Load management for fleet charging: the three architectures

This is the simplest thing that works and it is often enough. You configure a budget — say 80 kW for a group of ten ports — and the chargers divide it, giving full power to a single vehicle and sharing it as more plug in.Two limitations. The budget is blind: the group knows what the chargers are drawing and nothing about the compressor, the lighting or the shop, so the budget has to be set conservatively against the worst case of everything else on the service. And it is normally single-vendor, because the local protocol between units is usually proprietary even when the charger speaks a standard protocol to the outside world.

This is what the industry calls dynamic or adaptive load management, and the difference from a fixed budget is real money. A blind budget has to assume the rest of the site is at its worst; a measuring controller gives the chargers the headroom that actually exists at 02:00, which is usually most of the service.It is also the architecture the electrical code contemplates. Where the relevant editions of NEC Articles 625 and 750 are adopted, an energy management system that limits load is a recognised way to size a service against the managed maximum rather than the connected total, which can remove a service upgrade from the project entirely. Adoption and interpretation vary by jurisdiction, so agree the approach with the authority having jurisdiction before the design is final.What it does not do is know anything about vehicles. It manages power, not schedules. It cannot tell that the van on port four leaves first.

This is the layer that knows about the fleet. Fed by telematics, it knows which vehicle is at which port, how empty it is and when it leaves, and it can sequence more vehicles than you have ports, prioritise the long route, and respond to price signals or a utility demand-response event.The cost is a dependency. The intelligence lives somewhere else and reaches the chargers over a link that will, at some point, not be there. That is not a reason to avoid it. It is the reason the next section exists.

The middle answer is a locally stored default. OCPP smart charging allows a default charging profile to live on the charge point itself, so a unit that loses its connection can keep applying the limit it was last given rather than reverting to its rated maximum. Whether a specific product does that, how long the stored profile persists, and what happens when it expires are product questions with product-specific answers.So ask for the behaviour in writing before you buy, and then test it during commissioning. The test takes ten minutes: plug in every port, pull the WAN connection, and watch the meter. If the site total climbs, you have found the thing that would otherwise have found you in February.The row in the table below about per-charger fallback is the one most often missed. Each charger has its own configured maximum current, which is what it falls back to when it loses the local link. That fallback value times the port count is the worst case your service will ever see, so set it to a safe share of the site limit rather than leaving it at the factory default.

The last one is the audit. Everything before it is a demonstration under supervision; the interval data is what the meter actually recorded when nobody was watching.

There is a difference between a convenience setting and a safety limit. A cap that exists so the bill is smaller can be changed by anyone with the app password. A cap that is the reason the service was sized as it was cannot, because changing it makes the installation non-compliant.Bring this up early. An inspector who first meets your energy management system at final inspection is being asked to accept an unfamiliar argument at the worst possible moment. An inspector who was consulted at design stage usually tells you exactly what documentation they want to see.

A yard of identical vans that all leave at 06:00 needs power management, not scheduling, and the difference is a recurring subscription. A yard with three shifts, mixed vehicle types and more vehicles than ports needs both, and the scheduler will pay for itself in ports you did not have to install.Charger-native group control is the right choice for a small single-brand depot where the service has obvious headroom and nothing else on it moves much. It stops being the right choice the moment the group budget has to be set so conservatively that you are wasting capacity you paid for.