Clean Power

Published on July 5th, 2016 | by Christopher Arcus

48

For Maximum Renewable Integration, Load Following Is King

July 5th, 2016 by  

A previous CleanTechnica article reviewed the basics of power generation. That article is a good primer for this follow-up article delving deeper into the subject.

How do we most efficiently integrate the maximum amount of renewables into the grid? Let’s start with understanding how things work. As first-order simplification, we can examine a renewable source like solar to understand some of its characteristics. At a simple level, in a sunny area like the desert, and depending on latitude, solar will be available some eight hours of the day. Other places, it may shine through to the panels only six hours a day.

The ratio of output to maximum output is expressed as capacity factor. For well-sited utility-scale solar, this translates to capacity factors around 30%. With other techniques, like tracking, capacity factor can be higher.

But analyzing wind or solar output by capacity factor alone misses the bigger picture. If we look at daily demand variation, we see that solar matches some of the daily load variation very well.

daily-demand-california-summer-caiso

solar production demand

What’s Wrong with Capacity Factor?

Looking at solar capacity factor without taking into account demand variation is like assuming demand is constant. That’s the problem with capacity factor. It doesn’t take into account load matching. Capacity factor only measures the amount of output available versus the maximum possible. The maximum possible is constant output. A constant output with no other sources could only drive a constant load.

“Looking at solar capacity factor without taking into account demand variation is like assuming demand is constant. That’s the problem with capacity factor. It doesn’t take into account load matching.”

By ignoring demand variation, we miss the essential problem. We must match demand all the time, but demand varies. We cannot match demand variation with fixed output sources like coal and nuclear alone. We must add flexible ones like gas to do that. The graph above shows that. That’s how we balanced demand variation in the past.

“We cannot match demand variation with fixed output sources like coal and nuclear alone. We must add flexible ones like gas to do that.”

That’s a source of the misunderstanding in the myth of baseload power. So called “baseload” power perpetuates the myth that there is a base “load” and that inflexible power plants are necessary to meet it. More on that later.

Inflexible power plants have a different problem. Their capacity factors are misleading because they cannot match load. A high capacity factor means being unable to match load variation. Capacity factor is a calculation as if the load never varies.

“Inflexible power plants have a different problem. Their capacity factors are misleading because they cannot match load.”

What about Wind?

Solar has a reliable daily variation, but wind is different. Wind can have a diurnal pattern near coastlines, as it does in California.

And wind can have other variations in different locations. A University of Wyoming study highlights the differences between California’s diurnal wind patterns that peak in summer months, driven by increased land–sea temperature differences.

“Analyzing this precise wind data over the course of days and over the course of a year, the UW researchers confirmed that Wyoming and California wind patterns are not only very different, but also very complementary. Based on a yearly average, California wind is strongest at night, while Wyoming wind is strongest during the day and peaks in the afternoon — coincident with the time when the sun is beginning to set while the electric load is still increasing into the evening hours.”

The study goes on to say:

“Although the benefits of geographic diversity to renewable energy have been suggested for some time, only recently have there been attempts to quantify these benefits,” says the study’s author, Jonathan Naughton, a UW professor of mechanical engineering and director of the Wind Energy Research Center. “The renewable energy quality metrics proposed in this study are a start at being able to characterize different combinations of renewable energy sources. The result of applying these metrics to energy produced from Wyoming wind and California renewables provides a quite compelling case for geographic diversity.”

… For example, the study looks at a scenario for adding incremental renewables to an existing portfolio and, when comparing Wyoming wind to more same-profile California solar, Wyoming wind would yield a 50 percent higher capacity factor; a 41 percent lower relative variability; and increase by 86 percent the amount of time in which power is producing at 25 percent or greater.

“Capacity factor alone does not fully describe a renewable resource or … the combination of renewable resources,” the report notes. “How these quality metrics would benefit California’s electrical grid are outlined in a separate companion report by Jim Detmers, former COO of the California Independent System Operator Corporation.”

“Capacity factor alone does not fully describe a renewable resource or … the combination of renewable resources.”

A Tale of a Different Metric

And there lies that tale. Capacity factor is not the most important factor in renewable integration.

Let’s take the case of solar. We can see the typical shape of the daily solar curve in the graph. A quick take on capacity factor says it’s about 30%, but what does that mean? It means that, compared to a flat power output all day long, solar is available for only those hours with its given curve. But demand is not flat all day long or over weekends, seasons, and so on. Demand has an annual pattern, with daily, weekly, and seasonal patterns. If we assume solar must have constant output, as the capacity factor metric implies, we will not realize that much more can be integrated than capacity factor suggests. That’s why load matching matters. Let’s take a look at those now.

