44results about How to "Improve the efficiency of fault identification" patented technology

The invention discloses a cable fault detection and analysis method. According to the method, an impedance frequency spectrum and a phase frequency spectrum of a cable are taken as basis, subsection characteristic impedance on an impedance frequency spectrum attenuation trend curve is calculated with a subsection interception method, an equation of equivalent impedance is established, R, L, G and C parameters of cable transmission impedance are calculated, a transmission impedance model based on R, L, G and C parameters is obtained, an error distance graph with a cable distance length serving as a variable is obtained through error calculation and comparison of the transmission impedance model and the tested impedance frequency spectrum and phase frequency spectrum, finally, local zero points on the error distance graph are analyzed, and fault detection and fault type identification are realized. Compared with the prior art, on the basis of a relative signal transmission rate calculation method with low calculation difficulty and high calculation accuracy, accurate fault positioning and multi-point fault positioning can be realized, and the fault type identification can be realized.

The invention relates to an engine fault identification method based on time-domain energy and a support vector machine. The method comprises the following steps of: (1), respectively installing an acceleration vibration sensor on each measuring point, wherein each acceleration vibration sensor is connected with a computer through a vibration signal acquisition instrument; (2) dividing time-domain vibration signals acquired from a top dead center into N sections, wherein each section comprises M vibration signal amplitudes; (3) evaluating the time-domain energy Ei of each of the N sections; (4) evaluating the sum ESUM of the time-domain energy of the N sections; (5) comparing the Ei with the ESUM, and constructing a feature vector after performing energy unitization to obtain N time-domain energy feature values; (6) performing classified counting by a support vector machine method to obtain classified parameters of running states of an engine; and (7) comparing to obtain an engine fault identification result. By the method, the fault identification speed of the engine is effectively improved; the fault identification time of the engine is shortened; the fault identification efficiency of the engine is improved; and the method is simple and is easy to operate, and facilitates the repair and the maintenance of the engine.

The invention discloses a system and method for recognizing a power cable fault based on fractal and wavelet transform. The system comprises a current detection circuit module, a signal modulation circuit module, a data collecting card and a master control computer which are connected in sequence, wherein the current detection circuit module comprises an A-phase hall current sensor, a B-phase hall current sensor and a C-phase hall current sensor, and the signal modulation circuit module comprises an A-phase I / V switching circuit module, a B-phase I / V switching circuit module, a C-phase I / V switching circuit module, an A-phase signal amplifying circuit module, a B-phase signal amplifying circuit module, a C-phase signal amplifying circuit module, an A-phase filter circuit module, a B-phase filter circuit module and a C-phase filter circuit module. The method comprises the following steps that signals are detected in real time and uploaded synchronously; the signals are collected and saved; the cable fault is recognized. The system and method for recognizing the power cable fault based on the fractal and the wavelet transform has the advantages of being novel and reasonable in design, high in detection precision, stability and reliability due to the fact that the hall current sensors are adopted to detect the cable current, capable of easily, rapidly, accurately and effectively recognizing the type of cable short fault by means of the combination of the fractal method and the wavelet transform method, strong in practicability, good in using effect and convenient to popularize and use.

The embodiment of the invention discloses a fault identification method and device. The fault identification method includes: obtaining network state data of the network service system in a preset time period, determining whether each preset index corresponding to any monitoring mode is in an abnormal state in a preset time period or not; and determining whether the monitoring mode is in a fault state in a preset time period based on the historical network state data, wherein each preset index corresponding to any monitoring mode is determined based on a preset index with an abnormal state inthe historical network state data corresponding to the historical fault scene. A monitoring mode and each preset index corresponding to the monitoring mode are extracted from historical network statedata corresponding to a historical fault scene; therefore, the abnormal condition of each preset index corresponding to the monitoring mode can identify the real fault condition of the monitoring mode, so that the fault condition of the monitoring mode determined based on the comprehensive condition of each preset index corresponding to the monitoring mode is relatively accurate, and the fault identification efficiency is relatively high.

