Problem 1 - Concepts and Applications of Finite Element Analysis

Question

Imagine that a redesign produces small changes in \([\mathbf{M}]\) and \([\mathrm{K}] .\) Hence, the natural frequency \(\omega_{i}\) of each mode is slightly changed, by an amount \(\Delta \lambda_{i}\), where \(\lambda_{i}=\omega_{i}^{2}\). Using the Rayleigh quotient and neglecting terms of higher order, derive an expression for \(\Delta \lambda_{i}\) in terms of \(\lambda_{i},\\{\overline{\mathbf{D}}\\}_{i},[\mathbf{M}],[\Delta \mathbf{K}]\), and \([\Delta \mathbf{M}]\).

Step-by-Step Solution

Answer

The required expression for \(\Delta \lambda_{i}\) is given by \(\Delta \lambda_{i} = \frac{\overline{\mathbf{D}}_{i}^T \Delta \mathbf{K} \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T \mathbf{M} \overline{\mathbf{D}}_{i}} - \frac{\lambda_{i} \overline{\mathbf{D}}_{i}^T \Delta \mathbf{M} \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T \mathbf{M} \overline{\mathbf{D}}_{i}}\).

Step 1: Rayleigh's Quotient and Definitions

Rayleigh's quotient \(R(x)\) for a vibration problem is defined as \(R(x) = \frac{x^T \mathbf{K}x}{x^T \mathbf{M}x}\). In our case, we have a small change in \(\mathbf{M}\) and \(\mathbf{K}\) matrices which results in a change in \(\lambda_{i}\), where \(\lambda_{i} = \omega_{i}^{2}\). We also have \(\overline{\mathbf{D}}_{i}\) which presumably refers to a vector associated with the mode shape. We need to derive an expression for \(\Delta \lambda_{i}\) in terms of \(\lambda_{i}\), \(\overline{\mathbf{D}}_{i}\), \(\mathbf{M}\), \(\Delta \mathbf{K}\), and \(\Delta \mathbf{M}\). Ignore terms of higher order.

Step 2: Apply Rayleigh's Quotient For Initial Case

Applying Rayleigh's quotient for the initial system state, we have: \(\lambda_{i} = \frac{\overline{\mathbf{D}}_{i}^T \mathbf{K} \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T \mathbf{M} \overline{\mathbf{D}}_{i}}\) where \(\lambda_{i}\) is the eigenvalue of the system.

Step 3: Consider Changes in Matrices

Consider the effect of small changes \(\Delta \mathbf{K}\) and \(\Delta \mathbf{M}\) in stiffness \(\mathbf{K}\) and mass \(\mathbf{M}\) matrices respectively. With these changes, the Rayleigh quotient becomes: \(\lambda_{i} + \Delta \lambda_{i} = \frac{\overline{\mathbf{D}}_{i}^T (\mathbf{K}+ \Delta \mathbf{K}) \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T (\mathbf{M}+ \Delta \mathbf{M}) \overline{\mathbf{D}}_{i}}\).

Step 4: Derive expression for \(\Delta \lambda_{i}\)

Simplify the expression for \(\Delta \lambda_{i}\) by subtracting the original eigenvalue equation from the modified equation. Resultant equation would be: \(\Delta \lambda_{i} = \frac{\overline{\mathbf{D}}_{i}^T \Delta \mathbf{K} \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T \mathbf{M} \overline{\mathbf{D}}_{i}} - \frac{\lambda_{i} \overline{\mathbf{D}}_{i}^T \Delta \mathbf{M} \overline{\mathbf{D}}_{i}}{\overline{\mathbf{D}}_{i}^T \mathbf{M} \overline{\mathbf{D}}_{i}}\) . This neglects the term of second order, \(\overline{\mathbf{D}}_{i}^T \Delta \mathbf{M} \overline{\mathbf{D}}_{i}\times \overline{\mathbf{D}}_{i}^T \Delta \mathbf{K} \overline{\mathbf{D}}_{i}\), as instructed in the problem statement.

