Geometry-based image modeling method for intelligent state identification and fault prediction of wind turbines

Just-in-time defect prediction based on AST change embedding

Deep Learning is the latest research achievement in the field of Artificial Intelligence. To detect and identify the state of circuit breaker in ambient air automatic monitor station, Deep Learning was proposed to realize the circuit breaker state recognition in real-time. The Deep Convolutional Neural Network was used to construct the state detection and identification model of the circuit breaker and provided the detection process. The remote monitoring system of ambient air automatic monitor station and inspection data were used to build the experimental data platform, with the circuit breaker state recognition data was acquired from the multiple ambient air automatic monitor stations. Through training the model with the Deep Learning, the experimental results showed that the circuit breaker state detection and identification method have a simple recognition process and high accuracy. The real-time Detection and Identification function of the circuit breaker state was realized, meanwhile, the assistant decision-making function was provided to the ambient air automatic monitor station.

Fault prognosis technology, as the key to realize autonomy support and carry out condition-based maintenance (CBM), has been fully developed and researched in recent years. In this paper, the existing fault prediction methods are briefly summarized and studied, and focused on the data-driven prediction method. In the light of the scale of data, all kinds of prediction methods are divided into small sample-based and large sample-based fault prediction methods. The basic principles, advantages and disadvantages of each type of typical analytical methods are summarized, and finally, we put forward the research prospect the development direction of future data-driven fault prediction.

In the traditional software defect prediction methodology, the historical record (dataset) of the same project is partitioned into training and testing data. In a practical situation where the project to be predicted is new, traditional software defect prediction cannot be employed. An alternative method is cross-project defect prediction, where the historical record of one project (source) is used to predict the defect status of another project (target). The cross-project defect prediction method solves the limitations of the historical records in the traditional software defect prediction method. However, the performance of cross-project defect prediction is relatively low because of the distribution differences between the source and target projects. Furthermore, the software defect dataset used for cross-project defect prediction is characterized by high-dimensional features, some of which are irrelevant and contribute to low performance. To resolve these two issues, this study proposes a transformation and feature selection approach to reduce the distribution difference and high-dimensional features in cross-project defect prediction. A comparative experiment was conducted on publicly available datasets from the AEEEM. Analysis of the results obtained shows that the proposed approach in conjugation with random forest as the classification model outperformed the other four state-of-the-art cross-project defect prediction methods based on the commonly used performance evaluation metric F1_score.

Recent studies of software defect prediction typically produce datasets, methods and frameworks which allow software engineers to focus on development activities in terms of defect-prone code, thereby improving software quality and making better use of resources. Many software defect prediction datasets, methods and frameworks are published disparate and complex, thus a comprehensive picture of the current state of defect prediction research that exists is missing. This literature review aims to identify and analyze the research trends, datasets, methods and frameworks used in software defect prediction research betweeen 2000 and 2013. Based on the defined inclusion and exclusion criteria, 71 software defect prediction studies published between January 2000 and December 2013 were remained and selected to be investigated further. This literature review has been undertaken as a systematic literature review. Systematic literature review is defined as a process of identifying, assessing, and interpreting all available research evidence with the purpose to provide answers for specific research questions. Analysis of the selected primary studies revealed that current software defect prediction research focuses on five topics and trends: estimation, association, classification, clustering and dataset analysis. The total distribution of defect prediction methods is as follows. 77.46% of the research studies are related to classification methods, 14.08% of the studies focused on estimation methods, and 1.41% of the studies concerned on clustering and association methods. In addition, 64.79% of the research studies used public datasets and 35.21% of the research studies used private datasets. Nineteen different methods have been applied to predict software defects. From the nineteen methods, seven most applied methods in software defect prediction are identified. Researchers proposed some techniques for improving the accuracy of machine learning classifier for software defect prediction by ensembling some machine learning methods, by using boosting algorithm, by adding feature selection and by using parameter optimization for some classifiers. The results of this research also identified three frameworks that are highly cited and therefore influential in the software defect prediction field. They are Menzies et al. Framework, Lessmann et al. Framework, and Song et al. Framework.

The small fault with a vertical displacement (or drop) of 2–5 m has now become an important factor affecting the production efficiency and safety of coal mines. When the 3D seismic data contain noise, it is easy to cause large errors in the prediction results of small faults. This paper proposes an intelligent small fault identification method combining variable mode decomposition (VMD) and a support vector machine (SVM). A fault forward model is established to analyze the response characteristics of different seismic attributes under the condition of random noise. The results show that VMD can effectively realize the attenuation of random noise and the seismic attributes extracted on this basis have a good correlation with the small fault. Through the analysis of the SVM algorithm and the fault forward model, it is proved that it is feasible to realize intelligent predictions of small faults by using seismic attributes as the input of a SVM. The fault prediction method using a SVM that is proposed in this paper has higher accuracy than the principal component analysis method, as the prediction results have important guiding significance and reference value for later coal mining. Therefore, the method presented in this paper can be used as a new intelligent method for small fault identification in coal fields.

