EV Battery Return Logistics: A Case Study

1. The Reverse Logistics Problem

Forward distribution of EV batteries is a well-behaved logistics problem. Packs leave a small number of factories in known condition, at a controlled state of charge, in standardised packaging, on planned routes, in predictable volumes. Everything that makes a supply chain easy to design is present: uniformity, forecastability and a single point of origin.

Battery returns invert every one of those properties. A returns flow is not distribution run backwards; it is a structurally different problem, and treating it as a mirror image of outbound logistics is the single most common planning error we encounter.

Four characteristics make the difference:

• Unknown condition on collection. An outbound pack has a build record and a factory test result. A returned pack has a symptom report written by a workshop technician under time pressure. It may be fully functional, latently defective, mechanically deformed or already venting. The condition has to be established, not assumed.
• Mixed chemistries and formats. A single manufacturer's installed fleet spans several model generations, several cell chemistries (typically NMC and LFP) and several pack architectures. A return network receives all of them simultaneously, which means packaging cannot be specified once and forgotten.
• Dispersed origin points. Outbound flows originate at a handful of plants. Returns originate at every dealership, independent workshop, accident-repair centre and dismantler in the territory, each generating a low and irregular volume. Consolidation is the whole game.
• Regulatory classification burden. Under ADR, a battery's condition determines its classification, its packing instruction, its documentation and in some cases whether it may move at all. That determination must be made by a competent person, recorded, and defensible months later.

There is a second asymmetry that planners underestimate: time. An outbound pack moves on a schedule. A returned pack sits. It waits for a warranty decision, for a claims adjuster, for a consolidation load to fill, for a recycler's intake slot. Storage duration at the edge of a return network is routinely measured in weeks, and a pack of unknown condition sitting in a dealer's workshop for three weeks is a storage safety problem as much as a transport problem. Packaging that is certified only for carriage, and not for static storage, leaves that interval uncovered.

2. The Scenario: A European EV OEM

The operator in this scenario is a European EV manufacturer. It is not a real named company, and the numbers attached to it are modelled rather than measured. What makes it useful is that its network shape is typical: most manufacturers with a mature European fleet converge on something very close to this structure.

2.1 Network shape

The operator sells through a franchised dealer network and supports an authorised workshop network alongside it. Returns are generated at several hundred customer-facing points spread across a dozen countries. Those points feed six regional consolidation sites, one per major market cluster, which in turn feed a single central assessment hub. The hub performs detailed diagnosis, decides the disposition of each pack (repair, remanufacture, second life or recycling) and dispatches onward to the appropriate destination.

This hub-and-spoke shape exists for a reason. Detailed battery diagnostics require equipment, trained staff and a safe working environment that cannot be economically replicated at several hundred sites. Centralising the diagnosis is correct. The consequence is that every returned pack must travel, in an incompletely characterised state, from a site with little expertise to a site with a lot of it.

2.2 Pack population

For the purposes of this scenario, the fleet is assumed to produce returns dominated by a 90 kWh traction pack, with smaller 50 kWh units from an earlier compact model and a small number of 100 kWh performance packs. Module-level returns, where a workshop has removed individual modules rather than the complete pack, add a further format. This spread is deliberate: it reflects the reality that no single container size covers a real return network.

2.3 Return drivers

Table figures are illustrative shares for a representative network, not measured volumes from any specific operation.

The mix matters more than the totals. Accident-damaged packs represent a minority of units but the overwhelming majority of the risk, and they arrive least predictably. A network designed around the average consignment will be wrong exactly when it matters.

3. The Classification Bottleneck

Everything downstream depends on one decision made early and often at the least capable point in the network: is this pack stable, damaged, or critical?

Under ADR, lithium batteries identified as damaged or defective fall under Special Provision 376. SP 376 draws a distinction that governs the entire returns flow. Batteries that are damaged or defective but not liable to rapid disassembly may be carried under P908 (for cells and batteries) or LP904 (for large packagings). Batteries assessed as liable to rapidly disassemble, react dangerously, produce a flame or a dangerous evolution of heat or a dangerous emission of toxic, corrosive or flammable gases must be carried under the higher containment tier: P911 or, for large packagings, LP906.

3.1 Why this is a bottleneck

The assessment is not a formality. It requires a competent person to consider the visible damage, the failure mode reported, the age and chemistry of the pack, any diagnostic data recoverable from the battery management system, and whether the pack has shown heat, swelling, odour or gas. In a dealer workshop, several of those inputs are unavailable. The safe response to uncertainty is to classify upward, but classifying upward is only useful if the higher tier of packaging is actually obtainable within a working timescale.

