A parameter-adaptive VMD method based on grasshopper optimization algorithm to analyze vibration signals from rotating machinery

A novel optimized SVM classification algorithm with multi-domain feature and its application to fault diagnosis of rolling bearing

Unsupervised feature learning from the raw vibration data is a great challenge for rolling bearing intelligent fault diagnosis. In this paper, a novel method called deep wavelet auto-encoder (DWAE) with extreme learning machine (ELM) is proposed for intelligent fault diagnosis of rolling bearing. Firstly, wavelet function is employed as the nonlinear activation function to design wavelet auto-encoder (WAE), which can effectively capture the signal characteristics. Secondly, a DWAE is constructed with multiple WAEs to enhance the unsupervised feature learning ability. Finally, ELM is adopted as the classifier to accurately identify different bearing faults. The proposed method is applied to analyze the experimental bearing vibration signals, and the results confirm that the proposed method is superior to the traditional methods and standard deep learning methods.

Intelligent fault diagnosis of rolling bearing using deep wavelet auto-encoder with extreme learning machine

Parameter-adaptive variational mode decomposition (VMD) has attenuated the dominant effect of prior parameters, especially the predefined mode number and balancing parameter, which heavily trouble the traditional VMD. However, parameter-adaptive VMD still encounters some problems when it is applied to the data from industry applications. On one hand, the mode number chosen using parameter-adaptive VMD is not the optimal. Numbers of redundant modes are decomposed. On another hand, parameter-adaptive VMD has much space for the improvement when it is applied to compound-fault diagnosis. To solve these issues and further enhance its performance, an improved parameter-adaptive VMD (IPAVMD) is proposed in this paper. Firstly, a new index, called ensemble kurtosis, is constructed by combining with kurtosis and the envelope spectrum kurtosis. It can simultaneously take the cyclostationary and impulsiveness into consideration. Secondly, the optimization objective function of grasshopper optimization algorithm is improved based on the ensemble kurtosis. The improved method chooses the mean value of the ensemble kurtosis of all modes rather than that of the individual mode as objective function. Thirdly, to extract all potential fault information, an iteration algorithm is used in the new method. Benefiting from these improvements, the proposed IPAVMD outperforms the traditional parameter-adaptive VMD and further expands the application to compound-fault diagnosis. It has been verified by a series of simulated signals and a real dataset from the axle box bearings of locomotive.

Identification of mechanical compound-fault based on the improved parameter-adaptive variational mode decomposition.

The safety and public health during nuclear power plant operation can be enhanced by accurately recognizing and diagnosing potential problems when a malfunction occurs. However, there are still obvious technological gaps in fault diagnosis applications, mainly because adopting a single fault diagnosis method may reduce fault diagnosis accuracy. In addition, some of the proposed solutions rely heavily on fault examples, which cannot fully cover future possible fault modes in nuclear plant operation. This paper presents the results of a research in hybrid fault diagnosis techniques that utilizes support vector machine (SVM) and improved particle swarm optimization (PSO) to perform further diagnosis on the basis of qualitative reasoning by knowledge-based preliminary diagnosis and sample data provided by an on-line simulation model. Further, SVM has relatively good classification ability with small samples compared to other machine learning methodologies. However, there are some challenges in the selection of hyper-parameters in SVM that warrants the adoption of intelligent optimization algorithms. Hence, the major contribution of this paper is to propose a hybrid fault diagnosis method with a comprehensive and reasonable design. Also, improved PSO combined with a variety of search strategies are achieved and compared with other current optimization algorithms. Simulation tests are used to verify the accuracy and interpretability of research findings presented in this paper, which would be beneficial for intelligent execution of nuclear power plant operation.

A hybrid fault diagnosis methodology with support vector machine and improved particle swarm optimization for nuclear power plants.

