Battery cell production: scrap in coating, drying and formation

Coating, drying and formation produce high scrap

What happens. An electrode is coated, dried, rolled, assembled, filled and then charged for the first time in formation. Faults made early, such as uneven coating weight or leftover moisture, show up only days later in the test of the finished cell. The line cannot see them while they happen.

What it costs. Scrap of costly material, cells that fail at the end of a long chain, and a ramp-up that is slow because causes are found by looking back.

Why the usual fix fails. Measuring every stage inline is costly and does not predict the end result. Tightening each stage by hand adds margin and energy. Statistics on the finished cells show a problem weeks after the cause.

How a fixed-price remote diagnosis checks it. From the line data (coating weight, oven temperatures, web speed, moisture, formation curves) and the test results of the cells I look at which early values explain late failures. The fixed-price remote diagnosis tells you whether computed process states can warn early and steer drying, and what it needs.

Background: Computed measurements

The carbon footprint of each battery must be computed from process data

What happens. European battery rules require a digital battery passport and a declared carbon footprint for many battery types, and set recovery targets for materials such as lithium. The footprint has to be known per product, from the energy and material used for it.

What it costs. Manual data collection from many sources, estimates with large uncertainty, and recyclers who have to raise yield under rules they cannot yet measure.

Why the usual fix fails. Average values per plant hide that the energy per cell depends on yield, line loading and product. A spreadsheet at the end of the year does not give figures per batch. Recycling control tuned on throughput ignores recovery.

How a fixed-price remote diagnosis checks it. From energy meters, line data and yield records I look at how well the footprint of a single product can be allocated, and where data are missing. The fixed-price remote diagnosis tells you whether a footprint per product can be computed automatically, and for recyclers whether control can raise the recovery.

Yield during ramp-up of a new cell line lags the plan

What happens. A new cell line takes years from announcement to full output. Early in the ramp-up the yield is low, the process windows are not yet known, and each change of a setting needs days to show in the cells.

What it costs. Scrap on the most expensive part of the ramp-up, capital that earns little, and energy spent on the formation of cells that will fail.

Why the usual fix fails. Copying the settings of the equipment maker’s reference line meets different materials and a different hall climate. Waiting for the quality results of each batch makes learning slow. More inspection stations add cost.

How a fixed-price remote diagnosis checks it. From the data of the first weeks I look at which inline values predict cell quality and how formation could be shortened where cells are clearly good or bad. The fixed-price remote diagnosis tells you whether inline quality models and process control can speed up the learning.