India: smart steps it can take

American grid expert weighs in

Phil Carson | Aug 01, 2012


Yesterday I plunged into the speculation around the blackout of mind-boggling scale in India. For the columnist, the formula runs: courage + recklessness = just enough knowledge to be dangerous. I suggested that the upshot was the need to make a greater effort to decentralize and add distributed resources and automation.

If you enjoy gloating, you can read my inexpert speculation in "Lights Out for 700 Million?"

Erich Gunther, an IEEE smart grid expert and co-founder, chairman and CTO for EnerNex, made himself available for a brief chat yesterday. 

(You may recall our last chat with Gunther in "Q&A With Cat Herder Erich Gunther" and "Herding Cats: Standards and Interop, Part II.") 

For Gunther, the Indian outage was "not surprising," given what he knew about the grid in India. And, while awaiting analysis and raw data on grid behavior to determine the sequence of failures that must have taken place, he identified immediate and longer-term steps that India might take to prevent a recurrence. 

No doubt that would be good news for the government official who on Monday famously declared the blackout affecting 300 million people to be a "one off," only to see Tuesday's blackout affect more than double that number. 

News reports suggested several possible causes of the blackout. Because finger-pointing immediately focused on India's northern states overdrawing their allotment of power, I was skeptical. And because the leader of an enquiry into the last historic blackout identified the lack of manual circuit breaking actions (possibly due to political backscratching) my inner cynic thought that reason passed the sniff test. 

Gunther declared that a supply-demand imbalance was at the heart of the event. Why?

"It's all related," he told me. "I like to start with the raw data." 

Let's just say a bird flew in the window and told Gunther that frequency variations on the order of one-half to three-quarters of a hertz were recorded across the Indian grid. In the United States, he said, frequency variations of one-tenth of a hertz are considered dicey. 

"The frequency data clearly showed the frequency going down on both days for both events," Gunther said. "And the frequency went down substantially before showing any `bounce back' produced by the blackout. That's reminiscent of the pattern we've seen in other large blackouts. 

"That data reveals that the frequency was allowed to decay quite a bit, indicating a load-generation imbalance before a variety of things tripped off-line. We don't have enough information to know what really started to trip off-line and cause the cascade. That takes a while to figure out. But we've known for some time that India is challenged in providing the load-generation balance. 

"In this case, you've got the hottest summer on record. I don't know if any contingencies were in play or not. But a problem they've had for a long time came to the ultimate breakaway." 

Aspects of the Indian grid that might have contributed, according to Gunther: a limited number of interconnections for the high-voltage transmission system and load growth from population growth and industrialization. The high degree of unmetered power, i.e., stolen power, contributes to a lack of situational awareness as well. 

Gunther then ticked off a list of actions that could help ameliorate India's grid vulnerabilities in the short term. He explicitly acknowledged that, without a detailed knowledge of aspects of India's grid and its management, that some actions he suggested might be long term in nature or mired in political inertia. (I thought it interesting that Gunther did not mention the mindless solution of building more coal-fired power plants.) 

First, because India's grid is unreliable, many customers use their own generators for backup power. That's common in the U.S. among large commercial and industrial concerns, but in India the use of backup generation is widespread among all customer classes. An inventory of distributed energy resources (DER) would provide the first line of defense in an overloaded grid, if owners would agree to switch to backup power upon request by grid operators, according to Gunther. Add to that inventory any and all parties that would agree to respond to a demand response event, possibly for incentives.

Non-essential public loads should be identified and prioritized for taking off-line in case of an event. So street lighting would be turned off long before, say, traffic lights lost power. 

"System separation" is a strategy that could limit the scale of a blackout, Gunther suggested.

"Call it macro-gridding, if you will," he said. 

These steps all might be automated for speed and reliability of response, he added.

Longer term, increasing distributed energy resources would help, but that raises other issues such as cost and complexity, Gunther said. That's true in the U.S. and more so in India. 

"The fundamental problem in India is not enough generation," Gunther concluded. "Disconnected generation is widespread. The technology they need—the controls and interconnections—is more expensive than the DER itself." 

It all makes for a fascinating case with a global audience. I'm sure we haven't heard the last of it, as analyses take place within India and by outside experts. Stay tuned!