The invention discloses a fault diagnosis method for a rotor system based on principal component analysis and broad learning. Principal component analysis is used for carrying out dimension reduction on a feature matrix formed after feature extraction, the linear correlation between the data is reduced, redundant attributes are eliminated, and a low-dimensional matrix capable of retaining the essential features is obtained; and then the matrix is input into a broad learning system for fault identification, and the fault classification task of the rotor system is completed. According to the method in the invention, the principal component analysis and the broad learning system are introduced into the fault diagnosis and identification of the rotor system, through the method, the fault classification complexity can be effectively reduced, the data modeling time can be greatly shortened, the fault identification efficiency of the rotor system can be improved, therefore, the fault diagnosis task of the rotor system can be efficiently completed, the practicability is good, and the method is worthy of popularization.

The embodiment of the invention provides a fault diagnosis method and device based on an action current curve of a switch machine. According to the method, feature data corresponding to current curvesamples marked with a fault type is pre-stored, target feature data of the action current curve of the switch machine about to be subject to fault diagnosis is acquired, target current curve samples similar to the current curve are screened out of the current curve samples according to the target feature data and the feature data corresponding to all the current curve samples marked with the faulttype, a target fault type is determined according to the fault type corresponding to all the target current curve samples, and the target fault type is the type of a fault of the switch machine. Through comparison with the current curve samples marked with the fault type, automatic recognition of the fault of the switch machine is realized, the accuracy of fault recognition is improved, the possibilities of missing judgment and misjudgment are lowered, and fault recognition efficiency is improved.

The invention discloses a method for recovering power supply after stripping started by direct-current ground protection, which includes the following steps: (1) failure detection: a direct-current ground protection device is connected with a power supply system, and is used for detecting the failure of the power supply system; the power supply system comprises two uplink grid contact lines, two downlink grid contact lines and a plurality of feeder cabinets, and any feeder cabinet is connected with the two uplink grid contact lines or the two downlink grid contact lines; all the feeder cabinets are connected with the direct-current ground protection device, and when failure is detected by the direct-current ground protection device, all the feeder cabinets are tripped; (2) response processing: according to whether the feeder cabinets are connected with the two uplink grid contact lines and the two downlink grid contact lines at the same time, processing is carried out separately according to two conditions. By improving the failure recognition theory of the power supply system, the method for recovering power supply after stripping started by direct-current ground protection and the like, the invention can achieve the advantages of high failure recognition efficiency and high power supply recovery speed.

The invention discloses a multi-fault mode identification method and device of a swashplate of a helicopter. The multi-fault mode identification method comprises the steps that analog quantity signals collected in real time are pre-processed to obtain real-time data corresponding to the analog quantity signals; one or more time domain features in the real-time data are extracted; by comparing each time domain feature with a time domain feather threshold range corresponding to the time domain feature, one or more suspected faults are determined; based on the real-time data corresponding to the one or more suspected faults, one or more fault information items are obtained through a pre-trained radial basis function neural network. The multi-fault mode identification method and device of the swashplate of the helicopter improve real-time performance and efficiency of fault monitoring, eliminate potential fault threats, improve safety of automatic inclination, improve the accuracy of fault identification, shorten the time for fault identification and improve efficiency of fault identification.

The invention discloses a fault detection method and device which are used for solving the problem in the prior art that faults occur in the an operation system installation process cannot be known in time. The method comprises the steps of: obtaining a screen image of a target host in the operation system installation process through an intelligent platform management interface of the target host; determining the installation state of the operation system of the target host according to the display content of the screen image; and outputting an alarming message used for prompting the current abnormal installation of the operation system when the installation state is determined to be abnormal.

The invention discloses an insulator fault identification method and device based on a feature pyramid. The method comprises the steps of obtaining a background image containing an insulator; and inputting the background image into a preset insulator fault recognition model, carrying out insulator fault recognition on the background image, and recognizing the faulted insulator in the background image. According to the method, the fault of the insulator can be quickly and effectively identified by constructing the preset insulator fault identification model, the fault identification efficiencyis improved, and the stable operation of a power system is ensured.

