Addressing Uncertainties Through Improved Reserve Product Design

Traditionally contingency reserves were defined to manage generation outage uncertainties. With the renewable penetration and more frequent extreme weather events, market operations require additional margins and flexibility to manage increasing uncertainty and variability. MISO implemented 10-min ramp capability products and 30-min short term reserve (STR) product. It becomes increasingly important to appropriately define reserve product requirements and demand curves on market wide and sub-regional basis. This paper introduces methods developed at Midcontinent ISO (MISO) to enhance uncertainty management through improved reserve product designs. A systematic method is developed to quantify net load ramping and capacity uncertainties across different lead time. Machine learning clustering method is applied to define seasonal hourly dynamic demand curves and requirements for market wide requirements. The method is also applied to define sub-regional uncertainty events to ensure appropriate amount of reserve is procured at the sub-regions.