envelope-analysis

MCP server for predictive maintenance and machinery fault diagnosis. Gives AI assistants evidence-based vibration analysis - FFT, envelope, bearing fault detection, ISO 20816-3 severity - with a measured, blind CWRU benchmark. Local-first: raw signals never leave your machine. Includes a Claude Code plugin.

転がり軸受の振動診断ケーススタディ。信号処理(FFT/Welch/Hilbertエンベロープ)を自前実装しscipyと照合、CWRU(正解既知)で手法を固定してNASA IMS run-to-failureで検出器を設計・評価。独立runへの転移テストと限界の開示まで含む。

Diagnose car vibration and road-speed hum with a smartphone. Order analysis, envelope demodulation and engine-lock testing on raw accelerometer data — no industrial sensor required. Tested on Audi Q7 air suspension; works on any vehicle.

Physics-guided machine condition monitoring system using vibration signal analysis, FFT, envelope analysis, bearing kinematics, anomaly detection, and health trend monitoring.