The invention discloses a power equipment fault intelligent diagnosis method based on single-order infrared image target detection, and the method comprises the steps: collecting a plurality of infrared images of power equipment, carrying out the preprocessing of the infrared images, and building an infrared image data set; importing the infrared image training set into a YOLOV4 convolutional neural network, and training by using a transfer learning method and a K-means++ clustering analysis method to obtain an infrared image detection network; fitting a temperature and gray scale function relationship of an infrared image temperature bar; manufacturing PC electrical equipment infrared fault diagnosis software, and embedding an infrared image detection network and the temperature gray scale function relation into the PC electrical equipment infrared fault diagnosis software, wherein target identification and analysis operation are carried out on an infrared image, so that the temperature of a corresponding point of the electrical equipment is extracted; and carrying out fault diagnosis on the equipment according to an equipment diagnosis standard in the DL / T 664-2016 electrified equipment infrared diagnosis application specification, and automatically judging whether the power equipment has a fault or not.

The invention relates to a converter valve sub-module controller and a driving fault discrimination method. When the received driving feedback signals are abnormal, within a set time, the driving feedback signals are re-received, if the re-received driving feedback signals are abnormal, it is determined that a sub module generates drive fault. When the received driving feedback signals are abnormal, the driving feedback signals are re-received after a period of time, if the re-received driving feedback signals are still abnormal, it is determined that the sub module generates driving faults soas to avoid the misinformation generation that driving feedback signals are only received once to confirm that the sub module generates faults, reduce the risk of system off-stream and improve the reliability of the operation of the flexible direct current transmission converter valve.

The invention discloses a rail wagon coupler knuckle pin fault detection method based on image processing, and relates to rail wagon coupler knuckle pin fault detection. The problem that the accuracyof a detection result is low due to the fact that whether the coupler knuckle pin is lost or broken or not is subjected to fault detection in a manual troubleshooting mode is solved. The method comprises the steps of calling a corresponding template image according to coupler types of two adjacent carriages; preprocessing the image of the coupler knuckle pin combination area, and matching and correcting the obtained preprocessed image with the called template image so as to obtain a coupler knuckle pin initial identification area image; obtaining two to-be-identified fault area images of the coupler knuckle pin according to the contour features of the initial identification area images of the coupler knuckle pin; and judging whether a pin hole in each coupler knuckle pin to-be-identified fault area image exists or not, if so, judging a fault, and if not, judging whether the coupler knuckle pin in the coupler knuckle pin to-be-identified fault area image is broken or not by utilizing the SVM separator model so as to finish fault detection. The method is used for detecting faults of the coupler knuckle pin.

The embodiment of the invention provides a porcelain post insulator fault identification method and device. According to the method, vibration data of porcelain post insulator vibration acoustic detection equipment is used as a data source; the method includes: firstly, converting collected time domain data into a frequency domain; then, constructing a Gaussian mixture model for the processed signal by utilizing an expectation maximum algorithm, and solving important parameters of the Gaussian mixture model by utilizing the expectation maximum algorithm, namely an average value mu, a standarddeviation sigma and a weight coefficient alpha, wherein the average value mu can reflect the central frequency position of each modal component, the standard deviation sigma can reflect the frequencybandwidth, and the weight coefficient alpha can be used for representing the proportion occupied by each modal component; and finally, importing the obtained important parameters as characteristic data into an ELM for training, judging the fault characteristics of the porcelain post insulator through machine learning, and compared with the traditional spectral analysis, the porcelain post insulator fault identification method based on the Gaussian mixture model and the expectation maximum algorithm improves the fault identification efficiency.

The invention discloses an automatic overhauling system for urban rail transit vehicles. The system comprises an inspection vehicle device, an automatic boarding ladder device and lifting type trenchplatform devices, wherein the inspection vehicle device comprises a multi-degree-of-freedom mechanical arm and imaging equipment on the mechanical arm; the inspection vehicle device is used for identifying the guide lines of the ground on the two sides of an inspection warehouse track and the bottom of a trench, and inspecting the bottom of the vehicle and the two sides of the vehicle body along the trench and the two sides of a rail; the automatic boarding ladder device is positioned on the side surface of the rail and is used for leading drivers and passengers to normally get on and off thecar when the automatic boarding ladder device rises, and being flush with the ground to avoid passing of the inspection vehicle device when the automatic boarding ladder device descends; the lifting type trench platform devices are respectively arranged at the two ends of the row position of the trench, are flush with the ground when being lifted and flush with the trench bottom when being lowered, and are used for conveying the inspection vehicle device between the trench bottom and the ground. According to the system, manual overhaul can be replaced, missed detection can be effectively prevented, the overhaul time is greatly shortened for being capable of live working, and the operation efficiency of the urban rail transit vehicle is effectively improved.

