Skip to content

Fleet demand forecasting

Fleet demand forecasting is the process of predicting future charging and energy needs for an EV fleet—so you can size depots, plan grid capacity, schedule charging, and control costs while keeping vehicles ready for operations. It estimates how many vehicles will need charging, how much energy (kWh) they’ll require, and when (time-of-day) that demand will occur, across sites and routes.

What is fleet demand forecasting?

In an EV context, “demand” usually means:
Energy demand: total kWh needed per day/week/month (by depot/site)
Power demand: peak kW required at the site (drives connection size and demand charges)
Charger demand: how many charge points/bays are needed to avoid congestion
Time demand: the charging window shape (arrival/departure waves, dwell time)

Forecasting can be short-term (next week/month) for operations, or long-term (6–36 months) for infrastructure expansion planning.

Why it matters

– Prevents under-sizing (missed departures, queues, public charging fallback)
– Prevents over-sizing (wasted CAPEX, expensive grid upgrades)
– Supports phased rollout plans (add bays/power at predictable milestones)
– Improves tariff and demand-charge strategy (peak control)
– Provides evidence for business cases, grants, and internal approvals

What to forecast

By site/depot
– Vehicles assigned now and planned growth
– Daily energy (kWh/day) and seasonal variation
– Peak load (kW) by hour, with and without smart scheduling
– Charger utilization and congestion risk (bays needed)

By fleet segment
– Vehicle class (cars/vans/trucks/buses) with different kWh/km and duty cycles
– Route types (urban stop-start vs highway)
– Shift patterns (single vs multi-shift)
– Public charging reliance vs depot charging share

Core inputs (practical minimum)

– Vehicle count by site and expected growth timeline
– Mileage or duty cycle per vehicle (km/day, route length, operating hours)
– Energy consumption rate (kWh/km) by vehicle class and season
– Arrival and departure times (or shift schedules)
– Charging power options (AC kW per bay) and dwell time availability
– Site constraints: available connection capacity and hard caps
– Tariff structure: TOU windows, demand charge rules

Common forecasting methods

Bottom-up (vehicle-based) — most accurate
– For each vehicle: forecast km → kWh needed → allocate to site/time window
– Produces: kWh/day and hourly load shape

Top-down (site-based) — fast for early planning
– Use historical fuel/telematics totals to estimate total energy needs after electrification
– Apply assumptions for charging split (depot vs public) and efficiency

Scenario-based planning — best for scaling
– Build cases: conservative / expected / aggressive EV adoption
– Include “grid delay” and “vehicle growth” risks
– Attach trigger points: when utilization hits X%, add Y chargers or increase capacity

Peak modeling (demand-charge focused)
– Simulate worst-case arrivals and simultaneous charging
– Compare unmanaged vs managed charging to quantify peak reduction value

What good outputs look like

– kWh/day and kWh/month per site (baseline + growth curve)
– Hourly load profile (unmanaged vs managed)
– Required site capacity (kW) and recommended phase plan
– Chargers needed now vs later (with utilization targets)
– Risk flags: queue risk, peak risk, grid constraint risk
– Assumptions log (vehicle efficiency, temperature factor, charging losses)

Best practices

– Separate energy (kWh) from power (kW): they drive different decisions
– Model both unmanaged and managed charging—this is where savings and grid-fit appear
– Use real arrival/departure patterns (even rough shift windows helps a lot)
– Add seasonality for heating/cooling and payload effects (especially vans/trucks)
– Forecast in phases: “add bays” and “add power” are different projects
– Track actuals monthly and recalibrate assumptions

Common mistakes

– Sizing based on total kWh only, ignoring peak kW and time windows
– Assuming every vehicle charges every night at max power
– Ignoring multi-shift operations (daytime demand appears unexpectedly)
– Not planning for downtime/maintenance bays and blocked parking realities
– No scenario planning → infrastructure lags behind fleet growth

Fleet charging scheduling
Depot power management
Demand charges
Charging capacity planning
Dynamic load management
Charger utilization rate