When the Grid Gets Charged: Tracking Batteries and Market Behavior at DWR

David Zhou
’28 Electrical Engineering
Undergraduate Fellow, California Department of Water Resources

This summer, I’ve been working with the California Department of Water Resources (DWR) through the Shultz Energy Fellowship. While DWR is best known for managing the State Water Project — the vast network of reservoirs, aqueducts, and powerplants that moves water across California — it also operates behind the scenes as the state’s largest single electricity consumer. With DWR’s Operations and Maintenance office, I’ve been exploring how the department can adapt to California’s rapidly evolving energy market. 

So far, my main project has focused on analyzing how utility-scale batteries in California have been operating in the wholesale electricity market, especially under the California Independent System Operator (CAISO). While batteries are often talked about as a critical tool for balancing renewable energy, I’ve been digging into the actual charge and discharge data to see how they’ve behaved in practice. 

A typical day involves sorting through large datasets — hourly and sub-hourly records of battery charging, grid demand, and market pricing — and trying to make sense of them using MATLAB. One project I worked on involved aggregating charge and discharge data across different months to look at how battery behavior shifted over time. 

Instead of seeing the expected back-and-forth charge-discharge cycles, I noticed that cumulative charging trends showed a steady net discharging pattern — meaning batteries, over the long term, were supplying more energy to the grid than they were absorbing. 

That surprised me. California’s market is designed to reward batteries for providing both supply and demand flexibility, but this pattern suggests that market signals might be pushing operators to treat batteries more like traditional generators. It also raised questions about how these operations are shaping net demand curves and resource adequacy planning. 

A big factor here is that much of the battery charging may be happening behind the meter — charging done off-market and invisible in public data. This would help explain how batteries can appear to be discharging more than they charge in CAISO records. It also underscores how the full picture of grid behavior is often hidden beneath layers of market structure. 

What’s stood out most this summer is seeing how energy market operations, regulatory structures, and engineering data all intersect in ways I hadn’t appreciated before. Working at DWR has been a reminder that technical analysis matters most when it’s grounded in real-world systems — and that understanding the grid is as much about people and policy as it is about circuits and code.