Check out this GitHub repo to learn how the Japan IoT Team used an accelerometer, a simple rotating fan, and an SVDD model to demonstrate anomaly detection capabilities. This demo was developed to be an way to demonstrate anomaly detection using a real-time vibration data and SAS ESP. Input data is flowing in via MQTT at a rate of 12.8 kHz with RMSA calculated every second. After that, the SVDD model for is used for scoring. (Note: The SVDD model was trained using SAS Studio)
As shown in the video below, we made a device with 2 rotating fans (one with normal fan and another with a blade broken) and used SVDD to detect when the condition of the machine (fan) changes. The output is SVDD distance (top graph) and SVDD score (bottom graph). When the normal fan is rotating, the score is -1 (normal state) but when the device is switched to the broken fan, the SVDD score changes to 1 (faulty state).
Below is an instructional video on how to built your own fan simulator device that you can switch between normal and failure modes. The team attached an accelerometer to gather the vibration data and send to an ADLink USB-2405 DAQ. (Note: It is also possible to use ADLink MCM-100 DAQ (edge device) which is basically a DAQ device and also a PC)
Whether you're already using SAS Event Stream Processing or thinking about it, this is where you can connect with your peers, ask questions and find resources.
Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.
Find more tutorials on the SAS Users YouTube channel.
-
1232 views
-
3 likes
-
-
