Noise / Vibration Based
Quality Control
How a leading truck engine supplier used Wisdom Test to detect manufacturing defects in real-time, and to improve production yield.
The Client
Tier-1 turbomachinery supplier producing engines for global OEMs, producing thousands of units each year
The Challenge
Quality Inspection is manual intensive, error-prone, no quantitative explanation, the inspection queue is long
The Solution
Integrate Wisdom Test to monitor the quality inspection, oversee the process and provide predictions
The Outcome
The quality process is automatic, explainable, adaptive, continuously learning and has no human dependency
Technical Solution
Wisdom Test makes use of production / test data for training learning models and deploys them in real-time for extremely accurate quality predictions.
After sensor installation, noise and vibration recorded during testing, auto-cropped for each test
Creation of Spectrogram with frequency and loudness information for quantitative comparison
Clustering over thousand of samples with calibration, detailed mapping and overview
Root cause analysis for improving diagnostics and design improvements
Multi-Sensor Data Support
Wisdom Test supports combinations of different types of sensor data, such as pressure, temperature, noise, vibration, vision and electricity
Semi-Supervised Training
Wisdom Test uses clustering algorithms to inspect the process and makes use of minimal number of human labeled instances
Adaptive and Continuous Learning
Wisdom Test provides the framework for adaptive learning against data drift and integrates AI tools for easy of use
Benefits
Measurable Benefits in Quality Control and Cost Reduction