An issue paper on the U.S. Northeastern Blackout, August 14, 2003
During my many years as an engineer working as a consultant with accountants, I have been amazed at the data systems that accountants have created to keep track of investments and expenses, especially the way that these data systems can capture minutiae that boggled my mind. And my MIT Sloan School Masters had a concentration in Information Systems.
As I read stories about the blackout, I am amazed that the engineers don’t have the same quality of data systems that the accountants have put together. Maybe the accountants deal with money all the time and can more easily justify the data acquisition and storage necessary to keep track of money. I guess money begets money. Maybe we can use that concept to lessen the extent of the next blackout, so that it is a whimper instead of a bang.
Paying for Power Surges Between New York and Canada
The New York engineers analyzing the events of August 14 have talked about the massive surges of electricity to and from Canada. Since 1989 I have written several published articles saying we should price unscheduled flows of electricity, including surges of power such as were reported between New York and Canada. By attaching dollars to these surges, that is, by having a payment mechanism for them, we would cause cash registers to ring.
With cash registers ringing, the New York and Canada revenue accountants would have had data systems in place. These data systems would capture the information necessary to charge Canada for the electricity surges to Canada. The data systems would also determine the amount of electricity for which New York should pay Canada for the surges of electricity from Canada. And the engineers could use the accountants’ data systems to know what happened.
Coordinating Payments in New York With the Midwest
Reading reports about Midwest engineers analyzing the events of August 14, I see a set of comments similar to those in New York, though expanded to include details of voltage problems, of lines tripping out, and of plants going down substantially before the blackout. Again, there were surges from and to Canada. Putting the two sets of reports together tells the engineer in me that the Canada payments to and from New York should be coordinated with the Canada payments from and to the Midwest.
My differing ordering of prepositions “to” and “from” in the previous paragraph is intentional. The surge to Canada from New York is likely to have occurred at the same time as the surge from Canada to the Midwest. Massive surges of power are usually the result of a transmission line opening (or closing.)
When a transmission line opens, the power that had been flowing on that line flows on other available paths, and it looks like a surge on those other lines. With power flows from the Midwest to New York, opening a power line elsewhere, in Pennsylvania for example, would cause power to surge from the Midwest to Canada and from Canada to New York. This surge is instead of the electricity traveling on its original path through Pennsylvania.
The sudden increase in power flowing through Canada would be matched by a sudden decrease flowing through Pennsylvania. But our reporting styles tend to be dramatic. Thus we tend to report “surges” or sudden increases with another nation instead “un-surges” or sudden decreases across the connection with another state. Or maybe it is just a deficiency of our English language.
Rewarding Transmission Owners Who Helped Keep the Lights On
In regard to these unscheduled surges, I say we need to reward those entities that help us keep the lights on. Presuming that reported sudden surges across Canada helped keep the lights on, the pricing mechanism should reward Canada for allowing the sudden surges to happen. That means coordinating the payments to and from Canada at the Midwest interconnection with the payments from and to Canada at the New York interconnection. Again notice the tricky change in the order of the prepositions.
The price for the surge out of Canada should be higher than the price for the surge into Canada. This allows Canada to be rewarded for keeping the lights on and to pay for the costs that Canada incurred in handling the surges. At a minimum, Canada would incur the cost of the incremental heating of its transmission lines, generally referred to as marginal line losses. Under extreme conditions Canada might have had to change the loading on its generating plants, which is a costly endeavor.
A set of prices that are different at Canada’s border with the Midwest versus at Canada’s border with New York is called locational marginal price or lmp. Most industry people use the capitalized LMP, but I am leery of doing so since these same people have a specific procedure in mind when they use the term LMP. Generally the procedure is a bid based auction for future delivery of electricity. I believe that the lmp system for unscheduled flows of electricity needs to be different from the LMP derived from a bid based auction for future deliveries of electricity. After all, unscheduled flows are not something anyone can anticipate and thus no one is in a position to bid to provide these unscheduled flows. We need a market system other than the bid based auctions that have been developed in the last decade.
Synchronized Data Systems
The reported surges of power lasted for short periods of time, perhaps a minute, perhaps only a second or so. In the Eastern Interconnection (which covers parts of the US and Canada), meters record data about every two seconds. Some of this information is archived. However, much of the information is immediately discarded. Also, as news reports indicate, this information is poorly synchronized. Not only do we need to synchronize the pricing, we need to synchronize the data and retain the data.
In contrast, news reports indicate that meters in the Western Interconnection record synchronized data as often as 30 times a second. This makes me think that the accountants are playing a bigger role in gathering information on the Western Interconnection. We need to determine an lmp system for similarly short periods of time. The prices from the lmp system may be averaged over longer periods of time, but these longer periods of time should be much shorter than the hour long periods that are used to report inadvertent interchange or even the five or ten minute periods being used by some Independent System Operators for their LMP forwards markets.
