Grid Security in India

On 2011 February 26, S.K. Soonee, CEO at India’s Power System Operation Corporation Limited posted on LinkedIn’s Power System Operator’s Group a link to a paper written by the staff of India’s Central Electricity Regulatory Commission.   “Grid Security – Need For Tightening of Frequency Band & Other Measure” can be accessed at  Through LinkedIn I provided the following comments.

 I am dismayed that the simple elegance of the UI pricing vector, as shown in the two diagrams for 2002-2004 and 2007-2009, will be replaced by the convoluted vectors on 2011 May 3. It seems that there is a potential for mischief with having the multitude of simultaneous prices, with an undue accumulation of money by the transmission grid as some UI out of the grid is priced at very high prices at the same instant that other UI into the grid is being priced at a lower price. This is an unwarranted arbitrage for the transmission system.

The HVDC links between S and NEW could provide a warranted arbitrage situation where the grid with lower frequency delivers to the grid with the lower frequency. The different frequencies would result in different prices, with the price differences providing some financial support for the HVDC links.

I was surprised that there was no mention made in regard to Figures 1 and 2 as to when UI pricing started, and how that UI onset resulted in a narrowing of the spread between daily high frequency and daily low frequency. I believe these figures could be well supplemented by a presentation of histograms of the monthly frequency excursions, and how those histograms have changed over time. A numeric approach would include monthly average frequency and monthly standard deviation from 50 Hertz, a statistic for which you have a special name that I forget.

Parts of the Agenda Note discuss the serious impact of very short periods of frequency excursions. These short periods of concern are much shorter than the 15 minute periods used for determining UI. The various parts of the Agenda Note could be harmonized by reducing the size of the settlement period for UI from 15 minutes to 5 minutes or 1 minute.

There is a discussion of limits on the amount of UI power that a participant can transact. I question the need for such limits. As a participant increases the UI power being transacted, the price will move in an unfavorable direction, providing an additional financial incentive for the participant to reduce UI power transactions. For example, a SEB that is short of power and is buying UI faces higher prices as the UI transaction amount increases. These higher prices provide a multiplicative incentive for the SEB to reduce its shortage and its purchase of UI.

Many systems plan for the biggest credible single contingency, which the report treats as the single largest unit. The report shows that entire plants have gone out at a same time, suggesting that the biggest credible single contingency is a plant not a generating unit.

As an aside, in the listing of the generating capacity by size of generating unit, my experience in the US suggests that the list understates the number of generators. There would be many times the identified number of plants if the list included captive generators such as backup generators, which may be as small as a few KW. Again, based on my experience in the US, the total capacity of those unidentified generators will rival the total capacity of the identified generators.

I wonder why the under frequency relays in the East are set lower than the relays in the other regions.

I don’t understand the terminology that “Nepal has several asynchronous ties with the Indian grid (AC radial links).” My interpretation is that Nepal has a disjoint system with each section tied synchronously to different locations of India, making the sections synchronous to each other through their links to India.

“Too Much of a Good Thing” Revisited

In “Renewable Electric Power—Too Much of a Good Thing: Looking At ERCOT,” Dialogue, United States Association for Energy Economics, 2009 August, I looked at the impact that wind was having on the dispatch prices in ERCOT, the Independent System Operator for much of Texas.  Prices were negative during about 23% of the month of April 2009 in West Texas, the region dominated by wind generation and during about 1% of the month in the rest of ERCOT, a region dominated by fossil generation.

This week my Dialogue article was brought back to mind by two messages I received, one on the IEEE list server PowerGlobe the second a ClimateWires article sent to me by a friend.  Both dealt with the issue of “grid operation during very high levels of wind energy”, the subject line of the IEEE PowerGlobe message.  The ClimateWires article deals with Bonneville Power Authority’s reaction to such situations.

My reaction to both messages is that we need a true spot price for electricity.  I once heard that a spot commodity price was for the commodity delivered on the spot out of inventory, before more of the commodity can be produced.  We don’t have an inventory of electricity, but we do have an inventory of production plant.  So, combining the concepts, the spot price of electricity would be applicable to deliveries made before we can change the operating levels of our production plants.  That may mean a different price for each second.  Certainly a different price for each minute.

But a spot price should apply to a different quantity than might the dispatch prices developed by independent system operators (ISOs) like ERCOT.  The dispatch prices should apply to quantity specified by bidders in the ISOs.  Any variation from that quantity, up or down, should be priced at the spot market.  Further, the spot price should be allowed to vary greatly from the dispatch price.  Otherwise the weighted average price of the total delivery might be seemingly insignificantly different from the dispatch price, as shown in the following table.

