Genset base frame vibration: design, failure modes, fix

A genset that hums in a basement is usually not badly assembled. It is mounted on a base frame whose first bending mode sits close to the engine firing frequency, or on mounts that are too stiff to isolate the shaft unbalance. Genset base frame vibration is a two-body problem: the set on its mounts, and the frame between the mounts, and both have a frequency that has to be moved away from the engine's.

The engine tells you the number. Firing frequency is rpm × cylinders / 120, so a four-cylinder set at 1,500 rpm fires at 50 Hz and a six-cylinder at 75 Hz. The same set also excites a first-order unbalance at 25 Hz from the crankshaft rotation. Those three frequencies, 25, 50 and 75 Hz, decide the mount grade, the frame section and the gap between the base and the canopy. This guide gives the numbers for each and the failures that show up when one of them is missed.

What sets the vibration a genset base frame has to carry?

The firing frequency sets the target, the set mass sets the mount rate, and the frame span sets the mode. Get the order right and the design is a calculation rather than a trial fit.

Take a 3,000 kg set, engine, alternator, radiator and base, on four mounts. Static load per mount is 3,000 × 9.81 / 4 = 7,360 N. For a mounted natural frequency of 12 Hz, the static deflection needed is g / (2πf)² = 9,810 / (75.4)² = 1.73 mm. Each mount then needs a rate of 7,360 / 1.73 = 4,250 N/mm.

Check the isolation at 50 Hz. The frequency ratio r is 50 / 12 = 4.17, and with 5% damping the transmissibility is:

T = √[(1 + (2ζr)²) / ((1 − r²)² + (2ζr)²)] = √[1.17 / (269 + 0.17)] = 0.066

So 6.6% of the firing force reaches the floor. On a set producing 2 kN of dynamic force at the base, that is 132 N into the slab, which is quiet. Now check the 25 Hz unbalance with the same mounts: r = 25 / 12 = 2.08, and T works out at about 0.30, so 30% passes. Mounts alone cannot isolate the first order; the rotating assembly has to be balanced, and the frame has to be stiff enough that 25 Hz does not excite it.

Genset base frame vibration limits and the values we hold

The table is the working set for a fabricated base frame in 6–10 mm MS plate or channel, four to six mounts, with a canopy bolted over it. Values outside the last column change the frame design, so they need agreement before the drawing is released.

Why the frame between the mounts has its own frequency

A base frame is a beam between mounts, and a beam has a first bending mode that has nothing to do with how soft the mounts are. On an 1,800 mm span carrying 20 kg/m of frame, with a second moment of area of 1.0 × 10⁶ mm⁴, the first mode is about 50 Hz. That is exactly on the firing frequency of a four-cylinder set at 50 Hz, which is the worst place to be: the frame amplifies every firing pulse instead of resisting it.

The mode scales with the square root of stiffness, so the fix is a section change rather than a plate thickness change. Doubling the second moment of area from 1.0 × 10⁶ to 2.0 × 10⁶ mm⁴ moves the mode from 50 Hz to 50 × √2 = 71 Hz, which clears 50 Hz by 42%. A six-cylinder set at 75 Hz needs a stiffer frame again, or a shorter span.

Three rules follow:

• Keep the first mode at least 20% above the highest firing frequency the set will run at, including a 1,100 rpm low-load idle if that is a specified condition.
• Put the mounts under the frame nodes, not the antinodes. A mount at mid-span of the frame rail does nothing for the mode and adds a point load where the frame is most flexible.
• Do not hang the radiator or the exhaust silencer off the base frame unless the frame was analysed with that mass. A 60 kg silencer at the end of a 1,800 mm cantilever drops the mode by roughly 20%.

The frame also has to carry the static load without bowing into the mounts. A 100 × 50 × 6 channel at 1,800 mm between supports, with a 7,360 N load at mid-span, deflects 7,360 × 1,800³ / (48 × 210,000 × 1.0 × 10⁶) = 4.3 mm. That is not acceptable: the mount pads then sit 4 mm out of plane before the engine is even bolted down. The frame needs a cross member at the mount line so the load reaches the mounts through shear rather than through bending of a single rail.

Choosing the mount: rubber, bonded or spring

Rubber mounts are the default for sets up to about 3,000 kg and 1,500 rpm. They give 1.5–3 mm of static deflection with integral damping, they are cheap, and they fail safe: when the rubber ages, the mount collapses rather than releasing the set. Their weakness is frequency drift with temperature and age. A rubber mount that gives 12 Hz when new can give 15 Hz after five years in a hot canopy, which is why the 20% margin matters.

Spring mounts give 25–50 mm of static deflection and isolate down to 3–5 Hz, which is what a 1,000 rpm set at 33 Hz firing frequency or a large 1,500 rpm set needs. They need a viscous damper to control the starting and stopping transient, and they need a vertical restraint because a spring mount is not a structural connection. On a set with a seismic requirement, the restraint has to be designed, not improvised from a chain.

