How to Prevent False Alarms

This article outlines best practices for keeping the FitMachine AI maintained to minimise the potential for false alarms.

How can it help you?

Ensuring your system is set up correctly and well maintained allows your staff to appropriately triage alerts and make suitable decisions with accurate data. Further, you will minimise the potential for false alarms and the risk of missing real, valid alarms.

How does it work?

Maintaining your system requires your attention across four main areas; installation, the running cut-off, the learning period and reporting.

Installation

Most importantly, you'll need to make sure that the FitMachines are installed correctly. This ensures that they can collect the best possible data, aren’t going to issue false alerts, and can potentially detect signs of equipment failure earlier. Installation of FitMachine in unideal locations, however, can reduce data quality and accuracy (e.g. through introduced noise into the vibration data).

Our article About FitMachine Sensor Installation outlines best practices for FitMachine installation.

The Running Cut-Off

A correctly set running cut-off [RCO] helps ensure that the FitMachine sensor is reporting at the correct intervals. If the RCO is set too high, then the sensor may miss data when your equipment is running, and if set too low, the sensor will be gathering non-running data and presenting it as running data. This can trigger alerts on the start-up of equipment and collecting lots of non-running data will drain the battery unnecessarily.

A correctly set RCO is especially important during the Learning Period, as this ensures the FitMachine is learning accurate running and condition data - otherwise, it will significantly reduce the effectiveness of the FitMachine AI and trigger false alarms down the line.

The best practice here is to review the set RCO for each asset 1 week after FitMachine installation, or a major change in the system set-up. For example,

After maintenance on the asset
After FitMachine replacements or if the FitMachine is moved, or
After major production changes.

Learn more about the running cut-off in our article here, and how to adjust the running cut-0ff here.

The Learning Period

Similar to the point about RCO, ensuring your learning data is valid will decrease the potential for false alarms. Two ways to improve validity include:

1. Restart learning if the FitMachine has been moved, and following maintenance events with significant changes.

With these events, the learning data becomes invalid, and false alarms can and likely will be triggered. Try to restart learning following these events, for instance, if a bearing or damaged roller is replaced, or structural improvements have been implemented. You do NOT need to restart learning following temporary speed changes or unbalance cleaning on fans.

2. Review the learning if there's a major operational change

Processes can change and Operations targets can too, for instance, needing to run a production line at 120% for months on end. These types of changes will make an asset vibrate in a different way and can result in a Condition change detected by the AI. If you're going to be operating in a new way for a number of months, or the process has changed permanently, restarting learning is recommended.

3. Review the learning start date if you've had FitMachine for a while

If the learning period took place several years ago, the equipment's 'normal' running conditions may have changed, even only in a minor way, due to equipment degradation or other factors. Thus, with time, the changes detected may not be an accurate reflection of the equipment's health, and false alarms may be triggered.

You can read more about the learning start date in this article here, and our article Restart Learning of a FitMachine outlines in more detail when and how to restart the learning period.

Coming Soon...

By working through this problem, we identified that we’re missing an important step; how to tell if the FitMachines have been moved. We have tools available to identify these moves in-house, however, they haven’t been exposed to the customer. We’ll be releasing a new feature that comes with Condition Alarms (Red Alerts and Yellow Advisories) which will show you the gravity vector outputs of the FitMachine around the alert time so you can quickly filter out problems. Keep an eye on the What's New Page for its release.

Reporting

Finally, the FitMachines must be reporting to be able to give you valuable data, and alert you when the equipment's behaviour changes. There are two places in the dashboard you can go to ensure your sensors are reporting:

FitMachine Admin Page: This page details each of the outages on site (if any), and provides some insight into the signal quality and battery health of each FitMachine. Learn more here.
Equipment List Page: The status column on the list page draws attention to not only condition changes, but also highlights issues with the deployment which may affect its performance and efficacy. Our article What is the FitMachine Status Column telling me to do? outlines what to do in response to an issue being identified.

What else should I look at?

Congratulations! By taking these steps, your equipment is in great shape.

Now that your system is properly maintained, an excellent next step is to ensure you fully understand the system's capabilities and how it can help you. Some significant articles we recommend you check out are:

If you have a plan for exactly how you’d like your team to use the system, let us know and we’ll provide training to suit your needs.

---

Thanks for reading! If you have any questions or concerns, please reach out to MOVUS Support here.