Description MWH Price Extension
Dispatch 100 $40.00  $     4,000.00
Spot -5 $30.00  $      (150.00)
Metered 95 $40.53  $     3,850.00


The basic assumption is that the generator committed to providing 100 MWH at a price of $40.00/MWH, and that the ISO accepted that price.  As it turned out, there actually was a surplus, such that the spot price was reduced to $30.00/MWH.  For some reason, which irrelevant for this analysis, the generator only delivered 95 MWH through the meter during this period.  Thus, the generator effectively bought 5 MWH in the spot market to achieve its dispatch obligation of 100 MWH.  The effect was that the 95 MWH that were actually delivered had a unit price of $40.53/MWH.  Some would say that the generator got lucky in this situation.  An arrogant generator might say that he was smart and dispatched down his generator.  The point that I am trying to make with the table is that the average price experienced by the generator is only 1.3% different from the $40.00/MWH dispatch price.

Effect on Average Price of Spot Volumes and Spot Prices

Given 100 MWH Dispatched at $40/MWH


  -$50 $30 $40 $200 $2,000
-10 $50.00 $41.11 $40.00 $22.22 -$177
-5 $44.74 $40.53 $40.00 $31.58 -$63.16
0 $40.00 $40.00 $40.00 $40.00 $40.00
5 $35.71 $39.52 $40.00 $47.62 $133.33
10 $31.82 $39.09 $40.00 $54.55 $218.18


The next table shows the effect of making a variety of spot transactions at a variety of prices, including negative prices and prices many times the dispatch price.  I note that the average price stays at the $40.00/MWH dispatch price when the spot price stays at $40.00/MWH or when the spot delivery stays at 0 MWH.  The average price from the first table appears in this table at the price of $30.00/MWH and a spot delivery of -5 MWH.

Generators prefer to be in the top left portion of the table or the bottom right, first where they are short when prices are low and second when they are long when prices are high.  Consumers prefer to be in the top right portion of the table or the bottom left, first where they consume less than the amount entered into the auction and the auction price is high and second where thy consumer more than the amount entered into the auction and the auction price is low.

The Oil Spill Commission–Lessons Learned

I attended a dinner last night that included a presentation by William Reilly, co-chair of the Oil Spill Commission, who discussed their report, which is available with the following URL

Reilly almost began with a comment that the Three Mile Island Commission report had effectively killed nuclear power in the US, in that the US has not built a nuclear power plant in over 30 years.  He didn’t want the Oil Spill Commission report to have a similar effect on drilling in the Gulf of Mexico.  We will see if it does.  I guess that the appropriate metric for that would be the number of new drilling rigs that are added in the Gulf, which would be the metric comparable to the metric of the number of new nuclear power plants added (or not added) over the last 30 years.

Reilly discussed a Report Card mechanism that the nuclear industry now has in place, with the power plant operators evaluating each other, a peer review program.  The executives of the plants that have the worst performance evaluation are effectively the subject of ridicule by the rest of the operators while the executives of the plants with the best performance evaluation are effectively given bumper stickers that say “My child is an Honor Roll student at . . .”  So they have created a system of bragging rights.

From what I have heard about the nuclear power industry, their performance metrics have dramatically improved over the last thirty years.  Plants are operating at much higher capacity factors.  That improvement could be attributed to the Report Card process.  I have also heard people say that the industry restructuring has given better (more) financial incentives to operators.  I am sure that the answer lies at neither extreme of this spectrum and probably includes other dimensions as well.  Whatever is the source of the improvement, the economy is better off with the improvement, even though we have not yet built a new nuclear power plant.

I commented during Q&A that the electric control area operators used to have such an Honor Roll system (they even identified utilities as being on the Honor Roll, using that term) for Area Control Error (ACE).  ACE is a metric brought about by the socialization associated with interconnecting competing power systems, about which I have written on my blog.  I then commented that the Energy Policy Act of 2005 instituted mandatory performance standards which are administered by NERC.  Reilly was unaware of the control area Honor Roll concept.  I haven’t participated in the process run by NERC, but have heard intimations of a peer review process for electric operations, which could be another Honor Roll.

I found interesting parallels between Reilly’s comments on the Mineral Management Service (now Bureau of Ocean Energy Management, Regulations, and Enforcement) and the three companies involved with the process.  MMS (now BOEMRE) did not have the budget to develop the expertise and manpower to do in depth inspections, seeming to accept the reports produced by drill rig operators.  Indeed, MMS (now BOEMRE) has many fewer inspectors per hundred wells in the Gulf than does California for wells on land in that state.  Similarly, BP seemed to accept the expertise of Halliburton and the drill rig owner. Both MMS (now BOEMRE) and BP relied heavily on the expertise of the group doing the work.  Afterwards an attendee commented to me that he was disappointed that Reilly did not identify a black hat or evil doer.  Maybe that reflected Reilly’s bent to avoid destroying the industry or major players in the industry.  I am reminded that the indictment of Arthur Anderson killed that company in the Enron fiasco of 2001.

During Q&A, Reilly compared prescriptive regulation with the peer review process.  After two or three years rules and regulation will be out of date and could actually be detrimental instead of just ineffectual.  In contrast, the continuing peer review process is always using best practices as those practices change continually.