The oil of wind turbine gearbox contains a lot of information related to wear state. Through the real-time monitoring of oil abrasive particle concentration, moisture, viscosity, density, temperature, acid value, dielectric constant and other parameters, we can get the relevant information about gearbox wear state. Thus, a modeling method of wind turbine gearbox fault prediction based on partial least squares regression analysis is proposed. This modeling method can establish the regression analysis between wear state and key oil parameters, which provides a scientific basis for fault prediction of fan gearbox and has broad application prospects.

In order to improve the accuracy of distribution equipment fault prediction, a distribution equipment fault prediction method based on multi-source data fusion and hybrid neural network is proposed. Firstly, after preprocessing the multi-source data such as equipment fault and meteorological information in the distribution network, the outliers of the data are removed by K-means method, and the feature vector of the data is obtained by wavelet packet decomposition; Secondly, the features of the preprocessed data are extracted by convolutional neural network (CNN), and the extracted features are processed by long short term memory (LSTM), and then the fault is predicted; Finally, the actual distribution network equipment operation data in Henan Province are analyzed. The results show that compared with the compared fault prediction methods, the proposed method can obtain more reliability parameters according to the equipment characteristics, significantly improve the recall rate, accuracy and F <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> value, and can predict the distribution network equipment faults more accurately.

Modern society is to a larger and larger extent dependant on electric energy, and hence the reliance on and utilization of the electric grid is increasing steadily. At the same time the production and consumption patterns are changing from large centralized generation of electric power and pure consumers to distributed generation (DG) and more complex consumers. This transition causes higher stress on an aging infrastructure and major investments are required over the coming years to maintain a reliable supply of electric energy. Better monitoring solutions and predictive methods can increase the possible utilization of the existing grid and reduce the fault frequency. This paper presents some current challenges in the grid and a possible monitoring solution and fault prediction method. This is exemplified with statistics and field-measurements from the Norwegian power grid.

The environment of the three-trans towers is complicated, with the influence of strong noise and non-deterministic interference, the features of the vibration response signal of bolts are very weak, which is difficult to identify the bolts’ state. In order to avoid man-made damage to the tower and realize the identification of bolt state under natural vibration conditions, a state identification method based on quadratic wavelet transform and KNN was proposed. First, the wavelet denoising method based on new threshold function was used to filter out the noise in original signal. Then, the dB10 wavelet function commonly used in fault diagnosis was selected to decompose and reconstruct the denoised signal, the dimensionless time domain characteristics of the low-frequency signal component and the information entropy of the high-frequency detail component were calculated respectively. Finally, the bolts state feature dataset was established. By comparing common classifiers, KNN with the best performance was selected. The experimental case analysis verified the effectiveness of the proposed method for bolts state identification. The overall research has certain practical and engineering value.

Viscoelastic sandwich structure is widely used in mechanical equipment. However, therein viscoelastic material inevitably suffers from gradual aging. To keep equipment running safety, it is urgent to perform the aging state identification of viscoelastic sandwich structure by vibration response signal analysis. Nevertheless, the structural vibration response signal is non-stationary and its variation caused by the structural aging state change is very puny. The vibration-based structural aging state identification has become a challenging task. Therefore, a novel method based on variational mode decomposition (VMD) with parameter optimization by sparrow search algorithm (SSA), and deep belief network (DBN) is proposed for this task. To extract sensitive aging feature information, the structural vibration response signal is processed by optimized variational mode decomposition with sparrow search algorithm (SSA-VMD), and multiple permutation entropy (PE) features are extracted from the intrinsic mode functions (IMFs) for reflecting. To attain intelligent and desirable aging state identification, using the extracted PE features as input, DBN is introduced for identifying structural aging states. The proposed method is performed on a viscoelastic sandwich structure to validate its feasibility and efficiency, and is compared with the traditional methods. The results show that the proposed method can obtain a better aging state intelligent identification result and has a good application prospect.

