Charging a fleet during a power outage

Twenty ports rated at 11.5 kilowatts is 230 kilowatts if every one of them runs flat out at the same instant. They never need to. What the depot actually has to do is put a fixed number of kilowatt-hours into a fixed number of vehicles before dispatch, and it has all night to do it. Divide the energy by the hours and you get the continuous power the backup source must sustain, which is a completely different and much smaller number.This is the same arithmetic that keeps a service upgrade off a normal depot project, applied to a generator instead of a transformer. If the site already has managed charging — and if it does not, that is the cheaper problem to fix first — then the backup source only ever has to match the managed profile, because the controller will hold the site to it.Two conditions make this real rather than theoretical. The charging management system has to keep working on backup power, which means its controller, its network path and any metering it depends on must sit on the backed-up side of the transfer switch. And the generator or inverter has to be told the limit, so that it is not asked to serve a load it cannot carry the moment twenty connectors go in at once.

Most depots have never written this down, and the answer is rarely all of them. A utility service crew, a medical courier, a route with a contractual delivery window and a route that can slip a day are different tiers, and the difference is worth a great deal of capital. If eight vehicles out of twenty genuinely have to roll, the resilience problem is a fifth of the size it looked.Tier on partial charge as well as on vehicle. A priority van that only has to cover a shortened emergency route does not need a full battery, and the energy required drops in proportion. Deciding in advance that outage-day routes are shortened is free, and it moves the sizing more than any equipment choice.Put the list somewhere a dispatcher can act on at five in the morning without calling anyone. A resilience plan that requires a decision from a person who is asleep is not a plan.

Start at the bottom of the list, because the cheapest interventions are the ones most often skipped. A written tiering plan protects nothing electrically and prevents a great deal of confusion. A small uninterruptible supply on the charging controller, the network gear and the site comms costs very little and means the row comes back automatically and correctly when power returns, instead of needing somebody to drive out and reset things.A permanently installed generator with automatic transfer is the answer most depots eventually reach, and the one most often oversized. Battery storage earns its place differently: it also works every ordinary day, shaving the demand peaks that drive commercial bills, which gives it a second revenue line the generator does not have. Which of those wins is a tariff question, and the tariff is specific to your utility and rate schedule.One thing that is not on the list as a standalone answer: a grid-tied solar array. A conventional grid-tied inverter shuts down in an outage by design, so that it does not energise lines a utility crew believes are dead. Solar contributes during an outage only if the system is specifically built to island, which normally means storage and an inverter selected for it. Assuming otherwise is the single most common misconception in depot resilience conversations.

Ask two questions of any generator proposal. What is the consumption rate at the load this site will actually place on it, from the manufacturer's data sheet rather than from a rule of thumb. And how many hours does the installed tank give at that rate. Divide one into the other and you have the real answer, which is often shorter than the outage the plan was written for.Then ask the question nobody asks: who delivers fuel during a regional event, and where does your depot sit in their priority list. A fuel supply contract with a stated priority is part of the resilience purchase. So is a service and load-bank testing schedule, because a standby generator that has not been exercised is a large object with an uncertain relationship to electricity.Fuel storage brings its own permitting and secondary containment requirements, and those vary by jurisdiction and by quantity. That is a conversation with the same authority reviewing the rest of the depot, and it belongs early rather than at inspection.

This is not a formality and it is not negotiable anywhere in the United States. A source connected without proper transfer means a utility crew can be working on a line they have every reason to believe is de-energised. The specific requirements, the acceptable equipment and the review process vary by jurisdiction and by utility, so raise it with both the permitting authority and the utility at design stage rather than presenting them with an installation.The utility conversation has a second purpose. Some utilities have views, rules or programs relating to on-site generation and storage at a commercial customer, including how it interacts with your rate schedule. That is a per-utility fact rather than a national one, and it is worth getting in writing before the equipment is ordered.On incentives, be blunt with yourself about what is available. The federal charger credit under Section 30C terminated for property placed in service after June 30, 2026, so there is no federal charger credit to net against any of this. Anything real will be a state or utility program with a named sponsor and a published date, and resilience and storage programs are a different set from charger programs. Look them up by sponsor rather than assuming a national number.

The tier list, in vehicle order. The charging limit the site runs at while on backup. Who confirms the transfer worked and where they check it. Which routes are shortened and who tells the customers. The fuel supplier and the account number. The number for whoever services the generator. And the trigger to move to public charging as a last resort, which for a small tier of critical vehicles is sometimes the cheapest answer of all.Test it once a year on a scheduled morning. The test is not whether the generator starts; that is the easy part. The test is whether the chargers come back under the correct limit, whether the management platform reconnects, and whether the person who was supposed to know what to do knew what to do.