Managing Startup Resonance In Variable-Speed Industrial Equipment

Variable-speed industrial equipment creates a specific vibration challenge: the machine must pass through, or operate near, natural frequencies during startup, shutdown, speed changes, and process transients. A design that performs well at steady operating speed can still produce excessive vibration during run-up. The objective is therefore not simply to isolate vibration, but to manage the complete duty profile: mass and center of gravity, foundation and frame stiffness, isolator stiffness and damping, speed trajectory, environmental exposure, validation evidence, and procurement controls.

This article outlines a practical engineering and procurement framework for managing startup resonance. It is intended for OEM engineers, integrators, and technical purchasing teams. Final designs should be checked by qualified engineers against applicable standards, manufacturer data, and site conditions. The equations presented are simplified engineering aids; they do not replace multi-degree-of-freedom analysis, modal testing, or configuration-specific validation.

Why Startup Resonance Is Different in Variable-Speed Equipment

A variable-speed machine does not operate at a single forcing frequency. Motor speed, process load, drive harmonics, and control-loop dynamics produce excitation across a range of frequencies. During run-up, the machine may momentarily pass through one or more system natural frequencies. If it dwells at or near one of those frequencies, vibration amplitude can rise sharply, leading to:

• fatigue of mounts, welds, bolting, and piping;
• bearing and coupling damage;
• loosening of fasteners and premature wear;
• isolator bottoming, creep, or overheating;
• foundation cracking and structure-borne noise;
• drive-train torsional oscillation or control instability.

Steady-state transmissibility curves are therefore necessary but not sufficient. Startup resonance is a transient event. The design must consider how quickly the machine passes through critical speeds, how much damping is available, whether the control system can avoid sustained operation in resonance bands, and whether the installed system has enough travel and restraint to survive transient peaks.

A useful hierarchy for resonance control is:

1. Avoid sustained excitation at natural frequencies.
2. Shift natural frequencies by changing mass, stiffness, or mount layout.
3. Damp the system to limit amplification during pass-through.
4. Manage the speed trajectory with controlled ramps and skip bands.
5. Isolate the machine from its foundation where appropriate.
6. Monitor actual startup response and update the design or operating limits.

This hierarchy matters because no single measure is universally sufficient. A soft isolator may improve high-frequency isolation but increase rocking or travel. A stiff mount may reduce motion but transmit more vibration. A skip band may protect one speed range but create control or process penalties elsewhere. The correct combination depends on the duty profile and the consequences of resonance.

Natural Frequency, Isolation, and Damping: Correct Conditions Matter

For a single-degree-of-freedom approximation, the installed system natural frequency is:

$$
f_n = \frac{1}{2\pi}\sqrt{\frac{k}{m}}
$$

where (f_n) is in Hz, (k) is the effective stiffness in N/m, and (m) is the supported mass in kg.

For steady-state vibration isolation, the forcing frequency (f) should generally be higher than the mounted system natural frequency. For a simple undamped system, transmissibility falls below unity when:

$$
r = \frac{f}{f_n} > \sqrt{2}
$$

In practice, a ratio of (f/f_n \ge 2) to (3) is often targeted for effective isolation, subject to stability, travel, and machine requirements. However, this condition applies to steady operating speeds. During startup from zero speed, the machine must pass through (f_n) if the natural frequency lies within the speed range. A blanket statement that the natural frequency should be below the operating frequency is incomplete for variable-speed equipment: the startup pass-through must also be managed.

Force or displacement transmissibility for a viscously damped isolator can be approximated by:

$$
T = \sqrt{\frac{1 + (2\zeta r)^2}{(1-r^2)^2 + (2\zeta r)^2}}
$$

where (\zeta) is the damping ratio and (r=f/f_n). At resonance, (r \approx 1), and for low damping:

$$
T \approx \frac{1}{2\zeta}
$$

Thus, a damping ratio of 0.10 gives a steady-state harmonic amplification of roughly 5 at resonance; a ratio of 0.20 gives roughly 2.5. Higher damping reduces run-up amplification, but excessive damping can increase high-frequency transmissibility, generate heat in elastomers, and reduce isolation efficiency. Damping must therefore be selected from measured or validated data at the expected amplitude, frequency, and temperature, not from a single universal target.