“If we assume solar must have constant output, as the capacity factor metric implies, we will not realize that much more can be integrated than capacity factor suggests.”

How Demand Works

daily-demand-new-england-iso

See how the peak demand varies over the year? The system operator must keep enough generators available for the annual peak demand, usually in July or August during hot summer air conditioning use.

Reserves to the Rescue

But that’s not all. There must be enough generation available so that unplanned generation failures can be quickly substituted with other sources. These are referred to as reserves. A look at this graph shows that, for much of the year, these reserves are much greater than the load or demand. The dashed line above the load curve represents the additional reserves available on a non-summer peak day.

daily-demand-california-winter-caiso

The graph does not show weekly and other variations, but this animation shows demand variation all year long in great detail, but quickly.

This is one reason renewables can be integrated to high levels, as much as 40%, even without storage. Reserves must be available for the annual peak demand during summer. But most of the year, they are available to accommodate more variation than load.

But there is more.

“This is one reason renewables can be integrated to high levels, as much as 40%, even without storage. Reserves must be available for the annual peak demand during summer. But most of the year, they are available to accommodate more variation than load.”

Matching Generation to Demand

For electric power systems, the one rule is that generation must always equal load. But how do we follow this kind of load variation?

This graph shows how daily demand variation is met by various sources:

daily generation by fuel source CAISO

Here is an illustration, showing more clearly how inflexible generation like coal and nuclear must be kept below the daily minimum to avoid reducing output:

US Power Supply Demand Load Curves

But how can we match variable renewables to demand?

Breaking the Storage Myth

Storage is not the only way to match variable renewables to demand. While storage has evolved rapidly, and is welcome, it’s need has been exaggerated. German experts don’t think large-scale storage will be needed for at least a decade, or until renewables reach 60%.

Other techniques such as demand management are cheaper and effective.

These two videos show some of the ways we can match renewable generation to demand without high levels of storage.

Busting the Capacity Factor Myth

The power system requires flexible generation to meet demand. Inflexible sources require flexible generation to meet demand variation like daily variation and peak demand in summer months. Reserves are necessary for generation failures and must be larger than the peak demand all year long, plus an extra amount for failures. Most of the year, these are available to accommodate other variation. Any amount of load following reduces the need for reserves or storage.

What happens if we size solar much larger than today’s amounts? Do we run into an integration percentage limit at capacity factor? Aside from using other techniques like demand management, transmission, geographic dispersal, and storage, have we accounted for renewable demand management? Not if we analyze renewables integration by capacity factor alone. What appears more obvious is that using all those techniques to match demand is a practical approach used even today. A simple approximation of demand versus solar output shows that solar integration could be higher than capacity factor because of load following. If we look at the summer demand curve and superimpose the solar output, we can see that more solar can be integrated than a fixed load implies.

But then the rest of the year, when demand is lower, solar output is reduced as well. Any load matching increases the amount of potential variable renewable integration beyond a simple capacity factor analysis because capacity factor inherently implies constant load.

“Any load matching increases the amount of potential variable renewable integration beyond a simple capacity factor analysis because capacity factor inherently implies constant load.”

One thing is clear: Any analysis of integration must take into account annual load variation and generation matching. Our exploration of the subject shows it’s too complex to accurately sort out from simple analysis and anecdotal information. That’s why NREL and other researchers run long simulations using years of wind and solar data to model generation and match them against load data to figure out renewable integration. We know how that came out.

What about Overcapacity?

Since demand reaches a peak in summer months, a variable renewable like solar sized to meet summer peak demand would have excess capacity the rest of the year. But that is what we have with today’s conventional generation. We see that, just as with today’s system, a future system of variable renewables will have excess generation, or reserves. We shouldn’t be afraid of overcapacity, since we are using it in today’s conventional generation system.

Conclusion

We don’t need to be concerned about arguments that renewable capacity factor is like falling off the edge of the world. A deeper understanding of the power system reveals what the real issues are and that the technical problems, while challenging, are not insurmountable. Simple hand waving is not the way to understand complex systems. An approach like that used by researchers at NREL is a more realistic way of understanding renewables integration. And any load matching that renewables provide reduces the need for flexible sources and other techniques to adjust to variable output and demand.


Check out our new 93-page EV report, based on over 2,000 surveys collected from EV drivers in 49 of 50 US states, 26 European countries, and 9 Canadian provinces.

