This paper proposes a new signal feature extraction and fault diagnosis method for of low-speed machinery. Statistic filter (SF) wavelet package transform (WPT) are combined with moving-peak-hold (M-PH) to extract features signal, special bearing diagnostic symptom parameters (SSPs) in frequency domain that sensitive defined recognize types. The SF is first used adaptively cancel noises, then detection performed by exploiting the optimum time identify normal or state. For precise diagnosis,...

An enhanced intelligent diagnosis method for rotary equipment is proposed based on multi-sensor data-fusion and an improved deep convolutional neural network (CNN) models. CNN LeNet-5 constructed which can enhance the features of samples by stacking bottleneck layers without changing size samples. A new conversion approaches are also converting vibration signals into color images, it refine enlarge differences between different types fault fused images transformed in red-green-blue (RGB)...

In the past few years, deep learning techniques for predicting remaining useful life (RUL) have shown remarkable advancements, but model prediction accuracy and generalization to different data still need be improved. Moreover, complex interactions among high-dimensional variables within multidimensional time series (MTSD) can an impact on predictive outcomes. To address these challenges, a RUL approach based spatiotemporal graph convolutional network nested parallel route (GCN-PR) is...

A step-by-step fuzzy diagnostic method based on frequency-domain symptom extraction and trivalent logic diagnosis theory (TLFD), which is established by combining the inference with possibility theories, proposed herein. The features for diagnosing a number of abnormal states are extracted sequentially from measured signals using statistical tests in frequency domain. parameters (SPs) that can sensitively reflect symptoms then selected to provide effective information discrimination each...

Rotating machinery is widely applied in industrial fields. However, it generally operates under tough working conditions, which leads to the weak fault features and renders diagnosis more difficult. In this case, an emerging method called sparse representation classification (SRC) proposed enhance identify status. typical SRC theory fails consider locality of test sample training sample, set contains much redundant information, may reduce recognition accuracy. Moreover, time-shift deviation...

Abstract As a vital constituent of rotating machinery, rolling bearings assume pivotal function in ensuring the stable operation equipment. Recently, deep learning (DL)-based methods have been able to diagnose bearing faults accurately. However, practical applications, severe data imbalance caused by limited availability fault compared abundance healthy poses challenges effective training DL models, leading decrease diagnostic accuracy. In this paper, diagnosis method with improved residual...

Online dictionary learning (ODL) is an emerging and efficient algorithm, which can extract fault features information of signals in most occasions. However, the typical ODL algorithm fails to consider interference noise structural signals, leads weak that are difficult extract. For that, a novel feature enhancement method based on improved constraint model (ICM-ODL) has been proposed this paper. stage learning, elastic-net used promote anti-noise performance atoms. sparse coding, l <sub...

When it comes to intelligent diagnosis, the transfer of knowledge and experiences across machines is crucial. By transferring learned fault features abstract representations from one machine another, diagnosis algorithm can adapt new scenarios, thereby enhancing its versatility robustness. However, for some learning-based methods, accuracy cross-working condition satisfactory, whereas cross-equipment disappointing. To address this issue, a semisupervised graph convolutional networks...

Since roller bearing is one of the most vulnerable components, faults usually occur in an unprepared situation with multiple faults, and quantity sensors limited real-time working environment, resulting underdetermined blind source separation (UBSS) problem to extract fault features. Because collected signals are not independent sparse enough, traditional methods separating cannot perform well. In this paper, optimized intrinsic characteristic-scale decomposition (OICD) method proposed solve...

Sparse signal is a kind of sparse matrices which can carry fault information and simplify the at same time. This effectively reduce cost storage, improve efficiency data transmission, ultimately save equipment diagnosis in aviation field. At present, existing decomposition methods generally extract characteristics signals based on orthogonal basis atoms, limits adaptability decomposition. In this paper, self-adaptive atom extracted by improved dual-channel tunable Q-factor wavelet transform...

Compound faults often occur simultaneously or successively due to the complexity of intelligent mechatronic systems. The generation such group will bring more difficulties fault diagnosis. To separate compound under complex condition and improve accuracy separated signal, a step-by-step diagnosis method for equipment based on majorization-minimization (MM) constraint sparse component analysis (SCA) is proposed in this article. can perform that measurements are not enough signal sparsity...

This paper developed the basic theories and techniques for a plant machinery diagnosis robot (MDR). The workplace of MDR is usually large-scale or place with dangerous environment. It must autonomously carry out condition monitoring sensors in order to detect faults securing safety plant. In this paper, function, concept structure are discussed. intelligent control method by rough set, fuzzy neural network (FNN) self-location-azimuth correction that patrols on inspection route proposed....

Difficulty in extracting nonlinear sparse impulse features due to variable speed conditions and redundant noise interference leads challenges diagnosing faults. Therefore, an improved spectral amplitude modulation based on feature adaptive convolution (ISAM-SFAC) is proposed enhance the fault under condition. First, optimal bi-damped wavelet construction method learn signal features, which selects parameters with correlation criterion particle swarm optimization (PSO). Second, a...