That is the bottleneck in practice. If LP906-capable packaging is not on site, an uncertain pack either waits, sometimes for weeks, or is moved under a classification the operator cannot fully defend. Neither outcome is acceptable, and both are common where packaging is procured reactively.

3.2 Assessment-to-dispatch workflow

4. The Old Approach and What It Cost

Before restructuring, the operator in this scenario handled critical and uncertain packs the way most networks did: filler-based packaging. A steel or composite box, the pack lowered in, and the void space filled with dry sand or vermiculite to provide thermal mass and separation.

Filler works, after a fashion, as a heat sink. The problem is that it is an operational liability at every stage of the process except the one it was designed for.

There is a compounding effect worth naming. Because the filler process is slow and unpleasant, sites avoid it. Avoidance shows up as packs classified downward to escape the higher-tier process, as units left unpackaged in a corner of the workshop awaiting a decision, and as consignments dispatched with documentation that does not quite match the physical reality. The packaging choice quietly becomes a compliance risk, and it does so through human behaviour rather than through any failure of the material itself.

5. The Containment Approach

The restructuring in this scenario rests on three decisions: standardise on a single certified containment architecture across the whole network, size it by pack energy content rather than by box volume, and eliminate filler from the process entirely.

5.1 One certification, three modes, plus storage

BUNCKER® units are certified to UN LP906, the highest packing-instruction tier for large packagings, covering all lithium battery types including stable, unstable, damaged and end-of-life. The same certification is valid for road under ADR, sea under IMDG and rail under RID, and additionally for static storage indoors and outdoors.

For a returns network this single fact removes a large amount of design complexity. A pack classified as critical at a workshop in one country can be sealed once, stored at the edge site for as long as consolidation requires, moved by road to a regional site, transferred to rail or short-sea for the trunk leg, and arrive at the central hub without ever being repacked or reclassified for a mode change. The chain of custody is unbroken because the packaging did not change hands in an uncontrolled state.

5.2 Matching models to the pack population

The BUNCKER® Pro Line Heavy range is indexed by energy content in kWh, which is the correct specification variable: containment must be dimensioned against the total energy that could be released in a full runaway, not against the physical size of the object.

In this scenario the network standardises on Model M at the dealer and workshop tier, because that is where the 90 kWh pack dominates, holds Model SDR at sites constrained by lifting height or door access, and concentrates Model XL and Model XXL at the regional and central sites where the unusual formats accumulate. Two sizes cover the long tail; one size covers the volume.

5.3 Why filler-free changes the operation

BUNCKER® units require no sand, vermiculite or filler material. The patented bell jar architecture performs the thermal work structurally: the pack sits on the pallet base, the containment body lowers over it and seals, and the multi-layer construction keeps the outer surface below 100 °C even during a full thermal runaway event. The gas hazard is handled chemically by the integrated Fibercon Altilium® CP Blanket (FACB), which neutralises hydrogen fluoride in the vent stream rather than simply releasing it.

The operational consequences follow directly. Loading becomes a lifting operation rather than a materials-handling operation. There is no granular material to store, keep dry, move or dispose of. The pack remains accessible for inspection up to the moment of sealing, and the unit can be opened and re-inspected under a controlled procedure if a borderline consignment sits in interim storage longer than planned. Nothing has to be emptied at the receiving end before the pack can be assessed. On top of this it remains possible to add another battery or battery pack up to the moment of sealing and transport, which is not possible in case granulate was added and first needs to be removed.

Featured product

BUNCKER® Pro Line Heavy M: 90 kWh Containment

The volume workhorse for EV return networks built around a 90 kWh traction pack. Pallet Loading System, patented bell jar architecture and integrated Fibercon Altilium® CP Blanket HF neutralisation. UN LP906 certified for all lithium battery types including damaged and end-of-life, valid for ADR, IMDG and RID plus static storage indoors and outdoors. No sand, vermiculite or filler material required.

View Model M →

6. Operational Outcomes in the Scenario

The figures in this section are illustrative. They are modelled to show the shape and direction of the change that a filler-free, single-certification approach produces in a network of this type. They are not measurements taken from a named customer deployment, and they should not be quoted as such.

Illustrative scenario indicators. Contact the BUNCKER® team for verified data relating to a specific operation.

