A load-duration curve is a useful tool for comparing the impacts of different renewable portfolios on the grid. In this Rocky Mountain Institute analysis of renewable adoption on the Electric Reliability Council of Texas (ERCOT) grid, a generation mix of 25% solar and 15% wind yields the flattest load-duration curve over the year.
A load-duration curve shows the percentage (or number) of hours of the year at which the load is at or above a given value or percentage of peak load. To make a load-duration curve, the 8,760 hours of the year (8,766 average including Leap Day) are sorted in decreasing order of their average (or peak) hourly load. The y-axis can then represent either the actual load levels (MW) or the percentage of the peak load over the course of the year. The load-duration curve for a particular system makes it easy to see, for example, that the total system load exceeds 90% of peak load in 200 hours out of the year, or that in 50% of the year, the load is at or above some level of capacity in MW. In general, a flatter load-duration curve is better for grid operation, allowing dispatchable generation to run at a higher capacity factor over the entire year and requiring less peaking reserves. The easiest system to operate, for example, would be one with a load-duration curve that is a horizontal line—indicating the demand is entirely constant for the whole year.
This example, modeled with the RMI hourly dispatch model, compares various portfolios of wind and solar generation to meet 40% of annual energy demand on the Electric Reliability Council of Texas (ERCOT) grid. Rather than show a load-duration curve for the entire system load (which would look the same regardless of the renewable portfolio), this chart shows the duration curves for the net system load—the total initial load minus the wind and solar generation. This is the remaining load that the system operator must meet with dispatchable generation, and again, a flatter curve is desirable.
In this case, a generation mix of 25% solar and 15% wind yields the flattest net load-duration curve, though very flexible mixes are possible. Increasing the fraction of either solar or wind causes the right tail of the load-duration curve to dip, meaning that more renewable energy must be curtailed (spilled when generated but not needed). In a microgrid, this would place a larger burden on being able to export power to the grid. And with lower penetrations of solar (whose output is highly correlated with Texas peak summer afternoon demand), the left tail of the load-duration curve climbs, increasing the level of peak load that must be met with dispatchable peakers.
RMI analysis using data from:
Electric Reliability Council of Texas. 2004. “FERC Form No. 714-ERCOT.” link