Build Home Batteries Like We Build Cars

How similar are automotive and home energy battery assembly lines?

For years we have been building battery module and inverter assembly lines for the automotive industry. Cell sorting and grading, busbar laying, laser and ultrasonic welding, thermal interface material application, insulation resistance and hipot testing, terminal torque, end of line functional test.

Increasingly, we are being asked to build the same lines for home energy storage. And the conversation is uncanny, because it is the same conversation.

Same cell chemistry. Same welded busbar. Same insulation resistance test. Same torque on the same terminal, with the same consequence if it is wrong.

Two industries, working on opposite sides of the world for entirely different customers, have arrived at nearly the same product. The car got a battery because we wanted to stop burning fuel. The house got a battery because the grid became unreliable, or expensive, or both. In South Africa we did not need a white paper to explain that second one.

What is the real difference between an EV pack and a home battery?

Ask most people what separates an electric vehicle pack from a home storage pack and they will talk about energy density, cooling, or vibration. Those are real engineering differences and they matter to the design.

But the difference that should keep a manufacturing engineer awake is simpler than that.

A car gets serviced. It goes into a workshop at intervals, a technician looks at it, faults get logged against the vehicle identification number, and eventually the whole thing is scrapped through a controlled channel with a paper trail.

A home battery is installed once, in a garage or a utility cupboard, usually under somebody’s bedroom, and then nobody looks at it again for fifteen years. There is no service interval. There is no technician. There is no annual inspection.

So the home battery has to leave your factory more correct than the car battery does, because the factory is the last competent pair of hands that will ever touch it.

That inverts the intuition. People assume automotive must be the more demanding application because a car moves, crashes and vibrates. In terms of what the assembly line has to guarantee, the unattended product in somebody’s house is the harder problem.

What actually causes battery energy storage failures?

This is not a theoretical worry, and the best available evidence points in one direction.

EPRI, the American electric power research institute, maintains a database of battery energy storage failure incidents. In their 2024 review of 81 incidents going back to 2011, 26 had enough information to classify a root cause. The most common root cause was integration, assembly and construction. Operation came second, design third, and manufacturing of the cells themselves accounted for the fewest. Only three incidents traced directly back to the cell.

I want to be fair about the sample. That data is drawn mostly from grid scale installations rather than home systems, and 26 classified incidents is a small base to argue from. The same paper also reports that the global grid scale failure rate fell by 97% between 2018 and 2023, which is a genuinely good news story and worth saying out loud.

But the direction is clear enough to act on. The cell is not what is failing. The way the system is put together is.

That is an assembly line problem. It is precisely the problem automotive spent thirty years and a great deal of money learning to solve.

Which automotive practices transfer to home energy manufacturing?

When we look at an energy customer’s line next to an automotive one, these are the five practices that move across with no translation required.

1. 1. Error proofing at the station, verified on a schedule. Not a poster reminding the operator. A physical or digital interlock that will not allow the next step until the current one is right, and a routine that proves the interlock still works. In automotive, one major customer requires that check at the start of every shift.

1. 1. Enforced sequence. The unit does not release until the steps have been done, in order, with the results inside limits. This sounds obvious. It is rare outside automotive.

1. 1. Every result bound to a serial number. Not a batch. Not a shift. The individual unit, so that a problem discovered in three years can be contained to the units that actually had it.

1. 1. A designed failure path. What happens when the weld is out of specification, how many retries, who authorises a repair, what gets re-verified, what gets recorded. I wrote about this at length last month and it is the single biggest gap we see.

1. 1. Guided operator instruction that adapts to the variant. Home energy manufacturers are scaling headcount fast, often faster than automotive ever did, with new starters on complex electrical assembly. Static work instructions on a laminated sheet will not carry that.

None of these five requires a robot. All of them are software and process discipline sitting on top of the line you already have.

Do UL 9540 and IEC 62619 cover how the product is built?

Here is the thing that surprises energy manufacturers when we walk through it.

Their product standards are genuinely demanding. UL 9540 for the system, UL 1973 for the battery, UL 9540A for thermal runaway propagation, IEC 62619 and IEC 63056 for the cells and the storage system, UL 1741 and IEEE 1547 for the inverter side, NFPA 855 for how it gets installed. That is a serious body of testing and nobody should dismiss it.

But look at what those standards are for. They establish that the design is safe. They are type tests, fire tests and grid behaviour tests, carried out on samples.

What they do not do is force you to prove that your factory can build that certified design, every time, on a Friday afternoon, with a temp on the station. IEC 62619 gives you a single clause asking for a quality plan. There is no equivalent of the automotive requirement to predict your failure modes before you build, to prove process capability before you ship, or to have somebody audit how you make it rather than what you made.

The certification proves the design. It does not prove the process.

How does an assembly line operating system apply to energy manufacturing?

After three decades and more than 3,500 automated assembly systems, most of them for automotive and driveline customers, we built ODIN Workstation to be the operating system of the assembly line rather than a reporting layer on top of it. It plans the work, runs the work and proves the work, and it does not care whether the pack is going into a car or onto a wall in Somerset West. The physics is the same and so is the discipline.

That is really the whole argument. This is not new technology being invented for the energy sector. It is proven technology being pointed at it.

What is the takeaway for an energy manufacturer?

The energy industry is where automotive was in the early 1990s, building a product that is about to be everywhere, with quality systems designed for a much smaller installed base.

Automotive served that apprenticeship the hard way, through recalls, through warranty bills, through standards written in the aftermath of things going badly wrong. That knowledge exists now. It is transferable, it is affordable, and it does not need to be relearned.

Nobody has to pay that tuition twice.

Related reading:

ODIN Workstation, the assembly line operating system: https://odinworkstation.io/

Efficiency Edge Edition 29: Your new model is idling while your assembly line catches up: https://jendamark.io/newsletter/your-new-model-is-idling-while-your-assembly-line-catches-up/

Efficiency Edge Edition 28: What really drives a smart factory?: https://jendamark.io/newsletter/edition-28-what-really-drives-a-smart-factory/

Resources:

For a look at how we test electric vehicle battery packs by destroying them, watch this https://youtu.be/to-HUSGJGdE

And a shorter one on weld strength testing for battery packs at our facility in India https://youtu.be/GKREOo_zk_o