Lightweight fault diagnosis (FD) models are essential for efficient inference on resource-constrained IoT edge devices and practical industrial deployment. While existing lightweight FD research has achieved notable progress in network architecture compression, the substantial benefits brought by quantization techniques remain largely overlooked. However, directly applying mainstream quantization methods to FD tasks is problematic, as they suffer from two main drawbacks: 1) Uniform quantization struggles to simultaneously maintain local resolution in dense regions and dynamic range coverage in sparse regions. 2) Mainstream quantization-aware training (QAT) methods rely on the straight-through estimator (STE) to approximate discrete gradients, introducing gradient bias that further limits model accuracy improvement. To address these issues, this paper proposes an Adaptive Quantization Error Minimization Method (AQEMM) for fault diagnosis, with the following main contributions: 1) An adaptive quantizer is introduced to replace the traditional fixed quantization step size, expanding the search space of quantization points from discrete integer grids to a continuous subspace, enabling quantization levels to adaptively concentrate in data-dense regions. 2) A block coordinate descent strategy is incorporated to alternately optimize quantization parameters, with quantization reconstruction error minimization as the direct optimization objective, fundamentally circumventing the gradient bias introduced by STE. Validation on FD tasks for two typical industrial devices, induction motors (IM) and permanent magnet synchronous motors (PMSM), demonstrates that the proposed 2-bit quantization model saves up to approximately 82% of hardware resources compared to full-precision (FP) deployment on an FPGA platform while maintaining competitive diagnostic accuracy. Furthermore, by optimizing the hardware design to leverage the released hardware resource margin, a significant inference speedup of approximately 20.6× is achieved over pure software inference.

The intricate multicomponent coupling dynamics inherent in complex electromechanical systems often result in the aliasing of motor health status signals, posing a significant challenge for fault diagnosis techniques. To tackle this issue, this article presents an interpretable leakage flux pattern method, which introduces the motor mechanism knowledge system into the image representation method and enables the data-driven fault diagnosis method to have common knowledge with humans. Initially, the relationship between the magnetic field inside the motor and the leakage flux outside the motor is analyzed. Subsequently, the leakage flux pattern method is proposed to reconstruct the distribution of the magnetic induction lines inside the motor with a noninvasive measurement method. Furthermore, a lossless data compression technique grounded in causality is proposed, alongside a semantic-aware matching method for causal features. Finally, an interpretable fault diagnosis framework is proposed. The experimental outcomes demonstrate the excellent efficiency and performance of the proposed method.

In most industrial applications, it is difficult to obtain complete demagnetization fault signals of all conditions with labels for permanent magnet synchronous motor (PMSM), and motors are not allowed to be disassembled, so non-contact diagnostic methods are essential. A non-contact fault diagnosis method using magnetic leakage signal based on wavelet scattering convolution network (WSCN) and semi-supervised deep rule-based (SSDRB) classifier is proposed. Through magnetic equivalent circuit model analysis, the magnetic leakage signal on motor surface is selected as fault signal. To avoid complex signal processing, the symmetrized dot pattern method is introduced to convert fault signals into two-dimensional images. Then, WSCN is applied to extract features from images, and SSDRB classifier is adopted to diagnose demagnetization fault. Finally, faulty motor prototypes are manufactured for experiment. By comparing with other methods, the superiority and effectiveness of the proposed method using a small number of labeled samples under different conditions are verified.

Generalization ability is a critical property for practical motor fault diagnosis (FD). By converting time-series to images, several studies have made certain achievements. However, they still have following limitations. First, multisensor information fusion is rarely considered. Second, it is time consuming. To deal with the abovementioned problems, a multisensor-driven FD method based on visual features is proposed. Specifically, a color symmetrized dot pattern method is newly designed to infuse three multisensor signals to image. Next, a coarse and refined diagnosis framework is designed. In the coarse part, the color histogram features and a support vector machine (SVM) are utilized, and a threshold is selected to decide the coarse diagnostic samples. In the refined part, the gist (GIST) descriptor and another SVM are used to diagnose remaining samples. The results on induction motor and permanent magnet synchronous motor show that the proposed method achieved reliable diagnosis with relatively efficiency, and can generalize to different working conditions and noise.

Fault prediction of electromechanical equipment can greatly reduce its maintenance cost and prevent catastrophic damage. In order to realize the accurate fault prediction of electromechanical equipment, a fault prediction method based on spatial-temporal graph information is proposed in this article. In the proposed method, the signal data of each intermittent monitoring period are expressed by Markov field graph information, and the spatial-temporal correlation features of graph information are extracted and studied by multivariate spatial-temporal graph neural networks. The effectiveness of Markov graph information for different state expressions is demonstrated by motor fault data from the motor fault experimental platform, and the bearing fault prognostics data is used to demonstrate the feasibility and accuracy of the proposed fault prediction method. The results show that based on the local spatial state information and global time correlation information of monitoring signals, this method could accomplish accurate long-term and short-term fault prediction tasks, respectively.

Traditional fault diagnosis methods are easy to be affected by different working conditions. This article proposed a motor fault diagnosis method based on visual knowledge, to reduce the impact of changes in working conditions and improve the feature extraction ability. The mapping relationship between actual faults and image intuitive features by symmetrized dot pattern and scale-invariant feature transform is established in this article. The fault state is obtained by statistics of the matching point with the dictionary templates generated from signals of normal and unnormal motors. Compared with other machine learning algorithms, this method does not need too much data training and learning. The efficiency of this method is validated by experiments, and the data image processing technology has great industrial application value in the field of motor fault detection or monitoring in the age of intelligence.

During the long-term operation of motors, their working conditions are changing due to the industrial demands or declining health status, and traditional diagnosis methods perform poorly in that case. This article proposes a fault diagnosis method based on graph cardinality preserved attention network (GCPAT), which can work under varying working conditions, and can be generalized to the transient state. Diagnosis results are obtained by analyzing signal-converting graphs, which are composed of nodes and edges. First, the vibration signals are converted into symmetrical snowflake images by symmetrized dot pattern (SDP) method. Second, SLIC is developed to make homogeneous super-pixels in SDP images as nodes, and form graphs according to color, texture, and distance features. Finally, the GCPAT is utilized to distinguish motor status. Compared with other state-of-art methods, the results show the out-performance of GCPAT under varying working conditions both in steady and transient state.