Hamburg Alsterdorf Depot: Finding the Real Peak Behind a 127-Bus Fleet

The Problem

Hamburg's Hochbahn transit authority faced the question every large transit agency eventually faces: when 127 buses all return to the depot in the evening and need to recharge before the next morning's service, what does the depot actually draw at its peak, and can the electrical service handle it?

Get the answer wrong in one direction and you've over-built a multi-million-dollar grid connection nobody needed. Get it wrong in the other direction and buses don't leave the depot on time.

The naive answer ("127 buses × 150 kW chargers = 19 MW, so build for 19 MW") is almost never the right one. It assumes every bus plugs in and charges flat-out the instant it arrives. Real depots don't work that way: buses trickle in across the evening, some need more energy than others, and a competent charge management system spreads that load out. The question isn't "what's the nameplate sum"; it's "what does managed charging actually require, and where does it bind?"

How the Depot Digital Twin Was Used

Step 1: Model the real fleet, not a simplification. The full 127-bus fleet was entered with its real battery capacity, dwell windows, and arrival pattern, with each bus assigned its own arrival and departure time across the evening and overnight window, exactly as the depot actually operates. This single step is what separates a useful peak estimate from a back-of-envelope guess: stagger the arrivals, and the peak tells the truth.

Step 2: Let the engine find the binding constraint. Rather than asking "will 20 MW of service be enough," the simulation was run against the depot's actual worst-plausible winter day: full cold-weather derate, every vehicle present, charging managed to stay under the service limit wherever the electrical topology allows it.

Step 3: Read the spare-capacity breakdown. The results page doesn't just report a peak number. It walks the full chain from nameplate service capacity, through any reserved base-load, to the modeled charging peak, to what's left over:

Step 4: Visualize the 24-hour load profile. The interactive timeline chart shows exactly how the 4,538 kW peak builds and recedes across the night, with a scrubbable playhead that reads out net load, headroom, and per-node utilization at any instant, turning a single peak number into a story a utility reviewer or capital committee can actually follow.

Step 5: Check every electrical node, not just the headline number. The per-node view confirms that neither the service entry nor the downstream circuit is breached at any point in the simulated day, with margin. Nothing in the depot's electrical chain is close to its rating.

Step 6: Benchmark against the industry, automatically. The results page surfaces published reference bands from real transit electrification studies alongside the depot's own numbers: in this case, a real-world managed-charging study (Clovis Transit) showing a ~40% peak reduction from smart scheduling, giving the depot operator immediate external context for what "good" looks like.

The Result

The Depot Digital Twin modeled Hamburg Alsterdorf's managed-charging peak at 4,538 kW, landing within striking distance of the peer-reviewed academic benchmark for this exact depot (Jahić, Eskander & Schulz, 2019), which reports a managed peak of 5.47 MW under equivalent winter conditions.

That means a 20 MW service connection carries this fleet with more than 3× headroom to spare, a clear, defensible answer to the only question that actually matters at the capital-planning stage: is the grid connection big enough, and by how much?