Istanbul PV Depot — Cutting a Charging Peak 40% Without Touching a Single Charger

The Problem

Istanbul's second-largest transit depot faced the electrification question every large operator eventually faces at scale: with 100 buses on-site and a 4,000 kW electrical service, what does the fleet's charging pattern actually cost every month — and is there a cheaper way to run the same fleet without buying more equipment?

The naive answer is to just build more service capacity. The better question — the one this depot's planners actually needed answered — is whether the scheduling of when buses charge, not the size of the connection, is the real lever worth pulling first.

How the Depot Digital Twin Was Used

Step 1 — Model the full depot, including solar. The 100-bus fleet was built against 150 kW chargers and a 4,000 kW service, with 2,333 kW of on-site solar PV added as a generation asset — reflecting the depot's actual planned photovoltaic installation.

Step 2 — Establish the baseline peak. Running the worst-plausible day produced a clear starting point:

Step 3 — Turn the peak into a dollar figure. The Demand-Charge Exposure module priced that 3,750 kW peak against an entered utility rate: a directional ~$450,000 per year — the number that turns an electrical engineering result into a business case a CFO can act on.

Step 4 — Run the charge-schedule optimizer. This is where the twin stopped being a diagnostic tool and became a design tool. Rather than simply reporting the peak, the optimizer searched for a per-vehicle charge schedule — spreading each bus's charging window across its full parked time instead of letting every bus draw power the instant it arrives — and re-simulated the result through the same physics engine to measure the real outcome, not a theoretical one.

Step 5 — Read the result, with the trade-off shown alongside it.

A 40% cut in coincident peak, achieved purely by timing — no new chargers, no larger service connection, no capital spend. At the same $10/kW-month rate, that peak reduction is worth roughly $180,000/year in avoided demand charges.

Step 6 — Adopt it with one click, or hold it for review. The proposed schedule can be written directly into the depot's vehicle charge windows in a single, reversible action — turning a modeled proposal into an operating plan the moment the fleet team signs off.

The Result

The Depot Digital Twin found that Istanbul's depot could cut its coincident charging peak by 40% — from 3,750 kW to 2,250 kW — purely through smarter charge timing, at zero infrastructure cost. That reduction is consistent with the direction and scale of managed-charging results reported in the published techno-economic study of this depot (Duman, Yılmaz & colleagues, 2025), which found PV-assisted charging management could cut the depot's 25-year net present cost by more than half when combined with charger-sharing strategies.