In the last few years, with the development of deep learning theory, researchers have tried to introduce the method of artificial intelligence into the field of software defect prediction (SDP) to improve its prediction effect. To be fed into the neural network, the sample codes are represented as an abstract syntax tree (AST), and the AST is encoded as real numbers. However, in most cross-project defect prediction (CPDP) task, the method for converting the AST into a real number cannot effectively estimate the semantic distance between the ASTs, resulting in a significant reduction in training effects. To solve that problem, we present a new encoding framework, tree-based-embedding (TBE), to convert AST into real vectors and make the semantic gap between the ASTs measurable. To estimate the effect of this encoding method, we promise a tree-based-embedding convolutional neural network with transferable hybrid feature learning (TBCNN-THFL) to perform the CPDP tasks. TBCNN-THFL is fed data encoded with TBE method for learning the transferable joint features between different projects; meanwhile, TBCNN-THFL introduces a transfer component analysis algorithm. Furthermore, the model combines the handcrafted and deep-learning-generated features and then feeds them into the classifier to train a defect prediction model. A sufficient number of experiments demonstrate that TBCNN-THFL is superior to referential models on 72 pairs of CPDP tasks formed by 9 open-source projects.

Cross-project defect prediction (CPDP) uses the labeled data from external source software projects to compensate the shortage of useful data in the target project, in order to build a meaningful classification model. However, the distribution gap between software features extracted from the source and the target projects may be too large to make the mixed data useful for training. In this paper, we propose a cluster-based novel method FeSCH (Feature Selection Using Clusters of Hybrid-Data) to alleviate the distribution differences by feature selection. FeSCH includes two phases. The feature clustering phase clusters features using a density-based clustering method, and the feature selection phase selects features from each cluster using a ranking strategy. For CPDP, we design three different heuristic ranking strategies in the second phase. To investigate the prediction performance of FeSCH, we design experiments based on real-world software projects, and study the effects of design options in FeSCH (such as ranking strategy, feature selection ratio, and classifiers). The experimental results prove the effectiveness of FeSCH. Firstly, compared with the state-of-the-art baseline methods, FeSCH achieves better performance and its performance is less affected by the classifiers used. Secondly, FeSCH enhances the performance by effectively selecting features across feature categories, and provides guidelines for selecting useful features for defect prediction.

At present, most transformer fault prediction methods mainly focus on the prediction of state parameters, and the accuracy of the prediction results depends heavily on the quality of the data that is used by them. Aiming at solving these problems, a fault prediction method of power transformer by association rules and Markov is proposed. Firstly, apriori algorithm is applied for mining the association rules between different states to establish the state transition matrix of transformer states. Then a multi-dimensional correction factor system including family defects, operating environment and overhauling records is established and the relative degree of degradation is introduced to calculate them. Besides, an array-based apriori algorithm is proposed to realize the weight of correction factors by combining the mining rules of correction factors and states. Finally, the state of the transformer is predicted with the existing transformer state probability as the initial vector and the modified state transition matrix. The results reveal that the established prediction model can predict the development trend of transformer faults effectively and the modified transition probability matrix can make full use of various state information of power transformers so as to improve the effect of transformer fault prediction.

Traditional fault prediction methods are difficult to meet the needs of modern communication networks because of the limitations of static model, dependence on a single data source and lack of real-time processing ability. Graph neural network (GNN) can effectively capture the complex relationship between network devices and improve the accuracy and efficiency of fault prediction by virtue of its topological awareness, dynamic modeling potential and multi-modal integration advantages. In this paper, a cross-modal fusion method based on multi-scale dynamic GNN (MSGNN-CM) is proposed. The communication network is modeled as a dynamic graph, and the local and global topological features are extracted through the convolution of multi-scale graphs. The time-series evolution of network state is captured by Gated Recurrent Unit (GRU), and the topology, time-series and alarm information are fused through attention mechanism, so as to realize node-level fault probability prediction and critical link risk identification. This method has the advantages of multi-scale perception, dynamic adaptability and multi-source collaborative integration, and can effectively cope with the dynamic changes and complexity of the network. The experiment uses six months' operation and maintenance data of three city base stations, involving 500 nodes and 2,100 edges. The characteristics of nodes include traffic, delay and CPU load, while the characteristics of edges include bandwidth utilization and packet loss rate. The results show that MSGNN-CM performs best in node-level fault prediction tasks, with F1-Score and AUC-ROC reaching 88.6% and 92.3% respectively, which are significantly better than traditional methods and existing GNN models. In addition, this method also performs well in link-level risk detection, with Precision@10 and Precision@20 reaching 53.8% and 45.2% respectively. The analysis of dynamic adaptability shows that choosing a suitable time window is very important to the performance of the model, and GRU mechanism has the ability to respond quickly to sudden events. Explanatory verification shows that the cross-modal attention mechanism can effectively identify key failure indicators and improve the transparency and practicability of the model.