Key Concepts

Rayleigh Quotient

Understanding the Rayleigh quotient is essential for predicting how changes in a system's parameters, such as its mass or stiffness, influence its vibrational behavior. In the context of finite element analysis, the Rayleigh quotient, denoted as R(x), provides a simple yet powerful way to approximate natural frequencies (or eigenvalues) of a structure. The Rayleigh quotient is formulated as the ratio of the potential energy to the kinetic energy of a mode shape.
In mathematical terms, for a given mode shape vector \( \overline{\mathbf{D}}_{i} \), the Rayleigh quotient is defined by \( R(x) = \frac{x^T \mathbf{K}x}{x^T \mathbf{M}x} \) where \( \mathbf{K} \) and \( \mathbf{M} \) are the structure's stiffness and mass matrices respectively. The beauty of this quotient lies in its ability to approximate the eigenvalue \( \lambda_i \) by using assumed mode shapes. This becomes particularly handy for analyzing the impact of small changes in the structure, leading to variations in natural frequencies.

Eigenvalue Calculation

When we talk about eigenvalue calculation in the realm of vibration analysis, we refer to the process of determining the squares of the natural frequencies \( \omega_i^2 \) of a system. These eigenvalues \( \lambda_i \) correspond to the characteristic frequencies at which the structure prefers to vibrate, and each is associated with a specific mode shape.
Finite element software employs computational algorithms to solve the generalized eigenvalue problem of the form \(\mathbf{K} - \lambda \mathbf{M} = 0\), where \(\mathbf{K}\) and \(\mathbf{M}\) are again the stiffness and mass matrices. The resulting eigenvalues represent the potential resonant frequencies, which are crucial in predicting dynamic behavior. Calculating the change in these eigenvalues, expressed as \( \Delta \lambda_i \), involves understanding how modifications to the system's mass and stiffness matrices influence the natural frequency, which we approached through the Rayleigh quotient in the given problem.

Vibration Analysis

Vibration analysis is a critical part of structural engineering, allowing us to assess how a structure reacts under various dynamic loads. It involves studying how the energies and dynamics distribute across the structure, which can affect its longevity and performance.
The key to successful vibration analysis is pinpointing the structure's natural frequencies and understanding how these frequencies change with modifications to the system. By performing such analysis, engineers can prevent resonance, reduce excessive deflections or stresses, and ensure the structural integrity under dynamic conditions. This analysis is highly dependent on both the stiffness and mass properties, which govern the structure's dynamic response.

Stiffness Matrix

The stiffness matrix \( \mathbf{K} \) is a cornerstone concept in finite element analysis. It represents the stiffness characteristics of a system, indicating the relationship between nodal forces and displacements. In essence, this matrix provides a means to quantify how resistant a structure is to deformation under applied forces.
Each entry in the stiffness matrix gives the force at a node due to a unit displacement at another node while all other nodes are held fixed. The collective properties of these entries enable us to model and analyze complex structures for various loading conditions. A system's stiffness matrix is also instrumental in solving for mode shapes and natural frequencies, encapsulating how the structure's geometry and material properties influence its ability to withstand vibrations.

Mass Matrix

Parallel to the stiffness matrix, the mass matrix \( \mathbf{M} \) is fundamental in vibration analysis and finite element models. This matrix represents the distribution of mass within the structure and plays a vital role in dynamic analysis. It essentially addresses how the mass is spread across the structure, which influences how the system will respond to dynamic loads.
Understanding the mass matrix allows for accurate predictions of a structure's vibratory characteristics, particularly in relation to its natural frequencies and modes of vibration. Any change in this matrix, denoted as \( \Delta \mathbf{M} \), can alter the dynamic behavior of the system, necessitating a recalibration of sorts to predict subsequent mode shapes and frequencies.

Mode Shapes

Mode shapes are the specific patterns of vibration a structure undergoes at its natural frequencies. They are an intrinsic aspect of a structure's dynamic behavior, reflecting the deformation configuration at respective resonant frequencies.
Each mode shape is associated with a particular eigenvalue, \( \lambda_i \), and has an identifiable form, such as bending or twisting. Knowledge of these shapes is not only crucial for understanding vibration patterns but also for making informed decisions during the design and modification stages of a structure. Small changes in a structure can lead to significant changes in its mode shapes, affecting its dynamic response, and as illustrated in the given exercise, the use of the Rayleigh quotient facilitates an analysis of how these changes influence the natural frequencies.