205 results about "Fault class" patented technology

The invention relates to a nonlinear fault detection method based on semi-supervised manifold learning, which belongs to the field of electromechanical equipment fault diagnosis. The method comprises the following steps that (1) vibration signal data acquisition and preprocessing are performed on monitored electromechanical equipment, and hybrid-domain feature extraction is performed to obtain an initial sample set which represents an operating state of the equipment; (2) a semi-supervised Laplacian Eigenmap algorithm is adopted to perform manifold feature extraction on an equipment sample, so as to obtain essential manifold features sensitive to faults; and (3) an intelligent diagnosis model based on an LS-SVM (Least Squares-Support Vector Machine) is established in low-dimensional manifold feature space, so as to realize mode recognition and diagnosis decision to the operating state of the equipment faults. By using a semi-supervised manifold learning algorithm adopted by the invention, nonlinear geometric manifold features of a vibration signal sample can be effectively extracted, the fault category of the equipment operating state is judged, and the fault detection pertinence and accuracy are improved. The nonlinear fault detection method can be widely used for fault detection and diagnostic analysis of all kinds of mechanical equipment.

The invention discloses a multilevel inverter fault diagnosis strategy based on a principal component analysis and multi-classification related vector machines (PCA-mRVM). The multilevel inverter fault diagnosis strategy includes: subjecting primary samples to dimensionality reduction through principal component analysis, and extracting multiple principal components with fault features so as to form training samples; subjecting the training samples to fault diagnosis through the multi-classification related vector machine, outputting probabilities of fault classifications, and taking the fault classifications with the maximum probabilities as diagnosis results. The multilevel inverter fault diagnosis strategy has the advantages in the fault diagnosis with larger sample space and more classifications, is high in model sparseness, low in computation complexity and the like; most importantly, the probabilities of classification members can be outputted through the mRVM, probability and statistic significance is achieved, and uncertain problems can be conveniently analyzed.

A machine tool system is diagnosed by identifying a fault class to which an input measurement vector belongs. The fault class corresponds to a group of weight vectors in a code book of a self organized map that describes the machine tool system based on training data. Probabilities that the input measurement vector belongs to a given class are estimated based on the posterior probability of the weight vectors of the code book corresponding to the given class given the input measurement vector. Training data to create the code book may be collected under a first operating condition while the input measurement vector is collected under a second operating condition.

The invention discloses a diesel generator set fault diagnosis and detection device and method based on deep learning. The device comprises a frame (1), a loudspeaker (2), a displayer (6), a memory (10), a CPU (11) and a data collection device (18), wherein a deep learning module (24), a self-adaptive integrated strategy module (20), a historical signal database (23) and a fault category expert system bank (19) are contained in the frame (1), the self-adaptive integrated strategy module (20) is provided with an integrated strategy generator (201), the fault category expert system bank (19) isprovided with a fault category database (191), a fault index database (192), a fault marking database (193) and a fault level database (194), the deep learning module (24) comprises a clustering algorithm, and a signal transceiver (5) is arranged at the middle position of the upper end of the frame (1). Therefore, it is more accurate and convenient for people to perform fault diagnosis and state online monitoring on a diesel generated set.

The invention discloses a rail transit fault identification method based on an association rule classifier. The method comprises the steps that (1), attributive characters and fault categories corresponding to the attributive characters are extracted from historical fault data, each fault datum is represented by a transaction, one or more association rules corresponding to each transaction are built for the corresponding transaction, and an association rule set is obtained; (2), the support degree and confidence coefficient of each association rule are calculated according to the number of the transactions, containing the corresponding association rule, in a transaction set, and a strong rule is obtained; (3) an association rule hard classification model is built according to the strong rule; the percentage of each non-strong ruler in the association rule set is calculated, and an association rule soft classification model is built; (4) the attributive characters of the fault data monitored in real time are extracted, and are classified through the hard classification model and the soft classification model. According to the rail transit fault identification method based on the association rule classifier, fault identification accuracy is improved, fault correction time is shortened, fault self-diagnosis is achieved for equipment, and driving safety is ensured from the two aspects of operation and maintenance and equipment.

