An electric bus can be fully charged and still not be ready for service. It may be waiting for a cleaning bay, a manual state-of-charge check, a release decision or an available path out of the depot. That is why depot performance is a flow problem, not simply a question of how many chargers are installed.
Value Stream Mapping (VSM) makes that flow visible. It maps the steps, information, processing time, waiting and work in process between a defined starting point and the customer outcome. In an EV depot, the outcome is not “a charging session completed”; it is the right vehicle, safely prepared and ready for its next shift on time.
The worked example below uses an illustrative 40-bus depot. Its figures are planning assumptions, not industry benchmarks, and should be replaced with time-stamped local data before making investment decisions.
Define the value stream: return to shift-ready release
Set the boundary from vehicle return at the end of its shift to release for the next shift. Include the physical path (yard movement, charging, inspection, cleaning and staging) and the information that triggers those steps, such as route assignments, target SoC, charger status and defect notifications.
Overnight dwell time is the constraint because each bus has a limited window between returning and needing to depart again. The depot must fit charging and other readiness work into that window, while working around charger capacity, arrival variation and departure priorities. A bus arriving earlier has more available dwell; one arriving late, with low SoC or a charging fault, has less recovery time.
For this example, 40 buses return between 17:30 and 21:30. Twelve chargers deliver an average 50 kW, and an average charging session lasts 3.5 hours. The depot needs all 40 buses ready by 05:00. The U.S. Federal Transit Administration’s best-practices report is a useful reference when considering the wider reliability and maintenance context.
Current state: a charger queue and hidden release delays
For planning, assume the 40-bus cohort has a 14.8-hour average return-to-release lead time. The map below separates process time (PT) from lead time (LT). WIP is an illustrative typical count at each step; values are not additive because the same buses move through the stream.
| Current-state step | PT per bus | LT per bus | Typical WIP |
|---|---|---|---|
| Return check-in | 5 min | 5 min | 1–2 buses |
| Wait for a free charger | 0 min | 4 hr | 6 buses average in queue |
| Connection and handshake | 8 min | 8 min | 1 bus |
| Charge session, 50 kW average | 3.5 hr | 3.5 hr | Up to 12 buses |
| Manual SoC verification | 6 min | 6 min | 1–2 buses |
| Defect-tag walkaround | 3 min average* | 3 min | 1 bus |
| Wait for cleaning bay | 0 min | 45 min | About 3 buses |
| Cleaning | 25 min | 25 min | 1 bus |
| Release checklist | 10 min | 10 min | About 8 buses awaiting handoff |
| Parking repositioning | 12 min | 12 min | 1–2 buses |
| Handoff and departure-sequence wait | 0 min | 5 hr 24 min | Variable staging WIP |
| Total | 4 hr 39 min (≈4.7 hr) | 14 hr 48 min | – |
*A five-minute walkaround is completed on 60% of vehicles, equivalent to three minutes per bus across the full cohort.
On this basis, process cycle efficiency (PCE) is about 31.8%: 4.7 hours of process time divided by 14.8 hours of total lead time. The example also shows why maps must include the whole release process: the assumed 5.4-hour handoff and staging delay is substantial, even though the charger queue is the primary resource constraint.
The capacity check reinforces the concern. At 3.5 hours per session, 12 chargers can provide about 42 charger-hours in an 8.5-hour window: before allowing for access delays, interruptions or uneven arrivals. Meanwhile, the planning requirement is a steady release cadence: 510 available minutes ÷ 40 buses = 12.75 minutes per bus. Treat this takt time as the rhythm required to make dispatch demand, not as a claim that an individual bus completes every depot task in 12.75 minutes.
Find the eight wastes in the depot flow
Use DOWNTIME to prompt observation, not to label necessary safety or compliance work as waste without checking its purpose:
- Waiting: buses queue for chargers, cleaning slots or release handoffs.
- Defects: failed charger handshakes can leave a bus short of its required charge.
- Inventory: charged-but-unreleased buses occupy space and conceal readiness problems.
- Transportation: repeated repositioning shuttles buses between charging, cleaning and staging.
