Compare two fault-detection methods on bearing data

Northwind Pumps (fictional) · verified

Apply envelope analysis and a spectral-kurtosis check to labelled bearing data, then recommend one for the monitoring dashboard.

Company: Northwind Pumps (fictional)
Mentor: Theo Brandt
Tier: Q2 Foundation
Effort: About 12 hours
Compensation: Stipend: €150 on accepted completion
Apply by: 2026-11-05 (UTC)

Discipline: Mechanical engineering
How applicants are chosen: Scored against the published criteria
Mentor time: About 4 hours with you, replies within two business days
Expenses: None expected.
Payment: Northwind pays by bank transfer within 30 days, through its supplier process. Program staff hold the agreement.
Software and equipment: Python 3 with pandas and numpy, or any tool you prefer.

BuildsSignal analysisTechnical writing

The problem

The reliability team must pick one fault-detection method for its dashboard and wants an evidence-based recommendation.

What you'll deliver

Notebook comparing both methods
A two-page recommendation memo

Milestones

1Method plan agreed with mentor · day 5
2Comparison and memo · day 20

Required skill evidence

Signal analysis: Emerging or higher

How your work is assessed

Each criterion is scored 1 to 4. Required criteria must reach their threshold to pass; a high average doesn't make up for a missed one. One revision round is included.

Criterion	Required	Threshold	Skill evidence
Method comparison Both methods applied correctly and compared on the same data.	Required	3 of 4	Signal analysis
Recommendation Clear choice with trade-offs and limits.	Required	3 of 4	Technical writing

Resources

Data and starter material are shared once you're enrolled.

Terms

Learning project. Company data stays private; the approved summary may go in your portfolio.

Portfolio: You may publish the approved summary, not the company data.

Backup contact if your mentor is unavailable: Sam Rivera, staff@demo.test