Part 3 – Review of Existing Online Systems and Methodologies

by | Aug 5, 2021 | Online Monitoring

In 2019, ATS contacted a number of the new online systems and assessed their capabilities. With a few exceptions, the majority of these products did not have the normal features or tools that are common in major portable systems.

Let us review the above statement and examine each point to see the real potential of the online systems.

It is true that any online system will give its owner the flexibility of data collection intervals, day or night without additional cost, especially for machines that were already identified as “bad assets”. The ability to modify and shorten data collection intervals is a huge advantage, especially when the asset cannot be scheduled for immediate repair due to production schedules.

  1. The argument against the online system is that most “bad assets” generate indicator(s) way before that critical stage. The proof of that is a successful PdM program utilizing portable systems where the data collection interval is usually a monthly activity. Therefore, if the analyst has the need to check the machine vibration data regularly right to the moment of break down, the plant culture would have to be modified and more plant wide training is needed prior to the implementation of online technology. The data can be reviewed at any time or anywhere by any trained analyst. This is certainly a big advantage of implementing an online system. Currently, if the database and vibration software is installed in a shared server, the same thing can be accomplished by a portable system. However, when new data is needed in order to check the latest status of an asset, the online system absolutely maintains superior advantage because new data can be collected without the requirement of on-site labor. Additionally, measurement specifications can be tailored on the fly to give the analyst better opportunities to pinpoint the machine fault.
  2. One of the most touted features of the current online systems is the ability of the system to warn users once the monitored asset trips an alarm. The most popular type of alarm utilized is the overall vibration where the system administrator would set overall amplitude thresholds. When the amplitude thresholds are reached, an alarm is triggered, and in most setups, emails can be sent out to a pre-determined group of people needing to be notified. However, an overall vibration alarm simply announces that the condition of the monitored asset has been changed and it does NOT provide the reasons or causes for the condition change(s). To figure out the WHY, most likely you will need to involve an analyst to look at the data or maintenance staff to look at the data of the asset in alarm. More importantly, deployment of alarms can be very tricky and often a long process. The alarm has to be just tight enough to allow the normal fluctuation of the asset amplitude, yet it has to be tight enough to be triggered by early indicators of most anomalies.
  3. There are a handful of online companies that claim to utilize AI (Artificial Intelligence) to analyze the data and their system will continuously learn as new data/events come in. These claims cannot be substantiated at this time because there is very little information disclosed by the OEM. Additionally, it is a very thin line separating AI and a Rules Based system. Either system, when designed and implemented properly, can greatly reduce human analysis error but both of them require a tremendous amount of time to get detailed information about the assets and their related process. That’s the hidden cost that has to be accounted for financially. In practice, there is not a single system that can be integrated in any factory and have it work right off the bat. Leo Tolstoy wrote in the Anna Karenina novel: “Happy families are all alike; every unhappy family is unhappy in its own way.” This observation also fits the mechanical behavior of most machines exceedingly well; and that is why regardless of what online system you choose, the need to have a skilled analyst to frequently audit and adjust the system settings is very critical for a successful implementation of any online system.
  4. Setting the alarms at ISO recommendations is another popular approach mentioned by online systems. This approach can be very effective but unfortunately, it is only effective for installation sites where they already have a successful PdM program. It is also a suitable approach for sites in highly regulated industries such as nuclear power plants, chemical or petroleum plants. These types of facilities are most likely to have their critical machines in good shape due to safety and regulations; therefore, the application of ISO standards will help to evaluate the general condition of the plant and to detect a small number of outliers that need some tweaking to be in compliance. For a large number of normal manufacturing and lightly regulated process plants, the ISO standards will result in a large number of alarms while the plant does not have the resources in work force and budget to correct the problems in such a short notice. Consequently, maintenance staff members will learn to ignore the alarms and in some cases, disable the alarms.

From a ten thousand foot view, there is no doubt that the online system is the future of vibration analysis; but for the time being, it still falls short from replacing a vibration analyst. Even for systems that claim to be AI driven, the chance for error is too big; especially, in the monitoring of very critical machines. Additionally, even if the acquisition of this newer technology is feasible, the plant still has to change its culture and procedures to ensure a successful utilization of the new technology. These topics will be addressed in the next article.

Categories

Subscribe to Our Blog

We regularly write a blog about Predictive Maintenance and the technologies used in PdM programs. Give us your email address and we’ll make sure you get a copy in your inbox each time we publish a new post.

 

We also post each blog article on Twitter and LinkedIn.

 Follow our LinkedIn Company Page      

Related Blogs

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *


The reCAPTCHA verification period has expired. Please reload the page.