Uncovering the Underlying Physics of Degrading System Behavior Through a Deep Neural Network Framework: The Case of Remaining Useful Life Prognosis
Authors:
Sergio Cofre-Martel,
Enrique Lopez Droguett,
Mohammad Modarres
Abstract:
Deep learning (DL) has become an essential tool in prognosis and health management (PHM), commonly used as a regression algorithm for the prognosis of a system's behavior. One particular metric of interest is the remaining useful life (RUL) estimated using monitoring sensor data. Most of these deep learning applications treat the algorithms as black-box functions, giving little to no control of th…
▽ More
Deep learning (DL) has become an essential tool in prognosis and health management (PHM), commonly used as a regression algorithm for the prognosis of a system's behavior. One particular metric of interest is the remaining useful life (RUL) estimated using monitoring sensor data. Most of these deep learning applications treat the algorithms as black-box functions, giving little to no control of the data interpretation. This becomes an issue if the models break the governing laws of physics or other natural sciences when no constraints are imposed. The latest research efforts have focused on applying complex DL models to achieve a low prediction error rather than studying how the models interpret the behavior of the data and the system itself. In this paper, we propose an open-box approach using a deep neural network framework to explore the physics of degradation through partial differential equations (PDEs). The framework has three stages, and it aims to discover a latent variable and corresponding PDE to represent the health state of the system. Models are trained as a supervised regression and designed to output the RUL as well as a latent variable map that can be used and interpreted as the system's health indicator.
△ Less
Submitted 10 June, 2020;
originally announced June 2020.
Neural Network and Particle Filtering: A Hybrid Framework for Crack Propagation Prediction
Authors:
Seyed Fouad Karimian,
Ramin Moradi,
Sergio Cofre-Martel,
Katrina M. Groth,
Mohammad Modarres
Abstract:
Crack detection, length estimation, and Remaining Useful Life (RUL) prediction are among the most studied topics in reliability engineering. Several research efforts have studied physics of failure (PoF) of different materials, along with data-driven approaches as an alternative to the traditional PoF studies. To bridge the gap between these two techniques, we propose a novel hybrid framework for…
▽ More
Crack detection, length estimation, and Remaining Useful Life (RUL) prediction are among the most studied topics in reliability engineering. Several research efforts have studied physics of failure (PoF) of different materials, along with data-driven approaches as an alternative to the traditional PoF studies. To bridge the gap between these two techniques, we propose a novel hybrid framework for fatigue crack length estimation and prediction. Physics-based modeling is performed on the fracture mechanics degradation data by estimating parameters of the Paris Law, including the associated uncertainties. Crack length estimations are inferred by feeding manually extracted features from ultrasonic signals to a Neural Network (NN). The crack length prediction is then performed using the Particle Filter (PF) approach, which takes the Paris Law as a move function and uses the NN's output as observation to update the crack growth path. This hybrid framework combines machine learning, physics-based modeling, and Bayesian updating with promising results.
△ Less
Submitted 26 April, 2020;
originally announced April 2020.