Phil Carson 
Intelligent Utility Daily 



Related Topics


Another take...


One point that you have not brought out is the unique form of grid control that is practiced in India.  

If  we look back in our history,  during the peak hours,  the state utilities used to overdraw and allowed the frequency to go down. Often the frequencies used to hover very close to the trip levels of 47.5 Hz.  The idea was that for every 1 Hz that you allowed the grid frequency to go down, you could accommodate some extra 2000 MW.  In order to curb this practice of overdrawing by the utilities, the regulator introduced what is called as “Unscheduled Interchange  Charges ”  or  (UI charges as it is popularly called) . This is a very interesting concept.  If the particular constituent in the grid were to overdraw more than his Scheduled drawal , he was  made to pay a sort of penalty which was in turn linked to the frequency.  (Each 0.02 Hz step is equivalent to 16.50 Paise/kWh in the 50.2-50.00 Hz frequency range, 28.50 Paise/kWh in 50 Hz to 49.8 Hz and 28.12 Paise/kwh in frequency in the below 49.8 Hz to 49.5 Hz range.) 

Detailed order on the UI charges is given in the link below 

Now this is where the whole thing becomes interesting.  While the intent of introducing UI was meant as a deterrant to the erring utility, however, over a period of time this mechanism became an alternative power exchange mechanism. As you may be aware, there are two well established power exchanges in India and whenever somebody wants more power than his schedule he can always go to the exchange and buy the extra power.   What the utilities do is that at any point of time, they are looking at the UI rate and also the power rates on the exchange and decide their course of action on whether to overdraw or to go to the exchange.  Ofcourse going to the exchange has another dimension attached to it. That is you have to put up the cash upfront when you place the buy order.  Since some of the utilities are more cash strapped than others, they prefer to overdraw and pay the UI charges since the settlement of UI is done  later and they get a breather. 

While bypassing of the underfrequency relay which prevent automatic tripping of feeders  is one of the primary causes,  the regulator also needs to review the UI charges.  As a start, the UI charges should be so high that the utilities are not tempted to use it as a power exchange mechanism. 

With best regards,

Managing Director,
Solartherm Energy Ventures Pvt Ltd


Power System Stability

Right in the begining of courses on power system stability is taught why and how a Generator could lose synchronism if there is a sudden fault or sudden load thurst on it.A grid tied very closely and operating at high level of loading is very much liable to see such a scenario.It is true that it is indeed a rare occurance but if a large machine trips then it is almost similar to a new large load on the rest and sequential tripping could result.It is also very true that Indian grids are mostly operated at large loads with low margin for stability and the frequency variation is of the range that has been mentioned.In fact what has not been realised till date that power shortage in India is not 25 % but 75% as there is lot of pent up demand.India is having about 75,000 MW of distributed power generation recorded and some un recorded. The cost of onsite generations has gone up very substantially so those who have exctreme requirement of uninterupted power use them or else they use the grid power. After the availability of gas the cost of onsite generation will go down and many could switch to self generation releasing load from the grid.Gas pipe lines are being laid and this could happen faster than expected.Also what is not realised that 60% of Indian economy is unrecorded putting it some times ahead of China in certain Spheres if you take PPC into account.From all accounts it appears that India has broken out of the developing country Frame in lot of the regions in India.

Grid Efficiency

Let's take some standard ordinary steps toward grid efficiency. India's losses are approaching 30% at peak (not considering theft).  This means that marginal losses are likely to be even higher (I-squared R).  As the system heads towards peak, for every kilowatt demanded by an end user (whether metered or illegal), a smaller and smaller bit gets to them and a vicious cycle develops where demand simply cannot be served, regardless of how much generation there is, because the system can't handle it.  There needs to be a systematic program of installing bigger distribution wires and possibly changing voltage levels.  Illegal connections can be removed as that is done (which would provide additional temporary relief , though in the real world, they are likely to come back).  New transformers (an investment for the next 30-40 years) should be bought based on lifecycle cost, not just first cost.  If there's not enough generation to begin with, reducing this vicious cycle of internal technical losses may be one of the more important steps toward sustainability.