Open research questions in Machine Fault Diagnosis Techniques

What the literature leaves open

• Future work will focus on experimental validation using a practical bearingless induction motor platform, systematic parameter selection, combined sensor-calibration-error analysis, and code-generation-based implementation on real-time DSP or FPGA hardware.

  Confidence-Fusion-Based Fault-Tolerant Displacement Measurement Method for Bearingless Induction Motor · 2026 · DOI

• To overcome this challenge, future work will focus on fault decoupling identification of the segmentation strategy and targeted feature indicators guidance for ODFB selection and demodulation, as well as developing more robust strategies for speed variation adaptation, limited-data conditions, and error accumulation during cascaded decoupling, to enhance the applicability.

  A Novel Narrowband Filtering Demodulation Method Based on Adaptive Multi-Level Spectra Segmentation Strategy and Its Application in Bearing Fault Diagnosis · 2026 · DOI

• Future work will focus on collecting experimental and industrial gearbox lubricant data, validating the model under different operating conditions, expanding the range of gearbox health conditions, and evaluating the model through real-time condition-monitoring applications.

  Machine learning based diagnosis of gearbox wear patterns through lubricant analysis · 2026 · DOI

• Future studies should consider incorporating pre-task resting EEG and individual baseline normalization, together with multiple physiological signals, such as electrocardiography (ECG) and EMG, to improve the comprehensiveness, robustness, and accuracy of fatigue assessment.

  Study on mental fatigue detection of crane operators based on frequency and channel attention spiking neural network with multivariate variational mode decomposition in real construction environments · 2026 · DOI

• Future work will focus on designing more computationally efficient semi-supervised graph learning methods, while also investigating temporal encoders or sequence-aware graph construction methods to determine whether temporal information can provide additional discriminative value. The additional discriminative value of the discarded temporal order remains to be evaluated.

  Measured multi-source fusion graph neural network for semi-supervised imbalanced working condition recognition · 2026 · DOI

• The generalization ability of the proposed method on other test rigs and other types of gearboxes remains to be verified. The trade-off between computational cost and performance gain remains to be systematically investigated.

  Physics-Guided Dual-Branch Domain Generalization with Confidence-Adaptive Decision Fusion for Cross-Speed Fault Diagnosis of Planetary Gearboxes · 2026 · DOI

• Further research can be conducted to improve the robustness of the proposed method under severe missingness. The proposed method can be applied to other fault diagnosis tasks. The effectiveness of the proposed method can be evaluated on more datasets.

  MAEFF: missing-aware adaptive expert fusion framework for multi-sensor bearing fault diagnosis · 2026 · DOI

• The lack of a reliable and robust bearing fault diagnosis method that can handle missing data and varying operating conditions. The need for a method that can learn discriminative fault signatures directly from sensor measurements. The limitation of prior work in handling missing data and varying operating conditions.

  MAEFF: missing-aware adaptive expert fusion framework for multi-sensor bearing fault diagnosis · 2026 · DOI

• Third, the experiments were conducted under controlled laboratory conditions with a single UAV platform, and the generalizability of the framework to different rotor configurations, environmental conditions, and in-flight scenarios remains to be validated. Several limitations of the present study should be noted.

  Signal Intelligence: Vibration-Driven Deep Learning for Anomaly Detection of Rotary-Wing UAVs · 2026 · DOI

• These results verify implementation behavior only; report utility, user acceptance, and field effectiveness remain to be evaluated independently. The empirical support for the post-classification procedure is limited to the 24-pair rule-matrix implementation test, predictioninvariance checks, and two illustrative cases.

  Vibration–current multimodal fusion for within-batch recognition of predefined component faults of a high-pressure abrasive waterjet machine tool · 2026 · DOI

• While the current methodology incorporates bootstrapping to characterize stability, further work is needed to translate this stability into calibrated uncertainty estimates and to determine whether additional recalibration methods are required. First, the influence of the masking (noise) ratio during training warrants further investigation, as previous studies report effective ra-Wind Energ.

