
🏭 Manufacturing · Data Flow
How vibration and temperature data from machines becomes a maintenance work order before the machine fails.
Drawing diagram…
Data flow for predictive maintenance: vibration, temperature and motor current sensors on pumps and motors stream data to an edge gateway that computes features (RMS, FFT peaks), sends them to the cloud, where an ML model estimates failure probability and remaining useful life. High-risk predictions raise an alert to the maintenance planner and create a work order in the CMMS (Maximo); completed work order data is fed back to improve the model.
flowchart LR S[Vibration, temperature, current sensors] -->|Raw signals| E[Edge Gateway] E -->|Features: RMS, FFT peaks| C[Cloud Ingestion] C --> H[(Sensor History)] H --> M[ML Model - failure probability] M -->|Remaining useful life| D[Maintenance Dashboard] M -->|High risk| A[Alert to Planner] A --> W[CMMS Work Order - Maximo] W -->|Completed work, root cause| L[(Labelled Failures)] L -->|Retraining| M
The states a production order goes through on the shop floor, from release by planning to closure, including material shortages, rework and scrap.
An incoming and in-process quality inspection flow: sampling, measurement against the spec, and what happens to a batch that fails.
A connected factory: machines and sensors on the shop floor, an edge layer, MES and ERP, and cloud analytics for OEE and predictive maintenance.
How production orders flow from ERP planning to the shop floor MES and how confirmations, goods movements and scrap flow back.
How materials flow from suppliers to customers in a manufacturing business: purchasing, inbound quality, production, warehousing and distribution.
A schema for multi-level bills of materials: items, BOM headers and lines, routings and work centres.