Design of Fault Diagnosis Algorithm for Electromechanical System Based on Artificial Intelligence

Bearing fault diagnosis is a critical component of the mechanical equipment monitoring system. In the complex and harsh environment in which bearings operate, the fault diagnosis approach of multi-source information fusion can extract fault features more stably and extensively than the traditional single-source fault diagnosis method. However, most existing multi-source fusion methods are in infancy, and there are a number of pressing issues to address, such as subjective elements having a significant impact, excessive data redundancy, fuzzy multi-source signal fusion strategy, and insufficient accuracy. As a result, a new multi-source fusion fault diagnosis method is proposed in this paper. First, the residual pyramid algorithm is utilized to fuse the acoustic and vibration signals of multiple spatial positions respectively and then form two fused acoustic and vibration signals. Second, two improved 2D-CNN are used to extract the fault features contained in the above two signals separately to form a multi-source fault feature set. Third, an AdaBoost algorithm with a dynamic deletion mechanism is designed to fuse multi-source fault feature sets and produce the fault diagnosis findings. Finally, six different experimental data sets are used to test the performance of the model. The results reveal that the model has better generalization, higher and more stable fault diagnostic accuracy, and stronger anti-interference capacity.

The fault diagnosis method based on neural network has many defects, such as complicated network, long training time and slow convergence speed. The particle swarm optimization and neural network integration fault diagnosis methods are proposed to improve the fault diagnosis capability. Firstly, the heuristic global optimization capability of particle swarm optimization is used to optimize the neural network connection weights; then the transformer fault samples are trained and tested by using non-linear processing capacity of neural network. The test results show that such algorithm can effectively avoid unstable neural network, easily falling into local minimum and lower diagnostic accuracy etc. and can effectively increase the convergence speed and fault diagnosis efficiency compared with traditional fault diagnosis method.

Fault diagnosis and prognosis are important tools for the reliability, availability and survivability of navy all-electric ships. Extending the fault detection and diagnosis into predictive maintenance increases the value of this technology. The traditional diagnosis can be viewed as a single diagnostic agent having a model of the whole system to be diagnosed. This becomes inadequate when the system becomes large and distributed as on the electric ships. For such systems, the software multi-agents may offer a solution. This paper first presents a brief review on the traditional fault diagnosis method with an emphasis on its application to electric motors as important components on the all-electric ship. The software agent technology is then introduced. The discussion is made about how this technology supports the drastic manning reduction requirements for the future navy ships. Examples are given on the existing naval applications of diagnostic and prognostic software agents.

To find out the hidden danger in the industrial production process in time, it is necessary to monitor the health condition of the key components of the mechanical system in operation. However, traditional fault diagnosis methods usually adopt manual feature extraction, which not only costs expensively and depends on prior knowledge. Therefore, it is of great significance to study the application of automatic fault identification based on the original vibration signals. Recently, existing studies have shown that most of fault diagnoses are implemented by using deep neural network. Although these methods have achieved satisfactory performances, there are obvious limitations in real applications, that is, the complexity of deep neural network requires a lot of hardware computing resources. This hinders the development of online fault diagnosis tools. To solve this problem, this paper proposes a fault diagnosis model based on lightweight convolutional neural network MobileNet, and realizes an end-to-end intelligent fault classification and diagnosis application. We evaluated the proposed method with the rolling bearing dataset from Western Reserve University. The best average precision, recall and F1 score of ten different bearing health conditions are about 96%, 82% and 88%, respectively. In addition, we also compare the accuracy of the rolling bearing fault diagnosis classification model under the standard ReLU and the improved ReLU. Experimental results show that both obtain good performances, but the improved ReLU reaches the over 96% accuracy more rapidly.

The products of a batch process have high economic value. Meanwhile, a batch process involves complex chemicals and equipment. The variability of its operation leads to a high failure rate. Therefore, early fault diagnosis of batch processes is of great significance. Usually, the available information of the sensor data in batch processing is obscured by its noise. The multistage variation of data results in poor diagnostic performance. This paper constructed a standardized method to enlarge fault information as well as a batch fault diagnosis method based on trend analysis. First, an adaptive standardization based on the time window was created; second, utilizing quadratic fitting, we extracted a data trend under the window; third, a new trend recognition method based on the Euclidean distance calculation principle was composed. The method was verified in penicillin fermentation. We constructed two test datasets: one based on an existing batch, and one based on an unknown batch. The average diagnostic rate of each group was 100% and 87.5%; the mean diagnosis time was the same; 0.2083 h. Compared with traditional fault diagnosis methods, this algorithm has better fault diagnosis ability and feature extraction ability.

