Obama administration energy policy targets a 500-percent increase over the next 15 years in solar, wind and tidal power’s share of the U.S. electricity system to 25 percent. If a way isn’t found to stop the North American power grids’ declining ability to respond rapidly enough to shocks, the increase in renewable generation already underway will accelerate stress on the North American power system to the breaking point.
No amount of money spent on a smart grid or on expanding the transmission system will address this problem, while consequent administration pressure on the Federal Energy Regulatory Commission to fast track a solution risks backfiring into no solution at all.
Solar, wind and tidal renewables add stress to a power grid in two ways, as if in exchange for their zero fuel cost. One is by making very rapid random changes in the power they produce. The other is by their physical inability to very rapidly supply energy to arrest the effect on all generators of very rapid random generator output shortages before the effect grows too big on the power system, and thereby to help maintain a more stable continuous balance between supply and demand on the grid.
At the same time, the ability of the three North American power grids (called “interconnections”) serving the United States to provide rapidly enough such arresting energy (i.e., within seconds of a disturbance) steadily has declined, on the Eastern and Western grids (which are roughly separated by the Great Plains) especially since these were opened in 1996 to wholesale electric power markets. This has been especially true on the Eastern grid (Figure 1) which includes Canada from the Prairies eastward but excludes Texas and Quebec. However the decline had begun years earlier on the grids when for plant safety reasons nuclear power plants were proscribed from suddenly increasing generation to respond to a sudden variation in energy imbalance.
How does an electric power system normally prevent higher levels of stress from reaching the breaking point? The biggest shocks on a grid consist of sudden rapid either loss or excess of power due to loss of a generator, a power line, or a load. An electric power grid defends itself proactively against these shocks by offsetting them as rapidly as possible, specifically by farming out the job to all the generators and loads on the grid—a practice unique to electric power systems among physical systems. The generators are thereby each obligated to provide simultaneously within seconds a small portion of the total offset needed to keep supply and demand steadily balanced on the grid.
One or a few of the generators alone could provide the total offset, but too slowly (by taking as long as ten minutes to ramp up a generator). That adjustment time is critical, because the more seconds it takes to offset the shock, the bigger the shock grows by speeding up or slowing down all the generators on the grid to the blackout point where loads or transmission lines are cut by automated frequency or transmission relays1, causing more shocks. Provided that doesn’t happen because rapid offset was provided, a few generators subsequently do provide such slow offset to permanently replace within about ten minutes the power or demand sudden loss of which triggered the shock. This slow recovery enables all the generators and loads on the grid to withdraw their rapid offsets over that time and stand ready rapidly to temporarily offset the next shock.
How has the power grid’s ability to rapidly respond deteriorated and how can the deterioration be stopped? Can technology both decrease renewables’ instability and also increase their ability to provide stability to the grid? Why are very rapid changes a challenge unique to electric power systems?
An answer to the last question begins by recognizing that an electric power generating system traditionally has been a collection of rotating mass inside turbines with a quality called inertia. Inertia serves to make slower any sudden turbine slowdown due to a sudden torque force or counter-spin due to sudden loss of a power generator on the electric grid and due to the very rapid need to spread less power over the same user load base namely at lower electrical frequency (i.e., at fewer cycles per second of rotating speed). Also, electric motors that are consuming electricity experience the same torque that slows them down and thereby reduces demand so as to permit the electric power generators to slow down somewhat less.
More important, a device called a governor triggers all the electric generators equipped with one to simultaneously and very rapidly produce their small share of the very swiftly produced counter-torque (i.e. governor response) needed to rapidly neutralize the torque that’s causing the generators to slow down and needed basically to arrest the decline in turbine rotation speed. This change in speed of the turbines is measured by the change in frequency of alternating current electricity (which is normally 60 cycles per second, also known as 60 Hertz or Hz), which is everywhere the same on the grid. Arresting, and thereby minimizing, the decline in electrical frequency to an equilibrium point where supply and demand temporarily are rebalanced, is a capability that each power generator supplies mainly for itself, but it’s measured as a flow of power to the place where the original loss of a generator occurred (see Figure 2).
Governors are gradually reset to their original positions as one or more generators take the time—about 10 minutes—required to ramp up replacement of the original amount of power loss, thereby generating new speed-up torque prompting the governors to reverse torque to their pre-shock positions. Frequency thereby returns from the post-shock extreme constrained by the governor response deployed, to its pre-shock position, from where the governors are ready to respond immediately to the next event. That ex post action by a few generators—especially some deployed by the control operator where the shock originated—to restore frequency to its pre-shock position is called “automatic generation control” (AGC) or regulation and is what most control automation software provides.
In answer to the first and second questions of how the grids deteriorated and what to do about it, no sustainable, preferably market method yet has been developed to properly value or pay generators for providing very rapid governor response offset service, nor to penalize the bad control behavior by generators and load that causes very rapid and big supply/demand imbalances that trigger deployment of this service. Without such a method, the past decade and a half of steady decline in rotating mass and governor response to increasing shocks on the electric power grid, relative to the size of the grid, will continue until the interconnection cannot arrest in time a rapid power loss big enough to trigger a wide-area blackout. This brings us to an answer to the third question of how to reduce the very rapid variability in renewables’ collective output and improve their collective ability to respond very rapidly to sudden shortages on the grid.
