1 Introduction
1.1 Machine learning and structural design
1.2 Scope of investigation
2 Method and approach
2.1 Autoencoders and convolutional autoencoders
2.2 Building the convolutional autoencoder with Keras and TensorFlow
2.3 Input data sets
2.4 Regression model
Natural seismic source | Rarity | Cause | Magnitude |
---|---|---|---|
Deep Moonquakes | Most common | Tidal forces | Maximum of 3 \(m_b\) |
Shallow Moonquakes | Rarest | Unknown | Maximum of 4.8 \(m_b\) |
Thermal Moonquakes | Unlikely | Extreme temperature variations on the lunar surface | Very small amplitude |
Meteorite Impacts | Abundant (but less than deep moonquakes) | Thin atmosphere | Amplitude and frequency vary greatly |
3 Structural analysis
3.1 Lunar ground motion
Material property | Units | Value |
---|---|---|
Density | kg/m\(^{3}\) | 2300 |
Young’s Modulus | MPa | 287.3 |
Compressive strength | MPa | 4.2 |
Tensile strength | MPa | 0.42 |
Poisson’s ratio | – | 0.2 |
3.2 Motivation of arch shapes
3.3 Contour plot requirements
3.3.1 Material properties
3.3.2 Loading conditions
-
Component 1: lateral component representing the seismic load applied on the arch originating from moonquakes, meteorite impacts, etc. A value of 0.2 g (1.962 m/s\(^{2}\)) is chosen in consistency with Málaga-Chuquitaype et al. [11] case study.
-
Component 2: vertical component representing the self-weight load on the arch. This component will therefore be equal to the Lunar gravitational field, which equals to \(-\) 0.166 g (\(-\)1.628 m/s\(^{2}\)).
3.3.3 Boundary conditions
3.4 Abaqus algorithms
3.5 Structural analysis
3.5.1 Maximum in-plane stress
3.5.2 Minimum in-plane stress
3.5.3 Horizontal displacement
3.5.4 Vertical displacement
4 Results
4.1 K-means clustering results
4.2 Latent space variation
4.3 Regression results
\(t_{1}\) [m] | \(1_{1}\) [m] | \(t_{2}\) [m] | \(r_{2}\) [m] | \(z[-]\) | |
---|---|---|---|---|---|
Arch 1 | 1 | 15 | 1 | 15 | 0.3 |
Arch 2 | 5 | 15 | 1 | 15 | 0.3 |
Arch 3 | 10 | 15 | 1 | 15 | 0.3 |
Arch 4 | 15 | 15 | 1 | 15 | 0.3 |
Arch 5 | 20 | 15 | 1 | 15 | 0.3 |