Operating; Training Anomaly Detection Models - Siemens SITRANS SCM IQ Operating Instructions Manual

Operating

8.1

Training anomaly detection models

Once a Insights Hub Asset has provided a sufficient amount of data (typically two weeks), you
can continue with the training of an anomaly detection model. After the trained model is
available, you can start the model for continuous anomaly detection.
For more information, watch this video: SITRANS SCM IQ: How to add Machine Data to
SITRANS SCM IQ (https://youtu.be/AG31J1v3gB8).
Required user roles
• Required user roles without Insights Hub Subtenants:
– Insights Hub user role: TenantAdmin or StandardUser
– SITRANS SCM IQ user role: sitransscmiq:admin
• Required user roles with Insights Hub Subtenants:
– Insights Hub user role: SubtenantUser
– SITRANS SCM IQ user role: sitransscmiq:admin
Procedure
1. Choose the card or menu "Data modelling".
In the list you can choose between two different asset types:
– SitransMS
– Machine Data Assets
2. To get further information select an available asset.
3. To view the data, click "See details".
You can change the visible time range and select one or more variables for a more detailed
view.
SITRANS SCM IQ: Smart Condition Monitoring with IIoT Sensors
Operating Instructions, 07/2023, A5E50573389-AD
Note: Do not use TenantAdmin when using Subtenants. Subtenant can not see devices
and assets of other subtenants.
Variables Description
Acceleration Vibration in x y and z axis seen in space.
Health status Signal strength for SITRANS MS200 in a scale of -100 to 0.
•
•
Temperature Temperature from the place where the sensor is mounted. The temperature
sensor is inside the housing of SITRANS MS200.
Deviation index Deviation of the machine condition from the normal state. If the deviation is
too large, your anomaly will be generated.
-100 means that the bluetooth standard is bad.
0 means that the bluetooth standard is best.
8
57