Monday 7 January 2019

COMPLEX SYSTEMS

A few days ago, Dr North at EU Referendum wrote a terrific post (HERE) about the way the Dover-Calais transport link had developed after the eastern dockyard, formerly a Royal Navy base, had been handed over to civilian interests in 1946. It's a quick tour de force in how complex systems come into being and a very useful reference. He has continued on the same theme for a day or two now. All well worth a read.

Brexit of any kind is going to throw a spanner into the works. The route will probably recover eventually, but it may take some time and may never again reach the volumes we see now. Just in time trade may vanish altogether.

Reading about it put me in mind of a disaster which struck the eastern seaboard of the USA in 1965 when there was a massive interruption in power supplies. I studied this as an Open University module around 1970. It is an example of a highly complex system going out of control.

MIT (Massachusetts Institute of Technology)  has an explanation of it (HERE). Power in the NE USA was provided by a complex grid that had grown over time - just like the Dover/Calais route:  MIT explain it here:

"Two major power grids-the Ontario-New York-New England pool (formally known as the Canada-United States Eastern Interconnection, or CANUSE area) and the Pennsylvania-New Jersey-Maryland pool (the PJM interconnection)-together made up the northeast power grid, which provided a flexible network of power suppliers that could quickly meet fluctuating demands within the many parts of North America's most densely populated region. While the grid, which remains intact today, has proven to be highly effective, the night of November 9, 1965 serves as a reminder of how thoroughly our electricity-dependent lifestyles are tightly interwoven with the complex workings of a massive but often overlooked technological system".

On the cold November night a single faulty relay at the Sir Adam Beck Station no. 2 in Ontario, Canada caused a key transmission line to disconnect because it detected the line was about to be overloaded. The intention was to protect the overhead line itself.

To maintain power to the areas affected they were switched to other inter connectors, which themselves were running near to the maximum load. They then in turn became overloaded and automatically tripped out, the system attempted to reroute power again and again, causing more overloading in a controlled cascade - all working exactly as intended:

"At 5:27 p.m., November 9, 1965, the entire Northeast area of the United States and large parts of Canada went dark. From Buffalo to the eastern border of New Hampshire and from New York City to Ontario, a massive power outage struck without warning. Trains were stuck between subway stops. People were trapped in elevators. Failed traffic signals stopped traffic dead. And, at the height of the Cold War, many thought Armageddon had arrived. One pilot flying over a darkened New York City stated, "I thought, 'another Pearl Harbor!'" By 5:40 p.m. that evening, 80,000 square miles of the Northeast United States and Ontario, Canada, were without power, leaving 30 million people in the dark".

Engineers watched this happening over a 13 minute period but we're unable to prevent it or to mitigate it. The system had become so complex that no one person understood it or the ramifications of a simple and routine circuit breaker operating exactly as it was designed to do.

It was not a failure of equipment but a failure of those responsible for the  development of a system over a long period, to understand how it might work under every potential circumstance. As I recall, the initial relay was actually set too low, and the power line was quite capable of carrying the load anyway!

This is also what happens when your computer 'freezes' or 'locks up'. The processor has got itself into a loop which the software engineer has not foreseen and it has no way of recovering from. Essentially, the computer doesn't know what to do.

All of this came to mind reading Richard North's blog on the Dover-Calais crossing.

The crossing has grown slowly in efficiency and complexity over a very long time, allowing it to cope with around four million commercial vehicles (not cars) a year, presumably close to its maximum possible rate. Each step being tested and modified as necessary before being implemented.

Now Brexit is going to change all that and probably nobody really understands what might happen. A no deal Brexit will change overnight what has taken seventy years to develop. We don't know what the Calais officials will do and we don't know what advice will be given or how all the myriad parties, the ferry companies, border officials, freight forwarders, haulage companies, drivers or exporters and importers will react or even what extra infrastructure is needed. Don't forget the authorities controlling the route have no idea what is in any of the trucks carrying UK-EU traded goods.

Unless everything and everyone behaves as expected and in a completely deterministic way, the result could be chaos for days or weeks with knock-on effects which no one can foresee. Indeed it is probable that some small unforeseen issues could creat problems like the CANUSE grid in 1965, where a single circuit breaker caused massive disruption.

I have been involved in technology of one sort or another all my adult life and anyone who thinks a single person understands these complex systems is fooling themselves.

The best advice is that if something can go wrong, it almost certainly will and at the most inconvenient moment when the system is under maximum stress.

I note Dr North this morning is now saying:

"And it all points to one thing: a no-deal scenario is not a credible option".

I agree with this. But if the PM's deal doesn't get ratified there is only one other possibility - No Brexit.