- Motion: staff walk to distant bays to inspect, tag or verify vehicles.
- Overprocessing: manual SoC checks duplicate information already available through reliable telemetry.
- Non-utilised talent: drivers or depot staff wait for the next task instead of completing suitable pre-trip preparation.
- Overproduction: charging beyond the next run’s requirement adds energy cost without necessarily improving readiness.
Build a future state around departure need
A future-state map should coordinate vehicle return, charger access and departure readiness rather than maximize charging in isolation. The design below assumes the depot tests these changes and validates the resulting times:
- Level return appointments across available capacity; assign a first-in-first-out charge lane while preserving an exception rule for buses with urgent departures or low SoC.
- Set a target SoC for the next route. In this example, assume the average planned charge falls from 3.5 hours to 50 minutes because buses receive only the energy needed for their next scheduled work. Confirm this using actual route energy data before implementation.
- Perform the defect walkaround during charging, with safe access and a clear process for tagging and escalating faults.
- Clean in parallel with charging where layout, safety and staffing allow.
- Use reliable SoC telemetry in place of duplicate manual verification, with a defined manual fallback for data or communication faults.
- Standardise the release checklist and handoff, separating “ready,” “maintenance hold” and “charger fault” statuses.
With parallel work, the future-state estimate is 7.9 hours from return to release. For a consistent comparison, its 1.35-hour PT represents the sequential critical-path processing time: check-in, connection, assumed 50-minute target charge, telemetry confirmation, release and repositioning. Walkaround and cleaning still require labour, but they overlap with charging and are not counted twice in elapsed critical-path time. PCE is therefore 1.35 ÷ 7.9 = 17.1% on this stated basis. The lower percentage does not mean the future state is worse: its lead time is shorter because unnecessary waiting and idle time are reduced.
Compare the illustrative current and future states
These figures are a pilot baseline and target scenario, not promised results. In particular, the target of 38 buses ready by 05:00 depends on confirming route energy needs, charger reliability, staffing and yard access.
| Measure | Current-state example | Future-state target |
|---|---|---|
| Return-to-release lead time | 14.8 hr | 7.9 hr |
| Sequential critical-path PT | 4.7 hr | 1.35 hr |
| PCE on stated basis | 31.8% | 17.1% |
| Average charger queue | 6 buses | 1–2 buses |
| Missed-charge rate | 30% | 5% |
| Buses ready by 05:00 | 28 of 40 | 38 of 40 |
Use a process cycle efficiency calculator to check the arithmetic, and adapt a formal project-scope boundary approach so the team agrees on what “ready” means. Record actual arrival and release timestamps, SoC, charge duration, queue length, faults, cleaning time and on-time readiness. Stratify results by route, day and charger type before attributing improvement to a countermeasure.
Sequence kaizen over 90 days
| Timing | Improvement focus | Checks before advancing |
|---|---|---|
| Days 1–30 | Level return schedules and pilot SoC-targeted charging. | Track queue length, charge target attainment, energy use and buses ready by 05:00. |
| Days 31–60 | Parallelise cleaning and conduct the defect walkaround during charging. | Confirm safe access, staffing coverage, fault escalation and cleaning quality. |
| Days 61–90 | Introduce telemetry-based verification and standard release-checklist work. | Audit data accuracy, fallback steps, release ownership and on-time dispatch. |
Review the map with drivers, technicians, cleaners, dispatchers and energy managers. Their observations help distinguish a true charger constraint from access, staffing, maintenance or information delays. For further context on planning and operating charging systems, consult the National Renewable Energy Laboratory’s transit user group resources.
Turn the map into a repeatable improvement skill
A well-built VSM gives the depot team a shared view of the constraint, a measurable future state and a sequence of experiments to test. Lean Six Sigma training can help professionals strengthen those skills, from defining a project and measuring process performance to analysing causes and sustaining results.
Explore CSSC-accredited, self-paced Lean Six Sigma certification with Lean 6 Sigma Hub, from White Belt to Master Black Belt. Learn by doing through real DMAIC case studies, worked examples and practical simulations.
Kaizen. Kai-Care. Kai-Done. Lean Six Sigma.
