Overview
A leading automotive OEM with a dominant market share in India and a production ecosystem spanning multiple engine variants required a robust, error-proof solution for one of its most critical engine sub-assembly processes — Journal Bearing Assembly.
The bearing assembly process involves the precise manual picking and placement of multiple bearings across upper and lower assembly stations, with strict adherence to sequence and variant-specific configurations. Any deviation — a wrong bearing picked, a missed bearing, or an out-of-sequence assembly — would result in engine failure downstream, costly rework, and compromised safety.
The manufacturer needed a system that went beyond passive monitoring. They needed one that could actively verify every pick, every placement, and every sequence step in real time — and immediately interlock the line when a violation occurred.
Opportunity
- Supporting multiple engine variants with different bearing counts increased the risk of wrong bearing selection, missed assemblies, and sequence errors.
- Existing poka-yoke relied only on visual indications, providing no automated verification of bearing pick or assembly sequence.
- Defects were identified only at downstream inspection stages, resulting in rework, quality escapes, and increased supervisory dependency.
- Lack of real-time productivity data and engine-level traceability limited process optimization and accountability.
Jidoka's Approach
AI-Driven Video Process Monitoring
- Real-time AI video analytics deployed across two dedicated assembly stations (upper bearing and lower bearing), each monitored by a dedicated camera connected to a common panel and computer.
- Engine variant and model data received directly from RFID via PLC, enabling automatic work order initiation and variant-specific verification without manual input.
- AI motion study verifies that each bearing is picked from the correct location as indicated by the pika-pika lamp system — not just that a bearing was picked, but that the right bearing was picked from the right bin.
- End-to-end assembly sequence verified to confirm all bearings are assembled in the correct order and that no bearing is missed, before an OK judgement is issued.
Real-Time Error Prevention & Line Interlock
- On any SOP violation — wrong pick, wrong sequence, missed bearing — Nagare instantly issues an electrical output to the PLC, triggering a line interlock or audio-visual alarm as per the manufacturer's requirement.
- The operator's OK button is activated only upon Nagare's OK judgement; the button remains inactive on any NG condition, preventing the part from advancing to the next process.
- Both pick sequence integrity and assembly placement sequence are covered — creating a 100% fool-proof dual-verification system at every cycle.
Productivity & Performance Intelligence
- Cycle time monitored per activity step and per full assembly cycle, with each bearing pick treated as a discrete measurable activity.
- A static visual display shows the journal bearing assembly layout with real-time indication of assembly zones: sequence order, zones completed, and zones pending.
- Real-time dashboards surface per-operator and per-station performance, including cycle time, deviation events, and quality violations.
- Full video traceability stored per cycle on the local computer and transferred to the manufacturer's server, with traceability linked to engine number and model.
- Digital work instructions embedded in the interface, reducing dependence on paper SOPs and supervisory guidance.
Big Wins
Nagare moved a high-risk manual operation to a fully traceable, error-proof assembly process.
Zero Errors — 100% Fool-Proof Assembly
AI-powered dual verification of pick sequence and placement sequence eliminated wrong bearing assembly and missed bearing incidents entirely. The system's interlock ensures no non-conforming engine advances to the next process.
<1 Year ROI
Projected savings from productivity gains, yield improvement and the elimination of downstream supervisory inspection delivered full return on investment within one year of deployment.
Increased Yield
Real-time operator analytics and reduced rework cycles — enabled by immediate error detection at source rather than downstream discovery — drove significant improvement in yield.
Reduced Supervisory Dependency
Automated real-time verification eliminated the need for manual supervisory oversight at the assembly station, freeing up quality personnel for higher-value activities.
The leading automotive manufacturer moved its Journal Bearing Assembly process to Jidoka's AI-powered Nagare system, achieving a 100% fool-proof verification system across pick and placement sequences, zero assembly errors, a 10% yield improvement and full process traceability — all with ROI delivered in under one year.






.avif)