Asset Management / Optimization Stands to Take Significant Leaps of Value with the Internet of Things

Last week I talked about the “smart plant”, one of the key areas that is changing and opportunity for a step level of output value is in the “Asset Reliability” / “Operational Continuity”.

The real opportunity in is increasing capacity of through :

• Increased flexibility in the existing assets to run more products, and we understand asset condition through pattern recognition
•  Improved preventive, and “awareness” of asset condition and capability of performing at optimum. The devices / assets are “self aware” and self learning on improvement and conditions so early detection of conditions are seen and corrected in a timely manner.
•  Improved planning and asset utilization through transparency across assets on a site and across sites,

There is a lot of talk around the internet of things (IoT) in the general world but in the industrial world there is huge opportunity just due to the significant number of devices.
Industry pundits predict that by 2020 over 50 billion everyday objects will be connected to the Internet. This does not even include the Industrial IoT and the entire M2M environment, much of which is already in place in our factories, plants, and infrastructure.The initial trend will be to establish one-way communication, ultimately migrating to "many-to-many" communications as more physical objects be-come connected. Connecting all the assets and devices in communities of active tuning, decisions and optimization, requiring a significant rethink and change to current operational management/ supervisory systems and information systems to take advantage, but it aligns with the workforce operational transformation.So if we look at the clear steps that can happen in Asset Efficiency:

1/ Increased information, data in a one way capture of asset information.
This step is the first one and is well under way where increased intelligent devices are monitoring / calculating their performance and the information is logged to an historian. As stated in past blogs we seeing the I/O count between control systems and historians increasing by over a factor of 10. (example a pump use to be 5 to 10 variables, now is 120 to 200, a well head was planned to 50 points now logging 690 points).
Once you have this data companies like Pattern Discovery Technologies (http://www.patterndiscovery.com/) produce solutions that used defined events to investigate through Big Data Techniques asset condition patterns, from this vast historical data, so that better prediction is possible of conditions earlier.

2/ Is by direction, and communities of devices “learning” together and tuning their performance.

So instead of a device/ asset just learning on it’s own, imagine a community or similar type devices learning and sharing their learnings between them. This is not a linear learning of optimization but an exponential learning. So the conditions for a type can be immediately picked up and used by a new device / asset of the same type. Machine learning and community “hub” learning is a powerful predictive capability coming into the market.

Companies like MTELL (http://www.mtell.com/) have introduced some powerful “Machine Learning” capabilities, that combine with their “Transfer Learning” capability. Key is this does not have to wait to new devices/ assets on the plant it can be applied to existing assets, and the “learning” will begin.

The concept of going to “Smart Machines/ assets” that are:

• Self-Aware”
•  Self and Tribal Learning, so improving in predictive understanding of behavior

  Notification and increased analysis capability through powerful tools for asset analysis from the dramatic increase in data available.Now that devices can be connected through wireless to internet, and therefore a “cloud historian that is managed” and these analysis tools can executed centrally, or devices discover each other and learn together provides the breakthrough in the Asset world from predictive to prescriptive.