CAISO and Sisyphus: The Challenges of Operating a Rapidly Modernizing Grid

Asia Zhang
B.S. ‘22 Energy Resources Engineering and Data Science
M.S. ‘26 Energy Science and Engineering
Graduate Fellow, California ISO

Managing the flow of electricity across high-voltage, long-distance transmission lines in California could be described, most simply, as a sisyphean task. Ensuring that the supply and demand of electricity is in constant balance, operating a competitive wholesale electricity market, and planning for future transmission needs, the California Independent System Operator (CAISO) boulders this balancing act to provide reliable, 24/7 energy to its consumers. 

As part of the effort in pushing this boulder uphill, one of my projects is focused on price validation. Specifically, using statistical modeling and machine learning to identify anomalous prices and their drivers. The CAISO operates the Day-Ahead and Real-Time market to determine optimal generation schedules and prices at the least-cost, while simultaneously ensuring system reliability and meeting forecasted demand. This involves a complex optimization process that outputs a market solution which includes market clearing prices that reflect system conditions. These market clearing prices can sometimes spike or be invalidated (anomalous behavior) due to several factors, like scarcity of resource bids, invalid market solutions, or data input failures, just to name a few. The process for identifying price anomalies is currently mostly manual and requires the decades of combined expertise of the Market Validation team to decipher the issue and provide corrections. The hope is to build and test a workflow that can help automate some of the current identification processes.

The other project that I’m working on involves the Extended Day-Ahead Market (EDAM). The proposed enhancement to the current market structure is designed to address the large imbalances that occur between the day-ahead and real-time markets, enhancing price convergence and reducing inefficiencies. The EDAM also extends the day-ahead market’s optimization across a wider geographic area, allowing participants to benefit from a footprint wide diversity. My role in this project is to perform market simulations and conduct sensitivity analyses for different scenarios in EDAM, to better understand the benefits and limitations of the new market products.

Working for the CAISO has been an enriching experience, exposing me directly to the daily challenges of working with an increasingly complicated and dynamic energy grid. With the guidance of my incredibly knowledgeable mentors, Dr. Guillermo Bautista Alderete and Dr. Kun Zhao, who have generously taken the time to explain the intricacies of price formation and the many markets that CAISO operates and pointed me towards helpful modeling tools and methods using R and Python, I’ve been able to dive into the CAISO’s vast vault of data and begin to untangle its complexities.

In the face of increasing extreme weather events (i.e. wildfires, heatwaves), electricity loads (i.e. EV charging, data centers), and regulatory and political pressure to generate 100% clean electricity by 2045 (SB 100), the CAISO upgrades forecasting tools, develops market mechanisms, and adapts; they roll the boulder back up, and keep the lights on. CAISO’s responsibilities resemble a sisyphean task not because they are futile, but because they require a tireless vigilance in evolving conditions and an excited passion to tackle each challenge. After all, “One must imagine Sisyphus happy.”