AI for Predictive Maintenance via Vibration Analysis

Freelance AI & Predictive Maintenance Expert (Vibration Analysis)

We are looking for a specialized Data Scientist or AI Engineer to develop a Predictive Maintenance solution for a critical production line. The project involves analyzing vibration data from a high-capacity main electric motor to identify early signs of mechanical failure and optimize maintenance cycles.

Technical Context
Data Source: High-frequency vibration data collected via IFM sensors.

Dataset: Approximately 300,000 rows of historical records stored in a SQL database.

System Focus: Rotating machinery (Main Drive Motor) within an industrial manufacturing environment.

Core Goal: Implement an anomaly detection and health-scoring model to prevent unplanned downtime.

Key Responsibilities
Data Engineering: Extract, clean, and preprocess 300k+ rows of SQL-based vibration data.

Feature Extraction: Transform raw signals into meaningful features (Time-domain: RMS, Kurtosis, Skewness; Frequency-domain: FFT, PSD).

Model Development: Build and train Machine Learning models (e.g., Random Forest, XGBoost, or LSTM) for failure prediction and anomaly detection.

Signal Processing: Apply signal processing techniques to distinguish between operational noise and actual mechanical degradation.

Validation: Evaluate model performance using precision/recall metrics focused on reducing false positives in a factory setting.

Required Qualifications
Proven track record in Predictive Maintenance (PdM) or Industrial AI.

Deep expertise in Python (Pandas, Scikit-learn, SciPy, or Signal Processing libraries).

Strong experience in Time-Series Analysis and vibration-based diagnostics.