Instantly compute matrix addition, subtraction, multiplication, determinant, inverse, transpose, eigenvalues/eigenvectors and solve Ax=b — all in your browser. 2×2 to 4×4 support with CAE applications.
Input Settings
Matrix Size
Matrix A
Matrix B (for add/subtract/multiply)
Vector b (for Ax=b)
Operation
Scalar value k
Presets
Results
—
det(A)
—
Rank
—
Trace
—
Frobenius Norm
Press "Calculate" to see the result.
CAE Connection
FEM modal analysis solves the generalized eigenvalue problem (K−λM)v=0 for stiffness K and mass M matrices. Principal stresses are eigenvalues of the stress tensor (symmetric matrix). Buckling analysis is formulated as an eigenvalue problem of the geometric stiffness matrix K_σ.
Theory & Key Formulas
Eigenvalue problem: $\det(A - \lambda I) = 0$ → roots of characteristic polynomial are eigenvalues
Trace equals sum of eigenvalues: $\mathrm{tr}(A) = \sum_i \lambda_i$
What are Eigenvalues and Eigenvectors?
🙋
What exactly is an eigenvalue? The definition $A\vec{v}= \lambda\vec{v}$ looks simple, but what does it mean?
🎓
Basically, it's a special scaling relationship. When you multiply a matrix $A$ by its eigenvector $\vec{v}$, the result is just the same vector, scaled by a number $\lambda$ (the eigenvalue). The direction of $\vec{v}$ doesn't change. Try it in the simulator above: set the operation to "Eigenvalues/Eigenvectors", input a simple 2x2 matrix, and see the output vectors.
🙋
Wait, really? So for a given matrix, there are only a few special directions that don't get twisted? How do we find them?
🎓
Exactly! We find them by solving $\det(A - \lambda I) = 0$, which is called the characteristic equation. This gives you the eigenvalues $\lambda$. Then, for each $\lambda$, you solve $(A - \lambda I)\vec{v}= 0$ to find the corresponding eigenvector. In the simulator, when you click "Calculate", it's solving these equations behind the scenes. Try changing the "Matrix Size" from 2x2 to 3x3 to see a more complex case.
🙋
That makes sense. But why is this so important in engineering? It seems like a neat math trick.
🎓
It's foundational for understanding system behavior! For instance, in structural vibration, the eigenvalues correspond to natural frequencies squared, and the eigenvectors are the mode shapes—the patterns in which the structure naturally bends. The "CAE Connection" note mentions this: $(K - \lambda M)v = 0$. Here, $K$ is stiffness, $M$ is mass, $\lambda$ is related to frequency, and $v$ is the deformation shape.
Physical Model & Key Equations
The core eigenvalue problem defines a special scalar $\lambda$ and vector $\vec{v}$ for a matrix $A$.
$$A\vec{v}= \lambda\vec{v}$$
Where $A$ is an $n \times n$ square matrix, $\lambda$ is a scalar eigenvalue, and $\vec{v}$ (a non-zero $n \times 1$ column vector) is the corresponding eigenvector. This equation asks: "For which vectors does multiplying by $A$ only result in a simple stretch or compression?"
To solve for $\lambda$, we rearrange the equation to $(A - \lambda I)\vec{v}= 0$. For non-trivial solutions ($\vec{v} \neq 0$), the matrix $(A - \lambda I)$ must be singular, meaning its determinant is zero.
$$\det(A - \lambda I) = 0$$
This is the characteristic equation. Expanding the determinant gives an $n$-th degree polynomial in $\lambda$. The roots of this polynomial are the eigenvalues. Each eigenvalue is then substituted back to find its corresponding eigenvector.
Frequently Asked Questions
If the problem is in the form of a generalized eigenvalue problem (K−ω²M)φ=0, then yes, it can be used directly. Since this tool solves the standard eigenvalue problem Aφ=λφ, to solve (K−λM)φ=0 from the stiffness matrix K and mass matrix M, a transformation such as computing the eigenvalues of M⁻¹K is required. Please be careful when interpreting the results.
