Why move from condition monitoring to predictive maintenance? – Part 2

By and Craig Truempi | 3 minute read | April 15, 2019

Learn how to make the most of your manufacturing assets at the Maximo Academy, part of the IBM IoT Exchange.

Several weeks ago, I wrote a blog in which I discussed the history of condition monitoring and the differences between condition monitoring and predictive maintenance. You’ll recall that we left off introducing the next evolution of maintenance: condition monitoring 4.0. The most current advancements in predictive maintenance require automation of the analysis process using AI models. This is a practice known as asset performance management (APM). In this post, I will highlight the value of APM using an example from an industrial manufacturing plant.

Condition Monitoring 4.0 – Industrial Artificial Intelligence

New and rich data streams from ultrasonic sensors, edge processing, and cloud connectivity are enabling these AI models. We can detect pre-existing conditions without the time, cost and frustration of baselining. Established AI models allow 90 percent or more of industrial assets to have alerts accurately established upon initial deployment. These models can amass leading indicators with sensors and connectivity.  They can also automatically detect patterns among various weak indicators.

The final step to realize success in any predictive maintenance program is to translate the alert or condition to a very specific or “prescriptive” maintenance task. For this application, a second AI model deploys to translates the alert and severity to a prioritized maintenance task. This is then emailed or integrated with an existing computerized maintenance management systems (CMMS), such as IBM Maximo®, for planning, scheduling and managing maintenance work orders. This AI model relies on metadata about the asset type and context of its service criticality to establish the  work order. It reminds me of the catch phrase is “It’s not rocket science, we’re just automating the mundane tasks”. This frees your people for higher level thinking. Focus on the right work or assets and do fewer things but at a higher standard by examining criticality and workflow processes.

A 60-day predictive maintenance example case study on a 17-year old legacy piece of equipment

Figure 1.  Generate work order and task based on alert

The deployment of ultrasonic/vibration sensors on industrial equipment yields powerful results. In July 2018, they were used to monitor a rooftop make-up air on the top of a Midwest industrial manufacturing plant. The sensors were battery-powered and easily connected to the bearings using a two-rail magnetic mount. With existing cellular, cloud and smartphone infrastructure, the hardware and system were all in place. Using alert thresholds that were pre-set using a proven AI model, the alert system was in place. The last step was clear. Concise instruction on what to do when an alert is generated. For this a second AI engine delivers specific and customized instructions for each alert and each severity level of the alerts. This model translates the alert to a prescriptive maintenance task with an understanding of severity and prioritization (see Figure 1).


Figure 2. Manual inspection reveals problems

Achieve the full value chain of predictive maintenance

Since there is a connection from the magnet to the bearing of the machine, we can issue a minor alert. The system sends an email to address the work order. A technician inspects the bearing within one day. This leads to identification of a 1/4” gap between the shaft and the inner bearing race (see Figure 2). The bearing had been slipping on the shaft and had worn ½” of the diameter off the shaft.

They schedule a repair for 60 days out. Because of the quick action, the overall vibration levels remained normal.  The plan included a non-OEM repair selecting to use a split bearing and moving the bearing location to an undamaged location of the shaft. This completed the full value chain of predictive maintenance from monitoring to alert to prescriptive task and ultimately a planned and schedule repair that avoided any collateral damage, loss of production and reactive maintenance.

To learn more about predictive maintenance and APM, join us at IoT Exchange in Orlando, 24-26 April

We will be attending the IBM IoT Exchange in Orlando to learn about new developments in IoT and to meet other professionals who are interested in predictive maintenance. I hope you will join me and many other experts. It would be a pleasure to hear your perspective on this topic.

If your organization is looking to understand how to take steps to further their APM journey, take this assessment.

Learn more about IoT Exchange

Also, be sure to watch for our interview on Quick Bytes Live at IoT Exchange. You may also enjoy our previous THINK Quick Bytes Live interview.

For more information, visit assetscan.com

Connect with Dan Yarmoluk at dyarmoluk@atekcompanies.com

Follow Dan on Twitter and LinkedIn.


Connect with Craig Truempi  at ctruempi@atekcompanies.com

Also, follow Craig  on LinkedIn.