Wuhan University

Jiajun Duan

Yigang He

Liulu He

Hui Zhang

Xiaoxin Wu

Bolun Du

Baiqiang Yin

Chaolong Zhang

Yaru Zhang

Lie Li

Guolong Shi

Bing Li

Xiaoyan Liu

Kaipei Liu

Zhaorong Zeng

Yuan Huang

Yuanxin Xiong

Liufei Shen

Jinguang Jiang

Jian Zheng

Qiwu Luo

A transformer fault diagnosis method and system using induced ordered weighted evidence reasoning is provided. The method includes the following steps. A typical data sample of transformer sweep frequency response analysis is loaded and a diagnostic label is set as an identification framework. Test data of a device to be diagnosed is loaded. Basic probability assignment is calculated and a reliability decision matrix is constructed. An induced ordered weighted averaging operator and its induction vector are calculated according to a sample source of the data. An index weight vector is calculated. All evidence is fused by the induced ordered weighted evidence theory and reliability of comprehensive evaluation is calculated, so as to determine a diagnosis result. The disclosure realizes fault identification, fault type distinction and fault position of power equipment by interpreting detection waveforms.

Transformer fault diagnosis method and system using induced ordered weighted evidence reasoning

A method and a system for fault diagnosis with small samples of power equipment based on virtual and real twin spaces are disclosed, which belong to the field of fault diagnosis of power equipment. The method includes: test samples containing different locations, types and severity levels of fault of power equipment are acquired to form a real physical space; a virtual mirror space is acquired by simulation according to a simulation model of the equipment to be diagnosed; the training set in the real physical space is spatially integrated with the sample set in the virtual mirror space to obtain a training sample set in the twin spaces; the training sample set in the twin spaces serves as the supplement to the training set in the real physical space, and the fault type and fault location serve as diagnostic labels to be input to the deep neural network for training.

Method and system for fault diagnosis with small samples of power equipment based on virtual and real twin spaces

A method and a system of predicting an electric system load based on wavelet noise reduction and empirical mode decomposition-autoregressive integrated moving average (EMD-ARIMA) are provided. The method and the system belong to a field of electric system load prediction. The method includes the following steps. Raw load data of an electric system is obtained first. Next, noise reduction processing is performed on the load data through wavelet analysis. The noise-reduced load data is further processed through an EMD method to obtain different load components. Finally, ARIMA models corresponding to the different load components are built. Further, the ARIMA models are optimized through an Akaike information criterion (AIC) and a Bayesian information criterion (BIC). The load components obtained through predicting the different ARIMA models are reconstructed to obtain a final prediction result, and accuracy of load prediction is therefore effectively improved.

Method and system of predicting electric system load based on wavelet noise reduction and EMD-ARIMA

The invention discloses a failure diagnosis method for a power transformer winding based on a GSMallat-NIN-CNN network. The failure diagnosis method includes: measuring a vibration condition of the transformer winding by using a multi-channel sensor to obtain multi-source vibration data of the transformer; converting the multi-source vibration data obtained through measurement into gray-scale images through GST gray-scale conversion; decomposing, by using a Mallat algorithm, each gray-scale image layer by layer into a high-frequency component sub-image and a low-frequency component sub-image, and fusing the sub-images; reconstructing fused gray-scale images, and coding vibration gray-scale images according to respective failure states of the transformer winding; establishing a failure diagnosis model for the transformer based on the GSMallat-NIN-CNN network; and randomly initializing network parameters to divide a training set and a test set, and training and tuning the network by using the training set; and testing the trained network by using the test set.

Failure diagnosis method for power transformer winding based on GSMallat-NIN-CNN network

The disclosure provides a deep parallel fault diagnosis method and system for dissolved gas in transformer oil, which relate to the field of power transformer fault diagnosis. The deep parallel fault diagnosis method includes: collecting monitoring information of dissolved gas in each transformer substation and performing a normalizing processing on the data; using the dissolved gas in the oil to build feature parameters as the input of the LSTM diagnosis model, and performing image processing on the data as the input of the CNN diagnosis model; building the LSTM diagnosis model and the CNN diagnosis model, respectively, and using the data set to train and verify the diagnosis models according to the proportion; and using the DS evidence theory calculation to perform a deep parallel fusion of the outputs of the softmax layers of the two deep learning models.

