EV Fleet Charging Management with IoT

Fleet electrification arrives with an electrical reality that most depots did not plan for: a few dozen vehicles plugged in simultaneously can demand more power than the site’s transformer was sized to deliver, and the charging bill — billed on both energy and peak demand — can become the largest single line on the depot’s utility statement. The difference between an EV fleet that is an operating advantage and one that is a capacity nightmare is charging management: scheduling and throttling chargers against the site’s real electrical limits, the vehicles’ actual needs, and the tariff’s price structure, using data from the chargers, the vehicles, and the site meters.

The Depot Capacity Problem

A delivery fleet of forty vans, each drawing up to 11 kW on a Level 2 charger, needs 440 kW if every charger runs flat out. But the depot’s transformer and service were sized for the old building load plus a parking lot — often 300–500 kW total, sometimes less. The naive behavior — plug in at the end of shift, charge at full rate — hits the transformer limit, trips the main breaker, or triggers a demand spike that doubles the monthly bill. The management task is fundamentally one of load balancing under a hard capacity constraint: 440 kW of possible draw must be scheduled into a 300 kW envelope while still getting every vehicle to its departure time with enough range.

What the Management Layer Controls

An EV fleet charging management system (CMS) coordinates three things in real time:

Charging current per charger. The standard communication is OCPP (Open Charge Point Protocol), which lets a controller read each charger’s status and command its current or start/stop state. A 32 A charger can be set to 8, 16, or 24 A to fit the remaining budget.
Vehicle departure needs. Integrating with the telematics or the drivers’ schedules gives each vehicle a required departure time and a target state of charge; a van leaving at 6 a.m. for a 200-mile route has a different priority than one not leaving until Tuesday.
Site electrical state. The main meter and transformer load feed the capacity budget; the controller subtracts current site load (building, compressors, lighting) from the transformer rating to find the headroom available to charging at each instant.

The algorithm that binds these together is a scheduling optimizer run on a rolling horizon: every few minutes it re-plans the charging order and power allocation for the next several hours, respecting the capacity envelope, the tariff price structure, and each vehicle’s deadline.

Rate Optimization: Buying Power at the Cheap Hour

The tariff structure is the second lever. Time-of-use rates can price off-peak power at a third of the on-peak price, and a fleet that charges intelligently moves the bulk of its energy into the cheap window:

Off-peak charging. Vehicles with slack until morning are scheduled into the cheapest hours, typically after midnight.
Peak avoidance. The controller refuses (or heavily throttles) charging during the site’s on-peak window unless a vehicle genuinely cannot wait.
Demand-cap management. The CMS holds the site’s total demand — not just the chargers’ — under the tariff’s demand threshold, because the 15-minute peak that sets the monthly demand charge includes the building and the chargers together.

A realistic example: a depot with a $12/kW demand charge and a 4-cent/kWh off-peak spread saves $3,600 a month by holding the demand peak down (300 kW × $12) and additional thousands by shifting energy into the cheap window. Over a year, charging management routinely delivers more savings than the cost of the chargers themselves.

The Scheduling Math in Practice

Consider a fleet of twenty vans, each needing roughly 40 kWh of charge, with chargers capable of 11 kW, a site capacity of 250 kW, and the existing building drawing a steady 120 kW during the evening — leaving about 130 kW for charging. If all twenty vans plug in at 6 p.m. and demand 220 kW, the naive approach trips the transformer. The optimizer instead does this:

Prioritizes the vans with the earliest departures and the largest range needs.
Allocates the 130 kW budget so that, say, six vans charge at full 11 kW, eight charge at 6 kW, and six wait — the “fastest first, fill the rest” allocation.
Swaps the allocations as vans complete or as the building load drops after closing, gradually raising charging headroom.
Ensures every van reaches its departure target, verified against the telematics read of battery state rather than assumed from plug-in time.

The visible outcome is a transformer that never trips, a peak that stays under the tariff threshold, and a yard where every vehicle leaves on time with the range it needs. When a breakdown or a late return changes the picture, the controller re-plans within minutes.

Vehicle-to-Grid and the Grid-Services Option

Once charging is managed, the same infrastructure can run in reverse. V2G-capable vehicles can discharge to the depot during peak demand or to the utility during demand-response events, using the same OCPP-based control path. The economics are attractive in markets that pay for capacity: a 100 kWh fleet battery can carry a depot through a 40-minute demand event that would otherwise set the peak demand charge, and in regions with ancillary-service markets the same battery can earn revenue that exceeds the value of the energy it holds. The management system must handle the two-way flow the same way it handles charging: track SoC on both sides, respect the driver’s minimum departure state of charge as a hard constraint, and verify the net energy and revenue through the site meters.

The Data That Makes It Work

The integration layer is modest but critical:

Charger telemetry (OCPP) — session start/end, delivered kWh, commanded current, faults. This is the control surface and the billing record.
Vehicle state (telematics or OCPP-adjacent APIs) — SoC, estimated range, location, departure schedule. Without real SoC, the optimizer is scheduling blind against a guess.
Site meter data — the transformer headroom computation needs live total load, typically from the main meter over Modbus or the utility interval feed.
The tariff and event stream — price schedules and any demand-response signals the depot is participating in.

A shared timestamp discipline across these three sources is what lets the system answer the question every depot manager actually cares about: “which vehicle is charged enough, at what cost, by when?”

Deployment Realities

Start with a hard capacity budget from the actual transformer rating and main breaker, not the nameplate of the chargers. The optimizer is only as good as the constraint it respects.
Instrument the main meter first. Charging management without live site load is guessing; the whole value depends on knowing the headroom.
Phase the rollout. Prove load balancing on a handful of chargers and two shift patterns before rolling out across the yard.
Plan for the departure constraint. The single most common failure is a scheduler that saves money but leaves a vehicle short of range; every schedule must verify against SoC at departure, every day.
Keep a manual override for the dispatcher who has to run a special load tomorrow.

The Operational Result

Managed charging turns the depot’s electrical capacity from a constraint into a predictable budget. The fleet gets its range at the lowest possible energy price and without tripping the transformer, the site’s demand charge stays under control, and the same control path positions the depot for grid services as V2G matures. The numbers — kWh shifted, peak held, dollars saved — are all measurable from the charger, vehicle, and meter data the system already collects. Integrar IoT’s platform integrates OCPP charger telemetry, vehicle state, and site metering into a single scheduling view, applies the tariff and capacity constraints automatically, and verifies the delivered energy and cost against the meter record.