Battery Leakage Detection Machine

Leak testing sits at a decision point in cell production. Everything downstream of it assumes the test caught what it was supposed to catch, which makes the station's capability a constraint on the whole line rather than a local concern.

The test was performed manually. Throughput was operator-limited, and the escape rate depended on operator attention sustained across a full shift rather than on the capability of the measurement itself — so the station's real performance was neither stable nor knowable.

• Automate the seal-and-test sequence to remove operator variability from the result
• Meet the line's cycle-time allocation for the station
• Accommodate an expected change of cell format without retooling
• Keep the test itself traceable to calibrated reference specimens

• Fixed cycle-time budget inherited from the line balance
• Footprint limited by the existing line layout
• Cell format expected to change during the machine's service life
• Safety requirements for automated motion in an operator-accessible area

1. 01Established the test method's sealing and dwell requirements as the fixed input
2. 02Modelled the mechanism and ran the tolerance stack-up on the sealing interface
3. 03Sized pneumatic actuation and selected the seal stack
4. 04Completed a robot reach and cycle-time study against the line's allocation
5. 05Designed the control and safety interlock sequence
6. 06Supported build, then verified at buy-off against calibrated reference specimens
7. 07Re-verified on-line during production ramp

No FEA was performed. Structural and actuation sizing was done by hand calculation against component catalogue ratings with margin — appropriate for the load case, and stated here so that analysis is not implied where it was not done.

Machined and fabricated fixture components, commercial pneumatics and sensing, integrated robot cell, assembled and commissioned in-house before installation.

Verified against calibrated reference leak specimens at machine buy-off, so the station's capability was established against a known artefact rather than against production parts of unknown state. Re-verified on-line during production ramp, with the safety interlock sequence proven before release to operators.

• Value not published

  leak-escape rate at the inspection station

  Basis: calibrated reference specimens at buy-off, then production comparison during ramp

  Withheld under NDA. Direction of improvement was confirmed at buy-off against calibrated references. Absolute values withheld under NDA.

• Value not published

  station cycle time

  Basis: measured against the line's cycle-time allocation at buy-off

  Withheld under NDA. The station met its allocated cycle time. The allocation itself is confidential.

Impact: Moved a quality-critical test from operator-dependent to machine-capable, and did it in a way that survives the cell format change the line was expecting.

• Designing for the expected format change was the decision that paid off. The flexibility argument was harder to justify at quotation than it was to defend afterwards.
• Seal maintenance had to become a documented control-plan item, not an assumption. A test whose validity depends on a consumable will drift silently unless the drift is somebody's explicit responsibility.
• The safety and interlock scope was larger than initially estimated. Robot motion in an operator-accessible area moves work from mechanical design into sequence design, and that should have been planned rather than absorbed.

• Log test data per cell so station capability is trended continuously instead of confirmed at intervals
• Instrument seal condition directly, rather than managing it by maintenance interval
• Reduce changeover to a purely software operation by parameterising the fixture interface

CAD

SolidWorks

Hardware

Pneumatic actuationIndustrial robotLeak-detection sensingSafety interlocks

Process

Tolerance stack-upCycle-time studyMachine buy-off