The invention discloses a fault diagnosis method based on a semi-supervised learning deep adversarial network. The fault diagnosis method comprises the steps of: acquiring vibration signals of a bearing under different operation faults, and subjecting a vibration time-domain signal of the faulty bearing to wavelet transformation to form a two-dimensional image; and conducting supervised learning on a small amount of labeled data by a generative adversarial network, training on a large amount of unlabeled data in an unsupervised manner, extracting high-dimensional features by means of a convolutional neural network to achieve data classification, and therefore identifying a fault category of the bearing. According to the fault diagnosis method, a high-precision fault diagnosis model is obtained by training under the condition of limited labeled data, and a more precise discriminator is obtained, so that precise fault diagnosis can be carried out based on the vibration signals of the rolling bearing.

The invention discloses a bearing fault diagnosis method based on a symbolic probabilistic finite state machine. The method comprises the steps that a one-dimensional time series with concentrated training data is converted to a two-dimensional symbol matrix; a probabilistic finite state machine model is constructed for the two-dimensional symbol matrix; an extracted left feature vector is used as a feature value to represent an original bearing signal; and finally an improved K- nearest neighbor classification (KNN) algorithm is used to learn the left feature vector of a bearing signal of a known failure category. For a bearing signal to be diagnosed, the same feature extraction method is used, and then the improved KNN algorithm is used to realize fault diagnosis. According to the invention, a K-means clustering method is used to improve the traditional K- nearest neighbor classification algorithm, and the computation efficiency and the fault diagnosis effect of the bearing diagnosis algorithm are improved.

A method and apparatus for diagnosing faults. Process data is analyzed using a first metric to identify a fault. The process data was obtained from a manufacturing machine running a first recipe. A fault signature that matches the fault is identified. The identified fault signature was generated using a second metric and/or a second recipe. At least one fault class that is associated with the fault signature is identified.

The invention relates to a method and a system for fault diagnosis of electrical equipment. The method comprises the following steps: obtaining a state parameter data and confirming each abnormal state parameter according to the state parameter data; confirming each fault class related to each abnormal state parameter according to a preset fault diagnosis model which comprises an incidence relation between the fault class and the state parameter and a fault influence weight value of each state parameter to the fault class; and performing fault diagnosis on each related fault class according to the fault influence weight value of each abnormal state parameter to the related fault class. According to the method provided by the invention, the incidence relation between each state parameter and each fault class is established, the fault class is possibly related to a plurality of state parameters, and the fault diagnosis is performed on each related fault class according to the fault influence weight value of each abnormal state parameter to the related fault class, so that the comprehensive assessment for the state of the electrical equipment is realized and the accuracy of the fault diagnosis for the electrical equipment is increased.

The invention provides a fault prediction method based on synthetic minority class oversampling and deep learning. The Means method is used for clustering a few types of samples in the sample set; deleting the noise class cluster after clustering; dividing the class cluster into noise class samples in each class cluster by using a KNN method; fault samples and risk samples, deleting the noise samples; and finally, inputting a random number into each class cluster, and selecting a certain sample as an output sample according to a proportional relation between the random number and the fault class sample and the risk class sample in the class cluster;realizing oversampling of the SMOTE method ; and then increasing the number of a few types of samples through multiplication operation, so thatthe types of the samples in the finally obtained fusion sample are more balanced, and the acquired feature data are balanced, thereby facilitating model training, maximally mining the law behind thedata, and realizing a better fault prediction effect.