It is also important to distinguish steady-state amplification from transient pass-through response. The actual peak during run-up depends on acceleration rate, dwell time, mode coupling, foundation impedance, nonlinear isolator behavior, and control dynamics. A system with adequate steady-state damping can still experience a high transient peak if it dwells near resonance or if lateral and rocking modes are not addressed.

Build a Critical-Speed and Duty-Profile Map

Before selecting isolators, define the machines operating envelope:

• minimum, maximum, and normal operating speeds;
• acceleration and deceleration rates;
• expected dwell points and process hold speeds;
• number of starts per hour and per day;
• load changes during startup and normal operation;
• motor, gearbox, coupling, and driven-equipment excitation frequencies;
• torsional natural frequencies of the drive train;
• foundation and frame natural frequencies;
• transport, installation, and seismic load cases.

For variable-speed drives, identify both mechanical and electrical excitation sources. VFD current harmonics, pulse frequencies, and control-loop dynamics can excite torsional or electrical resonances that mechanical isolation does not solve. A critical-speed map should compare speed-dependent excitation orders with the mounted systems natural frequencies. Where continuous operation near resonance is unavoidable, use skip bands, controlled ramps, or revised stiffness/mass to move the critical speed outside the operating range.

The critical-speed map should be a controlled engineering document, not a single spreadsheet tab. It should record the assumed mass, stiffness, damping, foundation conditions, excitation orders, and operating limits. When the machine configuration or process changes, the map should be reviewed.

Mass, Center of Gravity, Foundation, and Frame Stiffness

The supported mass and center of gravity determine static load distribution and rocking modes. For a simplified static load estimate, the reaction at each isolator can be written as:

$$
R_i = \frac{W}{n} + W\left(\frac{e_x x_i}{\sum x_j^2} + \frac{e_y y_i}{\sum y_j^2}\right)
$$

where (W) is the equipment weight, (n) is the number of mounts, (e_x) and (e_y) are center-of-gravity offsets from the mount centroid, and (x_i), (y_i) are mount coordinates. This expression assumes a symmetric mount layout, an origin at the mount centroid, and negligible cross-coupling. For general layouts, the static reactions should be solved from force and moment equilibrium. Dynamic load terms from modal or transient analysis must be added separately; they should not be represented by an undefined CG offset term.

A high center of gravity increases rocking and coupling between vertical and horizontal modes. Mounts should be arranged to provide stable support around the CG, avoid large overhangs, and share load predictably. The mount layout should also consider maintenance access, bolt clearances, thermal expansion, and isolator travel.

Foundation inputs matter as much as isolator selection. A flexible floor, inadequate grout, or uneven anchor points can introduce local modes and misalignment. The foundation should be stiff and massive enough for the intended machine, with proper leveling, anchoring, and isolation of adjacent structures where necessary. In some installations, the foundation impedance is a dominant factor in the installed natural frequency and should be included in the dynamic model.

Frame stiffness also affects the installed natural frequency. A flexible frame can introduce local bending modes that are not captured by a simple rigid-body model. Mounts should be placed near stiffened regions, such as ribs, bulkheads, or reinforced pads, rather than on unsupported panels. Maintenance clearance must be preserved for inspection, bolt access, thermal expansion, and isolator travel.

Selecting Isolator Load, Stiffness, Damping, and Travel

Isolator selection is a multi-degree-of-freedom problem, even when simplified calculations are used. Specify vertical, lateral, and rotational stiffness separately. For each direction, verify:

• static load capacity with an appropriate margin;
• dynamic load capacity during startup, shutdown, and fault conditions;
• natural frequency relative to steady operating speeds;
• expected amplification during run-up;
• maximum deflection and travel;
• bottoming and snubbing requirements;
• stability under CG offset and wind or process loads;
• fatigue and creep over the service life.

Travel is often overlooked. An isolator that is correctly rated for static load may still bottom out during startup resonance or transient torque. Specify maximum vertical and lateral deflection, and include snubbers or fail-safe restraints where excessive movement would damage piping, couplings, or adjacent equipment.

Stiffness and damping must be evaluated together. A very soft isolator may provide good high-frequency isolation but allow excessive motion or rocking. A very stiff isolator may reduce motion but transmit more vibration. The correct choice depends on the duty profile, foundation, and acceptable vibration limits.