The invention discloses an application method of a wavelet singular entropy in periodic fault detection of a stacker. Wavelet transform is performed on a signal from a stacker to obtain wavelet coefficients at different wavelet scales corresponding to different wavelet frequency bands, and the wavelet coefficients form a wavelet coefficient matrix; singular value decomposition is performed to obtain a series of singular values capable of reflecting basic features of an original coefficient matrix, so that distribution of all components is realized and the distribution of all components is obtained; and then on the basis of statistical properties of an information entropy, uncertainty analysis is performed on a singular value set to provide a certain measure for the original signal and themeasure is used as a fault signal feature. With the disclosed method, the signal characteristics of periodic faults can be extracted from the original signal quickly; and the fault feature extracted based on the wavelet singular entropy is independent of the fault amplitude.

The invention discloses a typical fault automatic identification method for a data-driven rotor system. An offline training module, a fault classification knowledge base module and an online fault automatic identification module are included. Adaptively decomposing the original vibration waveform signal into a series of intrinsic mode function components by adopting optimized empirical mode decomposition; a feature conjoint analysis method is proposed to screen sensitive IMF components for signal reconstruction, and a multi-scale dispersion entropy value calculated by a reconstructed signal is screened as a feature value; constructing an LSSVM classifier to adaptively determine a penalty factor C and a kernel parameter sigma; the distance between the source domain data and the target domain data is minimized through transfer learning, a constructed transfer feature vector matrix serves as the input of a model, and fault mode recognition of the rotor system is achieved. According to the method, the original time sequence vibration waveform data is adopted as input, the fault identification conclusion can be automatically output, and the method has high identification accuracy and good generalization for rotor system fault data under different equipment and different working conditions.

The invention discloses a fault recognition model training method, a fault recognition method and device and electronic equipment. The fault identification model training method comprises the steps ofconstructing at least one positive sample and at least one negative sample corresponding to a set fault based on time series data corresponding to each first time period in at least one first time period, based on the at least one positive sample and the at least one negative sample corresponding to the set fault, and training a set binary classification model to obtain a fault identification model corresponding to the set fault, wherein the time series data comprises first time series data and second time series data, the first time series data represents the log information amount output ateach moment of the corresponding first time period, and the second time series data represents monitoring data of each performance index in the at least one performance index at each moment of the corresponding first time period.

The invention discloses a transformer winding deformation identification method based on an LSTM neural network. The method comprises the steps of 1) carrying out data cleaning on vibration data collected by a transformer; 2) distinguishing direct current magnetic bias and winding deformation, and calibrating winding deformation data; 3) designing an LSTM neural network; 4) adjusting the parameters of the neural network, and training the adjusted parameters of the neural network; and 5) starting verification by adopting the trained neural network, and performing fault identification on the transformer needing fault identification through the trained neural network. Under the condition that transformer vibration lacks a time domain signal analysis tool, by means of the LSTM long short-term memory neural network technology, time domain signal analysis can be rapidly and effectively carried out; and the method is different from a traditional transformer winding deformation identification method, and after neural network training is completed, the method has the advantages of convenience in neural network establishment, high fault identification efficiency and the like.

The fault detection method of railway wagon knuckle pins based on image processing involves the detection of faults of railway wagon knuckle pins; it solves the problem of using manual inspection to detect whether the knuckle pins are lost or broken, and the accuracy of the detection results is low question. The present invention retrieves corresponding template images according to the coupler types of two adjacent carriages; preprocesses the images in the coupling area of ​​the couplers, and matches and corrects the obtained preprocessed images with the retrieved template images, thereby obtaining knuckle pin initial recognition area image; according to the contour features of the knuckle pin initial recognition area image, two knuckle pin fault area images to be identified are obtained; whether the pin hole in each knuckle pin fault area image to be identified exists is judged, If the result is yes, the fault is determined; if the result is no, the SVM separator model is used to determine whether the knuckle pin is broken in the image of the fault area to be identified for the knuckle pin to complete the fault detection. Used for knuckle pin fault detection.

