Pump Bracket Resonance After Strainer Relocation
A 75 kW vertical inline pump operates at 1,780 rpm on a steel skid in a water treatment plant. A 0.9 m cantilevered bracket made from a 150 × 150 × 10 mm square tube supports a 180 kg Y-strainer and piping. After startup, vibration at the bracket measures 11.2 mm/s rms, above the 7.1 mm/s alert limit in the plant’s ISO 10816-based program. A bump test gives a first natural frequency of 27.5 Hz in the vertical direction. The original design calculation assumed 32 Hz, before the strainer was relocated onto the bracket. The pump 1× running speed is 29.7 Hz. The bracket is bolted to a 200 mm concrete pedestal with four M16 grade 8.8 anchors; no cracks are visible, but the anchor nuts have begun to loosen. Operations wants to keep the pump running until the next shutdown in six weeks, and maintenance proposes adding a second strainer in parallel on the same bracket.
1. Assess the resonance risk using the measured natural frequency and running speed. What separation margin is normally targeted, and does this installation meet it?
Read the worked answer
The measured natural frequency is 27.5 Hz and the 1× running speed is 29.7 Hz. The separation is only (29.7 − 27.5)/29.7 = 7.4%, which is well inside the usual ±20% avoidance band. The installation therefore sits close to resonance, and the 1× unbalance force is amplified.
For a single-degree-of-freedom system, the dynamic amplification is approximately 1/|1 − r²|, where r = f/f_n. Here r = 29.7/27.5 = 1.08, r² = 1.166, and the amplification is about 1/0.166 = 6.0. That is consistent with the measured 11.2 mm/s and with the loosening anchors, which indicate cyclic loading.
The normal target is to keep the natural frequency at least 20% above or below the running speed. The lower limit would be 0.8 × 29.7 = 23.8 Hz and the upper limit would be 1.2 × 29.7 = 35.6 Hz. The 27.5 Hz natural frequency is between those limits, so the bracket does not meet the separation criterion. The situation is more serious because the pump may operate over a speed range and because 2× or vane-pass frequencies may also be present.
2. Estimate the stiffness increase required to move the natural frequency to at least 20% above running speed, and evaluate the maintenance proposal to add a second strainer.
Read the worked answer
For a fixed mass, natural frequency is proportional to the square root of stiffness: f_n ∝ √k. To reach 35.6 Hz from 27.5 Hz, the required stiffness ratio is (35.6/27.5)² = 1.68. The bracket therefore needs about a 68% increase in stiffness, before accounting for the added mass of any reinforcement. A practical fix is to add a gusset or knee brace from the bracket to the pedestal, increase the tube section, or shorten the cantilever; the design should be checked by a structural engineer and the new natural frequency confirmed by bump test.
Adding a second 180 kg strainer is not a good remedy. If the bracket stiffness stays the same and the supported mass roughly doubles, the natural frequency falls to about 27.5 × √(180/360) = 19.4 Hz. That is below 23.8 Hz, so it is outside the ±20% band on the low side at running speed, but the pump must pass through 19.4 Hz during startup and shutdown. The added mass also doubles the static load and increases the fatigue stress range on the bracket and anchors. The proposal may reduce vibration at running speed but creates a transient resonance and higher anchor demand, so it should not be accepted as the permanent fix.
3. What immediate actions are appropriate for the six-week run, and what should be checked on the anchors?
Read the worked answer
The pump should not simply be left running at 11.2 mm/s with loosening anchors. Immediate actions should include daily vibration monitoring at the bracket and anchor locations, a check of pump-to-motor alignment and balance, and confirmation that the pump is operating at the intended speed. If a variable-speed drive is available, the speed should be moved away from the resonant band, but the entire operating range must be checked because the natural frequency is close to the normal speed. Temporary stiffening, such as a clamped brace, may be used only if engineered and inspected.
The anchors should be checked for preload torque, elongation, thread damage, and evidence of fatigue. The anchor nuts loosening under vibration suggest that the preload has been lost or that the dynamic load is exceeding the joint’s capacity. The embedment depth, edge distance, and concrete condition should be verified against the anchor manufacturer’s data. If the anchors cannot maintain preload, the bracket should be shut down and repaired rather than retightened repeatedly. A vibration limit and a shutdown criterion should be written into the temporary operating procedure, and the permanent repair should include a revised dynamic analysis and a post-repair bump test.