Best Practice

Distinguishing Operational Metrics from Mechanical Data

Saar Yoskovitz, CEO and co-founder of Internet of Things (IoT) mechanical diagnostics platform Augury, discusses the data required for predictive maintenance (PdM) in facilities management.
0 comments
P. Aleksandra: (Photography by P. Aleksandra).
(Photography by P. Aleksandra).

Saar Yoskovitz, CEO and co-founder of  Internet of Things (IoT) mechanical diagnostics platform Augury, discusses the data required for predictive maintenance (PdM) in facilities management.

 

"Uptime". "Energy consumption". "Throughput". These familiar, useful operational indicators tell us when machines are experiencing  a malfunction. And therein lies a problem. For broken machines are much more expensive to fix than those that are still in operation. New Predictive Maintenance (PdM) technologies have given rise to analyzing sources of mechanical data, that, in lieu of more traditional operational data, can be used to avoid downtime and reduce unnecessary breakdown incidents.


 

Where Operational Data Falls Short

For years, technicians have manually recorded operational conditions of machines, in order to make sure they are running as designed. Today, we see more and more machines being connected to computerized facility management systems, that sift through this operational data in search of patterns and irregularities. While important for monitoring the health of the production line and flagging emergencies for first respondance, operational data are lagging indicators, and that they don’t uncover the whole truth.

 

As an example, take a gearbox that has an oil leak. For months, the machine will gradually lose oil, and the gearbox will start grinding itself to a halt. By the time the machine’s temperature rises enough for the technicians to notice, severe mechanical damage has already happened. Operational indicators are lagging indicators, and can rarely be used to detect malfunctions early on and predict them before they happen.



Where Mechanical Data Comes From and How to Use It

Preventive Maintenance programs rely on statistical assumptions about machine wear and tear as well as total uptime hours. When analyzing mechanical data, a technician has visibility to a machine’s current condition and can prescribe maintenance and switch from schedule based maintenance to condition based maintenance..

 

In the past, this data was often collected by engineers trained in ultrasound and vibration analysis techniques that would take readings with various sensors and then go back into their offices to make calculations and create a report of their findings. This lengthy and expensive process, critical to early PdM approach, is now being replaced by technological advances in the field.

 


Combining the Best of Both Worlds

Modern PdM techniques now only require a technician and a smartphone. Using the smartphone as a connectivity point, a machine’s mechanical data can be transmitted over the Internet to servers where it is analyzed instantaneously. With big data, machine learning and unprecedented mobility, facilities can eliminate catastrophic equipment failure, reduce downtime, decrease costs of parts and labor and streamline efficiencies—all efforts that benefit the bottom line.

 

Saar Yoskovitz

About Saar Yoskovitz

Saar Yoskovitz is the CEO and Co-Founder of Augury (www.augury.com) which builds mechanical diagnostics platforms for the Internet of Things (IoT) and enables facility owners and service companies to deploy scalable predictive maintenance strategies that reduce environmental impact, energy usage and operational costs.

Article Rating

Vote Data