It is difficult for the traditional photovoltaic array fault feature extraction method to include as much fault information as possible, which leads to the problem of low accuracy. This paper proposes a fault diagnosis method for photovoltaic arrays based on Dropout optimized Stacked autoencoder (D-SAE). Since the deep-stacked network structure of SAE superimposes multiple hidden layers, the model parameters are increased, resulting in overfitting. To prevent the model from overfitting and improve the generalization ability and convergence speed of the model, this paper introduces the Dropout method and the Adam algorithm to optimize the stacked autoencoder. In this paper, the I-V curve is used as the input of the fault diagnosis model, and the fault diagnosis model is established to realize the end-to-end fault diagnosis of the photovoltaic array. The experimental results show that, compared with the traditional fault diagnosis method, the D-SAE diagnosis model can not only automatically extract the fault characteristics of the photovoltaic array, prevent the occurrence of overfitting, but also accelerate the convergence speed of the model and improve the generalization ability of the model, the accuracy rate is higher.

Parallel sparse filtering for intelligent fault diagnosis using acoustic signal processing

Rolling bearings are vital components in rotary machines where even minor faults can lead to the collapse of the entire mechanical system, resulting in significant economic losses. The condition monitoring and fault diagnosis of rolling bearings are crucial for safe operation of complex industrial systems. Traditional fault diagnosis methods mainly depend on expert experience, which leads to obvious deficiencies in diagnostic efficiency, especially in terms of accuracy. To address the issues of traditional bearing fault diagnosis relying on expert experience and low efficiency and accuracy of fault diagnosis, a bearing fault diagnosis model combining synchrosqueezed wavelet transform (SWT) and SE-ResNet is proposed. Firstly, the one-dimensional non-stationary bearing vibration signal is converted into a high-frequency two-dimensional time-frequency image through SWT, which serves as the input for the convolutional neural network; Secondly, an attention mechanism is introduced to construct an SE-ResNet model for extracting two-dimensional image features, which enhances the attention of important features and thus enhances the network's representation ability; Finally, the results of fault diagnosis are output through Softmax layer. To validate the capacity of the proposed model, experiments with the bearing dataset collected from Case Western Reserve University (CWRU) were conducted. The results show that the transfer learning accuracy of the CWRU dataset is 86.25% to 99%. This method has good feasibility, good convergence performance, and has high practical application value.

The electrohydraulic servo valve plays crucial roles as key component in combustion control system of an aeroengine, and its fault forms are diverse, nonlinear, and closed. At the same time, limited installation space makes it difficult to ensure the installation of a large number of sensors, and traditional neural network fault diagnosis methods cannot meet the demand for fault data. Therefore, in order to accurately diagnose and predict electrohydraulic servo valve faults, a fault diagnosis method based on digital‐analog fusion based on the state space model of electrohydraulic servo valve is innovatively proposed. First, electrohydraulic servo valve state space model is established by the method of mechanism modeling. Then, the Kalman filter model is established to determine the operating state of electrohydraulic servo valve based on the pressure, current command and other signals. Combining the hybrid deep learning models of BiLSTM (Bi‐directional Long Short‐Term Memory)‐ANN (Artificial Neural Network) to identify fault forms of electrohydraulic servo valve, the fault diagnosis strategy with the fusion of Kalman filter and artificial neural network is obtained. Finally, based on the experimental platform of electrohydraulic servo valve performance accelerated degradation, the fault diagnosis accuracy proposed in this paper is verified to be 97.5%. This paper ensures the accuracy of fault diagnosis for electro‐hydraulic servo valves under limited data, and simultaneously achieves the explainability of faults, providing a favorable reference for the research of fault diagnosis algorithms of hydraulic components in practical engineering applications.

Satellite attitude control system (SACS) is prone to faults due to its complexity and harsh working environment. If the fault is not detected in a timely manner to take appropriate countermeasures, it may lead to severe consequences. Traditional fault diagnosis methods have certain limitations, such as the need for accurate mathematical models or enough experience of fault diagnosis. Therefore, this paper proposes a fault diagnosis method for SACS sensors that integrates feature extraction with an optimized back propagation (BP) neural network. Firstly, the features of simulation data are extracted by wavelet packet energy (WPE), which can effectively highlight the fault characteristics. Then the osprey optimization algorithm (OOA) is used to optimize the initial weights and biases of BP neural network, so as to solve the problem that BP neural network is easy to fall into local minimum. The simulation results show that the proposed method can accurately detect and identify various fault types of sensors in SACS, thus providing a new and effective way for fault diagnosis of SACS.

