that is being undertaken.

There is an issue of professionalism here, if you're asked to design and implement something that you don't feel has the necessary reliance considering the potential SPF consequences, don't do it, don't complain about budget constraints - yes of course that's difficult.

Yes, the bigger you build it, the harder it falls.

See my comments earlier in this post about the power supply arrangements where I was based.

As I understand it, the track sections would show as if there's a train is there if there's no power, but when the power comes back, they will go back to clear if there's no train on it.
Is that right?

Nothing will show if the Control System has no power! The screens will be totally blank, no track layout, zilch.

I think you are thinking of a local trackside power failure, when what you say would be correct. But this was a failure of the control centre. Once it comes back up, it should show the state of the track circuits as currently reported by the interlockings. And that will depend on the type of interlocking and indication system, and the state they are in. And even once the track circuits are being indicated, the system may have lost all the train descriptions, in which case the TD displays will be blank, and so the signallers will have no clue what, if any, train or trains is in each section showing as occupied.

Further to what Belperpete says...

Track circuits will return to working normally when power returns (if power has even been lost, some remote areas will be powered by different supplies, not everything is powered from the same supply as the ROC/signalling centre). But if axle counters are being used, if their battery backup could not sustain their "state" (section clear or if occupied, the number of axles counted in), they will forget their state and will show the section as occupied until such time that they are manually reset.

You are assuming that the system would automatically switch back to the DNO supply as soon as it comes back on. On systems that I am familiar with, you don't switch back until you are sure that the supply is stable and can be relied on. If so, the system would switch to the generator when the DNO supply failed, and would remain switched to the generator even though the DNO supply came back on, until a technician manually triggered switching back. So how long the DNO supply was off for would be irrelevant.

The issue would be how long it took to get the generator up and running, and how long the UPS maintained the supply for. Did it take the generator too long to get on line, and/or did the UPS not maintain the supply for as long as it should have?

The control systems for railway standby diesel generators will indeed not automatically switch back to the mains (DNO) supply. They have to be manually switched back. And anyway, transferring back requires synchronising the phase of the generator to the exact same frequency and phase of the mains supply. You can't just parallel two different AC supplies. If the frequency and phase are different, there will be a big bang and the generator will stop dead and all power to the load will be lost...

Having said that, if the standby diesel generator develops a fault, some control systems will automatically attempt a switch over back to the mains (DNO) supply. But this is not seamless as the load loses power before the switchover occurs.

 

[Belperpete]

The ROC had a back up generator. However for whatever reason, it didn't function as intended and hence no power

Backup generators do seem to be a very unreliable form of backup. Unless you have the generator running continuously, you can never be sure it is actually going to work when you need it. Surely something as critical as a ROC should have something better. Such as a feed from the traction supply.

 

[cin88]

Liverpool workstation was also badly affected, no services passed Huyton Jn for a number of hours. (Hence two TPE class 802 units being sent back to Earlestown and Eccles loops after detraining passengers).

Signalling was only restored at 2138.

It was a similar situation in the fringe boxes. It was havoc where I was for about 90 minutes (unable to contact anyone for 30 of those), then just an awful lot of "yeah sorry drive, still no update".

 

[sharpener]

You are assuming that the system would automatically switch back to the DNO supply as soon as it comes back on. On systems that I am familiar with, you don't switch back until you are sure that the supply is stable and can be relied on. If so, the system would switch to the generator when the DNO supply failed, and would remain switched to the generator even though the DNO supply came back on, until a technician manually triggered switching back. So how long the DNO supply was off for would be irrelevant.

The issue would be how long it took to get the generator up and running, and how long the UPS maintained the supply for. Did it take the generator too long to get on line, and/or did the UPS not maintain the supply for as long as it should have?

No such assumption. Only that there would be a functional UPS that would keep all the plates spinning meanwhile, which clearly didn't happen. Once that isn't the case every piece of equipment is going to need re-booting, ideally automatically on restoration of supply (but I expect a lot of intervention would be needed in practice). It doesn't matter if it is restored by SPENW or your own genny or a link from the traction supply, you still need a working UPS to bridge the gap as @Annetts key makes clear.

And the days when the phone system had a big 48V battery are long gone which won't have helped.

I see the BBC web news also quotes this:

SP Electricity NW released a statement, detailing a power outage which had affected "around 5,000 customers in the Manchester area at 1:40pm today - all power was fully restored within 90 seconds," the company said.

so it looks as though they restored the supply without any swapping of the ROC feeders from the substation, it kind of implies the fault was further back.

 

[Annetts key]

Backup generators do seem to be a very unreliable form of backup. Unless you have the generator running continuously, you can never be sure it is actually going to work when you need it. Surely something as critical as a ROC should have something better. Such as a feed from the traction supply.

In my experience, railway standby diesel generators are reliable as long as they are properly maintained and the maintenance staff are granted permission by signallers to carry out the routine crash tests (see my earlier post to understand what I mean).

The public is completely unaware of how often some of these actually run. On my area, one standby diesel generator auto-started every time there was a lightening storm in the winter, because the remote interlocking it supplied was fed via an overhead 11kV pole route.

Only a couple of times did one of the standby diesel generators fail to auto-start when I was on duty. Many get replaced before they have even run for 2000 hours.

Of course, there will always be an exception...