Dynamics: Viscous Damping Models

In engineering dynamics, predicting how a vibrating structure settles is as crucial as predicting how it oscillates. Energy dissipation, or damping, determines the amplitude of resonant vibrations, the rate of decay after an impact, and the long-term fatigue life of components. While many physical mechanisms cause damping—friction, air resistance, material internal friction—the viscous damping model is the most analytically tractable cornerstone for vibration analysis, providing a foundation upon which more complex real-world behavior is often mapped.

The Viscous Damper Constitutive Law

The fundamental assumption of a linear viscous damper is that the damping force is directly proportional to the relative velocity across the damper and opposes the direction of motion. This is expressed by the constitutive law:

$$F_d=c\dot{x}$$

Here, $F_d$ is the damping force, $c$ is the viscous damping coefficient (with units of N·s/m or lbf·s/in), and $\dot{x}$ is the relative velocity. This model is often visualized as a piston moving through a perfectly Newtonian fluid, or a "dashpot." Its primary advantage is its mathematical convenience; when incorporated into the equation of motion for a single-degree-of-freedom (SDOF) system, $m\ddot{x}+c\dot{x}+kx=F(t)$, it yields linear differential equations with well-known solutions.

The character of the system's free vibration response is dictated by the damping ratio $\zeta$, a dimensionless parameter comparing the actual damping coefficient $c$ to the critical damping coefficient $c_c=2\sqrt{km}$. The three regimes are:

• Underdamped ($\zeta <1$): The system oscillates with an exponentially decaying envelope.
• Critically damped ($\zeta =1$): The system returns to equilibrium in the shortest possible time without oscillating.
• Overdamped ($\zeta >1$): The system returns to equilibrium slowly without oscillation.

For an underdamped system, the free vibration solution is $x(t)=X{e}^{-\zeta {\omega }_{n}t}\sin ({\omega }_{d}t+\varphi )$, where the damped natural frequency is ${\omega }_d={\omega }_n\sqrt{1-{\zeta }^2}$. The damping ratio $\zeta$ thus becomes the key parameter describing decay rate.

Equivalent Viscous Damping for Non-Viscous Dissipation

Real-world energy dissipation rarely follows the ideal linear velocity-dependent law. Mechanisms like Coulomb (dry) friction, aerodynamic drag, or material hysteresis exhibit nonlinear force-velocity relationships. To leverage the powerful analytical toolkit of linear vibration theory, engineers often calculate an equivalent viscous damping coefficient $c_{eq}$.

The most common method is based on equating the energy dissipated per cycle. For a system undergoing harmonic motion $x(t)=X\sin (\omega t)$, the energy $\Delta E$ dissipated per cycle by the actual mechanism is calculated or measured. For a true viscous damper, the energy dissipated per cycle is $\Delta E_{viscous}=\pi c\omega X^2$. By setting $\Delta E=\Delta E_{viscous}$, we solve for the equivalent coefficient:

$$c_{eq}=\frac{\Delta E}{\pi \omega X^2}$$

This equivalence is strictly valid only for steady-state harmonic vibration at a specific amplitude $X$ and frequency $\omega$. It is a powerful simplification for estimating resonant response amplitudes, but it does not perfectly model transient decay, as the equivalent $c_{eq}$ may itself be a function of $X$ and $\omega$.

Comparison with Structural (Hysteretic) Damping

A particularly important non-viscous model is structural damping (also called hysteretic or material damping). This model observes that for many materials, like metals under cyclic stress, the energy dissipated per cycle is often independent of frequency and more closely proportional to the square of the vibration amplitude. In this model, the damping force is modeled as being in phase with velocity but proportional to displacement, leading to a complex stiffness formulation.

The equation of motion for a harmonically excited system is often written as $m\ddot{x}+k(1+i\eta )x=F_0{e}^{i\omega t}$, where $i$ is the imaginary unit and $\eta$ is the loss factor. The complex stiffness $k(1+i\eta )$ implies the damping force is $k\eta x/\omega$, which is inversely proportional to frequency. This contrasts sharply with viscous damping, where the damping force $c\dot{x}$ is directly proportional to frequency.

Key Comparison:

• Viscous: Damping force $\propto$ velocity ($c\dot{x}$). Energy dissipated per cycle $\propto$ $\omega X^2$.
• Structural: Damping force $\propto$ displacement/$\omega$ ($k\eta x/\omega$). Energy dissipated per cycle $\propto$ $k\eta X^2$ (frequency-independent).

Structural damping is often a better model for the inherent material dissipation in structures. However, viscous damping leads to much cleaner time-domain transient analysis, which is why the equivalent viscous damping concept is used to translate a loss factor $\eta$ into an equivalent damping ratio for a specific frequency: ${\zeta }_{eq}=\eta /2$.

