Industrial Vibration Monitoring

The Aspinity analogML^TM core delivers a more intelligent, lower-power, always-on edge processing strategy for handling the massive amounts of vibration data that are collected for accurate predictive maintenance and condition monitoring of many different types of equipment such as:

 Current vibration monitoring systems digitize and process 1000’s of FFT datapoints to see if a failure has occurred, but in fact, many types of faults can be detected using just 10’s of relevant datapoints (e.g. fault-frequency/energy pairs, RMS levels, etc.). The analogML core is able to extract those important datapoints directly from raw analog sensor data and continuously monitor for faults at near-zero power.  When the first sign of a failure is indicated, the analogML core wakes up the downstream digital system so additional actions can be taken, such as performing a detailed FFT for a more deeply informed decision or requesting that maintenance staff are sent to the field.

By intelligently monitoring the most important datapoints while they are still analog and waking the higher-power system only when failures are detected, the analogML core keeps the power-on and processing time of the digital system to a minimum, significantly extending the battery life of remote sensor nodes used for vibration monitoring. 

With an integrated analogML core – you decide how much data are transmitted for operator review. The analogML core can either:
 

• Collect and send all extracted spectral peak information in order to follow trends and monitor industrial health.

      OR
 

• Wake the system to transmit data only when a condition fault is detected.

Either way, integrating the analogML core reduces the amount of data used to detect faults by 100x or more. It also produces more usable, real-time data for operators from a smaller, more power-efficient array of sensor nodes.