CODES

Open source codes:

1. Project 1: Matlab FEM Damage code - local and gradient non-local models
2. Project 2: Integrated Finite Element Neural Network for non-local damage
3. Project 3: Analysis of error convergence and computational cost of PINNs in non-local damage I-FENN
4. Project 4: Integrated Finite Element Neural Network for thermoelasticity

Description:  Matlab code for FEM models of local and non-local gradient damage models. Three solvers included: Displacement-based arc-length, Force-based arc-length, and Newton-Raphson. All solvers work with both analytically derived tangent and numerically calculated tangent matrices.

Papers:

1. Saji, R.P., Pantidis, P., & Mobasher, M. E. (2024). A new displacement-controlled Arc-length method for damage mechanics problems, Computational Mechanics (Link)

Open source code:

1. Matlab code (serial solver)

Description:  The codes used in this study to train and test a neural networks for non-local damage mechanics I-FENN. The damage mechanics solver is provided in Project 1.

Papers: 

1. Pantidis, P., & Mobasher, M. E. (2023). Integrated Finite Element Neural Network (I-FENN) for non-local continuum damage mechanics. Computer Methods in Applied  Mechanics and Engineering, 404, 115766. (Link)
2. Pantidis, P., Eldababy, H., Abueidda, D., & Mobasher, M. E. (2024). I-FENN with Temporal Convolutional Networks: expediting the load-history analysis of non-local gradient damage propagation. Computer Methods in Applied Mechanics and  Engineering, 425, 116940. (Link)

Open source code:

1. Sequential MLP model (Pantidis & Mobasher, 2023)
2. PI-TCN model (Pantidis et al., 2024)

Description: Simulation data and Python code used in the below study to assess the performance of PINNs used for non-local damage I-FENN model. The analysis considers maximizing accuracy, decreasing computational cost, and avoiding trivial solutions. The damage mechanics solver is provided in Project 1.

Papers: 

1. Pantidis, P., Eldababy, H., Tagle, C. M., & Mobasher, M. E. (2023). Error convergence and engineering-guided hyperparameter search of PINNs: towards      optimized I-FENN performance. Computer Methods in Applied Mechanics and      Engineering, 414, 116160. (Link)

Open source code:

1. Model, analysis, and data (Pantidis et al., 2023)

Description:  The codes used in the below study to train and test a neural networks for thermoelasticity I-FENN. 

Papers: 

1. Abueidda, D. W., & Mobasher, M. E. (2024). I-FENN for thermoelasticity based on physics-informed      temporal convolutional network (PI-TCN). Computational Mechanics. (Link)

Open source code:

1. PI-TCN model (Abueidda & Mobasher, 2024)