Predict Failure in IoT Devices
One of the greatest values of having connected devices is the ability to track and monitor those devices remotely. With the Internet of Things (IoT, and particularly with the Industrial Internet of Things (IIoT), connected devices can transmit detailed logs of the various sensor inputs, the devices response, and detailed status updates back to a centralized data repository for analysis. And one of the most common efforts behind these connected devices is to use this
Typical IoT data is a mess of disorderly and poorly structured log data. When one customer wanted to understand the recurring failures of devices deployed in the field, they first attempted to understand device failure using open source statistical tools. The project was conducted over a period of months and produced limited results, in part because of the complexity of the data.
They were then able to use the Emcien suite to analyze that same data. Looking across multiple device types and configurations, the software automatically found the combinations of data points that were unique to device failure. The analysis revealed not only combinations of device components, but also combinations of sensor and hardware readouts that predicted a high likelihood of failure.
To further improve the customer’s ability to understand and predict on their connected devices, the company created a recurring analysis with Emcien’s APIs. Integrating analysis into the existing infrastructure gave them an always up to date view of device performance. When a device begins to fail the predictions then identify both the individual device and the reasons they are predicted to fail. With that advanced knowledge the company can schedule maintenance before the device can fail.
Preventative maintenance improved device uptime.
Prescriptive outputs identified the hardware configurations that failed most often.
See a Demo
Want to see how this solution can help your team? Enter your contact information in the form below and an Emcien representative will reach out to show you a demo tailored to your use case.