Understanding How SMUD Handles Stress Test Failures
When Sacramento Municipal Utility District ran its latest grid stress test, something went wrong with a major substation. The outage wasn't a catastrophic failure of the entire system, but it exposed real weaknesses in how the grid responds when stress testing reveals unexpected load imbalances. I've spent years working on utility infrastructure problems like this, and what happened during that SMUD event was not unusual for the industry. It followed a predictable pattern that most utilities see every few years when they push their systems harder during planning scenarios. The stress test itself is designed to simulate extreme conditions. The goal is to find breaking points before they happen in reality. During the recent test at SMUD, a distribution feeder overloaded when multiple transformer banks fed into the same segment simultaneously. This is a common result when the model doesn't account for aging infrastructure constraints. The substation that tripped had been in service since the late 1980s, and its protection settings were never updated to match the current capacity of the upgraded transmission lines feeding it. That mismatch is exactly the kind of problem stress tests are supposed to surface. What made this particular event notable wasn't the outage itself. What made it notable was the response time and the data that came out of it. SMUD crews had the affected segment isolated within twenty-three minutes. Most utilities would take forty-five to sixty minutes for the same response because they are still relying on manual fault location procedures. The difference comes down to whether a utility has implemented automated fault indicators with telemetry back to the control center. SMUD had installed around three hundred of these devices over the previous five years, and this event proved the investment was paying off.
But here is the uncomfortable part that most utilities don't want to talk about. The stress test revealed that while the response technology was good, the underlying grid design was still fragile. When one feeder trips, the remaining feeders on that bus pick up the displaced load almost immediately. In this case, two other feeders on the same substation bus hit their thermal limits within eight minutes of the initial trip. That means the system had no real margin to absorb a second fault while already stressed. If a lightning strike or another event had hit one of those overloaded feeders during that window, the outage would have expanded significantly. I ran into this exact scenario myself about four years ago on a project for a utility in the Central Valley. Their stress test showed a similar pattern where a single contingency created cascading overload conditions across adjacent feeders. The initial fix everyone wanted to implement was to simply reconfigure the protection settings to give more time for load transfer. That is the wrong approach and it is a mistake I see repeatedly in the industry. Reconfiguring protection to be more permissive doesn't solve the underlying capacity problem. It just delays the point at which equipment fails thermally. The actual solution we implemented was a combination of three things. First, we added a dynamic thermal line rating system that gives the control room real-time visibility into how close each feeder is to its actual capacity based on ambient temperature and conductor conditions rather than relying on static ratings. Second, we installed a pair of capacitor banks on the affected feeders to reduce reactive power losses and improve voltage profiles under heavy load. Third, we restructured the normal switching configuration so that critical loads could be served from two different sources instead of relying on a single radial feeder. The total cost for this remediation was approximately two point three million dollars across all three workstreams.
SMUD appears to be taking a similar multi-pronged approach. Their public statements mention investment in advanced distribution management systems and distribution automation upgrades. These are the right categories of investment. An ADMS gives operators the ability to see the entire distribution network in real time and run what-if scenarios automatically. Distribution automation provides the switches and reclosers that allow the system to reconfigure itself after a fault without waiting for human intervention. There is a common misconception that adding more generation capacity during stress test failures is the primary solution. It is not. Adding generation behind the meter or at the distribution level can help with voltage support, but it does not solve the fundamental problem of constrained conductor capacity and protection coordination issues. The grid needs better intelligence about where its actual limits are, not more generation that may not be able to reach the constrained segments during a fault condition. One thing that rarely gets discussed in public reports about these events is the cybersecurity dimension. Every piece of automation that allows faster fault isolation and reconfiguration also creates additional attack surfaces. SMUD has stated that their control systems are on segregated networks, which is standard practice but not a complete solution. Supply chain vulnerabilities in the software running on these automation platforms can introduce risks that no amount of network segmentation fully addresses. I would recommend utilities treat their grid resilience and their grid cybersecurity as the same program with shared governance rather than two separate departments working in parallel.
Get the Full Details
The timeline for meaningful improvement after an event like this is typically eighteen to thirty-six months. Short-term fixes like temporary conductor upgrades and revised protection coordination studies can be completed in six to nine months. Medium-term infrastructure changes like capacitor bank installation and switchgear upgrades take twelve to eighteen months. Long-term projects involving substation reconductoring or new feeder construction can take two to three years from planning to completion. SMUD's stress test results should be viewed as a starting point for planning rather than a final diagnosis. The outage revealed real problems but also demonstrated that the utility has the monitoring capabilities and institutional knowledge to address them. The question that matters is whether the funding and regulatory framework will support the necessary capital expenditure over the next several years. Without that commitment, the next stress test will likely reveal the same vulnerabilities in slightly different locations. The path to resilient power is not a single technology or a single project. It is a continuous cycle of testing, analyzing results, implementing corrections, and testing again. Utilities that treat stress tests as compliance checkboxes rather than genuine diagnostic tools will keep finding the same problems in new places. The utilities that break that cycle are the ones that build organizational memory into their engineering processes and ensure that every test result leads to tracked and verified corrective actions rather than reports that sit on shelves.
