Predictive Maintenance

关于

Predictive Maintenance is an open-source framework that empowers Large Language Models (LLMs) like Claude with industrial machinery diagnostics capabilities through the Model Context Protocol. It transforms raw vibration sensor data and machine manuals into actionable insights, enabling AI-powered vibration analysis, bearing fault detection, and predictive maintenance workflows via natural conversation. This system helps democratize access to expert-level machinery diagnostics, augmenting human engineering work and making industrial AI more accessible for Industry 4.0 applications.

主要功能

• 0 GitHub stars
• Analyzes real bearing fault vibration data with 15 production-quality signals included.
• Offers advanced diagnostics including FFT spectrum analysis, envelope analysis for bearing faults, and time-domain feature extraction.
• Assesses vibration severity against the industry-standard ISO 20816-3 compliance.
• Performs ML anomaly detection using unsupervised/semi-supervised models (OneClassSVM/LOF) on healthy baselines.
• Generates professional, interactive HTML reports with Plotly visualizations for FFT, envelope analysis, and ISO compliance.

使用案例

• Automate bearing fault detection and diagnosis from vibration signals with LLM assistance.
• Evaluate machine vibration levels against ISO 20816-3 standards to assess operational health.
• Integrate machine manuals to extract specifications, calculate bearing frequencies, and diagnose faults using conversational AI.