It’s all about the things.

In IoT jargon, a thing is an entity or physical object that has a unique identifier, an embedded system, and the ability to transfer data over a network. Therefore lots of things are things! For IIoT we tend to think of things being, machines, equipment, vehicles, buildings, appliances, computers, and even people.

These are things that you can observe and control. IoT facilitates monitoring health, safety, and security of things; see status, helps to make plans based on how healthy they are, or where they are, and to compare and contrast allowing insightful decision-making. IIoT technologies allow Mtell to learn normal and failure behavior on a machine and transfer those signatures to a similar machine across the globe. Also, IIoT technology is the foundation for Mtell’s remote monitoring (datacenter) Reservoir and Summit products, allowing things to be viewed managed remotely, to be gathered into similar groupings for federated viewing across sites and potentially across customers, and for more intense and more accurate learning of behaviors of similar things regardless of disparate locations.

Mtell Reservoir

Bigger is Better

You can observe and control many disparate things in big and bigger groupings. Bigger needs bigger computers, and Mtell can migrate to very large clusters of Hadoop machines that scale with almost linear performance improvements to permit analysis on large asset groups. We call this population-based learning, where each thing donates its pieces of learning on its operational duty cycles and the specific modes of failure it has encountered. The group computation can then process all the things and their behaviors in one collection. The remote learning on collections provides the fastest, most accurate learning of the most behaviors and earlier higher fidelity warnings of issues, permitting wider levels of efficiency and optimization across pools and fleets.

Sharing is Caring

One of the strongest features of the Mtell application is the agent methodology that engages with the underlying machine learning technology. That design permits learning of discrete, unique patterns for normal operation, plus particular failure causes and failure modes. This assures agents can work independently, with minimal human supervision, but it also assures agents can share specific learned behaviors. A bearing failure signature learned on one pump is readily shared with similar pumps in the same location or different locations. But, one pump is unlikely (or unlucky) to witness all possible failure scenarios.

Sharing assures all similar equipment receive the signatures of all failures that affect any member of the group assuring the same safety and risk protection applies to them all.

I want to know more about