Presentation by Vlad Bacalu, CMRP, CAMA, SMRP Past Chair
The presentation, titled Artificial Intelligence & Increasing Reliability of Assets, first covers the evolution of maintenance practices as they apply to facility management, and then discusses the application of artificial intelligence and predictive analytics to transform vast amounts of data into information – empowering managers to make strategic decisions regarding their assets.
The Maintenance 4.0 concept is used when discussing the evolution of maintenance. This concept looks at the evolution of maintenance in the same way as the well-known Industry 4.0 concept deals with the industrial revolutions. The four components of the Maintenance 4.0 concept are: reactive maintenance with a run to failure strategy, time or calendar based preventive maintenance, condition based maintenance activities based on the implementation of IoT sensors and predictive maintenance technologies. The fourth element is where the assets are interconnected and predictive analytics are used to analyze the assets as a whole and make appropriate maintenance decisions. In this fourth phase, assets can self-diagnose, call for maintenance and even order spare parts.
The second part of the presentation will address what artificial intelligence is today, giving a few examples of commonly used AI applications, and will focus on the application of artificial intelligence in maintenance, specifically in the predictive maintenance area. The benefits of augmenting predictive maintenance with AI when dealing with assets are: ensuring peak performance of the assets, reducing maintenance costs by performing maintenance when needed, extending the useful life of the assets while reducing asset downtime.
In closing we will review some examples on how artificial intelligence and predictive analytics are changing the tasks we do to monitor and maintain our assets.
By using AI to analyze the condition of an asset, it allows AI to dramatically improve the availability of critical assets, as well as lower overall maintenance costs. AI creates meta learning data models out of months or years of captured equipment data, and it is up to 300% more accurate than human terms.