  Sensor-error-robust normal-behavior modeling for wind turbine drive train failure prediction using a masked autoencoder · 2026 · DOI

• The separate effects of entropy families, scales, and feature-control stages remain unknown. Limitations This study has several limits. These results indicate that the relative performance of an entropy family depends on the diagnostic task and operating condition and that results obtained from a single machine or operating regime are insufficient to identify a uniformly preferable entropy family.

  Dual-View Complementary Multiscale Entropy Fusion for Rotating-Machinery Fault Diagnosis · 2026 · DOI

• Future work will focus on (i) systematic parameter-sensitivity analysis across turbines and operating regimes, (ii) benchmarking against independent tacholess estimators such as Vold–Kalman filtering and multi-order probabilistic speed tracking, and (iii) real-time embedded deployment and robustness evaluation under severe noise and missing-data scenarios.

  Tacholess Order Analysis of Gearbox Vibration for Planetary-Gear Fault Diagnosis in Variable-Speed Offshore Wind Turbines · 2026 · DOI

• The limited dataset size and lack of real industrial operating conditions pose challenges to the robustness and scalability of the proposed framework.

  Multiphysics-informed predictive maintenance of continuous polymer gear systems via triboelectric–vibration-heat fusion and recurrent machine learning · 2026 · DOI

• By incorporating transfer learning with a pre-trained ResNet50 model, the method addresses the challenges of limited data and accelerates model training and optimization.

  Rolling Bearing Fault Diagnosis Method Based on SWT and Improved Vision Transformer · 2025 · DOI

• Future research will focus on two key aspects: first, exploring more advanced feature fusion techniques and incorporating additional domain knowledge to further enhance model performance; second, optimizing computational efficiency by exploring lightweight model architectures or hardware-accelerated solutions, enabling the proposed method to achieve faster training and inference times while maintaining accuracy.

  A Deep-Learning Method for Remaining Useful Life Prediction of Power Machinery via Dual-Attention Mechanism · 2025 · DOI

• Given that non-stationary conditions can significantly alter vibration signal characteristics and pose major challenges in real-world fault diagnosis tasks, future work will explore extending the application of the AWLT-GAN framework to more diverse operating conditions, including varying load levels and speed changes.

  Comprehensive Exploitation of Time- and Frequency-Domain Information for Bearing Fault Diagnosis on Imbalanced Datasets via Adaptive Wavelet-like Transform General Adversarial Network and Ensemble Learning · 2025 · DOI

• This paper introduced a novel spectral interpretable bearing fault diagnosis framework that synergistically integrates advanced signal processing, a deep learning-based Fault Classification Network (FCN), and a fine-tuned large language model (LLM) to address the critical challenge of limited interpretability in conventional diagnostic systems.

  A Spectral Interpretable Bearing Fault Diagnosis Framework Powered by Large Language Models · 2025 · DOI

• Future research should focus on hybrid modeling, advanced AI-driven classification techniques, and enhanced uncertainty quantification in order to further optimize predictive maintenance strategies.

  Machine Health Indicators and Digital Twins · 2025 · DOI

• Future research should focus on refining these AI models using real operational data, improving signal processing techniques, and optimizing computational efficiency to facilitate their deployment in real-time wind turbine monitoring systems.

  Artificial Intelligence in Wind Turbine Fault Detection and Diagnosis: Advances and Perspectives · 2025 · DOI

• Future work will focus on analyzing additional fault types, including combined or compound defects, and exploring alternative optimization algorithms within deep learning to enhance diagnostic performance. One limitation of the present study is the use of fixed wind turbine angular speeds centered on steady-state conditions, which do not always reflect the variability of real-world operating environments.

  Detection of Cracks in Low-Power Wind Turbines Using Vibration Signal Analysis with Empirical Mode Decomposition and Convolutional Neural Networks · 2025 · DOI

• This study focuses on the problem of class-incremental fault diagnosis based on nuclear power plant operational data and proposes an integrated “feature–temporal–time– frequency fusion” diagnostic framework that jointly considers feature construction, class imbalance, and temporal dependencies.

