Saturday, January 21, 2012

Rail security matters for Singapore: Questions to mull over

There must be many anxious hearts in SMRT hoping that the electric mass rapid transit (MRT) train system will behave itself over the four-day Lunar New Year long weekend from today till Tuesday.

The holiday season from pre-Christmas till now has not been a happy one for Singapore's largest train and bus operator, SMRT Corporation Ltd.

Hapless train commuters who experienced the pre-Christmas SMRT train breakdowns could say the same.

As we wait for the Committee of Inquiry (COI) set up by the Singapore government to tell us more, all parties involved in the matter must align themselves to the common objective of ensuring better transportation safety, security and reliability for the city-state's MRT system.

Any attempt at grandstanding the media and public to score points (political or otherwise), corporate bullying and selective release of facts and figures to gain public relations (PR) mileage will do nothing to help us pin down the root cause(s) of the train breakdowns. We must work towards the objective of making Singapore's transportation security protocols not failure-proof (because this is an ideal state that cannot realistically be achieved), but less failure-prone. We can only do so if we focus time, energy and resources to serving commuters better, rather than setting up others for the fall.

Key questions that we need to ask include:

1. How are our transportation security protocols calibrated to handle civil emergencies?
The Singapore Armed Forces (SAF), Singapore Civil Defence Force (SCDF) and Singapore Police Force (SPF) have put their forces through dress rehearsals involving a terror attack on Singapore's MRT system.

If the causal factor(s) leading to a train outage is not scripted in the CE response protocol, who makes the decision whether SAF, SCDF and SPF forces should be marshaled and deployed to handle the kind of transport chaos we witnessed 10 days before Christmas?

If the end result is the same, i.e. trains stopped in their tracks and people unable to get anywhere, would it be prudent to ensure that the steady flow of people milling in and around MRT stations does not result in people ponding which could endanger lives? It is a stampede risk, not to mention a safety hazard to vehicles when the people ponding spills onto busy roads during rush hour.

There is concern whether our CE protocols are so finely defined that the absence of a clear and present danger to lives/property, explosive devices and/or fire risk makes a train breakdown - however severe - unworthy of being deemed as a civil emergency. Perhaps mindsets need to be rewired?

2. What should SMRT and commuters do in future situations?
In my mind, SMRT's crisis management could have been better executed.

But as we castigate the company for not doing things better, we should realise that it would take an inordinate number of buses activated for bridging services to move every passenger affected by a system failure.

Let's work out the numbers. It has been estimated that 127,000 commuters were affected during 15 December 2011 system failure. Buses activated by SMRT to provide bridging services to move commuters between stations with stalled trains could each carry 80 to 120 pax. Even with around 90 buses plying the bridging service, this leaves a backlog of tens of thousands of commuters running out of time and patience. Mind you, they were stranded not on any Thursday night but during a holiday period 10 days before Christmas.

Evening rush hour, plus pre-Xmas shopping crowd, plus larger population thanks to the surge in immigrants plus more people obeying the call to use public transport either by choice or because of sky high car prices, plus a train system that is a radial line along a one-way street (Orchard Road) - SMRT couldn't ask for a more perfect storm.

Sure, more buses could be activated if we were playing a table top exercise. But how realistic is this? Would it not end up choking Orchard Road - a one-way street already congested during the evening peak hour - with even more vehicles?

For commuters to help themselves in future situations by walking away - as many Londoners did after the Tube bombings - people need accurate, relevant and timely information on the extent of the delay, the number of stations involved and the estimated duration to remedy the situation.

Fog of war - Dealing with imprecise information
Alas, the fog of war also affects peacetime situations and SMRT grappled with understanding the magnitude of the breakdown.On Thursday 15 December 2011, the first SMRT train was unable to move after losing power at around 6:47pm. By 6:55pm, three trains had stalled. Trains stalling are not unknown and it takes around five to 10 minutes for the train operator (TO) - usually the sole SMRT staff onboard a train that can carry 1,200+ pax - to run through safety checks and train restart protocols before the six-carriage train can resume its journey.

It is worthwhile assessing how TOs and controllers in the operations control centre, where train movements are directed and monitored, can build a better rail situation picture from fragmented and sometimes conflicting information trickling in from stalled trains. It appears some TOs radioed in to report stalled trains as the emergency lighting kicked in. But there was at least one train with three train cabins fully lighted (because the shoes could still collect power from the third rail) and the last three darkened cabins lit by emergency lighting. As luck would have it, the powered cabins were closest to the TO, so he reported that the train had stalled but was still drawing power. Other than the fact that the train was unable to move, all looked fine.

Another train is said to have limped from the Orchard area to Braddell, with internal lights flashing on and off as the train shoes struggled to maintain contact with the third rail which supplies the train with electricity. Power supply eventually died near Braddell. So this sort of conflicting information does nothing for OCC controllers who were saturated with data while trying to figure out what was happening on the ground.

This incident makes a compelling case for SAF information managers to collate, sort, prioritise and make sense of tactical information from frontline units to form a comprehensive, best effort picture to penetrate the fog of war.

