India's #electricity grid set several record highs Tuesday - there's a heatwave going on.
Per @GridIndia1's MERIT data (via @CSEP_Org's carbontracker.in): 1) Record demand met: ~216 GW 2) Record #coal generation: over 166 GW
📢 3 issues going forward...(short 🧵)
Issues: 1) This isn't over - typical annual peaks are in June(ish). 2) The peak was not AM any more, rather, closer to 3 PM 3) Coal's output has been steady near max, with long durations of highs.
These 3 pts have implications as expanded below.
1) It will take planning, luck, AND ₹ (or $ = forex!) to prevent blackouts.
Short-term issue is fuel - gas prices ⬆️ & coal imports also ⬆️. Railways is stretched so coal stockpiles at plants are ⬇️.
@IMDWeather forecasts decent monsoon (96%), but the issue is also the timing.
2) Till now this year, peak demand was closer to noon, nicely matching solar. 3PM is also daytime, but less so.
BUT, even the evening peak is now ↑ (and likely to rise). Note #RE has limited evening output.
If we measure "net demand" (= demand - variable RE), it peaks evening.
3) Generation is busbar (net), post auxiliary consumption. Thus, gross coal gen (capacity) was prob. ~180 GW, ominously close to the total installed capacity.
Coal output was ~steady high. This means future RE+storage would need long duration storage (and not a "4-hr battery").
Policy recommendations:
Some are obvious - get more fuel to plants.
Less obvious: a) more demand response & a smarter grid. We need time of day pricing, both wholesale and for retail.
b) Add more RE (saves coal) but esp. wind (helps evenings).
c) Better manage (& grow) hydro.
Will India face rolling blackouts? Government has been planning, scrambling to find fuel, pushing stockpiles. But will it be enough with 🔥wave?
If more imports are used, how much will that hit prices?
What is being done for more FIRM power at scale? Storage is still expensive.
How bad is India's coal-stockpile-based power crisis? NO ONE KNOWS!
The coping mechanism is brownouts ("load-shedding"), but these are not properly measured.
I've been suggesting & helping design proper feeder measurements since 2014-15.
A short 🧵 on what's happening vs. not.
Govt. is putting out more data than ever before...but we are missing some key data.
Here, daily shortfall (GW) data (from grid operator @PosocoIndia) are based on a notional peak and "as reported" or estimated instead of "as measured".
(usually) 11kV feeders are where load-shedding happens. These aren't fully instrumented (metered). Where metered, it's mostly for energy accounting purposes, & not real-time/compiled.
India doesn't bid for power, it bids for power plants.
You may get the cheapest coal plant, but not the cheapest power. RTC (“round the clock”) bids indicate trends but are also misleading. India needs Time of Day WHOLESALE pricing to improve generation portfolio signaling.
Far more worrying is whether such a long term PPA is needed, given power purchase agreements (PPAs) by the state (MP) exceed visible needs. Problem is a legacy mindset, including financing (who funds without a PPA? – same for #RenewableEnergy). Fuel is also linked to PPAs.
(1/n) #9pm9minutes PRELMINARY Analysis carbontracker.in timestamps are limited!
Key points:
1)Total demand fall DUE TO EVENT was ?~29 GW.
2)Demand fall started before 9 pm
3)Most balancing was done by hydro [BIG TIME]
4)Grid rise (aka “recovery”) finished ~10:30 pm
(2/n) Demand fall was not just due to lighting.
Too much fall, and some of load fall began as early as 8:45. All evenings “normally” decline 3.5-5.5 GW after peak ~7:30. Fall after 8:45 was measurably more than normal trend. See carbontracker.in for other days.
(3/n) [speculation] Folks were switching off loads to protect their devices. Really doubt so much load is lighting! That too participating.
Fall during the 9 minutes was less than the total event - if one wants to estimate "household lighting" in India.
(1/n) Thread on 9 PM lights-out & candles-on plan by @narendramodi
TL/DR – it should *likely* not be a problem because (1) we can plan; (2) the lighting load is modest (maybe 10% of total grid at most. (3) We have enough fast acting supply options that can manage.
(2/n) SCIENCE: 3 factors that matter; 1) how much switches off/on? 2) How fast does this happen? 3) What does the grid look like to be able to cope? Oversimiplification – we’ll only look macro level (means frequency, not voltage), and ignore transients.
(3/n) Earth day is usually not as high participation in India – but it also involves switching off more loads than just lights. We handle that easily. (NOTE in graph (from carbontracker.in), as I predicted before, after stabilizing, daily high is trending up – summer/ACs!)
(2/n) [DISCLAIMER: I helped set up India's smart grid space. Am ~evangelist, but also a realist]
Discoms have high losses, partly due to theft/poor billing (and collection); other reason is tariffs < costs. A smart meter (esp. in prepaid mode) offers promise to cut down losses.
(3/n) But it's wrong to claim ONLY smart meters cut down losses - Haryana, Ajmer, etc. prove so. Smart meters can't detect theft alone - requires analytics and political will. 1st Q: are discoms harnessing the data they already have? 2nd Q: what of theft "above" the meter?