This paper proposes a hybrid intelligent method for multi-fault detection of rotating machinery, in which three methods, i.e. including the redundant second generation wavelet package transform (RSGWPT), the kernel principal component analysis (KPCA) and the twin support vector machine (TWSVM), are combined. Firstly, RSGWPT is used to extract feature vectors from representative statistical characteristics in the decomposition frequency band, and then the KPCA in the feature space is performed to reduce the dimension of features and to extract the dominant features for the following classification. Finally, a novel support vector machine, called twin support vector machine is used to construct a multi-class classifier. Inputting superior features to this classifier, the condition of the monitored machine component can be determined. Experimental results demonstrate that the proposed hybrid method is effective for multi-fault detection of rotating machinery. The TWSVM is also indicated that has better classification performance and faster convergence speed than the normal SVM.

A hybrid intelligent multi-fault detection method for rotating machinery based on RSGWPT, KPCA and Twin SVM.

Variational time-domain decomposition of reciprocating machine multi-impact vibration signals

Due to the importance of bearings in modern machinery, the prediction of the remaining useful life (RUL) of rolling bearings has been widely studied. When predicting the RUL of rolling bearings in engineering practice, the RUL is usually predicted based on historical data, and as the historical data increases, the prediction results should be more accurate. However, the existing methods usually have the shortcomings of low prediction accuracy, large cumulative error and failure to dynamically give prediction results with the increase of historical data, which are not suitable for engineering practice.To address the above problems, a novel RUL prediction method is proposed. The proposed method consists of 3 parts: First, the multi-scale entropy-based feature – namely “average multi-scale morphological gradient power spectral information entropy (AMMGPSIE)” – from the rolling bearings as the Health indicator (HI) is extracted to ensure all the fault-related information is well-included; Then, the HI is processed with the enhanced Hodrick Prescott trend-filtering with boundary lines (HPTF-BL) to ensure good performance and small fluctuation on the HI; Finally, the deterioration curve is predicted using an LSTM neural network and the improved Particle Filter algorithm that we proposed. The proposed method is validated using the experimental bearing degradation dataset and the casing data of a centrifugal pump bearing from an actual industrial site. Comparing the results with other recent RUL prediction methods, the proposed method achieved state-of-the-art feasibility and effectiveness, conform to the needs of practical application of the project.

https://ieeexplore.ieee.org/ielx7/6287639/6514899/10006785.pdf

Remaining Useful Life Prediction for Rolling Bearings With a Novel Entropy-Based Health Indicator and Improved Particle Filter Algorithm

SI is a relatively recent technology that was inspired by observations of natural social insects and artificial systems. This system comprises multiple individual agents who rely on collective behavior in decentralized and self-organized networks. One of the biggest difficulties for existing computer techniques is learning from such large datasets, which is addressed utilizing big data. Big data-based categorization refers to the challenge of determining which set of classifications a new discovery belongs to. This classification is based on a training set of data that comprises observations that have been assigned to a certain category. In this paper, CIN-big data value calculation based on particle swarm optimization (BD-PSO) algorithm is proposed by operating in local optima and to improve the operating efficiency. The convergence speed of the particle swarm optimization (PSO), which operates in the local optima, is improved by big data-based particle swarm optimization (BD-PSO). It improves computing efficiency by improving the method, resulting in a reduction in calculation time. The performance of the BD-PSO is tested on four benchmark dataset, which is taken from the UCI. The datasets used for evaluation are wine, iris, blood transfusion, and zoo. SVM and CG-CNB are the two existing methods used for the comparison of BD-PSO. It achieves 92% of accuracy, 92% of precision, 92% of recall, and 1.34 of F1 measure, and time taken for execution is 149 ms, which in turn outperforms the existing approaches. It achieves robust solutions and identifies appropriate intelligent technique related to the optimization problem.

/pdf/big-data-value-calculation-method-based-on-particle-swarm-2id81md1.pdf

Wensheng Ma

Papers

A hybrid intelligent multi-fault detection method for rotating machinery based on RSGWPT, KPCA and Twin SVM.

Variational time-domain decomposition of reciprocating machine multi-impact vibration signals

Remaining Useful Life Prediction for Rolling Bearings With a Novel Entropy-Based Health Indicator and Improved Particle Filter Algorithm

Big Data Value Calculation Method Based on Particle Swarm Optimization Algorithm