Induction motor failure in industrial environment is a nuisance for the maintenance team. Typical failures of induction motors are:
Bearing failure which accounts over 40% of all induction machine faults, which occurs due to vibration, inadequate/contaminated lubrication, improper mechanical loading, etc.
Winding failure or stator coil failure by open circuit/short circuit of coil due to weakened insulation varnish from overheating.
Rotor failure due to bending of motor shaft, broken rotor bar and end ring.
Other failures like motor cooling fan failure, failure due to moisture, oil/chemical vapor can damage motors in harsh industrial environment.
Why SmartEdge Agile?Most motor failures happen with certain tell-tale signs beforehand, such as motor heating up abnormally, making weird loud noise, vibrating violently or drawing more current.
SmartEdge Agile is packed with sensors which can monitor real-time data. Using those data set AI can be trained to detect normal vs abnormal situation and inform maintenance team to take necessary action before failure happens. This is Meta Sensing, where AI model is developed by detecting pattern in sensed data. It's Edge AI because AI algorithm runs on the device.
Additionally, Brainium portal allows monitoring real-time data from the sensors.
SmartEdge Agile is a combination of IoT + Edge AI + Meta Sensing, enabling the way to monitor industrial processes like never before.
Getting Started Resources
- Brainium Quick Start Guide for SmartEdge Agile Hardware configuration
- Brainium Portal Setup Guide for Real-Time Sensing and AI deployment
SmartEdge Agile is a self containing hardware with all the sensors, cpu and battery in a box with BLE connectivity. Unfortunately, there are no exposed I/Os for interfacing external hardware.
It connects to Brainium Gateway App running on a phone over BLE, from there it connects to Brainium Cloud through WiFi or Cellular data connectivity of the phone.
Here is the single phase induction motor, that will be monitored remotely -
Well, any motor can be monitored by attaching SmartEdge Agile on it with couple of cable ties.
Motor Vibration Pattern Sensing for Normal/Abnormal OperationWhen motor starts from stop condition, there is a large bump of vibration for a moment followed by smooth normal operation condition vibration during full load operation. Any time while motor is running, sudden change of speed/load will also generate a short peak vibration.
When something goes wrong with motor bearing or shaft alignment, this normal operating condition vibration will increase to an abnormal level.
Edge AI can be trained with normal (operational) vibration pattern model of a specific motor. So, when the vibration is not normal, maybe something is wrong.
Along with other data like sound(noise) level, temperature, humidity, phase current, fault can be detected.
Motor Sound Pattern MonitoringHealthy Motors have a smooth humming sound/noise of a specific frequency, which depends on the motor rpm and mechanical load. For example an induction motor driving a water pump will generate smooth "hummmmmm" like sound with certain dB level.
The frequency and loudness will change if any of the following :
- Failing Bearing (e.g. harsh startup noise and louder operation)
- Alignment Problem (e.g. metal scratching noise)
- Loss of Mechanical Load ( e.g. pumping no water)
- Damaged Cooling Fan (e.g. fan blade scratching on metal cover)
SmartEdge Agile can be programmed to detect these anomalies with the help of onboard microphone.
Right now, Sound AI is not available, therefore only sound level threshold can be used to detect failing machines and send alerts.
Once, the Sound AI is developed, specific sound level, frequency and pattern can be used to detect different faults.
Motor Surrounding Air Temperature SensingDuring normal operation, any induction motor will produce heat. This heat is usually cooled with the help of a cooling fan attached on the back of motor's shaft. That is why, air temperature surrounding the motor is has a higher temperature that the ambient air temperature.
SmartEdge Agile can be used to sense this temperature for real-time remote monitoring.
This temperature will rise when:
- Motor is consuming more current
- Motor is mechanically overloaded
- Failing mechanical parts
Since additional hardware cannot be interfaced with SmartEdge Agile, I could not include some features like current sensing, motor body temperature sensing. But, it is possible to get custom hardware with SmartEdge Agile by requesting Avnet/Brainium team for your problem specific solutions.
As for AI based fault detection, it is necessary to create faulty situation on the motor and train the AI, That would require to intentionally (partially) damage the motor and record the vibrations, noise, temperature etc. Unfortunately, I can't do that right now, so I skipped the AI part (only telemetry data).
Comments