IoT predictive analysis system for rotating machines
This vibration analysis system produces conventional business data and predicts defects using algorithms running right
next to the sensor (Edge Computing). This information allows to smoothly scale up from preventive to predictive
Conventional on-board processing: rotational speed, temperature, RMS, OVL, OAL, characteristic frequency amplitude (low
frequency range), frequency band amplitudes, ISO 10816 / ISO 2372 threshold overruns. On-board predictive analysis
(experimental): classification of defects, vibratory energy drift by frequency bands (self-learning), expected uptime
before a threshold overrun (arbitrary or ISO).
Based on the analysis of acoustic signals, this reference design allows the detection of sound events and their
filtering to effectively discriminate sounds produced by a firearm from any other similar sound. The object includes a
patent-pending algorithm assuring extremely short reaction times and high reliability.
Data is securely transmitted over IoT LPWAN networks. The recipient is informed of where and when an assault involving a
firearm took place and what kind of weapon was used, to effectively decide on the response strategy.