Jiancheng Gong

This paper presents a precise time-frequency positioning method that assists in the fault diagnosis of each piston in piston pumps. Firstly, in response to the short-time broadband characteristics of the piston pump vibration pulse signal, the Local Maximum Transient Extraction Transform (LTET) is proposed to provide a time-frequency representation with more concentrated energy. Secondly, the parameter setting of LTET is discussed, and the improved LTET method (ILTET) is proposed to achieve the precise positioning of the 2D GD of piston pump vibration pulses. Finally, by combining ILTET with the time-frequency envelope and time-frequency mask method, a Periodic Pulse Screening method based on ILTET (ILTET-PPS) for piston pumps is proposed, which can effectively screen the piston vibration pulse, filter out random pulses and noise interference. Experimental validation using vibration data from six piston pump operating conditions demonstrates that the Rényi entropy value of the TFR obtained by ILTET is the smallest, and the non-zero values at the center frequency slice can accurately indicate the piston pulse components while disregarding noise components. In contrast, the TFR center frequency slices of the STET, TET, TMSST, and STFT methods exhibit a significant amount of noise components. This proves that ILTET exhibits superior energy concentration and noise robustness compared to other methods. The PPS method can more accurately extract high-quality piston pulses based on the TFR provided by ILTET, while the TFR from other methods can lead to the extraction of noise components in the PPS process. This demonstrates the effectiveness and superiority of ILTET-PPS.

This article proposes a time–frequency localization-assisted method for extracting fault features from piston pumps. This method can optimize the sample setting of intelligent fault diagnosis models, enhance diagnostic effectiveness, and reduce model complexity. First, ILTET-PPSE method is proposed, which integrates the improved local maximum transient extraction transform (ILTET) and periodic pulse screening and extraction (PPSE) methods. This approach enables precise localization, segmentation, extraction, and combination of vibration signals from a single piston. Subsequently, considering the nonlinear and nonstationary characteristics of the vibration signals, a novel fault feature extraction algorithm termed fractional-order bubble entropy (FBE) is proposed. Furthermore, a fault feature extraction method, PPSE-FBE, assisted by ILTET-PPSE, is introduced. PPSE-FBE demonstrates significant advantages in terms of parameter setting, noise robustness, and fault feature extraction sufficiency. The fault diagnosis model trained with the features extracted by PPSE-FBE can achieve higher recognition accuracy.

Traditional fault diagnosis methods often require extracting features from raw vibration signals based on prior knowledge, which are then input into intelligent classifiers for pattern recognition. This process is prone to information loss and can be inaccurate when relying on human experience for fault identification. To address this issue, this paper proposes an intelligent fault classification and diagnosis model for rolling bearings based on Fast Fourier Transform (FFT) combined with a time convolutional network (SE-TCN) incorporating an attention mechanism, with a Support Vector Machine (SVM) used as the classifier. First, the FFT is applied to transform the collected raw time-domain data of bearing faults into the frequency domain, obtaining the sequence information in the frequency domain. Second, the frequency–domain sequence data are fed into the SE-TCN model, which uses multiple convolutional layers and a channel attention mechanism to extract deep fault features. Finally, the extracted feature vectors are input into the SVM classifier, and the Particle Swarm Optimization (PSO) algorithm is used to optimize the SVM parameters. The optimal separating hyperplane is obtained through training to classify the fault types of the rolling bearings. To verify the effectiveness and diagnostic performance of the proposed method, experiments are conducted using bearing fault datasets from Case Western Reserve University (CWRU) and a laboratory self-built fault diagnosis experimental platform. The experimental results show that the classification accuracy of the proposed method exceeds 99% on the CWRU test dataset, and it also demonstrates advantages in handling small sample data, with an accuracy of over 90%. Additionally, it exhibits good diagnostic performance on the bearing fault data collected from the laboratory self-built platform. The results validate the effectiveness of the proposed classification model in bearing a fault diagnosis.

As a valuable method for quantifying irregularity and randomness, multivariate multiscale permutation entropy (MMPE) has found widespread application in feature extraction and complexity analysis of synchronized multi-channel data. Nonetheless, MMPE fails to consider the amplitude information of the data, and its coarse-graining process possesses inherent flaws, resulting in inaccuracies in evaluating entropy values. To address these issues, a novel nonlinear dynamic characteristic evaluation index, named refined composite multivariate multiscale weighted permutation entropy (RCMMWPE), has been developed. This index aims to comprehensively rectify the shortcomings of disregarding amplitude characteristics and incomplete coarse-graining analysis in MMPE, thereby preserving crucial information present in the original time series data. Through the analysis and comparison of multi-channel synthetic signals, the efficacy and superiority of RCMMWPE in assessing the complexity of synchronized multi-channel data have been confirmed. Subsequently, an intelligent fault detection framework is introduced, leveraging RCMMWPE, multicluster feature selection (MCFS), and kernel extreme learning machine optimized by the particle swarm optimization algorithm (PSO-KELM). The proposed fault detection scheme is then applied to test gearbox fault data and extensively benchmarked against other fault detection schemes. The results demonstrate that the proposed gearbox fault detection scheme excels in accurately and consistently identifying fault categories, outperforming the comparison schemes.

Rolling bearings usually works in the condition with high speed, high temperature and high pressure, so their faults need to be detected and judged in time. This paper presents a new multi-scale entropy measurement algorithm called time-shift multi-scale amplitude-aware permutation entropy (TSMAAPE), combining time-shift coarse-grained method and multi-scale amplitude-aware permutation entropy (MAAPE) algorithm together creatively, making the extraction of useful information in time series more sufficient. In this paper, this new entropy algorithm is combined with uniform phase empirical mode decomposition algorithm, improved max-relevance and min-redundancy (ImRMR) algorithm and random forest algorithm, which are used to implement signal preprocessing, select features for dimensionality reduction, execute fault recognition and classification respectively, forming a new comprehensive detection method of bearing fault. The new comprehensive method includes two crucial parts: fault pre-detection and fault identification. Firstly, the threshold set based on TSMAAPE method is used to judge whether the bearing is in fault or not, if there is a fault, the subsequent steps are performed to identify and classify the fault. Compared with other traditional feature extraction methods, the proposed method is more effective and robust. The experimental results of two groups of bearing data show that the average accuracy of this method can reach 98.6%. The comprehensive detection method is of great significance for the rapid and accurate detection and identification of bearing faults in practical engineering applications.