The invention discloses a fault diagnosis method for an electric actuator of a gas turbine, and belongs to the technical field of fault diagnosis of gas turbines. Comprising the following steps: step 1, carrying out standardized preprocessing on an acquired flow signal sample; 2, working condition identification and preliminary clustering are carried out on the data set in the step 1, and then a Gaussian probability density function model is trained; 3, constructing a self-learning network model, training the self-learning network model, extracting data features of the traffic signal samples, and storing network weights of the self-learning network model; 4, constructing a state classification and recognition network model of the electric actuator, and loading network weight parameters; and 5, carrying out fine adjustment on the network weight of the GMM-1DCNN of the electric actuator. The method has strong advantages in the aspects of diagnosis precision and diagnosis efficiency, and avoids the defects of large data processing workload, difficulty in effective feature extraction and the like caused by direct data processing.

The invention discloses a discharge current wave recording device with trigger enable and a fault identification method, which are used for collecting pulse current of a pulse xenon lamp power supply and fault judgment, and relate to the technical field of pulse power. The device includes an n-channel synchronous acquisition module, a data processing module and a trigger communication module. First, before the discharge of the pulse xenon lamp power supply, the trigger enable signal and experimental parameters are sent to the data processing module through the communication module to calculate the set current characteristics, and the trigger enable signal is released; then, the external trigger signal is sent to the pulse xenon lamp power supply to start Discharge, at the same time, the external trigger signal triggers the communication module to start data collection; finally, the data processing module performs fault judgment on the collected discharge data according to the set current characteristics, and sends the result to the host computer. The invention can avoid the risk of false triggering caused by electromagnetic interference, realizes automatic interpretation of the discharge current based on the discharge current calculation and feature extraction of the nonlinear characteristics of the xenon lamp, and can greatly improve the efficiency of discharge current fault identification.

The invention discloses a defect detection method and system for key components of a power transmission line. The method includes: constructing a defect semantic set of key components of the transmission line; the key components of the transmission line include power lines and insulators; training a semantic representation extraction model according to the defect semantic set; using the trained semantic representation extraction model to extract semantic representation; An image library of defects in key components of transmission lines; training an image feature extraction model based on the multi-category defect images in the image library; using the trained image feature extraction model to extract image features; based on the semantic representation and the image features, construct an embedding mapping matrix. The invention realizes the detection of the defect of the key components of the transmission line with zero samples through the semantic description information of the historical defect (defect that has been obtained or is easy to obtain samples) samples and the target defect (zero sample defect).

The invention belongs to the technical field of hydroelectric statistics. An intelligent operation analysis report system for a hydraulic power plant comprises a single-chip microcomputer control module, an input touch screen module, a data access module, a platform server, a platform retrieval module, an online comparison module, a power generation quantity detection module, a local memory, a local comparison module and a table drawing module. When the system operates, the power generation quantity detection module can collect the power generated by the hydraulic power plant, the power is stored in the local memory through the data access module, and when the system operates by adopting a local area network, the local comparison module compares the collected data with the data of the same period in the previous year in the local memory; when the system runs through an online network, the online comparison module calls data in the platform server through the platform retrieval module for further comparison, the online comparison module and the local comparison module draw comparison results into charts through the table drawing module, and the charts are stored in the local memory.