For critical applications, request a stiffness matrix or direction-specific data rather than a single vertical stiffness value. Elastomeric isolators in particular may show amplitude-dependent, frequency-dependent, and temperature-dependent behavior. Metallic springs offer more predictable stiffness over a wide temperature range but generally require separate damping elements.

Material, Temperature, Corrosion, and Service Life

Isolator materials must survive the actual environment, not only the laboratory condition. Consider:

• ambient and surface temperature range;
• oil, coolant, solvent, washdown, and chemical exposure;
• humidity, salt spray, UV, and ozone;
• expected creep and aging;
• fatigue from repeated startup cycles;
• replacement access and service life.

Elastomeric isolators can provide good damping and compact installation, but their stiffness and damping vary with temperature, frequency, and amplitude. Metallic springs offer predictable stiffness and wide temperature range but require separate damping. Viscoelastic materials can be effective for specific frequency and temperature bands, but must be validated for the application.

Corrosion-resistant hardware, coatings, and material compatibility should be specified in the procurement documents. For washdown or outdoor installations, verify that the isolator will not retain water, corrode at the mounting interface, or degrade from cleaning chemicals. Maintenance instructions should include inspection intervals and replacement criteria, especially for elastomeric components exposed to heat or chemicals.

Validation: Installed Deflection, Sweep Testing, and Functional Checks

Validation should demonstrate that the installed system behaves as predicted. A practical validation sequence includes:

1. Installed deflection measurement under known static load.
2. Natural frequency check by impact test, shaker test, or controlled run-up.
3. Frequency sweep across the operating range, using order tracking if available.
4. Startup and shutdown transient tests at representative load and temperature conditions.
5. Clearance and travel checks to confirm no bottoming or interference.
6. Repeatability checks over multiple starts.
7. Vibration acceptance against applicable standards or OEM limits.

Installed static deflection is related to natural frequency by:

$$
f_n \approx \frac{15.76}{\sqrt{\delta}}
$$

where (\delta) is static deflection in millimetres and (f_n) is in Hz. This relationship is useful for a quick check, but it assumes a simple vertical mode and linear stiffness. It should not replace modal testing or manufacturer data for complex installations.

Validation evidence should be configuration-specific. Generic catalog curves are not enough for a critical startup-resonance application. Request test reports that include the test setup, load, temperature, frequency sweep, damping estimate, travel limits, and measurement uncertainty. Where the machine has a variable-speed drive, validation should include representative acceleration and deceleration ramps, not only steady-state speed points.

Illustrative Sizing Example

Consider a 2,000 kg machine supported on four isolators, with the CG approximately centered. The static load per isolator is:

$$
F_i = \frac{2000 \times 9.81}{4} \approx 4.9\ \text{kN}
$$

Suppose the target vertical mounted natural frequency is 8 Hz. The total required vertical stiffness is:

$$
k = m(2\pi f_n)^2 = 2000(2\pi \times 8)^2 \approx 5.05 \times 10^6\ \text{N/m}
$$

Per isolator, the required stiffness is approximately (1.26 \times 10^6\ \text{N/m}). The corresponding static deflection is:

$$
\delta = \frac{W}{k} = \frac{2000 \times 9.81}{5.05 \times 10^6} \approx 0.0039\ \text{m} \approx 3.9\ \text{mm}
$$

If the lowest normal operating speed corresponds to 20 Hz, then (f/f_n = 20/8 = 2.5), which is favourable for steady-state isolation. However, the machine still passes through 8 Hz during startup. With a damping ratio of 0.10, the approximate steady-state resonance amplification is about 5; with 0.20, it is about 2.5. The startup ramp, damping, and travel must therefore be validated together. If the machine must dwell near 8 Hz, the design should be changed by shifting the natural frequency, adding damping, or introducing a skip band rather than relying on steady-state isolation alone.

This example is illustrative only. Actual design must include CG offsets, foundation impedance, frame flexibility, dynamic loads, multi-directional stiffness, and torsional effects.

