Rotating Machinery Fault Diagnosis Method Based on Improved Semisupervised Generative Confrontation Network

Error diagnosis and detection have become important in modern production due to the importance of spinning equipment. Artificial neural network pattern recognition methods are widely utilized in rotating equipment fault detection. These methods often need a large quantity of sample data to train the model; however, sample data (especially fault samples) are uncommon in engineering. Preliminary work focuses on dimensionality reduction for big data sets using semisupervised methods. The rotary machine’s polar coordinate signal is used to build a GAN network structure. ANN and tiny samples are utilized to identify DCGAN model flaws. The time-conditional generative adversarial network is proposed for one-dimensional vibration signal defect identification under data imbalance. Finally, auxiliary samples are gathered under similar conditions, and CCNs learn about target sample characteristics. Convolutional neural networks handle the problem of defect identification with small samples in different ways. In high-dimensional data sets with nonlinearities, low fault type recognition rates and fewer marked fault samples may be addressed using kernel semisupervised local Fisher discriminant analysis. The SELF method is used to build the optimum projection transformation matrix from the data set. The KNN classifier then learns low-dimensional features and detects an error kind. Because DCGAN training is unstable and the results are incorrect, an improved deep convolutional generative adversarial network (IDCGAN) is proposed. The tests indicate that the IDCGAN generates more real samples and solves the problem of defect identification in small samples. Time-conditional generation adversarial network data improvement lowers fault diagnosis effort and deep learning model complexity. The TCGAN and CNN are combined to provide superior fault detection under data imbalance. Modeling and experiments demonstrate TCGAN’s use and superiority.

IDENTIFICATION OF INCOMPLETE PENETRATION WELDING DEFECTS ON DIGITAL RADIOGRAPHIC FILM FIGURES WITH THE GEOMETRIC INVARIANT MOMENT

The examination of defects in radiographic films necessitates specialized knowledge, as indicated by an expert radiographer (AR) degree, yet the subjectivity of AR in identifying defects is problematic. To overcome this subjectivity, an automatic welding defect identification is needed. This is executed by using Matlab to create artificial neural networks, which is beneficial for users with the graphical user interface (GUI) feature. One of the breakthroughs in the figure extraction into seven feature vector values is the geometric invariant moment theory. This prevents translation, rotation, and scaling from changing the figure's characteristics. Therefore, a welding defect identification system with a geometric invariant moment was created in the digital radiographic film figure to overcome the reading error by AR. The identification system obtained an accuracy rating of 89.9%.

The Fingerprints of Resonant Frequency for Atomic Vacancy Defect Identification in Graphene

The identification of atomic vacancy defects in graphene is an important and challenging issue, which involves inhomogeneous spatial randomness and requires high experimental conditions. In this paper, the fingerprints of resonant frequency for atomic vacancy defect identification are provided, based on the database of massive samples. Every possible atomic vacancy defect in the graphene lattice is considered and computed by the finite element model in sequence. Based on the sample database, the histograms of resonant frequency are provided to compare the probability density distributions and interval ranges. Furthermore, the implicit relationship between the locations of the atomic vacancy defects and the resonant frequencies of graphene is established. The fingerprint patterns are depicted by mapping the locations of atomic vacancy defects to the resonant frequency magnitudes. The geometrical characteristics of computed fingerprints are discussed to explore the feasibility of atomic vacancy defects identification. The work in this paper provides meaningful supplementary information for non-destructive defect detection and identification in nanomaterials.

Building Façade Protection Using Spatial and Temporal Deep Learning Models Applied to Thermographic Data. Laboratory Tests

This paper proposes a methodology that combines spatial and temporal deep learning (DL) models applied to data acquired by InfraRed Thermography (IRT). The data were acquired from laboratory specimens that simulate building façades. The spatial DL model (Mask Region-Convolution Neural Network, Mask R-CNN) is used to identify and classify different artificial subsurface defects, whereas the temporal DL model (Gated Recurrent Unit, GRU) is utilized to estimate the depth of each defect, all in an autonomous and automated manner. An F-score average of 92.8 ± 5.4% regarding defect identification and classification, and a root-mean-square error equal to 1 mm in the estimation of defect depth equal to 10 mm as the best defect depth estimation, are obtained with this first application of a combination of spatial and temporal DL models to the IRT inspection of buildings.

SRAM Bitcell Defect Identification Methodology Using Electrical Failure Analysis Data

Static Random Access Memory (SRAM) has long been used for a new technology development vehicle because it is sensitive to process defects due to its high density and minimum feature size. In addition, failure location can be accurately predicted because of the highly structured architecture. Thus, fast and accurate Failure Analysis (FA) of the SRAM failure is crucial for the success of new technology learning and development. It is often quite time consuming to identify defects through conventional physical failure analysis techniques. In this paper, we present an advanced defect identification methodology for SRAM bitcell failures with fast speed and high accuracy based on the bitcell transistor analog characteristics from special design for test (DFT) features, Direct Bitcell Access (DBA). This technique has the advantage to shorten FA throughput time due to a time efficient test method and an intuitive failure analysis method based on Electrical Failure Analysis (EFA) without destructive analysis. In addition, all the defects in a wafer can be analyzed and improved simultaneously utilizing the proposed defect identification methodology. Some successful case studies are also discussed to demonstrate the efficiency of the proposed defect identification methodology.