  A Time-Frequency Fusion Fault Diagnosis Framework for Nuclear Power Plants Oriented to Class-Incremental Learning Under Data Imbalance · 2026 · DOI

• The complexity and interconnectivity of turbofan engines, - The need for accurate detection of change points in the degradation trajectory, - The challenge of integrating Bayesian inference with adaptive change point…

  Remaining useful life prediction and maintenance schedule designing for turbofan engines considering multistage change points and uncertainty awareness · 2026 · DOI

• WindEurope: Wind energy in Europe: 2024 Statistics and rep., WindEurope, 2025-2030, Tech.

  Condition monitoring of wind turbine drivetrains: state-of-the-art technologies, recent trends, and future outlook · 2026 · DOI

• The current study is based on a limited dataset of 30 experimental cases, - The study does not consider real industrial operating conditions, - The study does not incorporate additional sensing modalities such as acoustic emission, friction force, and wear debris analysis, - The study does not evaluate the scalability and practical applicability of the proposed…

  Multiphysics-informed predictive maintenance of continuous polymer gear systems via triboelectric–vibration-heat fusion and recurrent machine learning · 2026 · DOI

Most-cited papers in Machine Fault Diagnosis Techniques

• Applications of machine learning to machine fault diagnosis: A review and roadmap · Mechanical Systems and Signal Processing · 2020 · 2,928 citations
• A Hybrid Prognostics Approach for Estimating Remaining Useful Life of Rolling Element Bearings · IEEE Transactions on Reliability · 2020 · 1,822 citations
• Deep Learning Algorithms for Bearing Fault Diagnostics—A Comprehensive Review · IEEE Access · 2020 · 833 citations
• A perspective survey on deep transfer learning for fault diagnosis in industrial scenarios: Theories, applications and challenges · Mechanical Systems and Signal Processing · 2021 · 789 citations
• Machine Learning in Predictive Maintenance towards Sustainable Smart Manufacturing in Industry 4.0 · Sustainability · 2020 · 760 citations
• Bearing fault diagnosis base on multi-scale CNN and LSTM model · Journal of Intelligent Manufacturing · 2020 · 734 citations
• Deep learning algorithms for rotating machinery intelligent diagnosis: An open source benchmark study · ISA Transactions · 2020 · 698 citations
• Bearing fault diagnosis based on vibro-acoustic data fusion and 1D-CNN network · Measurement · 2020 · 629 citations
• Machine Remaining Useful Life Prediction via an Attention-Based Deep Learning Approach · IEEE Transactions on Industrial Electronics · 2021 · 602 citations
• Signal based condition monitoring techniques for fault detection and diagnosis of induction motors: A state-of-the-art review · Mechanical Systems and Signal Processing · 2020 · 590 citations

Most recent work

• Spatial-channel collaborative multi-scale graph interaction deep transfer learning for unsupervised rotating machinery fault diagnosis · Engineering Applications of Artificial Intelligence · 2026
• An Improved Variational Autoencoder and Graph Attention Network Method for Wear Prediction of Aerospace Self-Lubricating Bearing Using Acoustic Emission Signal · IEEE Sensors Journal · 2026
• A Hybrid Physical Damage Neural Network for Wear Prediction of Self-Lubricating Bearings · Journal of Computing and Information Science in Engineering · 2026
• Self-supervised learning for train bearing fault diagnosis based on time–frequency dual domain prediction · Structural Health Monitoring · 2026
• Digital twin-assisted graph contrastive domain adaptation for small-sample bearing fault diagnosis · Structural Health Monitoring · 2026
• Robust Open-Circuit Fault Diagnosis for PMSM Drives Under Unknown Operating Conditions · IEEE Transactions on Instrumentation and Measurement · 2026
• A liquid-impulse neural network model based on heterogeneous fusion of multimodal information for interpretable rotating machinery fault diagnosis · Mechanical Systems and Signal Processing · 2026
• PSRONet: A Deep Reinforcement Learning-Based Sensor Configuration Framework in Railway Point Machines Fault Diagnosis · IEEE Transactions on Instrumentation and Measurement · 2026
• Source-free domain adaptation for cross-domain remaining useful life prediction: A distributed federated learning perspective · Reliability Engineering & System Safety · 2026
• A digital twin guided physical-virtual denoising method for early fault detection of rolling element bearings · Mechanical Systems and Signal Processing · 2026

Related topics in Engineering