Measuring Damping Properties

Since damping is a system property that is difficult to predict from first principles, experimental measurement is vital. Two common methods for light damping ($\zeta <0.1$) are the logarithmic decrement (time-domain) and half-power bandwidth (frequency-domain) methods.

1. Logarithmic Decrements: For an underdamped free vibration record, the ratio of any two successive peak amplitudes decays exponentially. The logarithmic decrement $\delta$ is defined as: $$\delta =\ln \left(\frac{x_n}{x_{n+1}}\right)=\frac{2\pi \zeta }{\sqrt{1-{\zeta }^2}}$$ For small $\zeta$, this simplifies to $\delta \approx 2\pi \zeta$. One can use peaks $N$ cycles apart for improved accuracy: $\zeta \approx \frac{1}{2\pi N}\ln \left(\frac{x_n}{x_{n+N}}\right)$.

2. Half-Power Bandwidth Method: This requires the system's frequency response function (FRF) magnitude plot from forced vibration testing. At resonance, the power is maximum. The half-power points are the frequencies ${\omega }_1$ and ${\omega }_2$ where the response amplitude is $1/\sqrt{2}$ (approx. 0.707) of the resonant peak. The damping ratio is then approximated by: $$\zeta \approx \frac{{\omega }_2-{\omega }_1}{2{\omega }_n}$$ where ${\omega }_n$ is the resonant frequency. This method is excellent for identifying modal damping in multi-degree-of-freedom systems.

Selecting Appropriate Damping Models for Engineering Analysis

Choosing the right model is a balance between physical accuracy and analytical necessity. Your selection guide should follow this logic:

1. General Transient/Dynamic Response Analysis: Use viscous damping. It is the standard for time-domain integration of equations of motion (e.g., seismic analysis, impact simulations). The linear model ensures numerical stability and clarity in interpreting the damping ratio $\zeta$.

1. Steady-State Harmonic Response (Forced Vibration): If analyzing response to constant-frequency excitation (e.g., machinery on a flexible floor), and material dissipation is dominant, structural damping may be more physically accurate. However, you can often use an equivalent viscous damping coefficient calculated for the excitation frequency $\omega$: $c_{eq}=k\eta /\omega$.

1. System Identification from Tests: Match the model to the measurement method. The log decrement from a pluck test yields a viscous damping ratio. The half-power bandwidth from an FRF can be interpreted as an equivalent viscous damping ratio for that specific mode.

1. Complex Systems with Multiple Sources: For a system with both aerodynamic damping (viscous-like) and material damping, you may use a combined model. A common approach is to use a viscous model but allow the damping ratio to be frequency-dependent, informed by test data.

Common Pitfalls

1. Assuming a Single Damping Ratio Applies to All Modes: In a multi-story building or a complex component, different vibration modes (bending, torsion) engage different materials and joints. Each mode will have its own effective damping ratio. Applying the same $\zeta$ to all modes can drastically over- or under-predict response.

1. Misapplying the Half-Power Bandwidth Formula: This formula $\zeta \approx ({\omega }_2-{\omega }_1)/(2{\omega }_n)$ is only accurate for light damping ($\zeta <0.1$) and for systems with well-separated modes. Applying it to a heavily damped system or a closely-spaced mode pair gives erroneous results.

1. Confusing Damping Types in Frequency-Domain Analysis: Using a constant viscous damping coefficient $c$ in a frequency sweep analysis is incorrect if the physical mechanism is structural damping. In such an analysis, the viscous damping force $c\dot{x}=i\omega cx$ grows with frequency, which may not be physical. For frequency-domain analysis of material systems, the complex stiffness ($k(1+i\eta )$) is often more appropriate.

1. Overlooking Amplitude Dependence: The equivalent viscous damping for Coulomb friction or material yielding is amplitude-dependent ($c_{eq}\propto 1/X$). Using a $c_{eq}$ calculated for a small test vibration to predict the response to a large earthquake or shock will severely overestimate damping and underestimate response amplitudes.

Summary

• The linear viscous damper model ($F_d=c\dot{x}$) is foundational for dynamic analysis, defining the critical damping ratio $\zeta$ which governs oscillation decay and resonant amplification.
• Equivalent viscous damping allows nonlinear or non-viscous dissipation mechanisms (like friction) to be approximated by a linear model by equating energy dissipated per cycle, though it is strictly valid only for steady-state harmonic motion.
• Structural (hysteretic) damping, modeled via complex stiffness, often better represents material internal friction, as it produces a damping force independent of frequency, contrasting with the frequency-proportional force of viscous damping.
• Damping is best measured experimentally: the logarithmic decrement method extracts $\zeta$ from free-vibration decay, while the half-power bandwidth method extracts it from a forced-vibration frequency response function.
• Model selection is context-driven: use viscous damping for transient/time-domain analysis, consider structural/complex stiffness for frequency-domain analysis of material systems, and always ground your chosen damping ratio or loss factor in physical understanding or test data.