FDanzi/Clustering-Composite-Materials
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
Repository files navigation
=======================================================================
Cracking Composite Design: Trace-Driven Scaling for Cluster-Based
Universal Design
=======================================================================
Author:
Francesco Danzi
Reference:
F. Danzi, "Cracking Composite Design: Trace-Driven Scaling for
Cluster-Based Universal Design," Composite Structures, 2025.
DOI: [insert DOI once published]
-----------------------------------------------------------------------
Overview
-----------------------------------------------------------------------
This repository contains MATLAB scripts accompanying the paper
"Cracking Composite Design: Trace-Driven Scaling for Cluster-Based
Universal Design."
The codes implement the π-space clustering, sensitivity analysis,
and validation procedures described in the manuscript. Together,
they reproduce Tables 1–3, Figure 1, and results reported in Appendix 1.
-----------------------------------------------------------------------
Workflow
-----------------------------------------------------------------------
To reproduce the results or apply the framework to a new dataset,
run the scripts in the following order:
1. Composite_Materials_Clusters.m
- Clusters the material systems using the K-means algorithm.
- Computes the cluster centroids and plots the clusters with their
respective centroids.
- Includes all materials listed in Table 1 and reproduces the results
shown in Table 3, Figure 1, and Appendix 1.
- The number of clusters can be modified in the input section
(see the line indicated in the script).
2. davies_bouldin.m
- Computes the Davies–Bouldin Index (DBI) to evaluate the
cohesiveness and separation of clusters.
3. my_silhouette.m
- Computes the Silhouette Index to assess intra-cluster compactness
and inter-cluster separation.
4. Sensitivity Analysis (run independently)
- Sensitivity_Trace_EigenvalueNorm.m
Computes sensitivity with respect to the Eigenvalue Norm
(Section 3.2 of the manuscript).
- Sensitivity_Trace_FrobeniusNorm.m
Computes sensitivity with respect to the Frobenius Norm.
- Sensitivity_Trace_SpectralNorm.m
Computes sensitivity with respect to the Spectral Norm.
-----------------------------------------------------------------------
Output
-----------------------------------------------------------------------
The scripts generate:
- Cluster plots and centroids (corresponding to Figure 1)
- Davies–Bouldin and Silhouette indices (Table 2 metrics)
- Sensitivity results (Section 3.2 of the manuscript)
-----------------------------------------------------------------------
License
-----------------------------------------------------------------------
All scripts are released under the MIT License for research and
educational use.
-----------------------------------------------------------------------
Citation
-----------------------------------------------------------------------
If you use this code, please cite:
F. Danzi, "Cracking Composite Design: Trace-Driven Scaling for
Cluster-Based Universal Design," Composite Structures, 2025.
DOI: [insert DOI once published]
-----------------------------------------------------------------------
Repository Link
-----------------------------------------------------------------------
GitHub Repository:
https://github.com/FDanzi/Clustering-Composite-Materials
-----------------------------------------------------------------------