Tags: , , , , , , , ,


About the Author

has studied wind, electric vehicles, and environmental issues. An electrical engineer familiar with power and electronics, he has participated in the Automotive X Prize contest. He is an avid writer, specializing in electric vehicles, batteries, and wind energy.



  • Hans the Elder

    “A quick take on capacity factor says it’s about 30%, but what does that
    mean? It means that, compared to a flat power output all day long,
    solar is available for only those hours with its given curve.”

    It is probably just sloppy formulation, but here you suggest that capacity factor equals relative operating time. A common confusion that often exploited by renewable energy opponents. A PV system will, averaged over the year, produce electricity for 12 hours a day, i.e. a relative operating time of 50%. Most windturbines will produce electricity for 70 to 90% of the time.

    • eveee

      Yes. It’s in exact. It’s hard to bust a myth without restating it. And yes, agreed that capacity factor doesn’t tell you how often output is available. See the U of Wyoming article for some better metrics they created.

  • Freddy D

    West is the new south for siting PV systems to maximize financial revenue

  • sjc_1

    Solar from 8 am to 6 pm, wind from 12 pm to 6 pm with fast reactors.

  • Kevin McKinney

    Good summary, in my not-so-expert opinion. At least, it made sense to me!

    But this bit, I think, deserves comment:

    “…demand reaches a peak in summer months, a variable renewable like solar
    sized to meet summer peak demand would have excess capacity the rest of
    the year.”

    That’s not universally true. In particular, northern (or should I be more global and say ‘poleward’?) jurisdictions may have the reverse pattern. Alberta is an example. The link below is old–the data goes to 2009, since which much has changed–but shows that there, *winter* peak demand is a tad higher:

    http://www.aeso.ca/downloads/AESO_Future_Demand_and_Energy_Outlook.pdf

    And, of course, insolation there is strongly affected by seasonal variation; for Calgary, in southern Alberta, daylight near the summer Solstice is about 16 1/2 hours, but near the winter Solstice, just under 8.

    http://www.timeanddate.com/sun/canada/calgary

    That’s not to say, as some Canadian ‘mitigation denialists’ do, that solar doesn’t have a role there. In fact, the resource is better than Germany’s.

    https://solaralberta.ca/content/faqs

    But it is to say that the numbers are significantly different. Presumably RE integration there will look a quite a bit different than in, say, San Diego.

    • vensonata .

      I am off grid at the same latitude as Calgary. The only resource is PV. We overbuild significantly for summer in order to have enough for winter. PV is cheap enough to allow a 50% overbuild in the panels. Inverters and storage don’t need to increase though. Across the prairies in Canada both solar and wind are abundant in winter. In BC the winter light is not as good…but Hydro is basically 100% of all needs, so no worries. BC is the “Norway of North America!”

      • But you use batteries too, right?

        How much of your annual electricity comes from batteries? And at what times do you use them?

        • vensonata .

          Virtually all electricity goes “through” the battery bank. From sunset to dawn we exclusively use batteries for electricity. At our latitude the night is 17 hours in December, that requires about 50% of daily electricity supply, the other 50% demand is during the 7 hours of daylight and during solar production. On the shortest day of the year the battery will be fully recovered from the night use by noon, and we will actually be over producing if…it is sunny. We can go about 2.5 days on battery alone on cloudy snowy days. A high efficiency diesel generator will kick on automatically if the PV fails to keep the batteries above 50% full. Last year the generator ran a total of 85 hours.
          We are merely waiting for the arrival of lithium batteries to come close to 100% electric year round. 5 days storage should do it. But no one should think there is scrimping or uncertainty in off grid electrical production. In 16 years we have had about 20 minutes of down time. And however much electricity you want you can have at anytime..just kick on the generator. We overproduce by large amounts with our 12 kw PV array and are moving to hot water production through PV and a hybrid mitsubishi suv as soon as it arrives in Canada, just to try to find uses for the juices!.

          • Kevin McKinney

            Great comments; I love the user experience! Particularly as I look towards our home build, to start next year. It’ll no doubt have a sizable carbon footprint for the build, but hopefully will be offsetting that through reduced consumption for decades to come. We’re going to be in South Carolina, so solar is pretty much a no-brainer there.

            By the way, and if you don’t mind my asking, are you in BC? My roots are Ontarian, but there’s a whole branch of my dad’s family in BC.

          • Mind saying where in South Carolina? Just curious.

          • Kevin McKinney

            Tit for tat question! On Lake Wateree, northeast of Columbia, the state capital. Historically, it’s hydroelectric territory–that’s why there IS a Lake Wateree–and it’s also not far from the Sumner nuclear complex.

