Belt Cover Resonance: An Unmonitored Component

The Measurement Setup

The visit was for routine vibration monitoring on a process pump that had shown slightly elevated overall vibration at its previous inspection. The camera was positioned to cover the pump and its immediate surroundings in a single frame — including the belt drive and its cover. The pump itself showed moderate, coherent motion in the amplified image, consistent with normal operation at that load point. The belt cover, in the same frame, was visibly active: cycling with an amplitude several times larger than anything observed on the pump.

The maintenance records noted repeated fastener replacements on the cover — every three months, consistently — but no connection had been drawn to the drive train. The cover had simply been treated as a maintenance item.

Frequency Identification: 2× Driven Pulley Speed

The periodicity of the cover motion was correlated against the known rotational speed of the driven pulley using the amplified video. The cover was cycling at exactly 2× the driven pulley speed — not at motor running speed, not at belt pass frequency, and not at a structural natural frequency driven by broadband excitation.

In a belt drive, a 2× pulley speed excitation most commonly indicates one of two mechanisms: pulley non-circularity or out-of-plane deformation, which generates two belt force pulses per revolution as the belt engages and releases at the two high points; or a periodic tension variation in the belt that recurs twice per revolution of the driven shaft. Both produce a cyclic radial force at the bearing at twice the rotational frequency. That force is transmitted through the bearing housing and machine frame to all attached structures, including the belt cover.

Cover Operating Deflection Shape

The operating deflection shape of the cover was clearly defined in the amplified video. One long edge remained attached to the machine frame and showed no measurable motion. The opposite long edge — where the fasteners were positioned — was swinging up and down at 2× pulley speed. The cover was rotating as a rigid body about the fixed edge, with maximum displacement occurring at the free edge.

This ODS directly accounts for the fastener failure pattern. The free-edge fasteners were subjected to cyclic bending with each revolution of the driven pulley. Over three months of continuous operation, that loading is sufficient to progressively loosen and eventually exhaust fasteners in that position. Re-installing the fasteners without addressing the dynamic loading restarted the same failure cycle — which is what had been happening on every maintenance interval.

Diagnostic Sequence

Full-Field Measurement and Unmonitored Components

The maintenance team was following the correct procedure: re-torquing the bolts on schedule, keeping records, escalating when the pattern persisted. The gap was observational — there was no instrument on the cover, and no way to connect what was happening to the cover between inspections to anything in the plant monitoring data.

The bearing housings on the pump showed nothing unusual. A point vibration measurement on the cover would have given an overall amplitude, but not the frequency identification relative to pulley speed, and not the spatial picture needed to connect the cover ODS to a specific mechanism at the driven pulley. That spatial picture — the whole installation simultaneously, under operating load — is what resolved the case.

Auxiliary components such as covers, guards, brackets, and pipe supports are rarely instrumented, but they carry dynamic information about the machines they are attached to. In this installation, the cover was the clearest available indicator of a drive train fault that no conventional sensor was positioned to detect.