7/9/26
Vibration Analysis Report | Machbase Neo LLM Chat
With a single equipment vibration table and the prompt "Create a vibration analysis report," AI automatically generates an HTML report with visualizations, in-depth diagnostics, and recommendations.
🎬 Demo |
🔗Source |
This report template uses AI to automatically analyze bearing condition from raw vibration acceleration data in the Machbase Neo time-series database (C1/C2 channels, 25.6 kHz, 10 million records).
Analysis Pipeline
Time-domain analysis: Rolls up four key metrics by second: RMS (energy), P2P (amplitude), Crest Factor (impulsiveness), and mean (DC offset).
Frequency-domain analysis: Uses FFT at 0.2 Hz resolution (12.8 kHz, 4,096 bands) to provide evidence for identifying fault frequencies such as rotational harmonics and bearing defects.
Severity assessment: Automatically assigns one of four ratings (Good, Warning, Danger, or Critical) using an adaptation of ISO 10816-3 criteria and provides the deciding factors for each rating.
Trend Ratio analysis: Provides early warnings by comparing RMS trends in the latter part of the period with the initial part.
In-depth LLM reporting: Automatically generates a narrative analysis covering the current condition, spectrum interpretation, impact analysis, root-cause estimation, ISO assessment, and fault-mode diagnosis, plus recommendations for immediate, short-term, and medium-to-long-term actions.
Interactions
Supports scroll to zoom, drag to pan, double-click to reset, and precise tooltips with millisecond resolution.
✔ One-line Summary
Raw waveforms → Four time-domain metrics → Frequency-domain FFT → Severity assessment based on ISO 10816 → Automated LLM diagnostic report
#Machbase #MachbaseNeo #EquipmentDiagnostics #PredictiveMaintenance #VibrationAnalysis #FFT #ISO10816 #TimeSeriesDB #llm #AIAnalysis #SmartFactory #BearingDiagnostics #predictivemaintenance #vibrationanalysis