Modal Extraction for Random Vibration

What Is It?

Modal extraction is the first and most critical step in random vibration FEA. It produces the natural frequencies, mode shapes and modal parameters that form the basis for the random response analysis. The adequacy of the modal extraction — whether enough modes have been extracted over the right frequency range — determines the credibility of the entire random vibration result. No amount of sophistication in the random response solution can compensate for an inadequate modal basis.

Why It Matters

If significant modes are missing from the modal basis — because the extraction frequency range was too narrow or too few modes were extracted — the random vibration response will be underestimated. The missing modes' contribution to the response is simply not counted. This produces non-conservative results that may lead to under-design and in-service failure. Modal extraction adequacy is the first thing to check when reviewing a random vibration analysis.

The modal basis is the foundation of random vibration FEA. Missing modes mean missing response. No amount of sophistication in the random response solution compensates for an inadequate modal basis.

Natural Frequencies

The natural frequencies from modal extraction determine which modes will be excited by the input PSD. Modes with natural frequencies within the PSD frequency range are directly excited. Modes outside the range may still contribute through their tails, but less strongly. The natural frequencies must be accurately predicted — they determine where the response PSD peaks occur. A 5% error in natural frequency can move a resonance peak away from or into a PSD peak, dramatically changing the response.

• Natural frequencies determine which modes are excited by the PSD
• Modes within the PSD frequency range are directly excited
• Natural frequency accuracy is critical — 5% error can dramatically change response
• Frequency prediction depends on mass and stiffness modelling accuracy

Effective Modal Mass

The effective modal mass is a measure of how much of the total structural mass participates in each mode. It is the key metric for assessing modal basis adequacy. The cumulative effective mass — the sum of effective masses of all extracted modes — should approach the total mass in the excitation direction. If it falls short, significant modes are missing. A common criterion is cumulative effective mass > 90% of the total mass in each excitation direction.

Effective modal mass:

m_eff,n  =  Γ_n² · (φ_n^T · M · φ_n)

where:
Γ_n  =  participation factor for mode n
φ_n  =  mode shape for mode n
M    =  mass matrix

Cumulative effective mass:
  Σ m_eff,n  →  should approach total mass

Criterion:  Σ m_eff,n  >  90% of total mass
in each excitation direction

Check cumulative effective mass in each excitation direction. If it falls short of ~90% of the total mass, significant modes are missing and the response will be underestimated.

Mode Count

The number of modes to extract depends on the frequency range and the modal density. For a structure with many modes in the PSD frequency range, a large number of modes may be needed. The mode count should be determined by the effective mass criterion, not by an arbitrary default. If extracting more modes increases the cumulative effective mass significantly, more modes are needed. If the effective mass converges, the mode count is adequate.

• Mode count determined by effective mass criterion, not arbitrary default
• Extract modes until cumulative effective mass converges (>90%)
• For structures with high modal density, many modes may be needed
• Check convergence — if effective mass is still increasing, extract more modes

Frequency Coverage

The modal extraction frequency range must cover the full PSD input frequency range. If the PSD extends to 2000 Hz, modes up to at least 2000 Hz must be extracted. In some cases, modes above the PSD range may be needed for adequate effective mass. The extraction range should be set to cover the PSD range plus a margin — typically 1.2-1.5 times the upper PSD frequency — to ensure adequate mass representation.

Frequency coverage:

Modal extraction range:
  f_max,extract  ≥  1.2 × f_max,PSD

Example:
  PSD range:  20-2000 Hz
  Extract modes up to:  ≥ 2400 Hz

Check: cumulative effective mass at f_max,extract
  should be > 90% in excitation direction

Local Modes

Local modes — modes where only a small part of the structure deforms significantly — can be important for random vibration even if they have low effective mass. A local mode at a stress concentration, a bracket, a panel or a component can produce high local stress even though it contributes little to the global effective mass. Local modes should be checked at critical locations — if a local mode falls within the PSD frequency range, it should be included in the analysis even if the global effective mass criterion is met.

Local modes can produce high local stress even with low effective mass. Check for local modes at critical locations — stress concentrations, brackets, panels. Include them in the analysis even if the global effective mass criterion is met.

Truncation

Modal truncation — omitting modes above a cut-off frequency — is always present because it is impractical to extract an infinite number of modes. The question is whether the truncation is adequate. For low-frequency response (displacement, acceleration at low frequency), truncation above the PSD range is usually acceptable. For high-frequency response (stress at stress concentrations, high-frequency acceleration), truncated modes may contribute significantly. Residual mass or missing-mass correction can approximate the contribution of omitted modes.

Key Takeaways

• Modal extraction is the most critical step — missing modes mean missing response
• Effective modal mass is the key adequacy metric — target >90% of total mass in each direction
• Mode count and frequency range determined by effective mass convergence, not arbitrary defaults
• Local modes can produce high local stress even with low effective mass — check critical locations
• Modal truncation is always present — assess adequacy based on the response quantity of interest