Multistage Pump - Misjudging Data Case

A pump, a bearing replacement, and vibration signals telling a story…

A few days ago, we returned a between-bearing multistage pump to service after replacing the DE-side bearing (NU314). But right after the MRT, the vibration signals began telling a very different story:

📉 In the time waveform, distinct impacts appeared at 1X rotational frequency.
🎯 More interestingly, these impacts were amplitude-modulated at around 3.13 Hz.
🔊 In the frequency spectrum, clear sidebands surrounded the rotational frequency.
🌀 In the circular plot, impacts were clearly visible — with strongly nonlinear behavior.

❓The big question was:
The bearing was brand new — so where were all these impacts and modulation coming from?

---
✅ But just one day later, after 24 hours of continuous operation and a new set of data:

❌ No impacts
❌ No sidebands
❌ No modulation

📊 The signals had fully returned to normal.

---
🔬 Our interpretation? This transient behavior could have come from a few potential scenarios:

1️⃣ Imperfect axial float adjustment:
With NU314 bearings, if axial clearance or shaft positioning isn't set correctly during assembly, it can cause initial dynamic imbalance — which stabilizes after components warm up and settle.

2️⃣ Preload or mounting stress in the bearing:
If the bearing was installed with excessive radial or axial stress, it may create transient impacts or modulation — which disappear as internal stresses relax during operation.

3️⃣ Settling of coupling or base mounts:
Slight misalignment in shimming or coupling can generate cyclic forces during startup — but stabilize after a few hours of operation.

4️⃣ Transient behaviors like cavitation or trapped air:
Temporary flow instabilities can lead to modal excitation or soft contact — disappearing once the flow becomes fully stable.

---
🔄 But what about that 3.13 Hz frequency?

It was most likely modulation caused by axial rotor oscillations or internal transient pressure variations — disappearing as the system reached dynamic equilibrium.

---
🛠️ Key takeaways from this case?

✅ A new bearing doesn't mean it's faulty — sometimes it’s just new.
✅ Not every peak is a defect — sometimes it’s just an unsettled setup.
✅ Most importantly:
In condition monitoring, rushing = misjudging.

🧠 A professional analyst knows that sometimes, the best response to a suspicious signal is patience and re-monitoring under more stable conditions.

---
📌 When the data seems to be screaming...
...the truth might be quietly waiting just a few hours ahead.