Not all power generators are equally good at providing very rapid always-available governor response and rotating mass. Nuclear provides rotating mass but no governor response to sudden shortages. Wind, solar, and tidal are the worst: Not only do they provide no such rotating mass or governor response to shortages, but instead they are randomly contributing to an increasing extent to the very rapid supply and demand imbalances that require more rotating mass and governor response provided exclusively by all the other generators and by loads. Technology to improve the very rapid reliability of wind, solar, and tidal generators is limited and very expensive. For example, storage devices like batteries and flywheels can be used to provide very rapid substitute power for a sudden drop in output, but only once since they take a long time to recharge.
So the wind, solar or tidal generator would need an enormous spare battery capacity on hand, enough to respond to all the sudden major frequency drops that can occur during the long recharge period of a battery required to respond to a single major very rapid frequency drop, a very expensive proposition. Alternatively, renewable generators could buy very rapid response capacity and rotating mass from traditional generators. In any case, wind, solar, and tidal generators need to be charged/penalized for stressing the grid in order to incent them to buy the huge amount of very rapid response capability they need to fairly share the burden of all the other generators of responding to the very rapid stresses on the grid, including those caused by the wind, solar, or tidal power generator.
Meanwhile, deadband control technology also has been steadily contributing to the decline in the North American power grids’ ability to respond to very rapid changes in electrical frequency. Deadbanding software has been increasingly applied to the operation of governors on generators to delay/eliminate their response to rapid frequency changes until the changes grow, and thereby to save wear-and-tear cost. By the same token, however, deadbands have been causing large frequency variability to increase on the interconnections by not responding to it immediately, and thereby have been increasing the need for more capability to respond to rapid frequency changes thanks to the reduction in that ability caused by the governor deadbands. This was demonstrated in a recent study of the Electric Reliability Council of Texas’ grid (see Figure 3).
Originally deadbands were justified when placed on mechanical governors controlling older power turbines to account for gear lash, which is the delay in gear re-engagement when the torque on the gears is countered or reversed. Most of those older turbines have since been retired, while current turbines don’t encounter the gear lash problem but use deadbanding software on their governors that is designed to reduce turbine maneuvering cost, not to compensate for mechanical characteristics. That application has furthered the commercial interest of the individual generator at the expense of the reliability of the grid.
To reconcile the coming crisis in North American grids’ ability to stop sudden rapid deterioration in electrical frequency with the aggressive Obama renewable energy agenda, FERC issued an order on March 18.2 That order required the continent’s legislatively designated electric reliability organization, the North American Electric Reliability Corp., to accelerate its current ANSI3-certified industry-consensus process of developing an operating standard that assures that the two largest North American electric grids serving the United States achieve a reliable level of very rapid power fluctuation and response. FERC ordered that the process be concluded in six months. That deadline for solving a definitely 15-year-old problem was appealed4 successfully, when on May 13, FERC ordered5 a rehearing on the matter to be preceded by a technical conference with the industry at a so-far unspecified date. FERC appears to be waiting for NERC to post a draft standard soon for industry comment.
Meanwhile, the market economics of renewables might be slipping away. FERC’s Summer 2010 Energy Market & Reliability Assessment released on May 20 forecasted just 12.1 percent of U.S. wind power will be available on peak, a huge 4-percent drop from last year’s 15.2 percent of a 17-percent smaller capacity. The drop in peak availability is due to more experience with wind output among grid operators and improved forecasting. Add carbon capture and a vast natural gas supply coming into place in the Northeast, and alternative generation does not have an economic fighting chance of succeeding as presently funded, with wind development winding up as a government subsidized effort paying for MW output only—not development on an economic and market basis. So we might witness wind towers rusting in the plains. Add to this having wind pay its fair share of the reliability cost of stressing the grid, and administration energy policy faces much higher costs.
Moreover, since the mainland Canadian provinces except Quebec share the two largest of the North American power grids with the United States, their governments too have something to say about the reliability of those two grids, and any NERC standard requires their approval too for it to be legal and effective on a power grid where stresses don’t recognize legal jurisdictions.
Cost and jurisdictional hurdles aside, until a technically-sound and economically-rational standard is developed to properly address the North American power grids’ declining ability to proactively limit the size of very rapid power fluctuations, Obama administration policy of increasing the share of wind, solar, and tidal generation on the U.S. portion of those power grids will definitely increase the likelihood of blackouts.
1. Automated frequency relays are safety devices to disconnect a turbine if frequency ever lowers near the point of causing turbine blades to vibrate enough that the turbine would otherwise explode and project pieces more than a mile away. Transmission relays disconnect power lines if frequency ever rises near the point of overloading transmission lines with “surge”.
2. FERC Docket No. RM06-16-010, Mandatory Reliability Standards for the Bulk Power System, Order Setting Deadline for Compliance, 130 FERC ¶ 61,218, March 18, 2010.
3. American National Standards Institute
4. FERC Docket No. RM06-16-010, Request of the North American Electric Reliability Corporation for Clarification and Rehearing of the Order Setting Deadline for Compliance, April 19, 2010.
5. FERC Docket No. RM06-16-010, Order Granting Rehearing for Further Consideration and Scheduling Technical Conference, 131 FERC ¶ 61,136, May 13. 2010.