A new hybrid deep learning architecture, combining one-dimensional convolutional neural networks (1D-CNN) with bidirectional long short-term memory (BiLSTM) networks to the problem of automatic detection and early forecasting of mechanical faults based on raw vibration signals is suggested.
Abstract
Abstract. Timely fault diagnosis in turbomachinery plays an important role in avoiding disastrous failures, reducing the number of unexpected downtimes, and optimizing the maintenance cycle in industrial applications that use a lot of energy. The paper suggests a new hybrid deep learning architecture, combining one-dimensional convolutional neural networks (1D-CNN) with bidirectional long short-term memory (BiLSTM) networks to the problem of automatic detection and early forecasting of mechanical faults based on raw vibration signals. The combination of a multi-domain feature extraction strategy, which covers time-domain statistical indicators, spectral descriptors based on the fast fourier transform (FFT), and continuous wavelet transform (CWT) scalogram representations into a single 128-dimensional feature vector is used as the input of the model. The suggested CNN-BiLSTM design effectively extracts both local spectrotemporal features and long-range sequential dynamics of the rotating machinery vibration data. Experiments on the CWRU Bearing Dataset and a custom turbomachine testbed dataset show that the proposed model can achieve a mean classification accuracy of 97.6% with clean signal conditions and 95.1% at a signal-to-noise ratio of 10 dB, and outperforms standalone CNN, BiLSTM, SVM, ANN, and Random Forest baselines by up to 10.2 The findings determine the appropriateness of the proposed framework on edge-deployable and real-time condition monitoring systems.
Abstract. High-performance mechanical component structural health monitoring (SHM) is a vital issue in contemporary engineering, especially in the aerospace, automotive, and industrial turbomachinery sectors where component failure may be disastrous. This article introduces a new AI-aided SHM framework with multimodal sensor fusion and ensemble deep learning architecture with one-dimensional convolutional neural networks (1D-CNN) and bidirectional long short-term memory (Bi-LSTM) networks, which can be used to detect fault, classify fault, and predict remaining useful life (RUL) in real-time. The proposed system takes in time-series streams of vibration, acoustic emission, and strain gauge data, runs them through an adaptive signal preprocessing pipeline, and derives hierarchical features of fault relevance without using manually specified features. The tests are done on two benchmark datasets, which include the CWRU bearing fault dataset and a custom gas turbine blade fatigue dataset that were obtained under controlled laboratory settings and a real gas turbine compressor testbed. The CNN-BiLSTM ensemble suggested has an accuracy of fault classification of 98.7 and a mean absolute percentage error (MAPE) of 3.14 to predict RUL with average inference latency of 12.3 ms, which is appropriate to be integrated into embedded systems in real-time. These findings constitute a statistically significant step forward compared to the current state-of-the-art baselines and they generalize and scale to provide an AI-SHM paradigm of safety-critical mechanical systems.
Jasjeet Singh· Materials Research Proceedin...· 0 citations
Reliable fault diagnosis of centrifugal pumps is challenging due to the nonstationary nature of vibration signals, weak early-stage laboratory fault signatures, and overlapping characteristics among different mechanical defects. This study proposes a wavelet coherence-aware multi-branch deep ensemble framework that integrates physically meaningful time-frequency coupling with complementary deep feature learning. Multi-channel vibration signals are transformed into two-dimensional wavelet coherence maps to emphasize localized inter-sensor phase-consistent structures induced by mechanical processes. Three lightweight and architecturally diverse convolutional neural networks are trained in parallel to extract fine-scale, global, and compact structural features. Their outputs are fused through a probabilistic soft-voting strategy to improve robustness and decision stability. The framework is evaluated on vibration datasets collected from a PMT-4008 centrifugal pump test bench under three operating pressures (3.0, 3.5, and 4.0 bar). The results demonstrate consistent and reliable fault discrimination across all investigated conditions, with strong class separability confirmed by Receiver Operating Characteristic analysis and feature-space visualization. These findings demonstrate the effectiveness of the proposed framework for centrifugal pump fault diagnosis within the investigated experimental setup and operating conditions.
F. Saleem, Muhammad Umar, Jong-Myon Kim· Scientific Reports· 0 citations
An ensemble attention-based residual convolutional neural network optimized by the vortex search algorithm can effectively overcome the limitations of individual models and achieve superior fault identification performance than existing methods under many types of severe conditions.
A rolling bearing fault diagnosis method based on multi-scale depthwise separable convolution (DSC) and a convolutional neural network-Transformer hybrid model (CNN-Transformer) is proposed to address the non-stationarity of fault signals and the difficulty of jointly capturing local and global features. First, continuous wavelet transform (CWT) converts one-dimensional vibration signals into two-dimensional time-frequency images to enhance fault representation. Then, multi-scale convolution (MSC) and DSC are introduced to extract local impulsive features and fault patterns at different scales with fewer parameters. A CNN-Transformer architecture is further developed, where convolutional neural network (CNN) captures local details and Transformer models global dependencies. In addition, pretraining-finetuning, data augmentation, label smoothing, and normal sample optimization are adopted to improve training stability and diagnostic performance. Experimental results show accuracies of 98.80% on the Xi’an Jiaotong University bearing dataset (XJTU-SY) and 100.00% on the Case Western Reserve University bearing dataset (CWRU), demonstrating strong discriminative ability, stability, and robustness.
Shuai Yang, Yan-Chao Chen, Yang Yu· Engineering Research Express· 0 citations
Multi-sensor configurations incorporating acoustic data demonstrated faster convergence while maintaining perfect classification performance, achieving high classification accuracy across multiple sensor configurations, including single-sensor and fused measurements.
Lucky Islam Pataky, Feruza A. Amirkulova, Sabrina Hsu et al.· INTER-NOISE and NOISE-CON Co...· 0 citations
Aiming at the non-stationary, nonlinear and noise-sensitive characteristics of rolling bearing vibration signals, as well as the low recognition accuracy of traditional deep learning in small-sample scenarios, this paper proposes a rolling bearing fault diagnosis method combining Continuous Wavelet Transform with Ridge Tracking (CWT-RT) and Multi-Scale Wavelet Scattering Network. First, Variational Mode Decomposition integrated with Cramer Von Misses statistic (VMD-CVM) is adopted to denoise the original signal and improve the signal-to-noise ratio. Then, CWT-RT is used to transform the denoised signal into time-frequency spectrograms for intuitive time-frequency feature representation. Multi-Scale Wavelet Scattering Network is further applied to extract multi-level structural features, which are fed into Multi-Layer Perceptron (MLP) to realize bearing fault identification. To eliminate data leakage, all original raw vibration files are split into training and test sets at a 7:3 ratio before sliding window sampling. Validation experiments on bearing datasets from South Ural State University, CWRU, and XJTU-SY show that the diagnostic accuracies on two small-sample conditions reach 98.78% and 98.09%, respectively. the 95% confidence intervals for the two accuracy values are [98.21%, 99.15%] and [97.43%, 98.67%], respectively. Across 10 repeated experiments, p-values < 0.001 confirm the statistical significance of the results. The model maintains high accuracy under different loads and noise levels (0/5/10/15 dB). Comparative and ablation experiments verify that the method has high diagnostic accuracy, strong noise robustness and superior small-sample learning ability, with each module effective, and the full-pipeline latency meets the real-time requirements of IoT edge deployment, providing support for industrial engineering applications.
Hai Ling, Lufan Wang, Wen Liu et al.· Information Technology and C...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
