Ensemble of recurrent neural networks with long short-term memory cells for high-rate structural health monitoring

Abstract

The deployment of systems experiencing high-rate dynamic events, such as hypersonic vehicles, advanced weaponry, and active blast mitigation systems, require high-rate structural health monitoring (HRSHM) capabilities in the sub-millisecond realm to ensure continuous operations and safety. However, the development of high-rate feedback systems is a complex task because these dynamic systems are uniquely characterized by (1) large uncertainties in their external loads, (2) high levels of non-stationarity and heavy disturbance, and (3) unmodeled dynamics from changes in system configuration. In this paper, we present a deep learning algorithm specifically engineered for HRSHM applications. It consists of an ensemble of recurrent neural networks (RNNs) constructed with long short-term memory (LSTM) cells with transfer learning capabilities to cope with the highly limited availability of training data as it is typical for high-rate systems. The use of an ensemble of RNNs empowers multi-rate sampling capability to capture multi-temporal features of the time series, thus enabling modeling of non-stationarities. Also, because the RNNs use short-sequence LSTMs and are arranged in parallel, the computation time is substantially reduced to the sub-millisecond range. The performance of the algorithm is investigated on experimental high-rate dynamic data produced from accelerated drop tower tests and benchmarked against results from a prior investigation using a purely on-the-edge algorithm termed variable input space observer (VIO) and its hybrid architecture (hybrid VIO) that incorporated physical knowledge and is considered as an upper bound on the algorithm performance. Numerical results show that the ensemble of RNNs, composed of five RNNs, significantly outperformed the VIO, with performance close to that of the hybrid VIO when it comes to the estimation error metrics, with an average computation time of 25 μs per prediction step. Yet, the ensemble of RNNs exhibits chattering for low-amplitude excitations, likely attributable to the behavior of RNNs sampling at small time delays. An examination of the RNN weights and hidden states confirms that the algorithm can capture multi-temporal features, and an investigation with respect to noise in the training dataset showed that the algorithm is robust for signal-to-noise ratios up to 20 dB.

High-Rate Structural Health Monitoring and Prognostics: An Overview

“…To address the second and third challenges of lack of physics and inconsistency of the data observations, approaches may include developing physics-enhanced machine learning (PEML) models, which integrates the learning model with any known physics. Machine learning using data-based methods alone currently prove very powerful tools for predicting remainable useful life for rotational machinery [18,19] and prediction of high-rate dynamics [20]. Additional current trends include PINN (physical-informed neural networks), physics-informed machine learning (PIML) [21], and digital twin interpretability [22].…”

Real-time state estimation using recurrent neural network and topological data analysis

“…A key challenge in data-based formulations is in data scarcity arising, because high-rate experiments are expensive to conduct and may only include limited dynamic responses. 7 To address this issue, several studies proposed the use of transfer learning to train algorithms. 8,9 Of interest, recurrent neural networks are promising for modeling and predicting time series due to their sequential organization, but they face the challenge of vanishing gradient issues.…”

“…11 In particular, the authors have previously shown promise in forecasting non-stationary time series in high-rate systems by decomposing the nonstationary system into stationary systems and combining predictions from an ensemble of RNNs using a weighted sum. 7 The focus of this paper is the real-time state estimation of dynamic systems using data-based techniques to map time series measurements to structural states using TDA features, emphasizing the novel application of these features as input to recurrent neural network (RNN) algorithms. The objective of this approach is to improve the predictive capabilities of the RNNs using physical insights provided by TDA.…”

“…This dataset is selected due to the wide availability of experimental datasets and the suitability of DROPBEAR datasets for validating and benchmarking the performance of highrate algorithms. 7 The rest of the paper is organized as follows. Section 2 provides background information on time series embedding and computational topology.…”

Convolutional Neural Network Denoising Auto-Encoders for Intelligent Aircraft Engine Gas Path Health Signal Noise Filtering

“…This family of algorithms has gained tremendous popularity due to prominent applications in language modelling and machine translation [25]. Basic RNN algorithms are notoriously difficult to train, and more elaborate algorithms are commonly used instead, such as the Long Short-Term Memory Networks (LSTMs) [26] and the Gated Recurrent Units (GRUs) [27]. In this research, sequence modelling algorithms are integrated into the encoder and decoder of the denoising autoencoder to build a deep denoising autoencoder used for gas turbine gas path measurement noise filtering.…”