Can Wind and Solar be significant contributors to a low emission electricity system?

...there is a substantial body of evidence that variable renewable integration costs are hugely dependent on the flexibility of the system to which they are being added.

So claims a new report from a UK Energy Research Centre, which lists many studies after opining on them in The costs and impacts of intermittency – 2016 update. Integration costs are important and I'll write on some of the content of the UK paper in paragraphs below, but first I want to discuss the bias of the work, and a great accomplishment in Ontario.

Taken together, the full range of impacts add weight to the message that electricity systems and markets need to adapt and be reorganised to incorporate large proportions of variable renewable generation most efficiently. 

systems and markets may not be what "need" to adapt.

The key challenge facing policymakers, regulators and markets is how to ensure delivery of a flexible, low carbon system that makes maximum use of variable renewable generation whilst minimising overall cost and enhancing security and reliability. 

It is wrong to state a low carbon system maximizes "use of variable renewable generation."

My estimates indicate in January - usually one of the highest demand months of a year - Ontario generated less electricity with fossil fuels than in any month since at least 1973. Probably the least of any month in my life (I was born the day Dylan shocked Newport with an electric performance).

The Ontario system operator's senior management talk a lot about points the government desires them to focus on, but I suspect the accomplishments of the operators at the IESO are far more related to altered handling of renewables within the system - "changes to system operation, regulatory frameworks and the design of electricity markets."

When I started writing on energy in 2010, the joke was people wanted cheap, green and plentiful, but they could only have 2 of 3. For real environmentalist, I think the joke is people still thinking the world will decarbonize without all 3 together. This study could be seen as U.K. apologists for wind and solar explaining they aren't necessarily expensive.

Estimates of additional costs based on assumptions of flexible systems can be several times lower than estimates of additional costs based on assumptions of inflexible systems. Linked to this is the very strong finding that additional costs will be minimised if electricity systems are optimised to facilitate the integration of variable renewable generation. This optimisation includes changes to both the technical and economic characteristics of electricity generating plant, potential contributions from flexible demand, storage and increased interconnection capacity, as well as changes to system operation, regulatory frameworks and the design of electricity markets.

The review summarizes categories impacting costs:

Capacity credit and costs...a peak of less than £15/MWh even if the capacity credit of the variable renewable plant is assumed to be zero.Reserve requirements and costs ...most analyses conclude that the additional cost of these reserves remain relatively modest...
Curtailment...the point at which curtailment becomes significant can vary dramatically, with some analyses finding the inflection point to be as low as a 15% penetration and others finding the inflection point not being reached until there is over a 75% penetration of variable renewable generation
Transmission and network costs...these costs do not appear to rise sharply as penetration increases ...Very little data was found for penetration levels above 30%.
Thermal plant efficiency and emissions ...efficiency and emission impacts are particularly dependent on the assumptions over the mix and operating characteristics of the thermal (and/or hydro) plant whose output is being varied to accommodate intermittent renewable generation...
System inertia...Of those studies that do address this issue, the typical conclusion is that it is likely to only become significant at high penetrations of variable renewables

I moved capacity credit to the top of the list because the cost of everything else depends on this factor. The study states:

...capacity credit is often expressed in terms of the conventional thermal capacity that an intermittent generator can replace while still delivering the same reliability of supply to energy users, so for example, if 100MW of notional wind farm capacity is calculated to have a capacity credit of 25%, then it can notionally replace 25MW of conventional capacity on the system without reducing that system’s ability to meet demand.

It's a UK paper and it does cite studies claiming that 25% capacity credit estimate for that environment, but I'm skeptical. In general the paper does display most studies it examined show variable Renewable Energy Systems (vRES) in northern climates have capacity values near zero - and declining with market penetration (in Figure 3.5).   The paper hedges on declaring the decline in capacity credit with market penetration because there's little data of high vRES market penetration systems. Another recent article, Limits to growth in the renewable energy sector, indicates that may remain the case. I won't dwell on this point beyond noting/recommending K.M. Korhonen's Stall warning for renewable energy post.

An example of near nil capacity credit is found in Germany's expansion of renewables - largely wind and solar. From 2006 to 2015, according to data from ENTSOE, Germany added almost 65,000 megawatts of renewable capacity to its electricity system between, and in both years book-ending the period annual consumption peaked at just under 78,000. Despite adding capacity equivalent to 83% of demand, Germany retained essentially the same capacity from traditional source (hydro, gas, coal and nuclear). That system has certainly behaved like the capacity credit of wind and solar is zero.