Commissioning Lessons and Common Pitfalls

In commissioning practice, startup-resonance problems often arise from a small number of recurring issues:

• Using only catalog vertical stiffness. Lateral, rocking, and rotational modes are frequently the actual problem.
• Ignoring the foundation and frame. A flexible base can shift the installed frequency and introduce local modes that were not modeled.
• Underestimating travel. The isolator may be adequate for static load but bottom out during a transient peak.
• Assuming damping is a fixed catalog value. Elastomer damping varies with amplitude, frequency, temperature, and age.
• Testing only at steady speed. A machine can pass steady-state acceptance while still showing excessive vibration during run-up.
• Setting skip bands without validation. Skip bands can help, but they must be checked against actual response, control stability, and process requirements.
• Treating procurement as a catalog purchase. A low-cost isolator can become expensive if it causes downtime, premature replacement, or secondary damage.

These lessons reinforce the need for a system-level review rather than an isolator-only selection process.

Procurement and Integration Checklist

Procurement teams should require suppliers to provide validation data for the actual configuration, not only catalogue ratings. The lowest-cost isolator may become expensive if it causes downtime, premature replacement, or secondary damage. A total lifecycle cost view should include installation, maintenance, spare parts, energy, and production risk.

A practical procurement and integration checklist includes:

• Duty profile: operating speed range, ramps, dwell points, starts per day, load changes, and transient cases.
• Critical-speed map: excitation orders, natural frequencies, torsional modes, foundation and frame modes, and skip-band requirements.
• Load data: static and dynamic loads, CG location, mount reactions, and load margins.
• Stiffness data: vertical, lateral, and rotational stiffness at expected amplitude, frequency, and temperature.
• Damping data: validated damping values or curves for the expected service conditions.
• Travel and restraint: maximum deflection, snubber or fail-safe requirements, and bottoming limits.
• Material compatibility: temperature, oil, coolant, washdown, UV, salt spray, and chemical exposure.
• Validation evidence: test setup, load, temperature, frequency sweep, damping estimate, travel limits, and measurement uncertainty.
• Acceptance criteria: applicable standards, OEM limits, site limits, and test conditions.
• Installation requirements: leveling, anchoring, torque values, clearance, and inspection access.
• Service life: inspection intervals, replacement criteria, spare parts, and maintenance access.
• Change control: notification and review requirements for material, stiffness, or configuration changes.

Limitations and Design-Review Notes

The methods in this article are intended for preliminary engineering and procurement guidance. They do not replace a full dynamic analysis for critical equipment. Single-degree-of-freedom equations are useful for understanding trends, but real installations are multi-degree-of-freedom systems with foundation impedance, frame flexibility, nonlinear isolators, and torsional or control-system dynamics.

A qualified engineer should review the final design against applicable standards and site conditions. Where the consequence of resonance is highsuch as loss of production, safety risk, or damage to critical equipmentmodal testing and configuration-specific validation should be required.

Conclusion

Managing startup resonance in variable-speed industrial equipment requires a system-level approach. The mounted natural frequency must be evaluated against the full speed and duty profile; steady-state isolation criteria must be supplemented by startup pass-through analysis; and isolator selection must consider load, stiffness, damping, travel, environment, and validation evidence. By combining critical-speed mapping, appropriate damping, controlled speed trajectories, robust installation details, and configuration-specific validation, engineers and purchasing teams can reduce resonance risk and improve equipment reliability.

Applicable Standards and References

• ISO 20816-1:2016, Mechanical vibration Measurement and evaluation of machine vibration Part 1: General guidelines.
• ISO 20816 series, Mechanical vibration Measurement and evaluation of machine vibration.
• ISO 21940 series, Mechanical vibration Rotor balancing.
• IEC 60034-14, Rotating electrical machines Part 14: Mechanical vibration of certain machines with shaft heights 56 mm and higher Measurement, evaluation and limits of vibration severity.
• API 610, Centrifugal Pumps for Petroleum, Petrochemical and Natural Gas Industries, current edition, where applicable.
• API 617, Axial and Centrifugal Compressors and Expander-compressors, current edition, where applicable.
• Local structural, seismic, electrical, and safety codes.
• Manufacturer installation, validation, and maintenance instructions.
• Den Hartog, J. P., Mechanical Vibrations.
• Harris, C. M., and Piersol, A. G., Harris Shock and Vibration Handbook.
• Blevins, R. D., Formulas for Natural Frequency and Mode Shape.