What is a Frequency Response Function (FRF)?
🙋
What exactly is a Frequency Response Function? I see the simulator plots "Amplitude" and "Phase" against "Frequency", but what does it actually tell me?
🎓
Basically, an FRF is a system's ID card for vibration. It tells you how a structure, like a car chassis or a building floor, responds when you poke it at different frequencies. In practice, if you apply a force at a certain frequency, the FRF predicts how much it will shake and whether it lags behind the force. Try moving the "Natural Frequency fn₁" slider above—you'll see the big peak in the amplitude plot shift left or right. That peak is the resonance, where the system shakes the most.
🙋
Wait, really? So the "Damping ratio ζ₁" controls how tall and sharp that peak is? What's the "Q factor" that's calculated?
🎓
Exactly! The damping ratio is like the system's shock absorber. A low ζ (like 0.01) means very little damping, so the resonance peak is extremely tall and narrow—that's a high "Quality" or Q factor. Q is basically the "sharpness" of the resonance. The simulator calculates it as Q = 1 / (2ζ). Try setting ζ₁ to 0.05 and see Q jump to 10. Then, drag it to 0.5 and watch the peak flatten and Q drop to 1.
🙋
That makes sense! But what about the second set of parameters? When I add the second natural frequency (fn₂), I get a second, smaller hump in the plot. What's happening there?
🎓
Great observation! Real structures rarely have just one resonance. They have many modes of vibration. The second set of parameters lets you model a 2-degree-of-freedom system. The "Mass ratio m₂/m₁" controls how strong that second mode is. A common case is a cantilever beam: its first bending mode is at a low frequency (fn₁), and its second bending mode is at a higher frequency (fn₂). Play with the frequency separation—set fn₂ to be 3 times fn₁—and you'll see two distinct peaks. This is crucial for predicting how a structure will behave under complex vibrations, like an airplane wing in turbulence.
Physical Model & Key Equations
The core model is a single-degree-of-freedom (SDOF) damped harmonic oscillator. The receptance FRF, H(ω), describes displacement output per unit force input as a function of excitation frequency ω.
H(ω) = (1/k) / (1 - r2 + 2jζr)
Where:
• k is the static stiffness (N/m).
• r = ω / ωn is the frequency ratio.
• ωn = 2πfn is the natural angular frequency (rad/s).
• ζ is the damping ratio (dimensionless).
• j is the imaginary unit. The denominator's imaginary part, 2jζr, is what creates the phase lag.
From the FRF, we derive key dynamic properties that describe the resonance behavior.
Q = 1 / (2ζ), Δf = fn · 2ζ
Where:
• Q: Quality factor. A higher Q means a sharper, more pronounced resonance peak.
• Δf: Half-power bandwidth. This is the width of the resonance peak (in Hz) measured at an amplitude of 1/√2 of the peak value. It's a direct measure of damping in the frequency domain. In the simulator, you can see how increasing ζ widens Δf and lowers the peak amplitude.
Real-World Applications
Structural Health Monitoring: Engineers regularly measure the FRF of bridges or wind turbine blades. A shift in the natural frequency or a change in damping can indicate developing cracks or fatigue, allowing for maintenance before failure.
Automotive NVH (Noise, Vibration, Harshness): Car manufacturers use FRF analysis to find and modify resonances in body panels, steering columns, and exhaust systems. For instance, they might add damping material to suppress a resonance at the engine's idle frequency, making the cabin quieter.
Aerospace Flutter Prediction: Aircraft wings and control surfaces are tested extensively for their FRF. Engineers must ensure that resonant frequencies do not align with excitation frequencies from airflow or engine rotation to avoid catastrophic flutter vibrations.
Consumer Electronics Design: The design of smartphones and hard drives heavily relies on FRF analysis. Engineers ensure the natural frequencies of internal components and the casing are away from the operating frequencies of motors or speakers to prevent annoying buzzes and potential damage.
Common Misconceptions and Points to Note
First, do not assume there is only one type of FRF. In reality, multiple FRFs are defined based on the input-output pair, such as displacement/force (receptance), velocity/force (mobility), and acceleration/force (inertance). The simulator shows the displacement-based "receptance." For instance, when using an accelerometer in a vibration test, you are dealing with inertance, which has different characteristics, such as more noticeable noise in the low-frequency range. Next, the damping ratio ζ and the resonance peak height do not have a simple proportional relationship. Even if ζ doubles from 0.1 to 0.2, the peak height decreases by slightly more than half. Specifically, the peak amplitude is approximately 50 (1/(2ζ)) for ζ=0.01 and about 5 for ζ=0.1. Finally, be careful when setting the "mass ratio" or "coupling spring" in a 2-degree-of-freedom system. For example, if you set an extremely large mass ratio (e.g., m2/m1=100), the lower resonance mode becomes a "local vibration" state where essentially only the smaller mass (m1) vibrates, while in the higher mode, the two masses vibrate in opposite phases. When setting parameters in practical work, you must accurately estimate the masses and stiffnesses of the physical components you are considering; otherwise, you risk analyzing modes that are completely different from reality.
Worked Example
Steel cantilever beam, length 0.5 m, cross-section 50×10 mm, E=200 GPa: first natural frequency fn1=42.3 Hz, damping ratio zeta1=0.05 (5% critical), mass m1=2.4 kg. Calculator yields Q-factor=10, half-power bandwidth=4.2 Hz, resonance peak=24 dB at 42.3 Hz. Adding second mode fn2=118 Hz (zeta2=0.08) reveals anti-resonance notch at 65 Hz, critical for vibration isolation tuning.