This tool is limited to square matrices of size 2×2 to 4×4. For matrices of size 5×5 or larger, please use MATLAB, Python's NumPy/SciPy, or dedicated CAE software (such as ANSYS, Abaqus, etc.).
Yes, complex eigenvalues can appear for non-symmetric matrices. In structural mechanics, complex eigenvalues arise in systems with damping or in vibration analysis of rotating bodies, where the imaginary part corresponds to the frequency and the real part corresponds to the damping rate. This tool only displays real eigenvalues for real matrices.
If the determinant is zero (singular matrix), the solution either does not exist or there are infinitely many solutions. This tool will display a warning. In such cases, please review the input values or check whether physical constraints are insufficient. In structural analysis, adding boundary conditions is effective.
Real-World Applications
Modal Analysis in Structural Engineering (FEM): This is the prime CAE application. The generalized eigenvalue problem $(K - \lambda M)v = 0$ is solved, where $K$ is the global stiffness matrix and $M$ is the mass matrix. The eigenvalues $\lambda$ relate to the structure's natural frequencies, and the eigenvectors $v$ are the mode shapes—showing how the structure vibrates at those frequencies. Engineers use this to avoid resonance.
Principal Stress Analysis: In continuum mechanics, the stress state at a point is a symmetric tensor (a 3x3 matrix). The eigenvalues of this stress tensor are the principal stresses—the maximum and minimum normal stresses. The eigenvectors point in the directions of the principal planes where shear stress is zero. This is critical for failure prediction.
Linear Buckling Analysis: Predicting when a structure (like a column or thin shell) will buckle under load is formulated as an eigenvalue problem. It involves the elastic stiffness matrix and the geometric stiffness matrix (which depends on the load). The eigenvalues represent the critical load factors, and eigenvectors show the buckling mode shapes.
Data Science & Principal Component Analysis (PCA): In data compression and feature reduction, PCA finds the eigenvectors (principal components) of the covariance matrix of a dataset. These eigenvectors point in the directions of maximum variance in the data, and the eigenvalues indicate the importance of each component.
Common Misunderstandings and Points to Note
First, note that eigenvalues are not necessarily the "frequency" itself. What you obtain from the generalized eigenvalue problem $$(K - \omega^2 M)\vec{\phi}= 0$$ is $\omega^2$ (the square of the angular frequency). The λ output by the tool corresponds to this $\omega^2$, so to find the actual frequency f, you need to calculate $f = \sqrt{\lambda} / (2\pi)$. For example, if λ=100, then f≈1.59 Hz.
Next, consider the scale of your matrix element inputs. If you enter off-diagonal elements that are extremely large (or small) compared to the diagonal elements, the numerical computation can become unstable, and the tool may fail to produce accurate eigenvalues. Similar numerical errors can occur in practical FEM software if the mesh or material constant settings are poor. Try creating a matrix where the diagonal elements are 1 but the off-diagonal elements are 1000 and see if the results become strange.
Also, are you under the impression that "eigenvectors have a magnitude of 1"? The eigenvector $\vec{v}$ is meaningful only in terms of its direction, so it's typically normalized to a length of 1 for display. But the essence is its "orientation." When you look at mode shapes in CAE, you also focus on the "relative pattern" of deformation. Check the ratios of the components in the vectors output by the tool.
What is Matrix Operations & Eigenvalue Calculator?
Matrix Operations & Eigenvalue Calculator is a fundamental topic in engineering and applied physics. This interactive simulator lets you explore the key behaviors and relationships by directly manipulating parameters and observing real-time results.
By combining numerical computation with visual feedback, the simulator bridges the gap between abstract theory and physical intuition — making it an effective learning tool for students and a rapid-verification tool for practicing engineers.