Heathrow Power Failure: A Lesson in Infrastructure

 

Around midnight Thursday evening, Mar. 20, a transformer at the North Hyde substation in west London caught fire and failed, interrupting power to some 67,000 customers.  Power outages are not that unusual, and this one would not have made the news except for the fact that one of those customers was Heathrow Airport.  Although airport authorities claim that the backup emergency-power systems worked as expected, they decided to close the airport, which is the fifth busiest in the world. 

 

It took seven hours for firefighters to quench the blaze, and another twelve or so before power could be fully restored.  Consequently, all flights into and out of Heathrow were diverted or canceled until around 6 PM Friday.  The knock-on effects from this major disruption will be felt for several more days as stranded travelers find alternate routes and the transportation system strives to return to normal.

 

One engineering expert interviewed about the incident expressed surprise that there were not alternative supply paths for such an important load as the airport.  Apparently one terminal did not lose power, but the expert said that the grid in that area of London is "highly constrained" and has recently been stressed by increased development in the region. 

 

Admittedly, the failure of a substation transformer is an unusual event.  Such transformers can be the size of a small bus and cost on the order of a million dollars.  Utility companies normally monitor their condition remotely and perform routine maintenance on them such as cleaning or replacing the cooling oil that bathes the tons of steel and copper inside the sealed container that is visible to the eye.  Dropping the ball on any of these precautions can lead to a dangerous situation.  Impurities such as water can get into the oil, weakening insulation and leading to a sudden arcover and failure.

 

We will have to wait for investigation results to discover exactly why the transformer caught fire.  But another question is: why did Heathrow not have sufficient backup power to continue normal operations? 

 

The best guess is simple economics.  A modern airport uses a great amount of power for moving sidewalks, elevators, security equipment, and lighting.  Maintaining enough backup generators to provide the entire normal load would be expensive, probably complex because of the load's distributed nature, and would show up as a dead loss on the books of the private company, Heathrow Airport Holdings, which runs the airport.  So the bottom line is probably that the firm decided the temporary fallout from a short-term shutdown would be less expensive than paying for a large number of emergency generators that might be used only once every few decades. 

 

No one died or was even injured in the Heathrow shutdown, and in retrospect the decision to limit emergency backup resources was probably a wise one.  Nevertheless, this incident brings up an interesting issue with regard to how seriously we should take preparing for unlikely infrastructure failures.

 

In cases where a power failure could be deadly or extremely costly, institutions and organizations usually buy enough backup power to keep things running almost without interruption.  For example, most hospitals have enough backup power to keep operating rooms running normally, although the lights may go out in hallways and patient rooms if the main power fails.  You don't want a power failure in the middle of your brain surgery, and so most responsible hospitals make sure this can't happen.

 

Huge semiconductor plants also typically have enough backup power to keep their essential processes running without a hitch, although the emergency power system forms a costly and seldom-used aspect of the installation.  A former student of mine is heading the electrical installations at a major semiconductor plant, and showed me a photo of the ranks of large backup generators that are being installed.  If power were to go out in an operating semiconductor fab facility, the entire inventory all along the process lines would have to be scrapped, and this multimillion-dollar potential loss justifies spending extraordinary amounts to ensure that the machines keep running. 

 

On the other hand, while Heathrow is a vital part of Europe's transportation infrastructure, closing it for less than a day has not had many permanent ill effects.  The systems that communicate with airliners were unaffected and flights were successfully diverted, so other than a lot of travelers whose plans were disrupted and deliveries delayed, the incident will have few lasting consequences.

 

Once the cause of the fire has been determined, we may learn something about ways of preventing such fires in the future.  If the oil was dirty and led to deterioration in insulation, better maintenance is called for.  If the substation was overloaded due to new construction, maybe another substation is needed, or the North Hyde facility needs to be expanded with a second transformer.  And if there was some sort of transient or network disturbance that led to a stress failure, new smart-grid technologies can be brought into the picture to alleviate such incidents in the future. 

 

A transformer fire is one of the worst things that can happen to a power grid, but it does seem like some fairly minor changes in the distribution infrastructure could keep this from happening again.  A friend of mine who used to work in an aerospace job uses a phrase that would apply to this situation:  "single-point failure."  If a system has one component whose failure brings down the whole system, that single point is a vulnerability that should be addressed.  And maybe last week's shutdown of Heathrow will motivate changes that will keep it from happening again. 

 

Sources:  I referred to an Associated Press article on the incident at https://apnews.com/article/britain-london-fire-heathrow-airport-6d63b2f6615e8ff39f2647641bfbc160, a website called Open Conversation at https://theconversation.com/heathrow-closure-what-caused-the-fire-and-why-did-it-bring-down-the-whole-airport-expert-panel-252834, and the Wikipedia article "Heathrow Airport Holdings."