Bonded mounts, where the rubber is vulcanised to two plates, carry shear as well as compression and are the right choice where the set sees horizontal loads from a generator short circuit or a road journey. The horizontal stiffness of a bonded mount is typically 3–5 times the vertical, so the horizontal natural frequency lands higher than the vertical one, and the two have to be checked separately.

Selection rules that hold across all three:

• Load each mount within 20% of the mean. A corner mount 30% over its partner tilts the set, loads the flexible coupling and puts the alternator bearing out of line.
• Check the horizontal mode as well as the vertical. A 12 Hz vertical mount with a 40 Hz horizontal mode will transmit the 25 Hz unbalance sideways into the canopy.
• Specify the mount by rate and deflection, not by part number. A part number from one set of conditions does not transfer to a set 400 kg heavier.

The canopy interface: where isolation is usually lost

The base frame and the canopy are two structures with different masses and different modes, and bolting them together rigidly makes one structure. That is sometimes deliberate, for a transportable canopy that has to survive a lorry, but it means the mounts no longer isolate anything above the canopy joint.

Where the canopy is a separate structure, the rules are simple. The canopy frame sits on its own feet or on a gasket-isolated base, the panel joints carry a 3–5 mm gap, and every service connection crossing the gap is flexible. That includes the exhaust, the fuel line, the cable harness, the earth strap and the fire suppression pipe if one is fitted.

Where the canopy has to be rigid for transport, the mounts are chosen for the transport case and the acoustic performance is accepted as lower. That is a legitimate trade, but it should be a stated one on the drawing rather than a surprise at commissioning.

Acoustic panels add a second problem: a 1.5 mm skin with 50 mm of mineral wool has a coincidence dip in the 2–4 kHz range and a panel mode in the 25–60 Hz range. A panel bolted to a frame that vibrates at 50 Hz radiates directly, and the sound level at the canopy face can be 3–6 dB higher than the same canopy on a quiet frame. Stiffener pitch of 400 mm or less moves the panel mode up.

What Ganesh runs for genset base frames

Base frames are cut on the 12 kW fibre laser with the 40 ft × 16 ft bed, which takes a full rail and its mount pad in one pass, and formed on the 200 T, 5-axis brake with the 4,450 mm bed. Frames are welded in fixtures on the three CLOOS robotic MIG cells, with the mount pads tacked last and checked for coplanarity before the final pass; the distortion control guidance is the sequence we run.

The mount pad is the part that decides whether the set sits flat. Pads are set and tacked in the fixture before the main weld passes, so the coplanarity is built in rather than corrected afterwards, and the frame is checked on a surface plate diagonally, not just along the rails. Where a drawing calls for 0.2 mm over 200 mm, that is a fixture and sequence requirement, not a tolerance the welder can hold by eye. Coating runs through seven-tank pretreatment and the 18 ft × 14 ft powder booth, with the pads, earth studs and mount bolt holes masked so no film ends up in the load path.

The genset parts range covers base frames, canopy frames and panels, and the genset and canopy hub collects the rest. The skid base design guide covers pad flatness, weld sequence and drip trays on the same component, and the acoustic enclosure design guide goes into panel mass, louvre area and sound level, which is where the vibration problem usually becomes audible.

What to put on the base frame drawing

1. Set mass and centre of gravity, with the engine, alternator and radiator masses separately if the set is assembled later.
2. Cylinder count and running speed, which give the firing frequency.
3. Mount rate, static deflection and natural frequency, or the mount drawing with those values on it.
4. Mount positions, with the frame nodes marked and the load per mount.
5. Frame section and the first bending mode the design has to stay above.
6. Mount pad flatness and coplanarity, with the inspection method.
7. Canopy interface: gap, isolation detail and the flexible connections.

Send the set data sheet with the frame drawing through the quote request route; the mount check is a five-minute calculation once the mass and the cylinder count are on the page.

When this route is wrong

A fabricated steel base frame with rubber mounts is right for a 20–3,000 kg set on a concrete slab. It is the wrong answer in four situations.

On a suspended floor or a roof, the floor itself becomes part of the vibration path, and the mounts have to be tuned to the floor mode rather than to the set alone. That needs a floor survey and, usually, a spring mount with a damper; a rubber mount tuned to 12 Hz on a floor that moves at 8 Hz makes the problem worse.

On a set above about 500 kVA, the base is often a structural steel fabrication with a designed transport and lifting case, and the frame analysis belongs to the set manufacturer, not to the sheet metal supplier. We cut, form and weld to that design and report the pad flatness.

Where the environment is coastal or a car park, the corrosivity class is C4 per ISO 12944-2 and a powder film at 60–80 µm is thin. Hot-dip galvanising or a zinc-rich primer under the powder is the honest specification, and the mount faces still stay bare.

Where the customer wants the canopy and the base as one rigid welded box for transport, the isolation has to be built into the feet instead of the base-to-canopy joint, and the acoustic panel design has to accept the frame mode. That is a different drawing, not a note on this one.

We do not select or supply the engine, the alternator or the flexible coupling, and we do not issue the acoustic or vibration test certificate for a completed set. What we control is the frame geometry, the pad flatness, the weld quality and the coating, and those are the four things a base frame drawing can actually fix.