Deep parallel fault diagnosis method and system for dissolved gas in transformer oil

A transformer failure identification and location diagnosis method based on a multi-stage transfer learning theory is provided. Simulation is set up first, a winding parameter of a transformer to be tested is calculated, and a winding equivalent circuit is accordingly built. Different failures are configured for the equivalent circuit, and simulation is performed to obtain a large number of sample data sets. A sweep frequency response test is performed on the transformer to be tested, and detection data sets are obtained. Initial network training is performed on simulation data sets by using the transfer learning method, and the detection data sets are further trained accordingly. A failure support matrix obtained through diagnosis is finally fused. The multi-stage transfer learning theory is provided by the disclosure.

Transformer failure identification and location diagnosis method based on multi-stage transfer learning

A method includes steps: 1) obtaining monitoring information of different monitoring points in normal state of power equipment; 2) setting faults and obtaining monitoring information of different fault types, positions, monitoring points of the equipment; 3) taking the monitoring information obtained in steps 1) to 2) as training dataset, taking the fault types and positions as labels, inputting the training dataset and the labels to deep CNN for training; 4) collecting monitoring data, performing verification and classification using step 3), obtaining probability values corresponding to each of the labels; 5) taking classification results of different labels as basic probability assignment values, with respect to a monitoring system composed of multiple sensors, taking different sensors as different evidences for decision fusion, performing fusion processing using the DS evidence theory to obtain fault diagnosis result. The invention can intelligently realize fault detection, fault type determination, and fault positioning of the power equipment.

Power equipment fault detecting and positioning method of artificial intelligence inference fusion

The disclosure provides an internal thermal fault diagnosing method for an oil-immersed transformer based on DCNN and image segmentation, including: 1) dividing an internal area of a transformer, and using fault areas and normal status as labels of DCNN; 2) through lattice Boltzmann simulation, randomly obtaining multiple feature images of the internal temperature field distribution of the transformer under normal and various fault state modes, and the fault area serves as a label to form the underlying training sample set; 3) obtaining historical monitoring information of the infrared camera or temperature sensor, and forming its corresponding fault diagnosis results into labels; 4) combining all monitoring information contained in each sample into one image, and then extracting the same monitoring information from the samples in the sample set to form a new image; 5) segmenting image sample and then inputting the same into DCNN for training to obtain diagnosis results.

Internal thermal fault diagnosis method of oil-immersed transformer based on deep convolutional neural network and image segmentation

A fault diagnosis method for power electronic circuits based on optimizing a deep belief network, including steps. (1) Use RT-LAB hardware-in-the-loop simulator to set up fault experiments and collect DC-link output voltage signals in different fault types. (2) Use empirical mode decomposition to extract the intrinsic function components of the output voltage signal and its envelope spectrum and calculate various statistical features to construct the original fault feature data set. (3) Based on the feature selection method of extreme learning machine, remove the redundancy and interference features, as fault sensitive feature data set. (4) Divide the fault sensitive feature set into training samples and test samples, and primitively determine the structure of the deep belief network. (5) Use the crow search algorithm to optimize the deep belief network. (6) Obtain the fault diagnosis result.

Power electronic circuit fault diagnosis method based on optimizing deep belief network

A method and a system for power equipment diagnosis based on windowed feature and Hilbert visualization are provided, which belong to the field of power equipment fault diagnosis. The method includes: obtaining an original data set of monitoring data containing power equipment fault features; introducing windowed feature calculation considering logarithmic constraints to process data to obtain a feature sequence; using Hilbert visualization method for further processing to obtain a Hilbert image data set used to train and verify a convolutional neural network; and finally directly inputting newly obtained test sample data after windowed feature calculation and Hilbert visualization processing into the trained network for fault diagnosis and location. The disclosure uses windowed feature calculation and Hilbert visualization to process the monitoring data of a power equipment to fully extract fault features and effectively improve diagnostic accuracy, and uses the convolutional neural network for diagnosis to improve the intelligence of diagnosis.

