Response Estimation and System Identification of Dynamical Systems via Physics-Informed Neural Networks

Marcus Haywood-Alexander,Giacomo Arcieri,Antonios Kamariotis,Eleni Chatzi

2024-10-30

Abstract:The accurate modelling of structural dynamics is crucial across numerous engineering applications, such as Structural Health Monitoring (SHM), seismic analysis, and vibration control. Often, these models originate from physics-based principles and can be derived from corresponding governing equations, often of differential equation form. However, complex system characteristics, such as nonlinearities and energy dissipation mechanisms, often imply that such models are approximative and often imprecise. This challenge is further compounded in SHM, where sensor data is often sparse, making it difficult to fully observe the system's states. To address these issues, this paper explores the use of Physics-Informed Neural Networks (PINNs), a class of physics-enhanced machine learning (PEML) techniques, for the identification and estimation of dynamical systems. PINNs offer a unique advantage by embedding known physical laws directly into the neural network's loss function, allowing for simple embedding of complex phenomena, even in the presence of uncertainties. This study specifically investigates three key applications of PINNs: state estimation in systems with sparse sensing, joint state-parameter estimation, when both system response and parameters are unknown, and parameter estimation within a Bayesian framework to quantify uncertainties. The results demonstrate that PINNs deliver an efficient tool across all aforementioned tasks, even in presence of modelling errors. However, these errors tend to have a more significant impact on parameter estimation, as the optimization process must reconcile discrepancies between the prescribed model and the true system behavior. Despite these challenges, PINNs show promise in dynamical system modeling, offering a robust approach to handling uncertainties.

Computational Physics,Machine Learning

What problem does this paper attempt to address?

This paper attempts to solve several key problems in dynamic system modeling, especially those encountered in fields such as structural health monitoring (SHM), seismic analysis, and vibration control. Specifically, the paper mainly focuses on the following three challenges:

1. **State Estimation under Sparse Sensing Conditions**: In practical applications, due to problems such as cost, installation complexity, and sensor failures, it is usually impossible to achieve comprehensive observation of the system. Therefore, how to accurately estimate the unobserved states of the system based on partial observation data is an important issue. The paper explores the method of using physics - informed neural networks (PINNs) to solve this problem.

2. **Joint State - Parameter Estimation**: When both the response and parameters of the system are uncertain, simultaneously estimating these unknowns is a more complex task. The paper studies the application of PINNs in dealing with this double - uncertainty problem, improving model prediction by combining data and physical laws.

3. **Parameter Estimation under the Bayesian Framework**: In order to quantify uncertainty, the paper also explores the method of using PINNs for parameter estimation under the Bayesian framework. This method not only provides point estimates of parameters but also gives the probability distribution of parameters, thus more comprehensively describing the uncertainty of the estimation results.

Through the work in the above three aspects, the paper demonstrates the application potential of PINNs as an efficient tool in dynamic system modeling, especially in the case of modeling errors, PINNs can still provide reliable results. However, the paper also points out that the modeling error has a great impact on parameter estimation because the optimization process needs to reconcile the differences between the preset model and the real system behavior. Nevertheless, PINNs show significant advantages in dealing with the uncertainty and complexity in dynamic system modeling.