Machine Learning

Machine Learning

Plan today for tomorrow’s environment

After decades of using human and statistical techniques, machine learning methods are now being applied to further improve efficiency and accuracy in the diagnosis and prognosis of potential failure modes of rotating machinery. Machine learning methods allow monitoring systems to react even faster and more precisely than standard, traditional tools and methods.  It also further enhances the capability of skilled human experts.  ITR Machine Learning services are available through ITR’s smart Condition Monitoring and Analysis System (CMAS) deployments and expert analysis services. 

Each CMAS has connectivity to the ITR cloud services.  When active, verified findings are collated over time and form the basis of training sets used by ITR data scientists to develop CMAS, site, and asset specific algorithms for advancing real-time exception tests and notifications.  This iterative process means each CMAS continually improves and evolves with the lifecycle of each monitored asset.

 Similarly, ITR expert analysis services also leverage machine learning to digitize knowledge so experts avoid repetitive tasks and focus on the complex analysis that only 20+ year experts can do.  Each machine and customer are unique, so individualized algorithms are developed.  When verified as effective, these algorithms are used to augment human experts and can be deployed to edge devices to advance portable/offline monitoring methods.


For many organizations, vibration analysis is the cornerstone of their predictive maintenance (PdM) efforts and reliability process. Of all PdM technologies, vibration analysis often provides the greatest ROI because it is the most comprehensive and most precise. The technology is ideal for many electro-mechanical assets, especially ones experiencing varying operating conditions.


Infrared (IR) thermography is a non-destructive testing method used to measure thermal gradients in energy emitted by electrical and mechanical systems. From these gradients, competent thermographers identify variations in energy emissions that are traceable to potential asset failure modes. IR emissions often enable analysts to identify problems that are concerns otherwise undetectable using visual or mechanical inspections


Lubrication analysis is a non-destructive test used to assess the condition of lubricants or power transmission (hydraulic) fluids and determine the type and amount of contamination present. Because of the criticality of these types of fluids to industrial operations, fluids analysis trending over time is one of the most powerful predictive tools for identifying potential failures. Fluids analysis looks for three basic categories of elements affecting the lubrication effectiveness: wear metals, contaminants, and additives.

Periodic monitoring may prolong fluid life and identify elements that may indicate component or lubricant degradation. Common fluid tests include: elemental spectroscopy, viscosity, acid number, FT-IR, base number, particle count, shape analysis and water analysis.


Motion Amplification is a video processing algorithm that detects subtle motion then amplifies that motion to a level visible with the naked eye which enhances the understanding of the components and interrelationships creating the motion. Motion Amplification is one more tool in ITR’s process and sytem approach to effective, affordable, proven PdM success.


One benefit ultrasound emission testing offers over many other PdM technologies is that direct contact is not always required and skilled testing professionals can scan large areas quickly from safe locations, in some cases, remotely. For electrical systems, ultrasonic testing is often an ideal technology when used in conjunction with IR thermography as together, the spectrum of potential failure modes is expanded exponentially.