導入事例

Mass Production Quality: Inspecting 100,000 Automotive Lead Screws Daily with AI

ケーススタディのダウンロード

Automotive Lead Screw Inspection Case Study | UnitX

Automotive Lead Screw Inspection Case Study | UnitX

CONTENTS

シェアする

Introduction: Precision in Motion

その 送りねじ is a fundamental mechanical component in any vehicle, responsible for converting rotational motion into linear motion. Whether in power seats, braking systems, or steering, the lead screw must be perfectly straight and threaded.

Defects like bending, cracks, or “bad threads” compromise functionality. They affect pitch accuracy, create stress points, and eventually lead to system failure.

However, lead screws are produced in massive volumes. For our client, the sheer scale of production made manual inspection ineffective. How do you check the microscopic thread quality of 100,000 parts every single day without slowing down the factory?

The Challenge: The Thread Nightmare

The project goal was to automate “closed-loop quality control” for a high-volume line producing diverse parts.

複雑さ:

  1. ねじ形状: Inspecting threads is difficult. The system must distinguish between a functional thread, a “rough thread,” and a “missing thread.”
  2. Defect Variety: The system needed to catch 12+ specific defect types, including thread bulging, rust, burrs, steel wire entanglement, の三脚と surface bruising.
  3. モデルの多様性: The line produces 3 different product models simultaneously, requiring the system to handle changeovers or mixed batches seamlessly.

The Thread Nightmare

Lead screws require inspection of the entire threaded length for cracking, burrs, and machining errors.

The Solution: Parallel Processing Power

To handle the massive volume, UnitX didn’t just build a faster camera; we built a parallel processing engine.

並列処理能力

UnitX’s integrated vision system is capable of processing 21 products simultaneously.

システム機能:

  • Simultaneous Processing: The system is designed to process 21 products simultaneously. This parallel architecture is the key to hitting the daily volume targets.
  • Rotational Inspection: As the lead screws move through the station, they are rotated. The UnitX OptiX system captures the entire circumference to ensure no blind spots on the threads.
  • リアルタイムトラッキング: Results are sent to the MES and equipment controls, ensuring that bad batches are flagged immediately.

Results: High Volume. High Precision. High Speed. High Stability.

  • 大音量: 1日あたり100,000個. Enables 100% inspection of total factory output.
  • 高精度: False Acceptance Rate:≤0.5%。 False Rejection Rate:≤5%。 UnitX solution reliably identifies “bad threads” and scratches at breakneck speeds.
  • 高速: 0.4s per piece (≤1.2s for 3 parts). Eliminates the QC bottleneck with real-time processing.
  • 高い安定性: It manages 21-part batches across three product models simultaneously, feeding results directly to the MES for full data traceability

Defect Visualization

The AI excels at identifying thread-specific issues. In the images below, you can see how the system identifies “Bad Thread Machining” (structural error) versus a surface “Scratch.”

Defect Visualization1

The AI model precisely highlights thread machining errors and surface scratches.

結論

High-volume manufacturing often forces a trade-off between quantity and quality. This case study proves that with the right AI architecture, you can have both. By inspecting 100,000 lead screws a day with sub-second cycle times, UnitX enables automotive suppliers to scale up production without scaling up risk.

Need to inspect high-volume components?
接触 UnitX to discuss our high-throughput solutions.

関連ケーススタディ

画像91(1)
グループ20
56画像
52画像
グループ9
グループ19(1)
グループ6(1)
グループ27(1)
59画像
60画像
上へスクロール