            BTW, SC is, according to the state, ~56% nuclear, ~25% coal, ~12% gas, and ~3% hydro. Despite the good solar resource, they needed 4 decimal places to right of the point to get to a ‘1’ for solar. Ouch! (But I doubt that includes distributed generation, for which there are some passable policies, I believe, though I think it’s early days in creating a strong SC solar ecosystem.)

            http://www.energy.sc.gov/files/view/2015%20South%20Carolina%20Energy%20Statistical%20Highlights.pdf

          • Mike

            Go solar!!

            When we built, we also specified steel roofing as normal shingles are not known to age as well under the panels.

          • vensonata .

            Location, south of Kamloops in the mountains. Back up diesel is trivial, about 2.5 litres per hour of run time. So, maybe $200 year. Divide by 15 residents about $13 year each!!. But we have spent years of conscious design and efficiency in the system, the results are satisfactory and still ongoing.

          • Kevin McKinney

            Thanks!

          • Excellent. Thanks for the details! Very interesting.

            And… how many people does this serve?

            (Just 20 minutes of downtime in 16 years beats any grid I’ve ever lived on. :D)

          • vensonata .

            Average adult population is 12-15. Peak accommodation of 24. One needs to realize that life off grid is completely comfortable and as elegant as one wishes to make it. We enjoy simplicity but not deprivation.

          • Mike

            Our average use is 15 kWhs per day. Based on your experience backed opinion, five days for us would be a 75 kWh LI battery pack (we have a n.g. Generac whole home generator to deal with the hydro outages here).

          • vensonata .

            For the first 14 years we used an 80 kwh lead acid battery bank. At present we have reduced it to 40 kwh because the PV input is large. I would say for a single family residence a 30 kwh lithium bank would meet about 99% of demand. When “load shifting” it is often better to use high efficiency NG directly than through a generator for such things as hot water or cooking. However we are on the edge of proper pricing for lithium batteries within a year or two…or possibly now with certain subsidies, so I expect to see some adventurous types going 100% solar/battery and zero generator soon.
            We will aim for that within 3 years.

          • Mike

            Please keep us posted with your move to LI in three years.

            Our situation, we use n.g. for radiant floor heat and supplement the solar water heater. Results in 3.5 tonnes of carbon a year (not including carbon burned to get n.g. to my meter).

          • eveee

            What’s your thoughts on Aquion?

          • vensonata .

            Aquion are promising. But they have stated that they will fall in price to $180 kwh (from $500 kwh now) after their factory is completed in a couple of years. Their only fault is their acceptance capacity. The thing about PV is that in the middle of the day we can generate 12 kw. And it is nice to have a battery bank big enough or flexible enough to actually accept that amount of juice. Alternatively we can store it in hot water, which we are looking to do on a significant scale soon. Right now we use it for electric cooking, water pump up from 300 ft to 1000 gallons of storage, and anything else we can think of. Soon we will purchase the Mitsubishi outlander PHEV which will use up some of our excess as well. It is a nice problem to have.

          • eveee

            Outlander, sweeeet. Stay tuned. Might have some info on battery pricing. They are starting to ramp up. It seemed like they were doing better than I initially thought. A pleasant surprise. They look much better than lead acid. 100% discharge, no problem. Partial state of charge, no problem, and much better cycle life. Then there is the low toxicity. All nice.

          • eveee

            Don’t know what your wind situation is like, but wind can potentially be higher in winter than in summer, balancing the solar winter decrease. In some northern latitudes, getting above the trees to the wind requires a tall tower. But that has to be balanced against the cost of alternatives.

          • vensonata .

            We measured wind for a full year. Neglible! So pure solar it was. At first it was expensive and we put up small ground mount systems with large battery back up and generator, then as prices fell it became obvious to just overpanel the PV. After 15 years on PV I still find it miraculous, completely reliable, and the only maintenance is occasional snow removal.

          • eveee

            Yup. Sometimes it’s just not possible.

    • Mike

      Good summary of the summary.

      As I’ve mentioned in the comments section before, my 5.7 kW system here in eastern Ontario has, for the past 6.2 years, consistently produced a surplus (on a yearly basis) of 1.2 megawatts over and above actual whole home consumption.

      I know my future ev will end up using that surplus……

      If/if the Canadian building standards require carbon neutrality by 2030 I see embedded solar as one great way to get closer to the goal.

