SimuStruct Dataset (0.1)





Thin structural plates with geometric discontinuities have wide applicability in the automotive, aerospace and maritime industries. Discontinuities such as holes, with different shapes and sizes, allow reducing the total weight of the structure, establishing joints (as bolted and /or riveted) and creating means for accessing other parts of the structure. Mechanical testing of these plates is essential.
The Simulated Structural Parts Dataset, SimuStruct, contains 1000 cases of such 2D rectangular plates with holes under load along with the measurements of respective von Mises stress.
Each sample of the dataset corresponds to a rectangular plate with six identical circular holes arranged in a rectangular pattern. The sizes and positions of the holes vary across different cases. All of the plates are modelled as being made of a steel alloy with linear elastic behaviour, characterized by a Young’s Modulus of \(E=210\) GPa and a Poisson Ratio of \(v=0.3\).
In all of the cases in the dataset, we simply support the plates on the bottom side, i.e., fixed in the y-direction, and subject them to uniaxial loading at the top with a stress of \(\sigma_1=100\) MPa. The measurements of the von Mises stress are simulated using the Finite Element Method.
A representation of the setup used in this dataset for one of the plates is shown in Figure 1.

Figure 1. Demonstration of an example plate in the dataset.
Each case in the dataset contains information about the plate geometry, i.e., characterization of the holes, the corresponding mesh data, and the von Mises stress for every point in the mesh, as illustrated in Figure 2.

Figure 2. The SimuStruct dataset includes: (a) geometric information, (b) mesh data, and (c) von Mises stress results.
To download the dataset, head to this link.
Dataset structure
For each sample of the dataset, the data is stored inside a folder named according to the respective sample ID.
In each folder, you will find 4 csv
files that describe the plate’s geometry, the corresponding mesh geometry and topology, and the von Mises stress measurements, as shown in Figure 3.
Description of each file:
-
circular_holes.csv
: This file contains information about the six holes in the plate. Each row represents a hole, with the first two columns being the x and y coordinates of the hole’s centre and the third column being the hole’s radius. There are six holes in total, so the shape of the array stored in this file is6 x 3
. -
mesh_geometry.csv
: This file contains the coordinates of each node in the mesh. The two columns represent the x and y coordinates, respectively. The shape of the array stored in this file isnumber of nodes x 2
. -
mesh_topology.csv
: This file contains information about the elements in the mesh. Each row represents an element and contains the IDs of the nodes that make up that element. Since only triangular elements are used, each element is defined by 3 nodes. The shape of the array stored in this file isnumber of elements x 3
. -
von_Mises_stress.csv
: This file contains the von Mises stress values for each node in the mesh. The shape of the array stored in this file isnumber of nodes x 1
.

Figure. 3: Dataset structure.
Along with the dataset, you can also find a jupyter notebook with the necessary steps to read and plot the SimuStruct dataset.
Ideas on how to use this dataset
This dataset can be used for different purposes, such as:
-
Machine learning experimentation in mechanical engineering scenarios: train machine learning models to compute the von Mises stress.
-
Material science research: study the effect of holes on the stress distribution and mechanical properties of a plate under uniaxial tension.
-
Design optimization: find the optimal configuration of a plate with circular holes subjected to uniaxial tension.
Acknowledgements
The authors would like to thank the collaboration of Dr. Faez Ahmed from Massachusetts Institute of Technology (MIT), Dr. Miguel A. Bessa from Brown University, Dr. Jorge Belinha from the Instituto Superior de Engenharia do Porto and Dr. Sérgio M.O. Tavares from the University of Aveiro.
🔩 Have fun! 🔩
If you have any doubts regarding the dataset, don’t hesitate to reach out at datasets@inductiva.ai!
Also, in case you’re using this data on your projects, we would love to know how!
Besides this, don’t forget to check out the other cool datasets!
Recent posts from our blog

Sofia Guerreiro

Cristiana Carpinteiro
In this series of blog posts we will explore a specific case of the use of AI in the pharmaceutical industry - using Graph Neural Networks for predicting binding affinity. But for now, let’s start by understanding the problem of drug discovery and some fundamental concepts like binding affinity.

Ivan Pombo

Luís Sarmento
Coastal engineering projects protect the coast from erosion, flooding and other events. Due to their cost, the design phase of these projects is heavily based on computational simulations. In this blog post, we enlighten how Inductiva API allowed researchers to scale their coastal engineering simulations.