Collected molecules will appear here. Add from search or explore.
Application of Physics-Informed Neural Networks (PINNs) for detecting damage and monitoring the integrity of physical structures (Structural Health Monitoring).
Defensibility
stars
7
forks
1
This project is a personal academic repository associated with a master's thesis. With only 7 stars and 1 fork over a 4-year period, it lacks any meaningful community adoption or ecosystem. Technically, it applies the standard PINN (Physics-Informed Neural Network) architecture—pioneered by Raissi et al. in 2019—to structural engineering problems. While the niche (Structural Health Monitoring) is specialized, the codebase is a reference implementation of known techniques rather than a defensible tool. Since its inception, professional-grade libraries like NVIDIA Modulus and DeepXDE have emerged, which provide much more robust, optimized, and extensible frameworks for solving the same physics-informed problems. The author's note about 'hiding details for privacy issues' further reduces its utility as an open-source resource. It is highly susceptible to displacement by any modern industrial IoT or engineering simulation software that incorporates physics-aware ML.
TECH STACK
INTEGRATION
reference_implementation
READINESS