The fundamental purpose of an independent system operator (ISO) or balancing authority is to maintain grid reliability by ensuring that energy demand is met and by implementing measures to ensure against energy shortfalls. The ISO does not produce energy, but instead facilitates and manages the market between energy producers and utility distributors. To ensure that energy is consistently and reliably delivered, the ISO must balance production and demand by informing markets and procuring energy reserves. Thus, it requires accurate estimates of energy production, energy consumption, and uncertainty to plan the procurement of energy reserves that can be quickly dispatched.
Over a day in advance, the ISO predicts the demand for electricity as well as creates an initial estimate of total (especially must-take) energy production. Based on this balance and on energy-transmission constraints, energy prices are established. The ISO can then create scheduling instructions for unit commitment and available reserves. In “real time” (minutes to hours ahead at up to 5 min resolution), conditions often do not match the DAM schedule because of inaccurate demand or renewables-production forecasts. It is the ISO’s responsibility to actively manage the grid in real time by operating the RTM, dispatching reserves, curtailing production, and maintaining up and down regulation. Accurate realtime estimates of production and demand are therefore also required.
For solar energy, the ISO requires an accurate estimate of power production for both the DAM and the RTM. In principle, aggregate energy production (i. e, all rooftop and utility-scale solar-power plants) at each LMP node must be predicted. However, in practice forecasts are provided for each utility-scale plant individually. Since solar forecasts are imperfect, a measure of forecast uncertainty is included in the form of an energy-production confidence interval. Using this range, the ISO can consider worst-case scenarios when determining energy availability and reserve requirements. Since intrahour scheduling does not currently occur in the DAM, an accurate characterization of intrahour variability is not essential. However, accurately predicting the timing of sustained, large-magnitude ramp events is critical. Though ramp events are partially reflected in the mean-power forecast, an independent ramp-probability time series is desirable to alert operators to potential risks and to describe uncertainty in ramp timing. Provided at an hourly resolution once per day, the mean-production forecast, uncertainty, and ramp-event probability form the ideal DAM forecast for the ISO.
In the RTM, the same forecast components are required but at a much finer temporal resolution (e. g., 5 min) and with frequent updates (generally once per hour). Since accuracy standards are higher, uncertainty bands must be narrower than in the DAM. Generally, intrahour variability is resolved directly (since the RTM forecast is provided at a high temporal resolution) and can warn of unexpected production shortfalls. For the ISO, a valuable solar forecast minimizes the probability of loss of load and economic losses that result from compensating for energy shortfalls in the RTM through either energy purchases (equation 14.4) or capacity procurement. Typical ISO forecast requirements are summarized in Table 14.1.