Paying Responsive Generators and Loads
lmp would reward Canada for having the transmission lines that kept the lights on in New York. At least the news reports suggest that Canada helped keep the lights on for the first few surges. However, lmp would also provide incentives for generators and consumers to react to the transmission constraints. For Canada to be rewarded for keeping the lights on, the pricing mechanism needs to be consistent on each side of Canada, at the interconnections with New York and at the interconnections with the Midwest. Expansion of this consistent pricing plan across the interconnection would provide the necessary incentives for generators and consumers to react.
Generators and consumers downstream of the constraint should face a high price, the same high price mentioned for the Canada interface with New York. Looking at the map, downstream of the interface includes not just New York but also New England. An appropriate lmp would thus reward people in New England that responded to the constraint and helped keep the lights on in New York.
The lmp incentive is not limited to those who are downstream of the constraint. Those who are upstream of the constraint should also have an incentive to help keep the lights on. The incentive for those upstream of the constraint would be in the form of very low lmp, the very low price mentioned for the Canada interface with the Midwest. Looking at the map again, upstream of the interface includes not just the Midwest but could potentially include the South. This depends on the where the constraint was located. For instance, if the constraint cut across Canada and Pennsylvania/New Jersey, a huge area would be upstream of the constraint.
How a Low Price Provides an Incentive to Generators
Owners of electric generators have contracts to provide electricity at fixed prices to their customers, a form of futures or forwards market. Indeed, a utility tariff can be considered to be a forwards market. Most generators burn fossil fuels to produce electricity. A low lmp for unscheduled flows of electricity can provide an incentive for generators to shut down. When then generator shuts down, its owner would then obtain electricity from the grid at the lmp price for unscheduled flows of electricity.
Consider a generator with a fuel cost of $30/MWH facing an lmp of $3/MWH. Any reduction in production level would save the owner of the generator $27/MWH. A large gas fired turbine generator might be sized at 100 MW. The owner of such a turbine generator could thus earn $2,700/hour by reducing production, reducing the use of fuel, and using the cheap power to meet its delivery obligation. Again looking at the map, upstream of the interface between Canada and the Midwest includes a lot of land.
There was extensive reporting in Washington, D.C., about a discussion between the Midwest ISO and a merchant power plant operator. MISO tried to jawbone the merchant power plant operator into reducing its generating level because of the location of the power just upstream of a heavily loaded power line. The merchant power plant operator wanted instead to increase power due to requests from its sales staff. A properly functioning lmp would have provided the merchant power plant operator an incentive to reduce generation while still providing the sales staff the electricity it wanted at very low prices to meet its sales obligation.
Pricing Reactive Power
The least understood aspect of electricity is reactive power. Most people know that our delivery system is AC, for alternating current. Our electricity is 60 Hertz, or 60 cycles per second. Most people also know that a substantial amount of our electricity is used for motors which contain electromagnets. Some of the electricity is used to “charge up” the magnets in the motor, turning a chunk of steel into an electromagnet. The electromagnet then uncharges. The charging and uncharging occurs 60 times per second.
Since the magnet must be charged before the motor can turn, the current and the voltage do not increase and decrease at the same time, they are slightly out of phase with each other. For magnets and other inductive devices, the current lags behind the voltage. For capacitive devices, the current leads the voltage. Power lines can have both inductive and capacitive effects depending on configuration and loading levels. Oh, yes, lagging power factors result in low voltages in an area. Leading power factors result in high voltages in an area.
One of the issues mentioned in some news reports is that some locations experienced unusually low voltages during the hours that preceded the actual blackout. These low voltages mean that the lagging load in the area was too great. Low voltages can often be solved by having local generators operate with a leading power factor and by operating devices known as capacitors. Or at least, some of the voltage problems can be solved this way.
lmp can be used to set prices for reactive power in much the same way that lmp can be used to set the price for active power. Again, the reactive lmp must be coordinated with the active lmp and with local conditions all in an effort to reward those whose efforts kept the lights on.
We need to install revenue accounting for unscheduled flows of electricity. Revenue accounting for unscheduled flows of electricity will provide a financial rationale for improved metering and data gathering systems. These data systems would have allowed the engineers to determine the details of the August 14 blackout much faster than existing data systems permit.
Revenue accounting for unscheduled flows of electricity would also encourage wider participation in avoiding blackouts. Presently, transmission operators attempt to jawbone individual plant operators to help with lowering the loading on transmission lines. Pricing unscheduled flows of electricity provides every plant operator with an economic incentive to help keep the lights on.
