IoT Manufacturing Machine Case - KNIME Error Detection and Prediction System

Case Background

  1. The manufacturing industry hopes to run the machines for longer times with less major parts breaking or being replaced. On the other hand, companies also hope to avoid machines to break completely less to avoid major losses. These reasons make error detection in machines very important.
  2. As Internet of Things (IoT) becomes a real concept, users can create operating logs on the machines and create large amounts of monitoring data to evaluate and optimize cost and efficiency. Users can also detect errors and predict future events.

Issues Faced

  1. How can past data be used to conclude if similar errors have not happened before to prevent production losses?
  2. How can error information be identified in monitoring data?
  3. How is error information delivered to management staff?

Solutions/System Screen

  1. The system uses "time line graph", "scatter graph", "correlation graph" and "thermal image" to observe and analyze the monitoring data to obtain the difference between "normal" and "abnormal" operations.






  2. It uses the “SPC” technology to define the time range of a normal operating machine. The range takes the average signal value as the center and the upper and lower limit is 2 standard deviations away. If the signal is over or under the limits, an error alarm should be issued.


  3. The second method is to use the monitoring data collected when no errors are detected to build an auto recursive (AR) model, like a control graph, and calculate the upper and lower limits. When the new monitoring data predicts that the operating value of the machine is about to exceed the upper of lower limit, the error alarm will be issued and the manager must check the machine.


  4. When an error alarm is issued, an Email is sent to the manager to request an inspection.


Relevant Cases:

  1. Foxconn- Introducing Smart Factory Management System
  2. LED Packaging Manufacturing case- Introducing the iPASP System
  3. Plastic Mold Industry Case- Introducing the iPASP System
  4. Semi-conductor IC Testing Case- Introducing the iPASP System
  5. Precision Technology Industry Case- Introducing the iPASP System
  6. Certain Manufacturer- MicroStrategy Sales Analysis Management Platform
  7. POSCO Korea Steel Plant- IBM SPSS Modeler Quality Prediction and Control System

AsiaAnalytics Taiwan Ltd.


5F, No. 356, Sec. 1, Neihu Rd., Neihu Dist., Taipei City 11493, Taiwan (R.O.C.)

TEL:+886 2 7728 7958 FAX: +886 2 2627 0667

e-mail:service@asia-analytics.com.tw