Method and system for power equipment diagnosis based on windowed feature and Hilbert visualization

The disclosure discloses a power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification, including 1) a winding equivalent circuit is established, and a transfer function thereof is calculated; 2) a sine wave excitation source is set at one end of the power transformer winding to obtain the amplitude-frequency characteristic curve of each winding node; 3) circuits under various fault statuses are subjected to scanning frequency response analysis to extract amplitude-frequency characteristics; 4) a feature matrix is established based on the obtained amplitude-frequency characteristics; 5) scanning frequency response analysis is performed on the diagnosed power transformer to form a feature matrix; 6) the feature matrix is converted into an image, simulation and historical detection data are used as a training set, and a deep convolutional neural network is input for training; 7) diagnosed transformer is subjected to fault classification and positioning.

Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification

The disclosure provides a silicon photovoltaic cell scanning eddy current thermography detection platform and a defect classification method. The technical solution adopted by the disclosure is: firstly, fixing the position of the electromagnetic inductive coil and the thermal imager, and using the main conveyor belt to carry the silicon photovoltaic cell to move forward on the production line to form a scanning eddy current heating of the silicon photovoltaic cell. Secondly, the defect temperature information is obtained through the thermal imager in terms of thermal image sequences. Thirdly, the feature extraction algorithms are used to extract the silicon photovoltaic cell defect features. Finally, the image classification algorithms are used to classify the silicon photovoltaic cell defects, and the sorting conveyor belts are used to realize the automatic sorting of silicon photovoltaic cells with different types of defects on the production line.

Silicon photovoltaic cell scanning eddy current thermography detection platform and defect classification method

A system for testing a Nakagami fading channel and a verification method thereof are provided. The testing system includes a signal generator, a Nakagami fading channel simulator, and a computer. The signal generator is used to output a sine wave signal whose frequency is f and transmit the sine wave signal to the Nakagami fading channel simulator and the computer. The Nakagami fading channel simulator is used to generate a Nakagami fading channel. The computer is used to perform data processing and analysis. In the verification method, time domain fading characteristics, first-order statistics characteristics, and second-order statistics characteristics of the Nakagami fading channel are respectively verified. Verifying the time domain fading characteristics is verifying a waveform fluctuation rate and a fluctuation range on a time domain under different Nakagami fading factors. Verifying the first-order statistics characteristics is mainly verifying amplitude and phase distribution statistics characteristics of the Nakagami fading channel by means of Kolmogorov Smirnov (KS) hypothesis test. Verifying the second-order statistics characteristics is mainly verifying the shape and bandwidth of a power spectrum density function. In the present invention, verification on performance of the Nakagami fading channel simulator or a simulation model has features of accuracy and feasibility.

System for testing Nakagami fading channel and verification method thereof

A Deep Belief Network (DBN) feature extraction-based analogue circuit fault diagnosis method comprises the following steps: a time-domain response signal of a tested analogue circuit is acquired, where the acquired time-domain response signal is an output voltage signal of the tested analogue circuit; DBN-based feature extraction is performed on the acquired voltage signal, wherein learning rates of restricted Boltzmann machines in a DBN are optimized and acquired by virtue of a quantum-behaved particle swarm optimization (QPSO); a support vector machine (SVM)-based fault diagnosis model is constructed, wherein a penalty factor and a width factor of an SVM are optimized and acquired by virtue of the QPSO; and feature data of test data are input into the SVM-based fault diagnosis model, and a fault diagnosis result is output, where the feature data of the test data is generated by performing the DBN-based feature extraction on the test data.

Deep belief network feature extraction-based analogue circuit fault diagnosis method

A space vector pulse width modulation (SVPWM) method for suppressing a common-mode voltage of a multiphase motor includes the following steps: (1) dividing all basic vectors of the multiphase motor into q types, and selecting therefrom x types having equal common-mode voltage magnitude of which an absolute value is smallest; (2) for each type in the x types of basic vectors, structuring y classes of auxiliary vectors according to an optimization model; (3) synthesizing reference vectors by virtue of the auxiliary vectors to obtain functioning time of basic vectors functioning in each switching period; and (4) obtaining an optimal functioning sequence of the basic vectors functioning in each switching period with fewest switching operations of a converter as a purpose. The present invention may effectively suppress a magnitude and frequency of the common-mode voltage of the multiphase motor without increasing calculation complexity or reducing other performance indexes.

Growing community of inventors

Hubei, China

Jiajun Duan