The invention relates to a library construction method for a transformer substation equipment fault case library, a fault diagnosis method and system, and the method comprises the steps: constructinga fault case which is expressed as {a fault category, a fault feature information set, and a solution}, wherein the fault feature information set is a set which is composed of various fault feature signals which can be collected, and the probability of occurrence of corresponding fault signals; And importing the constructed fault cases into a database to form a transformer substation equipment fault case library. According to the invention, a large amount of secondary equipment overhaul data accumulated in the transformer substation is utilized; and constructing a fault case according to the fault type of the equipment, the fault feature information set and the solution, and storing the constructed fault case in a database to form a transformer substation equipment fault case library so asto prepare for pre-diagnosis of a transformer substation secondary equipment fault.

The invention provides a fault tree-based numerical control machine tool fault removal scheme judgment indication method. According to the method, an input target character string used for describing a fault phenomenon is subjected to fault state description word extraction so as to search for potential target fault event nodes from a fault tree; various fault monitoring index parameters corresponding to potential fault bottom event nodes are compared with input various current running state parameters of a numerical control machine tool; potential fault bottom event nodes corresponding to obtained abnormal running state parameters are determined as confirmed fault bottom event nodes; fault removal scheme texts of corresponding fault types of the confirmed fault bottom event nodes are extracted according to a corresponding relationship in the fault types and used for indication, so that the fault type identification efficiency is improved, and the fault type possibility identification comprehensiveness and the fault type confirmation accuracy are ensured; and the method can be used for assisting in improving convenience of assisting in field repair or remote repair of the numerical control machine tool.

The invention relates to a method for fault classification of a smart electric meter based on cluster analysis and a cloud model. The method comprises a step 1) of obtaining historical smart electric meter fault data sample points, and adopting a K-means algorithm to divide historical smart electric meter fault data sample points into K large fault classes and central values corresponding to each large class; a step 2) of taking the central values corresponding to each large class as sample means, taking smart electric meter fault data sample points contained in each large fault class as data points, and generating a corresponding K-class electric meter fault cloud model; a step 3) of adopting a reverse normal cloud generator to calculate the electric meter fault cloud model, and obtaining qualitative cloud characteristics of the electric meter fault cloud model; and a step 4) of subdividing the K large fault classes into a plurality of small fault classes according to the qualitative cloud characteristics. Compared with the prior art, the method has the advantages of qualitative analysis and fine classification.

The invention belongs to the technical field of intelligent control, and discloses a distribution network single-phase earth fault identification method based on a long-short memory network, which comprises the following steps: performing classification identification of arc-containing fault class and arc-free fault class and classification identification of earth type on a single-phase earth fault; acquiring zero sequence currents of continuous multiple periods during a fault steady state period, taking the waveform of the zero sequence current of a period time slot as a sample point, and performing classification identification of the arc-containing fault class and the arc-free fault class on the zero sequence currents of continuous multiple period time slots by using an LSTM deep learning classifier; single-phase earth fault of the arc-containing fault class and the arc-free fault class calculating unit period earth transition resistance of continuous multiple period time slots by using instantaneous voltage and current of continuous multiple period fault phases respectively; and carrying out classification identification on the unit period earth transition resistance of the continuous multiple period time slots by using a multilayer neural network by taking a change curve of the unit period earth transition resistance of one period time slot as a sample point to finish theidentification of the single-phase earth fault.

The invention discloses a method for diagnosing a fault of an oil-immersed transformer on the basis of a rough set and a bayesian network. The method comprises the following steps that (a) the type of the fault is determined, as much as possible input fault characteristic vectors are selected in an original sample set, and an input attribute set is determined; (b) discretization processing is carried out on a fault data set through a data discretization method in the rough set theory, and a discretization decision table is established; (c) establishment of the bayesian network is carried out through Matlab; (d) a conditional probability table is initialized, wherein all the possible conditional probabilities of each node relative to the father node of the node and the quantitative description of the corresponding problem domain are listed in the conditional probability table; (e) parameter learning is carried out, and a deduction engine is established to carry out deduction after the bayesian network is established; (f) a test sample set is input, the posterior probability is solved, and the type of the fault is judged. The method for the oil-immersed transformer on the basis of the rough set and the bayesian network can simplify the scale of a diagnosis network, enhance the anti-interference performance of the network, diagnose various faults of the transformer rapidly, and reduce the outage rate of the transformer greatly.

