Depot sizing, measured
How much charging does the fleet actually need?
Every fleet is asking it. Almost none can answer it. DeltaSoC measures the answer from the fleet's own operation - vehicle by vehicle, across seasons, for trucks that have never been electric.
Daily demand across a recorded fleet. The connection, the charger count and the storage are all decided by the right-hand tail - and the only way to find that tail is to measure every day rather than a representative one.
The problem
The sizing is an estimate. The estimate gets capitalized.
A depot's grid connection, its charger count and its storage are all sized before anyone knows what the fleet will actually draw. The number used is an estimate, built from vehicle class and a rule of thumb.
Then it gets built, and it gets capitalized. The fleet pays for it in every kWh for the life of the site.
An overbuild is not a one-time cost. It is a permanent one, built on a guess. An underbuild is a truck that does not leave in the morning.
The real answer is already in the fleet's own operation. It has simply never been measured.
What it measures
Movement in. Energy at the battery out.
DeltaSoC takes how a vehicle actually moved - speed, distance, load, ambient temperature - and returns the energy a battery truck would have drawn doing the same work.
It needs no access to a battery management system, no manufacturer permission, and no electric vehicle in the fleet. A diesel truck running its normal route is enough to answer whether it could run electric, on which days, and what the depot would have to deliver overnight to make that true.
The inputs are ordinary, and in most fleets they already exist - in a telematics platform, in an OEM's connected-vehicle data, in the operator's own system, or in a device fitted for the purpose. DeltaSoC reads whichever is there. The engine is source-agnostic by design, because a mixed fleet is the normal case and no single source covers one.
In, per vehicle
- Pseudonymous identifier
- Speed and distance over time
- Ambient temperature
- Axle load or gross weight, where the vehicle reports it
Out, per vehicle
- Energy per trip and per day
- Wh per km
- Share of the usable battery window
- Whether one overnight charge covers the day
- The same again under the hardest day, not the typical one
Out, per depot
- Load across the night
- Connection size required
- Charger count
- Cost to run under the site's own tariff
- Which vehicles can convert now, which are borderline, which cannot yet
Evidence
Measured operation, not a simulation.
The engine runs on continuously recorded commercial operation - real vehicles, real routes, real weather. Not a synthetic duty cycle, and not a representative week.
Every day, not a sample
Nothing is selected out, so the hard days are still in the data. A representative week is precisely the thing that hides them.
Per trip, not per month
Each trip is priced on its own, so the daily and seasonal spread survives rather than being averaged into a number that sizes nothing.
Checked against held-out trips
Accuracy is tested by removing a trip, predicting it, and comparing against what was recorded - not by fitting the same data twice.
What is measured, and what is not
Energy is computed from physics rather than metered at a battery, because there is no battery in the source data. Where an operator already runs electric vehicles, those become the ground truth and the model calibrates against them. That belongs here rather than being discovered later.
See it running
One recorded weekday at a real four-truck depot, anonymized. Energy per trip, the load across the night, which trucks the day could not have been run on, and the connection it all needs. Including the trace the energy was computed from.
Specifics of the dataset behind any given assessment come with that assessment.
Where to start
Two ways in.
If you operate a depot
Size the site on what your own trucks actually do, before you sign anything with the utility.
- Which vehicles can convert now, and which cannot yet
- Connection size, charger count and storage, from measured demand
- Cost to run under your own tariff
- Then operate it so no truck leaves short and none is overcharged
If you build platforms or fund infrastructure
The layer is source-agnostic. It runs on the operating data a telematics or OEM platform already carries, and returns the specification to the customer where they already are.
- Integrates against an existing platform rather than replacing one
- Converts a conventional fleet into a specified electric one
- For an investor, the sizing case is underwritten by measurement rather than estimate
- Scoped as a window, with no permanent dependency
Who is behind it
Nils Dullum
Seventeen years in charging infrastructure and energy, on the operator side, the OEM side and the interoperability side. DeltaSoC is the layer that kept being missing from every one of them.
Based in Portland, Oregon, working across the United States and Europe.
ChargeAdvice
Founder and CEO. A wide range of international clients.
Volvo Energy
eMSP build across the EU and the US. BESS launch and partnerships.
IONITY
Commercial network build.
Hubject
Co-CEO.
CleanCharge Solutions
Founder and CEO.
Get in touch
Tell me what you are sizing.
A reply within two business days. If there is a depot, a platform or a portfolio behind the question, say so - it changes the answer.