A large cement-plant induced-draft fan, a shutdown's worth of balancing that never took, and the coast-down that found the real culprit.
A boiler induced-draft (ID) fan on a cement plant was shaking hard enough to threaten a shutdown — 7.5 to 10 mm/s at the bearing housings, right across the 50% to 80% speed band it ran in. The obvious suspect was a dirty or unbalanced wheel, so the plant did all the obvious things. None of them worked, because the wheel was never the problem.
Inside a planned shutdown, the team worked through the full mechanical checklist. Both drive-end and non-drive-end bearings and their plummer blocks were changed. Coupling bushes were replaced. The impeller was cleaned and balanced on site. The shaft was ultrasonically tested for cracks. The VFD was checked, and every bolt, component and sub-assembly was checked for looseness.
The fan came back and still read 7.5 to 10 mm/s. That persistence is itself the diagnosis. When a fan has been re-bearinged, cleaned, balanced and tightened, and the once-per-revolution (1x) vibration still will not come down, you are no longer looking at the rotor — you are looking at what the rotor is bolted to.
Imbalance is a force that rotates with the wheel at 1x. You measure it, add or remove weight, and it goes away. Resonance is different in kind: the supporting structure has a natural frequency, and when an excitation — here the 1x running speed — coincides with it over part of the operating range, the structure amplifies a force that may already be within spec. On this fan the elevated concrete pedestal's natural frequency sat inside the running band. Balancing a resonant machine chases a ghost: the wheel is fine, and the reading refuses to follow the correction.
The mistake is expensive because it is a loop — balance, still high, re-balance, still high. Before you touch balance weights, run cheap tests that tell resonance and imbalance apart:
| Test | Imbalance | Structural resonance |
|---|---|---|
| 1x after a clean balance | Drops to spec and stays there | Barely changes — the clearest tell |
| Coast-down (cut power, watch it slow) | Amplitude falls smoothly as speed drops | Spikes sharply at one speed, then falls away |
| Bump / impact test (fan stopped) | Structure rings well clear of running speed | Natural frequency lands on or beside running speed |
| Phase (wheel vs. pedestal) | Pedestal barely moves; readings uniform | Large phase shift through the band; the base amplifies |
| VFD speed sweep | Vibration tracks speed smoothly | One narrow speed band lights up |
The single strongest clue is the top row: a 1x that will not come down after a genuine balance. Confirm it with a coast-down — imbalance force scales with speed squared and fades smoothly as the fan slows, whereas a resonance flares at one speed and then releases. A bump test on the stopped machine finds the structure's natural frequency directly; if it sits on the running speed, you have your answer without a single balance weight.
Once resonance was identified as the driver, the correction targeted the support, not the rotor: the drive-end pedestal was strengthened with an 18 mm MS plate welded to the existing channel support, raising the support's natural frequency clear of the running range. At 522 and 531 rpm the readings fell to 1.4 to 3.6 mm/s in auto cycle — well inside the ISO 20816 assessment. Just as telling, the readings were now steady. A number that will not settle is a resonance signature; a stable, low reading says the structure is finally stiff enough to stop amplifying.
Every one of these tests is cheaper before handover than after a plant trip. Take the vibration baseline at both bearing housings (drive and non-drive end), in horizontal, vertical and axial directions, on the housing itself — never the base frame. Assess it against the ISO 20816 zones: a new fan belongs in Zone A or B, and a Zone C reading at commissioning is an installation problem to be found and fixed, not a “monitor it later” note. On variable-speed boiler ID and forced-draft (FD) fans, sweep the full speed range once and watch for any speed where vibration jumps — that is a resonance. Program it out as a skip band and record it in the commissioning report. Two minutes on the day beats a shutdown in the year.
Talk to us about fan vibration diagnosis →
Jitamitra Electro Engineering · Fan-engineering notes, written for the engineer.
Sources & basis. Based on a real Jitamitra field-service engagement — a large boiler ID fan at a cement plant vibrating at 7.5 to 10 mm/s that was diagnosed with structural resonance (an elevated concrete pedestal whose natural frequency fell inside the running band) and corrected with a drive-end pedestal stiffener, taking it to 1.4 to 3.6 mm/s. Read alongside Jitamitra's internal fan-commissioning checklist and site-commissioning training module. Customer and job identifiers withheld; vibration assessed to the ISO 20816 method, with numeric zone limits held to the equipment manual.
Flow, static, gas temperature, application — or attach a spec, GA drawing or a multi-fan schedule. Engineer to engineer.
ISO 9001:2015 quality system · performance-tested to IS 4894 / ISO 5801 / AMCA 210 method · witnessed FAT on request, at no cost.
*For our standard range, additional days required for special projects
