<aside> 🔖

Contents

</aside>

Project Description

This project involved the application of cubic and quadratic splines for shape optimization, specifically designing a concept car case for a vehicle. The project analyzed how splines can be used to achieve a smooth and accurate shape by comparing different spline types and evaluating their performance in various interpolation scenarios. The study aimed to determine which spline method (cubic or quadratic) is optimal for achieving precise and efficient interpolation in the design process.

Methodology

A cubic spline is a piecewise function composed of third-order polynomials, each defined between two consecutive nodes or data points. The cubic spline equation used in this project was:

$$ S_i(x) = (z_i / 6h_i) * (t_{i+1} - x)^3 + (z_{i+1} / 6h_i) * (x - t_i)^3 + ((y_{i+1} / h_i) - (z_{i+1} * h_i / 6)) * (x - t_i) + ((y_i / h_i) - (z_i * h_i / 6)) * (t_{i+1} - x) $$

MATLAB was used to determine the spline coefficients $z_i$ using Gaussian elimination.

On the other hand, a tension spline modifies the traditional spline by adding a tension parameter (τ) which controls the sharpness of the spline curve. The tension spline equation is:

$$ f(x) = {z_i * sinh[τ * (t_{i+1} - x)] + z_{i+1} * sinh[τ * (x - t_i)]} / [τ^2 * sinh(τ * h_i)] + ((y_i - z_i / τ^2) * (t_{i+1} - x)) / h_i + ((y_{i+1} - z_{i+1} / τ^2) * (x - t_i)) / h_i $$

The tension parameter was varied to observe how different values affect the spline’s shape and sharpness.

The project compared cubic splines with quadratic splines to evaluate which method provides better accuracy and smoothness. Quadratic splines were generated for the same set of nodes and analyzed for their ability to replicate the vehicle's shape.

Results

Discussion

The project demonstrated that cubic splines are better suited for smooth shape optimization compared to quadratic splines, which fail to capture the necessary detail, especially at higher node counts. Additionally, tension splines provide flexibility in controlling the sharpness of the curve, making them highly applicable for specific design needs. However, increasing the node count and tension parameter comes at the cost of computational efficiency, which must be considered in real-world applications.

Gallery

Cubic Splines Interpolation for the 41 points

Cubic Splines Interpolation for the 41 points

Quadratic spline with 41 Points

Quadratic spline with 41 Points