Another example in a much different climate: California. The L.A. Times recently ran an article attributing high electricity bills in that state to a bias towards building new power plants - and James Bushnell responded to the article with an excellent, mostly alternate, explanation in a blog post at the Energy Institute at Haas site. I went to the data, from the U.S. EIA, and we see the same situation as Germany - since 2006 demand is unchanged, non-vRES generation capacity is up slightly, and all the new wind and solar is therefore implied to have near-nil capacity credit.

Having established there is no capacity credit for vRES (wind and solar) in many locations, it's only fair to say they require neither additional generation capacity, nor less. They are simply additional capacity and all transmission and network costs incurred due to them are entirely due to them (the U.K. paper provides a range of £5-£20/MWh for this cost). Costs due to reductions of thermal plant efficiency and emissions the study notes as being very small, but these costs are not so much small as intertwined with curtailment costs.

Near-nil capacity credit values of vRES imply a simple way to estimate the value of vRES generation as the fuel cost of a thermal plant generation displaced when the sun, and/or wind, allow.  To demonstrate this, consider a situation where a 100 megawatt natural gas-fired power plant is operating at 50% capacity factor as an intermediate generator, with a fixed cost of $15 million a year and fuel cost of $40/MWh. The math works out to an average unit cost of $74/MWh over a year. Now add a 100 megawatt wind facility guaranteed the same rate, $74/MWh, for each megawatt it can generate - assume a 35% capacity factor but remember only 50% of the time can its output displace gas. The math shows 153,000 megawatts of wind generation wasted (either curtailed or, often in the real world, dumped on neighbouring grids); the reduced usage of the gas generator raises it's average price to $93/MWh, and the curtailed wind makes the effective price of a useful megawatt from the turbines $148/MWh.

This simple model indicates cost increase by $16.6 million dollars, which equals the $22.7 million cost of wind less the fuel cost of displaced gas generation (153,300 MWh at $40/MWh = $6.132 M). While the U.K. study shrugs of LUEC's decreasing utility, this simple scenario shows adding $75/MWh wind to $75/MWh gas increases costs by $38/MWh. This is a result of the near-nil capacity credit of wind and solar (in may environments) - whether the increased costs are attributed to decreased usage of the gas generator or curtailment is not relevant. On a positive note, assuming an emissions intensity from the gas-fired power plant of 420 kilograms CO2equivant per megawatt-hour (kg CO2e/MWh), 64,386 tonnes of CO2 equivalent (tCO2e) are avoided - unfortunately with $16.6 million of added cost implying a carbon cost of $258/tCO2e.

Other scenarios, as the study notes, make the addition of wind more financially attractive. Using the same assumptions as the preceding example, assume the replacement of a 100 MW nuclear power plant with both the 100 MW gas power plant, and a 100 MW wind facility. This time we assume 85% capacity factor for the nuclear, and 85% of wind's output capable of replacing needed generation - with the remainder of the removed nuclear output to come from the natural gas power plant.

Replacing 100 megawatts of nuclear capacity with both 100 MW of wind and 100 MW of gas hardly costs anything in this scenario, where wind and nuclear are both prices at $75/MWh for all they could produce, and natural gas-fired output would be to if it operated at a 50% capacity factor - but it operates above that in meeting demand. Wind appears to provide much better value when displacing baseload nuclear - along with gas. However, emissions rise 203,276 tCO2e, and the emission intensity of generation rises from near nothing to 273 kgCO2e/MWh.

This is not an unexpected result. Neither California nor Germany have reduced emissions notably while adding renewables, and removing nuclear - nor has/will Japan. It is possible that environments wealthy in reservoir hydro generators could find great utility in increasing renewables, but it is highly unlikely variable renewable energy systems will contribute to significantly lower emissions elsewhere.

Ontario hasn't seen 273 kg CO2e/MWh emissions intensities in many years. Based on the January gas-fired generation of 572,341 MWh I calculated from IESO Generator and Output Capability reporting, at 420 kg/MWh the emissions of the IESO system's generation of 13.27 TWh in January had an emissions intensity of about 18 kgCO2e/MWh. There's a lot of actions Ontario could take to get that up to 273 kgCO2e/MWh, but maybe the zealots who believe the future has revealed itself to them in the shapes of solar panels and industrial wind turbines should shop for new churches.

It must be noted that many commentators, many quoted by me, have been skeptical about the ability to operate the grid with little of the gas generation (and previously coal) online, and ready for ramping. I believe Ontario's actual operators of the system - the ones who do it on a daily basis - have a story to tell. That story would likely include improved forecasting, revised wind turbine (and perhaps solar panel) regulation to provide a programmed reactive power element, and rationalized market bid rules forcing the curtailment of wind and solar output prior to impacting nuclear units.

It is variable renewable energy systems that need to be flexible.



worksheets on monthly/annual generation
worksheet with supply mix scenarios