Leveraging advanced FFT data gathered by wireless sensors to make continuous monitoring economically feasible

CASE STUDY

Leveraging advanced FFT data gathered by wireless sensors to make continuous monitoring economically feasible

IMI Publications

Written By IMI Publications

May 22, 2025

TotalEnergies is one of the leading oil and gas companies in the world. With 25 refineries running 24/7 across the globe, it is essential to have a complete overview of the current operational status of all their assets to maintain an efficient and safe environment.

One of the challenges is to manage the thousands of items of rotating equipment installed on each site. These assets vary from pumps, gearboxes, and fans, to static equipment with resonance issues. Historically, assets are monitored at intervals, often monthly, with manual vibration measurement to verify that an asset is correctly functioning and possibly detect a failure early.

During this snapshot, vibrations are measured over a broad spectrum and then analysed by an expert to determine whether further action is needed. While this process helps to manage unexpected downtime, it has a number of significant downsides:

  • It is extremely labour intensive;

  • Substantial gaps between snapshots may not catch all potential issues in time, risking possible catastrophic failure taking place between the checks carried out.

Total Refinery Antwerp has tested IMI's NEON Vibration Sensors to address and improve on the typical downsides of vibration measurement.

Use case

The NEON Vibration Sensor was installed on the bearing housing of the pump block within a pump installation. This device sends an RMS velocity value at 15-minute intervals, which creates excellent insight into the overall condition of the asset and, as such, is a significant upgrade from monthly monitoring.

Quote
“IMIs NEON Vibration Sensor gives us better use of our time. We don’t need to go out when the sensor indicates all levels are OK; Spending time only on assets with actual problems.” - Franky Oste, TotalEnergies Reliability Engineer
Quote

In addition to the single value, the sensor also transmits a full FFT measurement every two days (which is only advisable for short periods in these maintenance use cases) for deeper analytics.

This full-spectrum measurement allowed TotalEnergies to benefit from a complete view of the ongoing status of assets in the field and resulted in detailed information.

Soon after installation within the TotalEnergies Refinery, there was a change in the usual behaviour recorded so far, with a distinct drop in vibration levels. The evaluation of the data from the RMS velocity and acceleration, combined with the detailed information from the FFT has now been recognised as an early warning signal and therefore been integrated into the day-to-day maintenance routine for TotalEnergies Antwerp.

Comparing graphs 1 and 2, we see three apparent changes between the two graphs. The first clear difference in the FFT is a decreased amplitude at around 1100 Hz. Next, the second indicator displays the same behaviour but at a higher frequency. Finally, the last and most significant difference is seen in a big bump, a ‘Haystack’ in the spectrum between 6000 and 6666 Hz.

Two line graphs comparing vibration acceleration versus frequency. The top graph (Graph 1) shows multiple peaks with a significant increase around 6000 Hz, while the bottom graph (Graph 2) has lower amplitudes overall, with isolated peaks near 1100 Hz, 4000 Hz, and 6000 Hz, indicating a change in vibration behaviour after maintenance.

NEON’s expanded frequency range versus that of competitors made this visible. After comparing these two measurements, it was apparent that something unusual was happening. Upon closer inspection, engineers realised this lower level of acceleration was occurring immediately after relubrication. Measuring low levels of acceleration would indicate correct lubrication. What was however seen was the sharp drop of vibration levels, followed by a quick rise of the vibration data collected, pointing to ‘over-lubrication’, which could cause long-term machinery damage.

Power your operations with industrial IoT

Get guidance on the right sensors, software, and integrations for your goals.
Engineer stands with arms crossed at IMI site in US office, with colleagues and technical posters in background.