On July 26th Alberta’s electricity market hit its regulated peak price of $1000 per megawatt-hour (MWh) and stayed there for hours 17, 18 and 19. The price soon dropped back down but the commentary continues.
Upon seeing there had been a price spike, I checked to how industrial wind turbines had performed and saw they’d performed exactly as I’d expected, with output dropping from hour 14 to hour 18.
I expected that as I’d seen it in 2014 and in 2012. I didn’t think this was a particularly big event. Prices have been very low in Alberta and this spike will do little to change the yearly average. Alberta is examining a capacity market, and the intent of those is to prevent high price hours – but Texas is an example of a jurisdiction thus far avoiding capacity markets/costs by upping the maximum peak market price. Theoretically, peak pricing can be healthy in encouraging new market entrants with peaking generation, or demand reduction, products. While seeing the one event as not particularly problematic, I did put some short thoughts up on twitter:
Capacity credit is an awkward term I’ll return to.
A response to my tweet tagged Andrew Leach who later put some other suspects for the cost spike up on Twitter, including:
Baseload is probably not the coal units cited by Professor Leach. Battle River 5 entered service in 1971 and the holders of power purchase agreements (PPAs) took advantage of a change in legislation to walk away from them last year. Milner entered service in 1972 and owners have some approvals to convert the site to gas. The owners of Keephills and Sundance units (which with only one exception entered service prior to 1983) have announced plans to either soon shutter or convert units to burn natural gas.
The units are nobody’s future, and no investments are being made to extend operations.
The term baseload could refer to a plant designed to run frequently. The best way to explain may be with natural gas power plants. A combined-cycle gas turbine generates power with gas driving a turbine, and then again with steam generated with that heat driving another. This is more expensive to construct than a single, or simple, cylce gas turbine (SCGT), but uses less fuel to generate the same quantity of electricity. Generally CCGT is considered intermediate power, and SCGT peaking power. With gas prices low, many CCGT units can now run as “baseload” – meaning at a steady output all the time. Theoretically one could add carbon capture to a CCGT plant – if they did it would be due to it being more economical to run most of the time (ie. due to a price on carbon emissions). This consistent, or baseload, mode is the cheapest to run units as it avoids the stresses of heating and cooling of equipment that comes with multiple stops and starts.
As coal units age, their mode can and does change. New units have been cleanest, and used less fuel, so they would run all the time. As the units age other, cheaper to run, options are constructed and the units would step down to intermediate options – and then peaking, or emergency reserve, status.
The 4 units being used to make big claims about baseload power are better described as old coal generators all politicians want shut.
Pembina’s blog post includes:
At 2 p.m. the Keephills 2 coal plant tripped offline, and at 3:45 p.m. a second coal plant followed.
Why did the second unit trip? Recently wind turbines were blamed for a major black-out in South Australia. There are lots of opinions about what occurred, but my favourite is that software shut down the turbines due to threats from power line changes from failed transmission lines. I like that explanation because people can learn from it.
others can’t learn from anything.
“if Alberta had significant solar PV generation installed this emergency could have been avoided completely.”
That is one possibility.
It’s also possible the grid would have been more stressed causing more units to fail.
Alberta is a winter peaking jurisdiction (one reason why some units would be on maintenance outages in the summer), so solar is unlikely to eliminate the need for other generation. Despite this distributed solar can have more value due to transmission lines, and thermal units, having lower capabilities in the summer – while demand is less in the summer, in reality generating units probably are less capable then too. However, if the solar panels are positioned for maximum output, solar output would typically be declining, in Alberta, throughout the price spiking period of July 26th.
Using Ontario reporting to illustrate a likely hourly output pattern for solar, I’ve pulled IESO forecast data for the one location reporting at latitude as far north as Alberta (Northland Power), and the data for the Northwest zone (for the same reason – much of Ontario’s northeast zone is south of Alberta).
The hours probably are offset by 1, with hour 14 in Ontario being comparable to hour 15 in AESO reporting, but even at that I think it’s clear most solar panels would have declining output during the hours Alberta pricing spiked.
If it was grid stressors, from rising demand and declining wind, that knocked a second coal plant off-line at 3:45 pm, solar would have accentuated that stress.
Pembina makes a bold claim that Renewables can provide backup as part of an integrated grid
I’ll manipulate you to read this carefully through formatting:
Myth: Renewables require special backup
Reality: Renewables can provide backup as part of an integrated grid
Renewables that will be built to meet the Alberta renewable energy target won’t require special backup. In fact, there’s no such thing as “special” backup.
There’s no such thing as what renewables can provide?
I said I’d return to “capacity credit” – and this is the time. A grid operator is expected to meet a strict reliability level (ie. outages less than 1 hour every 10 years). Capacity value basically measures a generator’s ability to contribute to meeting this standard. It is a complicated task but if it were not: considering system peak occurs in the dark of winter solar’s capacity credit, or value, in Alberta would be expected to be nil. This might not be the case – one would have to model and test. The point is system capacity is not based on infallibility of baseload generators, it is based on capacity credit.
Perhaps the capacity credit of 100 megawatts of 40-year-old coal generation is 50% in the summer heat.
If coal units failed it would make more sense to add 40-year-old coal units than wind or solar – because 50% is still higher than nil (wind) or 45% – which is a guess of where solar would be by 6 pm, in the middle of Alberta’s July 26th price spike.
Baseload has nothing to do with it.
Unloved coal units failing on a hot afternoon does not fit a narrative of any procurement strategy – especially one of contracting comically poor ‘green’ consultants.