The invention discloses a transformer fault diagnosis method based on a random forest. The method comprises the following steps of collecting fault gas concentration data in insulating oil in a transformer and a corresponding fault type as a training sample; according to the training sample, based on the generation steps of a decision tree, establishing a fault decision tree; according to the fault decision tree, establishing a random forest model; and collecting the fault gas concentration data of a unknown fault type, inputting into the random forest model so as to acquire the fault type through the random forest model. The fault gas concentration data in the insulating oil in the transformer is taken as the training sample so as to establish the random forest model, a whole transformerfault can be accurately diagnosed, stability is high, and the method can be applied to the transformer diagnosis technology field.

Disclosed is an excavating diagnosis method of failure data based on a bayesian network. The method is used for diagnosis of equipment failures. Parameter points, capable of being obtained, of existing devices with failures are monitored in advance, and a monitoring signal and a failure type are used as input variables of the bayesian network; then, nodes are sorted at random, side adding or changing operation is carried out in the search problem, the generated bayesian network is graded and stored in a database, finally, the bayesian network with the highest score in the database is used as an equipment failure diagnosis inference graph, equipment data needed to be diagnosed are input into the diagnosis inference graph, and the diagnosis result can be obtained. The search density can be set according to the actual condition, and can be prevented from being caught in endless loop; a penalty term can be adjusted so as to prevent excessive fitting; an initial sequence is generated at random to enable the system to have a certain roughness; a user can see the equipment failure diagnosis inference graph, and therefore logic analysis can be carried out on the failure happening mechanism.

The invention discloses a fault diagnosis method based on an improved generative adversarial network for small sample features, and solves problems that in a fault detection and diagnosis process, cost of manually adding labels to network data is too high, a convergence fluctuation of a generative adversarial network is large, and label loss in actually collected network data is serious. Firstly,reasons of network faults are analyzed, and a semi-supervised fault diagnosis model is provided by improving a loss function of a generator network and an output layer of a discriminator network; andsecondly, the model is further optimized, an algorithm combining a generative adversarial network and a convolutional neural network is provided, the generative adversarial network is responsible forgenerating data of various fault types, and then the convolutional neural network is trained through the data to complete diagnosis of network faults. According to the method, accurate diagnosis of the network faults can also be realized under the condition of a small amount of labeled data.

A method and apparatus for diagnosing faults. A fault is detected. One or more process variables that contributed to the fault are determined. A relative contribution of each of the one or more process variables is determined. A determination is made as to which fault signatures match the fault, a match occurring when the relative contributions of the one or more process variables are within relative contribution ranges of the matching fault signature. Each fault signature is associated with at least one fault class.

The invention discloses a multi-label rolling bearing fault diagnosis method based on meta-learning, and the method comprises the steps: constructing a multi-label fault data set of a rolling bearing,and dividing the multi-label fault data set into a training set and a test set according to fault types; extracting a time-frequency signature matrix feature T-FSMs of the fault signal; establishinga multi-label convolutional neural network model MLCML based on meta-learning; training an MLCML model by using the training set sample; verifying the trained MLCML model by using a test set sample; and carrying out fault diagnosis on the small-sample multi-label rolling bearing by utilizing the trained model. According to the method, multiple semantics contained in the rolling bearing fault sample are fully utilized, the fault diagnosis result is more accurate, meanwhile, the problem of small samples in actual fault diagnosis of the rolling bearing can be better solved through the time-frequency signature matrix characteristics and the meta-learning strategy, design is reasonable, operation is easy and convenient, and wide application value is achieved.

