Course Description
This course develops deeper machine learning capability specifically for vibration-based fault diagnosis. Where general condition monitoring AI covers many technologies, this course focuses entirely on vibration — the richest and most widely used condition monitoring signal.
Participants learn the machine learning techniques best suited to vibration diagnostics: signal processing for feature extraction, supervised classification of fault types, unsupervised detection of novel faults, and deep learning on raw vibration signals. The course is aimed at those who want to build, not just use, vibration diagnostic models.
What you will achieve
By the end of this training course, participants will be able to:
- Apply signal processing to extract vibration features for ML
- Build supervised models to classify vibration faults
- Apply unsupervised learning to detect novel vibration patterns
- Use deep learning on raw vibration signals
- Handle imbalanced and limited fault data
- Evaluate vibration diagnostic model performance
- Combine physics-based knowledge with data-driven models
- Deploy vibration diagnostic models
How the course is delivered
The course combines vibration signal processing with hands-on machine learning on vibration datasets. Participants build and evaluate diagnostic models from feature extraction through deep learning.
The programme balances signal processing, classical ML, and deep learning, grounded in real vibration fault data.
Designed for
This training course is ideal for:
- Vibration analysts moving into data science
- Industrial data scientists focused on vibration
- Reliability engineers building diagnostic models
- Condition monitoring specialists
- Technical staff developing vibration AI
Daily programme
- Vibration signal characteristics
- Signal processing for feature extraction
- Time, frequency, and envelope features
- Building vibration feature sets
- Supervised learning for fault types
- Model selection and training
- Handling imbalanced fault data
- Evaluation and validation
- Unsupervised novelty detection
- Clustering vibration patterns
- Deep learning on raw vibration signals
- CNNs and autoencoders for vibration
- Combining physics and data-driven models
- Model robustness
- Deploying vibration diagnostic models
- Case study workshop and review
Certification & accreditation
BII Certificate of Completion
BII Certificate of Completion Upon successful completion, participants receive a BII Development Institute Certificate of Completion with a unique reference code that is independently verifiable. Our certificates are recognised internationally and reflect successful completion of your chosen programme.