Four qualitative changes are more important than any of the numbers, and they are the ones that transfer reliably to other networks:

• Loading becomes a repeatable task at low-capability sites. When the process is place, lower, seal, label, a dealer workshop can execute it correctly with brief training. When the process involves several hundred kilograms of granular material, it cannot. Compliance at the edge of a network is a function of how easy the compliant path is.
• Mode flexibility becomes real rather than theoretical. A consignment sealed under a certification valid for ADR, IMDG and RID can be rerouted mid-journey without repacking. In a network consolidating irregular low volumes across a dozen countries, the ability to put a unit on whatever trunk leg is available next is worth more than any single-mode optimisation.
• Interim storage stops being a regulatory gap. Because the same units are certified for static storage indoors and outdoors, a pack awaiting a warranty decision is in approved containment for the whole waiting period, not merely for the hours it spends in motion.
• Uncertain packs can be classified upward without penalty. This is the most significant behavioural change. When the higher containment tier is on site and quick to use, an assessor faced with an ambiguous pack takes the cautious decision. When it is slow, scarce or unpleasant, the incentives push the other way.

7. Compliance and the Audit Trail

A return network is audited on its records, not on its intentions. In this scenario the documentation design is treated as a first-class part of the logistics design rather than as paperwork bolted on afterwards.

7.1 What has to be defensible

For every returned pack, the operator should be able to reconstruct, at any later date: who assessed it, when, on what evidence, what classification they assigned, which packing instruction that triggered, which certified packaging was used and what its certification status was, where the unit travelled, and where the pack finally went. Each link in that chain is individually unremarkable. The failure mode is that one link is held in a different system from the rest and cannot be joined up under audit.

7.2 The EU Battery Regulation dimension

The EU Battery Regulation adds a reporting and traceability layer on top of the transport rules, with obligations around collection, treatment and recycling of industrial and EV batteries, producer responsibility, and the progressive introduction of a digital battery passport for EV and larger industrial batteries. The practical implication for a returns network is that the pack-level identity established at the point of return is the same identity the regulation expects to see carried through to treatment.

That argues strongly for one design choice: bind the record to the pack serial number at intake and keep every subsequent record attached to it, including the identity of the containment unit it travelled in. Operators who instead key their records to a consignment number, a job number or a transport order find that the thread breaks the first time a consignment is split or consolidated.

It is worth noting how much of this is made easier simply by not repacking. Every repack is a point at which the physical consignment and the paper record can diverge. A single sealing event at origin, under a certification valid for every mode the unit will subsequently travel on, removes most of the opportunities for divergence before any information system is involved.

8. Lessons and a Transferable Checklist

The specifics of this scenario are illustrative. The structural lessons are not, and they apply to any operator designing a battery return network: manufacturers, fleet operators, leasing companies, dismantlers and recyclers alike.

8.1 Five lessons

• Design for the worst consignment, not the average one. Accident-damaged and critical packs are a minority of units and the majority of the risk. A network optimised for the warranty return will be improvising when the difficult unit arrives.
• Put the highest tier of packaging at the edge. Classification happens where expertise is thinnest. If the cautious decision is also the slow and inconvenient one, it will not be taken consistently.
• Specify containment by energy content. A 90 kWh pack needs containment rated against a 90 kWh runaway event. Volume is a fitting constraint, not a safety parameter.
• Count the interim storage interval. Packs wait longer than they travel. Packaging certified only for carriage leaves the longest part of the timeline uncovered.
• Treat repacking as a defect. Every mode-driven repack costs labour, creates a handling risk with an uncertain pack, and opens a gap in the audit trail. One certification covering road, sea and rail removes the need.

8.2 Network design checklist

1. Map your actual return drivers and their condition profiles. Separate warranty, accident, end-of-life and campaign volumes. They behave differently and they need different planning assumptions.
2. Profile your pack population by kWh, not by model name. Include module-level returns. This is what determines your containment model mix.
3. Identify every point where a classification decision is made. Confirm that a competent person is designated at each, and that they have a written decision procedure covering the SP 376 criteria.
4. Establish how BMS data will be captured at sites where the pack is still readable, and what the assessment procedure is when it is not.
5. Position LP906-capable containment at every site that can generate a critical pack, not only at the hub. Availability at the edge is what makes upward classification practical.
6. Check that your packaging approval covers static storage, indoors and outdoors, for the full interim period, and not only the transport leg.
7. Eliminate filler from the process to remove the handling, mass, inspection and waste-disposal penalties, and to make the compliant path the fast path.
8. Bind all records to the pack serial number from intake onward, including the identity of the containment unit used.
9. Track the inspection and service status of your containment fleet as an asset register, so that certification validity on any given date is provable.
10. Plan for surface modes only where damaged and defective batteries are involved, since carriage by air is forbidden for that category.
11. Test the network with a scenario exercise before you need it: an accident-damaged pack arriving at your least-equipped site on a Friday afternoon is the case that reveals whether the design works.