However, revenue accounting for unscheduled flows of electricity requires prices for these unscheduled flows of electricity. Only FERC can rule on that issue. The North American Energy Standards Board is now investigating the issue in its Inadvertent Interchange Payback Task Force. Previously all but two of NAESB’s standards for gas have been adopted by FERC as rules. We shall see what happens with electricity, especially with regard to inadvertent interchange.
| Ravinder Singh
| Dear Mark, It
seems you have wasted you time at MIT and you
were attached to some accountants than some engineering
company. Magnets are charged/discharged 2 times/
cycle and every second about 120 times not 60.
You have advocated handing over or give more controls
to accountants replacing engineers in organisation
of electricity system. This is the real cause
of worsening power system in North America.
You have even proposed pricing of power surges and tried to legalise abnormal rare conditions. Tommorow you will start Traing Program also on pricing of surges. American power system will not improve until power plant engineers and electrcal engineers are brought back as key decision makers and planners.
You already had accountants / MBAs in control of ENRON. Why do you want repeat of ENRON everywhere? -----firstname.lastname@example.org
| Mark Lively
| Margnets are
indeed charged 2 times a cycle, once with one
polarity and once with a reverse polarity, about
60 times each way per second, a total of about
120 times in total. Sometimes the shorthand of
writing leaves out the distinction that each of
the two polarities occurs about 60 times a second.
As far as handing over control of the system to accountants, I just want to give the engineers operating the system better incentives to do the right thing to keep the lights on. There have been several articles about how merchant power plants don't get paid for reactive power and thus have no incentive to produce such power. Indeed, since reactive power can be used to fill up a transmission line, the absence of a price for reactive power provides merchant power plants a way to clog the transmission line with bad reactive power and prevent others from competing for the real power that has a price. My concept of getting better revenue accounting to set an appropriate price for reactive power does not give accountants control of the power system, merely gives the engineers more incentives to do the right thing.
As far as the role of engineers as key decision makers and planners, everyone should be involved. Since someone has to pay for the plan, the accountants and MBAs must also be involved. And the lawyers.
| Peter Manos
| Ravinder, yes
we all have job titles, and some people are labelled
engineers and others accountants. I wouldn't blame
one group any more than another, as in my opinion
they are all victims of a transmission system
being forced to perform in ways it was not originally
designed to perform. As for Enron, a thief is
a thief in any profession...but you are certainly
entitled to your opinion too.
I think the truth lies somewhere in between. Utilities are keeping thousand of turbine-generator sets rotating in synchrony at 3600 RPM. It was hard enough to do this when things were more islanded. Whether Mark's pricing proposals will work or not, it is of value to ask new questions like this. The key incentive that is already there, however, is to protect the equipment, which in the case of the 8/14 blackout required shedding of the load....and which is why circuit breakers may be monitored 30 times a second (because a momentary spike of several 1/60th of a second could burn them out).
What I do not understand is why NY went down. At the time NY isolated itself from Canada, it was (paradoxically for the summer) exporting power to Canada. So why couldn't some subset of the NY state generating units have been shed after that point, run in a no-load state, and keep the state as a whole powered up?
Regards, Peter Manos
| Len Gould
| Peter: I'm guessing
re NY going down even though it was in an export
position before isolating, but I think NY's export
surplus probably came from it's share of the Niagra
hydro plants, which are (?) in the same facility
as Canada's share and can't be isolated from each
other. NY isolating from Canada means it also
cuts off a large share of its own power. Ditto
for Ontario, on a peak hot day can't isolate from
NY without loosing it's Niagra gen, critical to
operation. Neither side can island itself without
loosing the large contribution from Niagra.
I agree, its a poor arrangement, and have previously advocated installing of load management units such as thyristors or even tap changers to fix the problem and allow islanding.
Of course the entire problem (and many others, e.g. VAR issues etc.) is a result of using vast interconnected AC grids. Before committing enormous sums to metering and fast recording/storage systems which essentially will contribute nothing physically to the solution, consideration should be given to converting strategic inter-region (at least) interties to DC transmission. The costs would not be that bad since existing cabling can easily carry more power with lower losses as DC than AC, and much of the existing transformer setups should be re-useable.
I think 8/14 was just a small foretaste of coming problems caused by underfinanced transmission trying to tune huge continental grids for unpredictable power flows. If it is given that merchant power gens, windpower, distributed microgen etc. are desireable, then a replacement needs to be implemented for the dwindling group of experts who used to "twiddle the dials" of the AC grid like an old tube radio. I think that's the top priority.
But that's just my opinion.
| Peter Manos
| Thanks Len--appreciate