Variables of Variability

I recently participated in a web discussion with proponents of nuclear power about the future of baseload power (as represented by nuclear plants that have to operate around the clock) and what they see as the problems with integrating large amounts of renewable power into the grid.

Their big bugaboo was variability – renewable resources do not operate 24/7. They fluctuate in output based on the windiness or cloudiness of the weather. But demand likewise isn’t consistent 24/7. The variability of renewable power, nuclear proponents claim, means we have to build large amounts of back-up fossil generation to balance the renewable output. But scads of recent research and operating experience across the nation’s electrical interconnections tell a different story. So far variability has been no big deal, and as we add more and more renewable power the integration challenge can be managed by better coordination of the system and prudent use of generating resources.

Simply put, variability is not the intractable problem baseload power promoters imply. It is an eminently manageable challenge.

There is plenty of research that shows variability can be routinely and effectively managed. If you have to balance resources over a small geographic footprint using bilateral contacts and all your renewable generation is in one place, then you may have problems. To offset variability one needs to have resources with uncorrelated variability. In other words, one needs resources that operate at different times from each other. Distance matters when you are dealing with variability. The more distant the resources are from each other, the less correlated their variability is. The wind is usually blowing somewhere, and a cloud doesn’t cover the entire world at once.

All generation is somewhat variable when you consider unforced outages, maintenance and the like. Most renewable variability is pretty predictable, as Xcel Energy routinely shows as they curtail coal generation in favor of wind power in Colorado based on forecasts. Northrop Grumman has also done work in this space developing a modeling platform to decide where to locate wind generation in order to match the generation shapes of wind from one area with that of another, using radar data developed for the Department of Defense. The National Renewable Energy Laboratory has done a number of wind integration studies in both the East and the West showing the value of grid coordination and exploiting uncorrelated renewable energy variability. The University of Wyoming also has done work comparing wind shapes from parts of Wyoming with wind from Colorado, Utah and California. The results show real value to the system from making both the uncorrelated variability and low cost, high capacity factor wind resources available to other western states.

If grid managers can draw from a bigger footprint and are able to aggregate uncorrelated variability, facilitating rapid dispatch via consolidated control in an organized energy market as Regional Transmission Organizations (RTOs) do right now, the problem is greatly diminished.

We can also use demand side resources to further assist with the task. If you can remove some of the unpredictability of demand, and match it to the expected and forecast performance of the variable generation, the problem becomes a non-problem and part of normal system operations. Grid operators are gaining the tools to regulate demand just like power plants, dispatching down load at peak times or turning demand up when the system has an oversupply of power, by, for example, charging electric vehicles.

If you think a smart grid is some exotic concoction for the future, you may want to guess again. There is a vibrant industry developing around system controls, advanced power electronics, electricity storage (as evidenced by the Tesla “gigafactory,” and driven in part by the California electricity storage mandate), and advanced system communications tools. The horse is well and truly out of the barn. We are not going to live with the grid structure of the last century. The system needs to be (and in fact is) being modernized (see the following: for a good overview of where things are headed from the DOE Electricity Advisory Committee).

Despite assertions to the contrary, renewable resources are not automatically a drain on the grid. They are quickly becoming the main source of power in many locations in the system, particularly in the western U.S. Xcel Energy has at times served over 60% of their load from wind, though so far this has been rare. Renewables provide enormous benefits, even on the distribution grid, where they are being accused of cost shifting, but where they actually can help avoid costly and unnecessary generation and transmission investments.

Activating the functions already inherent in many microinverters on photovoltaic panels, for example, can provide such grid services as frequency response, low voltage ride through (they can keep working if the system falters as voltage drops) and reactive power (power needed to keep the system energized so active or “real” power can flow to consumers). This obviates the need for new gas generation for these purposes, both saving money and reducing all types of air pollution – not just carbon dioxide.

As the inverters already have this ability, the only cost is in turning it on. And turning it on can be a relatively simple task. A few weeks ago, some of this solar panel functionality was remotely switched on to more than 800,000 microinverters in Hawaii from Petaluma California in one day. The California Independent System operator is currently considering requiring all new large scale solar to provide this functionality going forward.

Welcome to the future. This is not your granddad’s grid. That one served us well – and so will the new one.


Carl Zichella is the director of western transmission for the Natural Resources Defense Council. 

Morning Consult