With fully interconnected embedded devices now the norm in many factory settings, industrial design teams are implementing predictive maintenance strategies to reduce downtime, lower personnel costs, and increase production. There are basically three different approaches to equipment maintenance in any setting:

  1. Reactive: Run equipment until it fails then repair
  2. Preventative: Periodic shutdown to test and replace worn parts
  3. Predictive: Monitor equipment continually and repair as needed

The goal of predictive maintenance is to pinpoint when a failure is going to occur so that repairs can be made at a convenient time before the breakdown actually happens. Successful predictive maintenance requires real-time monitoring and analysis of important equipment parameters via remote sensors, management tools, and diagnostics along with universal connectivity.


To simplify this remote data collection process, Intel has developed a number of technologies that can be implemented in embedded systems supporting industrial automation applications.  One of the most recent introductions is the Intel® Intelligent Systems Framework (Intel® ISF) to simplify connecting, managing, and securing embedded devices. Intel® ISF combines processor architecture, operating system software, and other tools to create secure, interoperable platforms for intelligent systems.  The framework is built around system processors with Intel® vPro™ Technology (Intel® vPro ) providing built-in hardware support for remote management, virtualization, and platform security functions that can be used to extend the uptime of industrial embedded systems.

Intel® Active Management Technology (Intel® AMT) is a key element of Intel® vPro that enables real-time data collection from production equipment sensors that can be used in predictive maintenance analysis. Intel® AMT delivers certificate-based security allowing remote access to the embedded system for management and security tasks even when the system is powered off. This technology gives device support personnel a low cost technique to monitor operation, perform diagnostics, deliver product training, and manage required software updates from a remote location. In the event of a software failure, Intel® AMT enables a remote boot from an external operating system image over the network even if the system kernel has been completely corrupted. Most of this reboot process can be done with automated scripting and little human interaction which can significantly lower system downtime.

Several members of the Intel® Intelligent Systems Alliance offer off-the-shelf Intel® AMT compatible processor platforms that designers can incorporate into industrial systems to collect the necessary data. For example, the CEQM77 COM Express module family from Intel® Intelligent Systems Alliance member Radisys combines a 3rd generation Intel Core i7 processor and the Intel QM77 Express chipset in a 95mm x 125mm form factor for high performance industrial applications (See figure 1). The CEQM77 supports DirectX 11 and OpenGL graphics, up to 16GB of error correcting code (ECC) memory plus one PCI Express 3.0 x16 PEG port and seven PCI Express 2.0 x1 ports.  The module provides Trusted Platform Module (TPM) support as well as support for Intel® AMT enabling remote access and diagnostics via the Radisys Embedded Software Platform (eSP).


The analysis portion of a predictive maintenance strategy can range from simply verifying that remote measurements are within pre-established limits to much more involved prognostic algorithms. For example, the Center for Intelligent Maintenance Systems (IMS) at the University of Cincinnati has developed a Watchdog Agent Prognostics Toolkit that includes algorithms for neural networks, logistics regression, Gaussian mixture modeling, and statistical pattern recognition that can be used to detect and predict faults in everything from critical factory machinery to human organs (See figure 2). The toolkit adds signal processing and analysis functionality to the LabVIEW system design software suite from Alliance member National Instruments. There are five basic steps in the IMS approach to prognostics: data acquisition, feature calculation, principal component analysis, fault classification and health monitoring, and health prediction.

NI IMS.png

With remote data collection enabled by Intel® AMT in operation, industrial designers can implement multiple condition-based maintenance strategies to foresee equipment failures and reduce downtime. If you are involved in a predictive maintenance project requiring remote equipment management and data analysis, please share your questions and comments with fellow followers of the Intel® Embedded Community. You can keep up with the latest technical articles and product announcements at the Embedded Computing Design archives on industrial systems maintenance.




Solutions in this blog:


Related topics (blogs, white papers, and more):


Warren Webb
OpenSystems Media®, by special arrangement with the Intel® Intelligent Systems Alliance

Radisys and National Instruments are Associate members of the Intel® Intelligent Systems Alliance.