In order to solve the problem that the effect of traditional neural network in fault data diagnosis of power system is not obvious, inspired by the application of deep learning image recognition, this paper combines immune algorithm and deep learning, and proposes a new smart grid fault data analysis method based on deep immune network. Deep immune network is an immune-improved deep learning network, which extracts features from data through the multi-layer structure of convolutional neural network, and inputs feature samples into the immune-optimized radial basis network to classify faults. By analogy with the immune process of biological immune system, a fault diagnosis immune model suitable for power grid fault data analysis is proposed and established. By establishing a one-to-one mapping relationship with biological immune system, the model structure was improved and optimized accordingly. Finally, the experiment proves that the fault diagnosis immune model has excellent learning ability and the fault recognition efficiency is improved compared with the traditional fault diagnosis method.

With the development of machine learning in recent years, the application of machine learning to machine fault diagnosis has become increasingly popular. Applying traditional feature extraction methods for complex systems will weaken the characterization capacity of features, which are not conducive to subsequent classification work. A reciprocating compressor is a complex system. In order to improve the fault diagnosis accuracy of complex systems, this paper does not use traditional fault diagnosis methods and applies deep convolutional neural networks (CNNs) to process this nonlinear and non-stationary fault signal. The valve fault data is obtained from the reciprocating compressor test bench of the Daqing Natural Gas Company. Firstly, the single-channel vibration signal is collected on the reciprocating compressor and the one-dimensional CNN (1-D CNN) is used for fault diagnosis and compared with the traditional model to verify the effectiveness of the 1-D CNN. Next, the collected eight channels signals (three channels of vibration signals, four channels of pressure signals, one channel key phase signal) are applied by 1-D CNN and 2-D CNN for fault diagnosis to verify the CNN that it is still suitable for multi-channel signal processing. Finally, further study on the influence of the input of different channel signal combinations on the model diagnosis accuracy is carried out. Experiments show that the seven-channel signal (three-channel vibration signal, four-channel pressure signal) with the key phase signal removed has the highest diagnostic accuracy in the 2-D CNN. Therefore, proper deletion of useless channels can not only speed up network operations but also improve diagnosis accuracy.

Digital twin is a significant way to achieve smart manufacturing, and provides a new paradigm for fault diagnosis. Traditional data-based fault diagnosis methods mostly assume that the training data and test data are following the same distribution and can acquire sufficient data to train a reliable diagnosis model, which is unrealistic in the dynamic changing production process. In this paper, we present a two-phase digital-twin-assisted fault diagnosis method using deep transfer learning (DFDD), which realizes fault diagnosis both in the development and maintenance phases. At first, the potential problems that are not considered at design time can be discovered through front running the ultra-high-fidelity model in the virtual space, while a deep neural network (DNN)-based diagnosis model will be fully trained. In the second phase, the previously trained diagnosis model can be migrated from the virtual space to physical space using deep transfer learning for real-time monitoring and predictive maintenance. This ensures the accuracy of the diagnosis as well as avoids wasting time and knowledge. A case study about fault diagnosis using DFDD in a car body-side production line is presented. The results show the superiority and feasibility of our proposed method.

With the development of wind power, the fault of wind turbine is increasing year by year. The gearbox is the fault multiple component of the wind turbine. The monitoring system gets massive data, and the health monitoring of the fan has also entered the era of “big data”. The new problem in the field of health monitoring of mechanical equipment is to excavate information from these large data and to identify the health status of equipment efficiently and accurately. Deep learning is an effective tool for large data processing. Based on the strong perception and self-learning ability of deep learning theory, a fault diagnosis method of wind turbine gearbox based on deep belief network and vibration signal is proposed in this paper. The original fault sample vibration signal is input to the deep belief network to get the deep learning sample model. At the same time, Batch Normalization is added to reduce the overfitting probability and improve the convergence speed of the network. Then the trained deep neural network model is applied to the fault diagnosis of the planetary gear box of the wind turbine gearbox, and the results show that the method is more accurate than the DBN, the BPNN algorithm and the traditional fault diagnosis method.

Modular multilevel converters (MMCs) have a complex structure and a large number of submodules (SMs). If there is a fault in one of the SMs, it will affect the reliable operation of the system. Therefore, rapid fault diagnosis and accurate fault positioning are crucial to ensuring the continuous operation of the system. However, the IGBT open-circuit faults in different submodules of MMCs have similar fault features, and the traditional fault feature extraction method cannot effectively extract the key features of the fault so as to accurately locate the faulty submodules. In response to this problem, this paper proposes a fault diagnosis method based on weighted-amplitude permutation entropy (WAPE) and DS evidence fusion theory. The simulation results show that WAPE has better feature extraction ability than basic permutation entropy, and the fused multiscale feature decision output has better diagnostic accuracy than the single-scale feature. Compared with traditional fault diagnosis methods, this method achieves the diagnosis of multiple fault types by collecting a single signal, which greatly reduces the number of samples and leads to higher diagnostic accuracy and faster diagnostic speed.