PHM Technology

Prognostics is the process of predicting the remaining useful life of an asset by assessing changes in its condition or the extent of its degradation. Health monitoring is the process of measuring and determining the extent of deviation or degradation of an asset from that expected in normal operation. Diagnostics is the process of detecting the occurrence and determining the type of a particular fault or condition. Health management provides the information required to obtain accurate diagnostics and prognostics.

 

History of PHM Technology

In the early 1990s, rapid advancements in sensor technology and computing power resulted in the collection of extensive operating data for critical equipment across a broad range of industries. Automated data collection programs were installed and enormous amounts of operating data have been collected. As maintenance, engineering, operations, safety and management personnel began to evaluate this avalanche of data it became obvious that automated tools to accurately monitor and extract information from the data were lacking. Health monitoring technologies were developed to address this need. These products provided maintenance alarms once a value was measured outside a predetermined range and were a vast improvement over previous techniques. However, by the time the alarm was raised the equipment was often failed or failing.

Advancements in pattern recognition and decision assessment technologies now allow early and accurate diagnoses to be made automatically from this health monitoring data as well as predictions regarding equipment remaining useful life. Pattern recognition algorithms are used to detect abnormal equipment behavior while operating data is within normal operating ranges–well before the abnormality can be detected using traditional means. This enables impending failures to be identified early before they disable productive equipment.

Our Approach to PHM

The key steps in setting up a PHM system consist of defining and calibrating the software models and then using the software models to predict behavior, detect differences in asset behavior, diagnose the detected differences for abnormal behavior and ultimately predict remaining useful life to allow corrective action to be taken before an asset fails. These steps are illustrated in the flow diagram below.

Prognostic and Health Management Software Set up and Use

PHM Software Setup and Use Flow Chart

Important Considerations in Selecting PHM Software

  • Software Sophistication and Performance
    		•
     
  • Ability to monitor many variables in real time
    		•
     
  • Accuracy and false alarm rates
    		•
     
  • Implementation and roll out time
    		•
     
  • Ease of use
    		•
     
  • Quality of reporting
    		•
  • Flexibility and Adaptability
    		•
     
  • Applicability to multiple equipment types and brands
    		•
     
  • Adaptability to in-house legacy systems
    		•
     
  • Ability to add future features and capabilities rapidly
    		•
     
  • Integration with in-house proprietary technology
    		•
  • Price and Risk
    		•
     
  • Payback time for software investment
    		•
     
  • Proven software and in use at other sites
    		•

     

    Evaluating PHM Solutions

    To compare PHM solutions, certain software characteristics should be evaluated:

     
  • Hardware platforms supported
    		•
     
  • Software operating systems supported
    		•
     
  • Data access formats supported
    		•
     
  • Data validation tools provided (clean up and eliminate bad data automatically)
    		•
     
  • Prediction models and algorithms employed
    		•
     
  • Fault detection models and algorithms employed
    		•
     
  • Diagnostic decision models and algorithms employed
    		•
     
  • Prognostic life estimation models and algorithms employed
    		•
     
  • Ability to integrate in-house proprietary algorithms and software
    		•
     
  • Number of variables that can be processed
    		•

     

    Third Generation PHM Software

    Many PHM solutions have been developed to address the increasing demand for condition based maintenance and equipment failure prediction tools. Most of these software products are based on First or Second generation PHM technology and are difficult to install and calibrate and suffer from high rates of false alarms. In conjunction with NASA and the Departments of Defense, Energy, and Homeland Security, Expert Microsystems has developed and patented Third generation PHM software that is now available for commercial applications. Contact Expert Microsystems to learn more about this innovative and revolutionary software suite.