3. How can mass media be better employed to share time critical information?
For reasons too technical to explain and too sensitive to mention, Singapore's short messaging system (SMS) run by local telcos cannot broadcast SMS alerts to everyone at the same time. We're talking about a subscriber base of several million mobilephone users.

A better way might be through radio and television. As with the CE issue raised above, newsrooms need to better calibrate their sensitivity to breaking news, particularly in situations with no deaths, no bomb or fire. The reality of the transport business is that mechanical things do break down. The trick is not making the broadcast decision regime too sensitive and crying wolf all the time.

4. What impact, if any, did SMRT's fourth generation trains have on the rail network?
The fourth generation trains of the C151A model used today are some 20 tonnes heavier than the lightest SMRT trains and around five tonnes heavier than the 3rd Gen Kawasaki trains.

The 4th Gen trains were made in China. As China-made products have a tarnished image for shoddy workmanship, poor QC and substandard materials, questions have been raised over the choice of the C151A rolling stock.

The fact that the 4th Gen trains came from Chinese factories concerns me less than the question of whether the heavier rolling stock was vibration tested? If not, then why not?

It is understood that in the immediate aftermath of the 15 December outage, China-made trains were withdrawn from the North-South Line as a precaution. The N-S Line is the stretch on which the trains were stalled. These C151A trains have since been reintroduced on the N-S and East-West Lines, but run at slower speeds over certain stretches of the tracks.

5. Give the public the background to all statistics.
Numbers released in Parliament point to an increase in disruptions on the SMRT North-South and East-West lines that were timed between five and 10 minutes long.

In 2007, there were 213 such incidents clocked. In the first 11 months of 2011, 217 disruptions were recorded. In and by itself, this rising trend does not look great. The conclusion that there is something inherently wrong with SMRT's maintenance regime is therefore natural and not unexpected.

I believe Singaporeans deserve to know just how many of these incidents were contributed by the installation of half height platform screen doors at elevated MRT stations. In my opinion, it is wrong to regard disruptions caused by lack of synchronisation between train doors and newly installed half height screen doors in the same way as train breakdowns arising from the loss of motive power.

The release of data sets without any context ends up tarring the reputation of SMRT employees, particularly its engineers who have been working behind the scenes to make things right. Worse, it does not make Singapore's transportation security any safer because people may jump to the wrong conclusions based on such data. Looking further downstream, which young engineer would want to join a company that is seen as technically slack? In the mid to long-term, SMRT's rail expertise will be hurt if engineers look elsewhere for employment, to the detriment of us commuters because rail engineers sit at the core of SMRT's system reliability, safety and efficiency as a people mover.

There are indications that SMRT lines that were shutdown for overnight checks after the second outage on Saturday 17 December 2011 could have been reopened earlier, if not for additional checks imposed on SMRT engineers.

Just what were these checks and were they in any way related to the dislodged claws holding up the third rail which trains tap motive power from?

In my opinion, the delay in reopening SMRT stations that Sunday ended up making the company's rolling stock team look inept and technically incompetent in delivering what they had promised. But in hindsight, is this impression fair?

I am hopeful the additional checks were introduced on the grounds of caution and that there was nothing sinister behind this.

There are many people tracking this investigation, not just the COI. All parties should therefore perform their duties honorably and professionally.


Hsien said...

Hi David,
Just a comment on Point number 3.
How can mass media be better employed to share time critical information?
I'm not sure tv and radio reception is available in the tunnels. But twitter has been a useful tool for emergency management. During the recent Queensland Floods. Twitter was used by the Queensland Police Service to broadcast updates. PUB has started broadcasting flood updates on twitter and Channelnewsasia and 90cents both have twitter feeds. Just looking around the MRT carriage, every other person has a smartphone and one in each carriage is bound to have twitter feed.

David Boey said...

Hi Hsien,
If protocols are followed, updates for commuters on station platforms or in trains should be handled by the rail operator.

In my opinion, mass media alerts help control the inflow on people into stations as this could contribute to further congestion.

re: Smart phones. True. But do these same folks subscribe to the twitter feeds that pump out time-critical messages? Chances are, many don't.

Queenslanders live in a different setting. They have cyclones there, wild fires and floods. People who live in a place beset by natural disasters are hard-wired with a survival mentality that makes them more self-reliant.
Singaporeans are not like that.

If you walk around the train cabin asking commuters if they have something as simple as the ICE (In Case of Emergency) contact keyed into their phone list, I bet many people would say no.

Many recreational joggers go for their runs with only their car keys or house keys in the pocket, little realising the danger they bring upon themselves if they collapse or are involved in an accident without any ID. To them, accidents only happen to other people.

So it's going to take time for something like twitter and the self-reliant, can-do spirit to filter down to the masses.

If and when SMRT shares its take on twitter and the survey numbers, I think this would help observers understand the constraints better.

Agree, twitter has its usefulness. But the message, not the medium used to communicate, is what matters in crisis comms situations like the one discussed here.

Best Regards,