sábado, febrero 04, 2012

Smart Grid: SoS "… interacting in unpredictable ways that regulators and investors cannot comprehend, far less control.”

Summary: Similar to financial markets, system crashes are expected in smart grid, because they have been though to be just complex technological systems, when they are in fact ultra large scale socio-technical systems. The difference between the two kinds of systems is told in “… the story of the London Millennium Bridge, which opened in June 2000 and two days later was closed for two years to remedy destabilizing swaying motions induced when groups of people walked over it.” As industry restructuring was flawed, legislators, regulators, and investors have a change to minimize the damage in the making on the power industry, by learning about their responsibility of the now known error of the Normalization of Deviance before it is too late.

---

At the very end of last year, the Financial Times published the article “Flash Crash Threatens to Return With a Vengeance.” That article, whose first sentence says “What should investors watch in 2012?,” introduces:

[A] fascinating transatlantic research paper published by the Bank for International Settlements. One of the paper’s co-authors is Dave Cliff, formerly a financial trader who now runs the UK government’s Large-Scale Complex Information Technology Systems project, an endeavour that analyses the risks of IT systems in sectors including healthcare, nuclear energy and finance. The other, Linda Northrop, runs a similar project at Carnegie Mellon University, which was initiated a decade ago by the US military.

For starters, the original report, “The global financial markets: an ultra-large-scale systems perspective,” gives a strong answer to the EWPC article Should the Smart Grid be a Technological Project to Address a Challenge Faced by Utility Executives? The answer shows the very large potential flaw introduced in the power industry architecture for the smart grid. The flaw is selecting the project as a technological system instead of a socio-technical system. The authors say:

Take, for example, bridge building. As an engineering activity this is something that dates at least as far back as ancient Rome (c.150BC) and so probably doesn’t figure as a risky technology for many people. Yet for decades, engineering students have been taught the story of the Tacoma Narrows suspension bridge, opened in July 1940, which collapsed four months later, where the designers did not anticipate the prospect of wind-flows over the bridge deck reinforcing the deck’s natural mode of vibrations, leading to the bridge shaking itself apart. Presumably, current and future students will also be taught the story of the London Millennium Bridge, which opened in June 2000 and two days later was closed for two years to remedy destabilizing swaying motions induced when groups of people walked over it. A significant difference between Tacoma Narrows and London Millennium is that in the latter case, it was the interaction of people, the users, with the engineered system that caused the problem. The Millennium Bridge on its own, as a piece of engineering, was a fine and stable structure; but when we consider the interaction dynamics of the larger system made up of the bridge and its many simultaneous users, there were serious unforeseen problems in those dynamics that only came to light when it was too late.

The Financial Times article says among other things:

In recent years, these two teams have used engineering and science skills to analyse what they call socio-technical risks, or the dangers that occur whenever complex technological systems proliferate, creating “systems of systems” that nobody understands. In early 2010, well before May 6, they released a brilliantly prescient report that predicted that a systems failure loomed.

Since then, they have continued their research, with sobering conclusions. Most notably, these researchers believe that the flash crash was not an isolated event; on the contrary, it was entirely predictable given how IT systems have proliferated to create a system of systems that is now interacting in unpredictable ways that regulators and investors cannot comprehend, far less control.



Usually, this danger is not visible to investors. After all, markets generally work well, leaving financiers in the grip of a phenomenon that Diane Vaughan, the sociologist, called the “normalcy of deviance” (based on her work on the Challenger 1987 space shuttle disaster): because people have sailed close to the wind and survived, they assume they can continue to do this — and turn a blind eye to anything that seems uncomfortably bizarre.

Is there any solution? Cliff and Northrop offer one idea that might help: regulators and bankers should replicate what some scientists have done elsewhere and join forces to create a cross-border computing centre that is capable of extremely advanced, large-scale financial simulation. Their idea is that this would essentially replicate what is done in meteorology or complex engineering to map the markets — creating the equivalent of wind tunnels to test new financial products and ideas, and warn of looming trouble.


As a result of research by others and myself, there is another idea that might be considered to introduce a shift from a technological to a socio technical systems architecture. The idea is the design of unprecedented systems, as suggested by Eberhardt Rechtin and Mark Maier in their book “The art of system architecting.”

Most people know that, according to Wikipedia, Murphy's law is an adage or epigram that is typically stated as: "Anything that can go wrong will go wrong". As can be seen below, that’s the case with the financial markets, but also the emerging electricity market under what’s been called the Smart Grid. But, according to Rechtin and Maier, Murphy’s law is a systems-level descriptive architecting heuristic, which leads to several descriptive heuristics, such as:

  • Simplify, simplify, simplify.

  • The first line of defense against complexity is simplicity of design.

  • The most reliable part of an airplane is the one that isn't there - because it isn't needed.


To see the progress already done, below can be found posts in the EWPC Blog that quote Rechtin. The data includes the title, the date posted, and the number of times it has been viewed, so far.