Tricks A Dumb Grid Can Play

Inauguration Day in Brookville, Pennsylvania arrived with a bang. Within minutes after Donald Trump swore to preserve, protect, and defend the constitution of the United States, the 911 calls began.  In one house, light bulbs were exploding.  Another resident reported that a power strip was smoking.  At another house, the siding was on fire, and at yet another the electric meter was engulfed in flames.  The main radio transmitter at the police station tripped out, so the 911 call center was unable to contact them until the dispatcher used his battery-powered radio to make contact.  In all, about 400 residents of the small western Pennsylvania town of 3800 suffered some type of damage, ranging from singed carpets to fried computers and exploding fluorescent light fixtures.

The power surge had nothing directly to do with Donald Trump.  A quick investigation by Penelec, the local electric utility, revealed that an insulator on a power line had failed.  What follows here is my extrapolation from the limited details in the Associated Press wire story, but represents what I think is a good guess.

Electric power networks are divided into transmission lines and distribution lines.  The transmission lines are the tall steel-framed towers that span many miles across the countryside, and are the interstate highways of the electric grid, transmitting megawatts of power from generating plants to substations many miles away.  To transmit this much power efficiently, the voltage of these lines is generally above 100,000 volts (100 kV).  For example, there is a 138-kV transmission line that connects Brookville with the rest of the power grid in western Pennsylvania, and there may be others at even higher voltages.

           

Once power arrives at a substation, it is stepped down in voltage with large, expensive units called transformers to a lower voltage suitable for distribution locally.  Distribution lines, the neighborhood streets of the network, carry voltages in the range of 12 kV to 35 kV, or occasionally higher in rural areas.  My guess is that Brookville has at least two or three separate distribution circuits with voltages in the 25-kV range.  The familiar wooden power poles that carry telephone and cable TV lines also support power-distribution cables, always suspended on the highest point of the poles.  Every few hundred feet, a "pole pig" (distribution transformer), usually a metal can a couple of feet tall, lowers the voltage still more to 240 V or less for delivery to commercial and residential customers. 

What probably happened was this.  One of the high-voltage insulators on a transmission line carrying in excess of 100,000 volts failed mechanically, dropping its conductor on or near enough to one of the town's distribution lines to allow a flashover (an arc) to jump from the 100+ kV transmission line to a 25-kV distribution line.  All power-line insulators are built with a safety margin.  That is, an insulator for a 25-kV line may be able to withstand 50 or even 100 kV, which can happen during situations such as lightning surges and so on.  This is good in normal circumstances, but in this case it backfired.

The insulation of the 25-kV distribution line held just long enough for the high voltage, four or five times normal, to get into the distribution transformers and ultimately the houses of about ten percent of the town.  So for a few seconds or maybe even longer, equipment designed for 120 V was receiving, say, 500 or 600 V. 

There are kooks on YouTube who delight in taking innocent electric appliances such as razors, clocks, toasters, light bulbs, and so on, and connecting them to high-voltage power sources just to see what happens.  They are never pretty.  Every electric appliance has a maximum rated voltage, and when you exceed it by 500% you either blow a fuse or, in the case of equipment that doesn't have fuses such as light bulbs, the excess heating makes something melt or vaporize or explode. 

Many power strips have surge-arresting devices in them meant to absorb fast transient surges caused by lightning.  But those surges usually last only milliseconds, and a surge of several seconds overheats such a device, making it smoke, which explains the reports of smoking power strips. 

Why didn't all the protective devices that a utility normally uses, such as fuses and circuit breakers, operate right away?  Because they are designed primarily for lightning strikes, not an overvoltage that lasts many seconds.  And the fuses probably didn't blow right away because a fuse only slightly over its rated current takes a considerable time to melt. 

Penelec has announced that they will compensate those who have suffered losses as a result of the surge.  Fortunately, no one was injured, but a considerable amount of property was damaged and the mess will take weeks or months to clean up. 

The "dumb grid" in the title refers to the fact that most electric utilities still use protective technology that was developed prior to World War I, for the most part:  fuses and electromechanical relays.  Innovative "smart grid" technology connecting the power grid to the Internet and replacing many electromechanical controls with faster-acting solid-state devices promises a number of good things, mainly pertaining to increased efficiency and reliability.  But it's also possible that if Brookville's grid was smarter—that is, if it could have figured out within milliseconds what happened and cut off the surge then—none of the bizarre damage might have occurred.

Admittedly, situations in which a transmission line arcs over to a distribution line are rare.  But as Brookville shows, they can happen.  If the new Trump administration wants to improve America's infrastructure, encouraging utilities to take the smart-grid path is one way to do it.  Whatever Washington does about it, though, it's too late for a few hundred residents of Brookville, who are still replacing siding, light bulbs, and computers as a result of a freak accident that shows we still have a ways to go in improving electric utility safety and reliability.

Sources:  The Associated Press report on the Brookville incident of Jan. 20, 2017 was carried by a number of news outlets, including  PressFrom.com at http://us.pressfrom.com/news/us/-21781-tiny-towns-power-surge-fries-computers-appliances-siding/ on Jan. 28, 2017.