    • eveee

      Quite clearly, Canada will be different from the US, with air conditioning loads lower. However for the 48 states, it’s surprising that even northern tier ISOs like ISO-New England and MISO show a distinct pattern of peak summer electrical usage. At least in the US, electrical heating is not a dominant factor in electrical loads. Air Conditioning dominates even in ISO New England.
      http://www.iso-ne.com/isoexpress/

      In Canada, a completely different story.

      The plus side to this for countries like Canada and Germany is that thermal space heating and water heating loads are less “pesky” and instantaneous. In fact, electric water heating is one load most manageable historically. That leaves the possibility that a source like wind that is strongest in winter can provide thermal space heating loads conveniently.

      The entire situation could change dramatically if electric vehicle use grows. That will provide a large buffer for daily electrical loads and possibly an increase in demand, smoothing the daily load cycle.

      • neroden

        Most of the Northeast still uses gas (NG) heating, or propane heating, or even worse, “heating oil” heating.

        We’re going to have to switch all of that to electric heating. That’ll move the peak from summer to winter. It’s a slow process.

        Switching to an electric heat pump is already extremely cost-effective for anyone using “heating oil” or propane, and close to breakeven for people using NG *at current prices*, which will go up.

        • eveee

          Yep. An opportunity for winter wind generation.

      • Freddy D

        The nice thing about AC loads is that they match solar production pretty nicely month to month and hour to hour. Obviously summer evening isn’t a perfect match but overall it’s very close. Decarbonizing space heating will be interesting. Heat pumps can go a long way but issues remain. There’s R&D on co2 heat pumps that will work more place.

        • eveee

          Great idea. Pass any info or links on CO2 heat pumps as you find it.

          • Freddy D

            Latest I heard was JB Straubel talk about improvements in CO2 heat pumps / AC in one of the videos posted here on Cleantechnica about a month ago. Almost as if he was hinting that was on the roadmap for tesla. I haven’t researched the latest for a while. Mercedes planned on going with co2 systems. The thermodynamics make perfect sense for building climate control. Companies like Carrier and others seem to move in geologic time as far as innovations however. Might take a company like Daikin or another to really commercialize.

    • eveee

      Thanks for the Alberta links. I checked it out and I will look at it in more detail. I wonder about the projections for increasing energy intensity.

      • Kevin McKinney

        Very welcome. I suspect those projections are completely toast–between 2009 recession and current state of the oil business, well…

        • eveee

          That’s why I welcomed your thoughts on energy intensity. It can go up as manufacturing intensifies. All the statistics are skewed by the tar sands project which consumes vast amounts of electricity. IMO, the EI projections are skewed by the expectation that tar sands operations increase. The opposite has been happening lately as oil prices plummeted.

          • Kevin McKinney

            Makes sense. And sounds like you’ve got more insight on it than I have.

    • First Officer

      My personal pattern is largest peak in winter with a smaller one in summer. I have an all electric house with a heat pump.

  • In those graphs I see words like “typical” and “average”. What happens when conditions aren’t typical or average?

    • Brent Jatko

      Wouldn’t engineering contingencies allow for such conditions?

    • CU

      In that case, you build a nuclear plant over night when demand appears, or tear it down. Or you may have some other reserve capacity.

      • eveee

        A nuclear plant cannot help for peak demand. Only flexible generation can do that. Nuclear plants want to run flat out 24/7. Useless for summer peak demand. If they run at less than full capacity, nuclear power plants become even less economic.

        • nitpicker357

          This was what is known as “sarcasm.”
          See the second sentence. Also France has nuke plants that they throttle, or at least they did.

          • eveee

            Duh. My bad. Sometimes my sarcasm detector gets loose. 🙂

    • Frank
    • eveee

      Those curves are used to show how it works. For example, NREL takes years of meteorological, solar, and wind historical data and runs computer simulations with that positioned against historical demand data and an uses algorithms with some Random Monte Carlo analysis to determine wind, solar, and other renewable output matching load. It’s even more complicated than I described, because they take grid congestion into account also, and vary many variables. Their conclusion is that using 2010 tech and reasonable evolution from them, 80% renewables by 2050 can meet demand 24/7.
      http://cleantechnica.com/2015/04/13/80-renewables-by-2050-in-us-says-nrel/

      That’s how an analysis is done scientifically. A cursory analysis using capacity factor is to crude to determine effects accurately.

    • Matt

      I will place a $1000 bet and pay 2 to 1 that part of the solution will be over capacity. Oh wait, we have always used one capacity. Ok 3 to 1.

  • Frank

    Nice article!

Back to Top ↑