It’s been a bad month for Network Rail. High profile failures over Christmas at King’s Cross and Paddington, followed by problems at London Bridge, have put plenty of pressure on Mark Carne’s team.

London Bridge is in the middle of a massive improvement programme to deliver a better Thameslink service from 2018. With platforms closed for rebuilding, NR and the station’s train operators had already decided to cut services. (Thameslink trains diverted away and Southeastern services to Charing Cross no longer calling.) This proved inadequate against the crowds using the station in the first working week of January.

NR Route Managing Director Dave Ward told the BBC: “We need to step back and see exactly how we have got into this situation – what has caused this. We believed, and the execution of the works kind of proved it, that our plans were very tight and were very thorough. We’d been modelling this a number of times.”

He added: “We thought we would be presenting a new station and a new method of operation after several months of misery leading up to it and that hasn’t happened and I’m deeply regretful for it.”

As he spoke, NR was preparing to publish a report explaining what had gone wrong at Paddington and King’s Cross, were passengers were disrupted on December 27 2014. It duly arrived on January 12.

This report explained that at King’s Cross a plan to renew points at Holloway quickly went wrong when untried equipment proved faulty. Problems then snowballed with the result that drivers needed to move trains of ballast and other materials ran out of hours. No replacements could be found.

The report also revealed that a plan agreed between NR and train operators was not implemented by local staff. The plan should have seen long-distance trains arrive in Platform 4 to drop southbound passengers before shunting to Platform 5 to pick up northbound passengers. A local plan decided by station staff and signallers at King’s Cross box used only Platform 4, removing the shunt. However, this meant that passengers struggled to leave incoming trains because the platform was already full with those waiting to board.

NR suggests there was a “short term breakdown in communications in how to handle longer distance trains”. It doesn’t make clear whether the NR-TOC plan to use two platforms was not communicated to station staff and signallers or whether it was communicated and then changed locally.

Details are sketchy of the problems at Paddington with NR’s report failing to get to the bottom of why signal testing expected to take two hours actually took ten. Unlike at King’s Cross (where there was warning of 14 hours that the line would not open on time), there was no warning at Paddington.

Signal testing around Old Oak Common (OOC) was reported as finished at 0330 with First Great Western’s first long-distance train timetabled to leave Paddington at 0730. Testing was in the hands of NR’s contractor SSL (a joint venture of Alstom and Balfour Beatty). SSL had a significant overrun at Poole last May when resignalling the line to Wool. A report into this found that there had been no senior SSL management on site, that communications with NR had been poor and that NR’s and SSL’s project offices were geographically remote.

Remedial action was taken for OOC’s work but NR now admits that it was unlikely the Poole investigation found the root-causes of that overrun. Another review of Paddington beckons with NR’s report saying: “The work management processes of SSL that led to the incorrect conclusion that the signalling testing of the main lines was complete at 0330 will be thoroughly reviewed by SSL and Network Rail staff.”

Checks of paperwork supporting the testing showed there was more to do. This was both physical work that needed to be redone or rechecked and inconsistencies in the paperwork that needed to be resolved. What is absent from the report is any explanation of why work had to be redone and why there were inconsistencies.

It’s clear that NR’s planning and management needs to improve. Both project passed NR’s risk tests as being 95% likely to be completed on time. There’s no mention in the report whether those conducting these tests knew, for example, that the King’s Cross contractor was using untried kit.

There’s now likely to be pressure to allow projects to have more time to complete work. This could mean more weekend closures or more weekday blockades of lines. This is an option that TOCs and most passengers are unlikely to welcome but there should be a clear debate.

There must also be a debate about the level of work being done. With two major overruns in the last year, NR could be very tempted to drop SSL as a contractor. However, with only a handful of signalling contractors to choose from, and a massive amount of work looming, that option seems untenable.

So it is with experienced project managers. NR has a laudable apprentice programme and puts considerable effort into management training. Despite this, it’s seen key staff leave for other companies, such as HS2. Are we now seeing the effects of the long-talked-about skills shortage?

A shorter version of this article appeared in RAIL 766.

By Philip Haigh

Freelance railway writer, former deputy editor at RAIL magazine - news, views and analysis of today's railway.