The invention discloses a power system fault monitoring and alarming system, and relates to a power system. The system disclosed by the invention comprises an identification parameter matrix determination module, a data acquisition module, an intermediate data solving module, an identification parameter matrix recursive solving module, an admittance matrix change solving module and a fault line determination module, wherein the identification parameter matrix determination module is used for determining an identification parameter matrix; the data acquisition module is used for acquiring a power supply voltage phasor on an engine node and a load current and current phasor on a load node; the intermediate data solving module is used for solving the estimated value of eta<n-1, add n> after the nth sampling data is added; the identification parameter matrix recursive solving module is used for solving the estimated value of the identification parameter matrix after the removal of the (n-p)th sampling data; the admittance matrix change solving module is used for solving the change rate of each item in a load-sending admittance matrix; and the fault line determination module is used forextracting fault items with the change rate being greater than a preset value in the load-sending admittance matrix and determining the fault line. According to the invention, the previous data is removed, so that the calculation processing amount is reduced. The data far away from the current moment is removed, so that the estimation precision of the system is effectively improved, and the faultline of the power system can be effectively identified.

The invention discloses a method for predicting the remaining life of a marine bearing based on migration learning and multi-time windows, belonging to the technical field of wear detection technology of mechanical parts. The method includes: training a CNN aging model and a multi-time window prediction model based on an LSTM neural network; The vibration signal of the predicted bearing is input into the CNN aging model to determine whether the bearing has an early failure; if an early failure occurs, the CNN depth features of various preset length windows corresponding to the vibration signal are input into the multi-time window prediction model, and various presets are obtained. The life prediction value corresponding to each length window; fuse multiple life prediction values ​​to obtain the target predicted life of the bearing to be predicted; the invention adopts the method of fusion of multiple preset length windows to solve the problem that a single window is difficult to adapt to multiple degradation modes, The multi-time window prediction model based on LSTM neural network estimates the remaining service life of the bearing to be predicted, and the prediction accuracy is high.

The present application discloses a speed reducer with automatic fault discrimination, which includes a reduction transmission mechanism, a transmission shaft, a transmission bearing, a box body, an automatic fault discrimination device and a plurality of first sensors, the first sensor is installed on at least one part of it, and is used to It is used to detect and upload the sensing signal of the corresponding part of the reducer; the automatic fault discrimination device is used to decompose, reconstruct and calculate the energy of the sensing signal of the training sample, and use the energy vector formed by the calculated signal energy as the training sample respectively Input the neural network of the corresponding signal type for training, and input the sensor signals of different reducer parts uploaded by multiple first sensors into the trained neural network of the corresponding signal type, and obtain the component part of the reducer output by the neural network where the fault occurs and its fault type information. The reducer improves the automation and intelligence of fault discrimination, and optimizes the weights and thresholds of the neural network.

The invention discloses a discharge current recording device with trigger enabling and a fault recognition method, which are used for collecting pulse current of a pulse xenon lamp power supply and judging faults, and relates to the technical field of pulse power. The device comprises n paths of synchronous collection modules, a data processing module and a trigger communication module. Firstly, before a pulse xenon lamp power supply discharges, a trigger enable signal and experimental parameters are transmitted to a data processing module through a communication module, set current characteristics are calculated, and the trigger enable signal is relieved; then, the external trigger signal is sent to a pulse xenon lamp power supply to start discharging, Meanwhile, the external trigger signal starts data collection through a trigger communication module; and finally, the data processing module performs fault judgment on the acquired discharge data according to set current characteristics, and sends a result to an upper computer. According to the invention, the risk of false triggering caused by electromagnetic interference can be avoided, the discharge current is automatically interpreted based on the discharge current calculation and feature extraction of the nonlinear characteristics of the xenon lamp, and the efficiency of discharge current fault recognition can be greatly improved.

The invention discloses a cable fault detection and analysis method. According to the method, an impedance frequency spectrum and a phase frequency spectrum of a cable are taken as basis, subsection characteristic impedance on an impedance frequency spectrum attenuation trend curve is calculated with a subsection interception method, an equation of equivalent impedance is established, R, L, G and C parameters of cable transmission impedance are calculated, a transmission impedance model based on R, L, G and C parameters is obtained, an error distance graph with a cable distance length serving as a variable is obtained through error calculation and comparison of the transmission impedance model and the tested impedance frequency spectrum and phase frequency spectrum, finally, local zero points on the error distance graph are analyzed, and fault detection and fault type identification are realized. Compared with the prior art, on the basis of a relative signal transmission rate calculation method with low calculation difficulty and high calculation accuracy, accurate fault positioning and multi-point fault positioning can be realized, and the fault type identification can be realized.