Augury’s powerful AI and machine learning algorithms were able to identify bearing wear on the water pump for a huge building materials company.

Industry – Building Materials //
Machine – Multiple cooling water pumps, cooling towers //
Fault Type – Electrical //

This Win of the Week features a leading American manufacturing company best known for its bath and kitchen plumbing products. The company’s manufacturing operations require multiple cooling water pumps and towers to ensure the normal functioning of its machines.

The company uses Machine Health for Critical Assets and is supported by Augury’s team of vibration analysts to monitor the health of many of its critical assets, including five cooling water pumps and multiple cooling towers, enabling it to predict and prevent pump failures.

Detecting higher vibrations on pumps

Recently, vibration increased in one of the company’s cooling water pumps—indicating potential issues such as misalignment, bearing failure, looseness, or pipe resonance. Augury’s powerful AI and machine learning algorithms were able to identify bearing wear on the water pump. An Augury vibration analyst informed the onsite team of the pump’s bearing wear status and escalated the pump’s status to Danger for Bearing Wear and Impacting—the most severe Machine Health classification—noting that the pump could fail unexpectedly due to bearing wear if not addressed.

Image 1:
Machine Health status and alerts are tracked, logged, and visible to the Augury vibration analyst and on-site teams in the Augury platform as issues are detected and addressed.

 

Augury’s system also detected a vibration anomaly on another pump at the same time as the first pump. This pump had the exact vibration signature as the first one, minus the high amplitude. As the two pumps were located close to each other, it was initially assumed that the vibration on the second pump was carry-over from the first asset. However, when Augury’s vibration analyst reviewed alerts on two additional assets, he recognized this was happening across multiple pieces of equipment. The two assets had the exact vibration signature and rose at the exact same time as the first one. He suspected it was more than just a bearing issue.

Image 2:
Augury’s algorithms analyze over 840 unique feature sets. This image depicts the velocity RMS and shows spikes in the motor velocity trends, indicating a serious electrical issue.

 

Electric faults that could lead to electrocution

Because Augury’s vibration analyst was monitoring multiple assets, he was able to compare and identify the same pattern across different pumps. The patterns helped him flag that the issue was more than just bearing wear. The vibration analyst identified a possible electrical issue that was spreading across the system and affecting all the assets. He quickly sent an urgent message to the onsite team on Friday night.

The company’s onsite team then reviewed a fifth cooling water pump for issues—a machine that was not monitored by Augury. The team discovered that the fifth pump had unbalanced phase voltages; Phase B was drawing 2.1 amps, while phase A and C were drawing 12 amps. All 5 assets were on the same electrical distribution system and displayed identical patterns of vibration increase at 120Hz. A bad power feed coming to the electric bus was subsequently identified, which impacted the entire electrical distribution system of the manufacturing unit.

Electric faults were diagnosed on all five cooling water pumps, which could have led to a possible electrocution hazard, as the machines could be experiencing a grounding issue.

By detecting the electrical issue early on and notifying the onsite team before the weekend, the company prevented a possible major electrical event, potentially saving the lives of workers on the plant floor.