The invention relates to a motor bearing fault diagnosis method, which comprises the following steps of S1, building a generative adversarial network of a small sample data category through a GAN training method based on a discrimination model and a generation model, and generating a data set conforming to small category features; S2, adding the generated data set into an original small class sample training set to form a balance data set; S3, constructing deep convolutional neural networks of the discrimination model and the generation model, wherein the deep convolutional neural network of the discrimination model comprises three convolutional layers and three corresponding pooling layers, two full connection layers being arranged behind the third pooling layer, and taking the optimizedbalance data set as a training set of the deep convolutional neural networks; and S4, for the training set, learning fault features from training data in a self-adaptive layer-by-layer mode, and carrying out diagnosis and recognition of different fault types through a classifier. Compared with the prior art, the method has the advantages of learning the fault features in a self-adaptive mode, andimproving the diagnosis and recognition rate of faults with small data volume and the like.

The invention provides a train control onboard device fault classification and recognition method based on a rough set-neural network model. The method comprises steps: according to a fault case library analyzed and sorted by a train control onboard device fault log file, a corresponding relationship between a fault class and a fault code is dug out, fault codes and fault classes in the fault caselibrary are coded, an initial decision table is generated, and a classification rule is determined; RST is used to carry out attribute reduction on the initial decision table, and a final decision rule is generated; and based on the final decision rule, a neural network model is built, and the neural network model is used to realize fault recognition on the train control onboard device. Accordingto the fault classification and recognition method with the neural network and the rough set theory combined provided in the invention, the problems of low fault recognition rate for text fault dataof a high noise-containing train control onboard device, poor incomplete knowledge processing ability and the like can be solved, and accuracy of the fault classification and recognition for the traincontrol onboard device can be ensured.

The invention discloses an air conditioner fault information display method and device. The method comprises the following steps: detecting air conditioner fault information; determining the fault class of the air conditioner according to the fault information; and displaying the fault information on different interfaces of the air conditioner display device according to the fault class. The method solves the technical problem of irrational fault prompting since all the fault information is displayed on one interface in the prior art.

The invention discloses a wagon coupler joist breaking detection method, and belongs to the technical field of wagon safety. The invention aims to solve the problem of low reliability caused by manualdetection of the breaking fault of the coupler joist of the existing rail wagon. The method comprises the steps: establishing a data set for training, marking identification boxes in fault areas or suspected fault areas of coupler joist fault samples in the data set in a blocking mode, and configuring category labels for all the identification boxes; building a Faster-Rcnn model based on a ResNetdetection model, and performing training to obtain a weight coefficient of classification; inputting an image to be identified into the Faster-Rcnn model loaded with the weight coefficient; and carrying out fault category prediction, firstly obtaining a fault initial judgment region in the fault prediction process of the to-be-identified image, then obtaining confidence corresponding to the faultinitial judgment region, determining the fault initial judgment region with the confidence greater than a preset threshold as a fault region, and carrying out alarming. The method is used for detecting the fracture of the coupler joist.

The invention discloses a fault diagnosis method for an AC/DC charging equipment power device based on wavelet packet analysis, and the method comprises the steps: analyzing a charging equipment powermodule, and determining a fault class; obtaining output signals of the charging equipment power module under the normal condition and various fault conditions; carrying out the high-low frequency decomposition of the output signals, carrying out the multilayer wavelet packet decomposition of the high-low frequency signals obtained through decomposition, and carrying out the reconstruction of a wavelet packet decomposition coefficient; calculating the extracted signal energy at all frequency bands, carrying out the normalized calculation of the signal energy, and finally determining a centralized frequency band range of the signal energy; carrying out the power spectrum analysis of the reconstructed wavelet packet decomposition coefficient, determining the feature frequency and power spectrum value of a frequency band signal, comparing the comparison with the fault analysis, and determining the type of a fault; and carrying out the fault type coding of the type of the fault. The methodimproves the time-frequency resolution of a signal, achieves the precise fault positioning through the spectrum features, and improves the precision of the fault diagnosis of the AC/DC charging equipment power device.