What actually happens when Sacramento's grid gets stressed
Sacramento Municipal Utility District, commonly called SMUD, has been dealing with recurring power outage events that reveal some real weaknesses in the local distribution network. The most common scenario plays out during late summer afternoons when ambient temperatures climb above 100 degrees and the grid is pulling near its maximum capacity. Transformers overheat, voltage drops across certain feeders, and protective devices begin tripping in sequence. This isn't theoretical. I have worked on several outages where the root cause traced back to aging underground feeders that were originally rated for far lower loads than they are carrying today. The core issue here is a combination of infrastructure age, load growth, and insufficient hardening against thermal stress. Much of SMUD's distribution system was built between 1955 and 1980. Those underground conduits and transformer pads were designed for peak demands that look negligible compared to what modern residential and commercial loads require. Air conditioning alone has shifted the load profile dramatically. When you add electric vehicle charging and heat pumps into the mix, the math gets worse quickly. I spent a week troubleshooting an outage in the East Sac area last July. Three separate blocks lost power over a six hour window. The initial assumption from field crews was a substation-level fault. It turned out to be a cascading failure originating from a single overloaded lateral transformer. The transformer's thermal protection relay tripped, which de-energized a feeder section, which pushed more load onto adjacent transformers, which then tripped in turn. The whole thing could have been contained if there had been better sectionalizing switches and a microgrid configuration at the critical nodes. Instead, operators had to manually isolate sections one at a time.
Understanding blackout resilience requires looking at this from two angles: preventing the outage from starting, and minimizing the damage when it does start. Most utilities focus heavily on the prevention side. Hardening infrastructure, upgrading conductors, installing smart reclosers and fault indicators, and using advanced metering infrastructure to detect anomalies before they become failures. SMUD has been doing some of this work. They have replaced a number of older distribution transformers with higher capacity units and deployed automated switches on key feeders. The results show up in reduced SAIDI and SAIFI numbers, but the improvement has been incremental rather than transformative. The second angle, which gets far less attention, is resilience planning. This means designing the grid so that when something goes wrong, the impact stays contained and recovery is fast. Passive resilience is what you get by default with a radial distribution layout. A single fault takes down everything downstream. Active resilience requires reconfigurable network topology, distributed energy resources that can island themselves, and storage systems positioned strategically to maintain critical loads. SMUD has piloted some microgrid projects near institutional campuses, but these remain small-scale experiments rather than a systematic upgrade path. One thing that catches people off guard is the relationship between underground and overhead systems during extreme weather. Underground lines generally perform better during wind events and falling tree incidents. But when soil temperatures rise due to prolonged heat waves, underground cables lose thermal dissipation capacity. The ampacity derating can be significant. I once calculated a derating of nearly 30 percent on a 15 kilovolt underground circuit after three consecutive days above 95 degrees Fahrenheit with minimal airflow through the conduit paths. That kind of reduction pushes marginal circuits right over their thermal limits without any increase in actual electrical demand.
Another counter-intuitive point involves distributed solar. People often assume that rooftop solar behind-the-meter reduces grid stress during peak periods. It does, up to a point. But when solar production coincides with low general demand, such as midday on a weekend in spring, the net load on distribution feeders can drop so low that voltage rises beyond acceptable thresholds. Transformers begin feeding power back toward the substation. This reverse power flow condition is not ideal for the equipment that was designed for unidirectional flow. Some feeders in SMUD's service area are now experiencing this issue regularly enough that it has become a scheduling and protection coordination problem. Here is a practical workaround I developed for dealing with voltage regulation issues on feeders with high solar penetration. Instead of relying solely on traditional voltage control devices like line regulators and capacitor banks, you can use inverter-based resources to provide reactive power support. Most modern solar inverters have the capability to absorb or inject reactive power through their power electronics. By configuring the inverters to operate at a leading or lagging power factor depending on local voltage conditions, you can stabilize feeder voltage without upgrading hardware. SMUD has begun experimenting with this approach on a handful of feeders. The results have been promising, but scaling it across the entire distribution network would require significant communication infrastructure and coordination logic. When it comes to actual outage response, the timeline matters more than anything else. In my experience, the first thirty minutes after a fault occurrence are the most critical. If automated switching can isolate the fault and restore power to unaffected sections within that window, the overall impact shrinks dramatically. If operators have to drive out, visually inspect, manually close switches, and wait for crew clearance, the duration stretches into hours. SMUD has invested in a distribution management system that provides real-time visibility into feeder status. The system works well under normal conditions. Under severe events where communications fail due to cell tower outages or fiber cuts, the automation degrades and operators fall back to manual procedures.
Get the Full Details
I would be remiss if I did not mention the limitations of the current resilience strategy. Upgrading distribution infrastructure is expensive and slow. A single transformer replacement on an underground circuit can cost anywhere from eight thousand to twenty five thousand dollars depending on the rating and installation conditions. Sectionalizing switches and smart reclosers run a few thousand each, but you need them at regular intervals, which multiplies the cost across hundreds of miles of feeder. The budget constraints mean that SMUD prioritizes based on reliability metrics, and some neighborhoods simply rank lower on the list. This creates an equity concern that goes beyond the technical problem. Another hard limitation is the weather dependency of many resilience solutions. Solar plus storage microgrids sound great on paper, but they depend on sunny weather to recharge. During prolonged storm events or periods of heavy smoke from regional wildfires, those systems can deplete their batteries and fail to sustain critical loads. Battery storage is improving, and costs are coming down, but the current generation of lithium ion systems still has finite capacity. For a facility that needs to survive a multi-day outage, you need more storage than most commercial offerings currently provide at reasonable cost. If you are looking at this from a policy or planning perspective, the path forward involves combining several strategies rather than betting on one silver bullet. Distribution system hardening through conductor and transformer upgrades should continue, but it needs to be paired with deliberate network reconfiguration to create more redundant paths. Distributed energy resources should be integrated with grid-support functions like voltage regulation and frequency response. Microgrids and islandable loads should be expanded beyond pilot scale. And communication systems need redundancy so that automation keeps working when primary networks go down.
On a personal level, dealing with SMUD outages has taught me that the grid is more fragile than most people realize and more robust than most experts predict. The failures are usually concentrated in specific weak points rather than systemic across the whole network. Identifying those weak points early through data analytics and targeted field testing is where the real work happens. It is tedious, unglamorous work. But it is also the part that makes the difference between a localized flicker and a multi-hour outage that leaves thousands in the dark. For anyone managing or studying grid resilience in the Sacramento area, the takeaway is straightforward. The technology exists to make meaningful improvements. The constraints are primarily financial and organizational. Progress will be incremental. The grid will continue to stress under growing demand and changing climate conditions. The question is whether the adaptations keep pace with the problems or whether the problems outpace the response. I have seen both trajectories in my career. Neither one is inevitable. Both require deliberate action and sustained investment.
