33results about How to "Enhanced fault signature" patented technology

The invention relates to a bearing fault diagnosis method under a strong noise variable speed condition based on an energy weight method. The method comprises the steps: extracting a vibration signalorder through employing a time-frequency ridge feature point linear interpolation and masking algorithm method according to a time-frequency representation graph based on Gabor transformation; performing instantaneous frequency estimation and secondary fitting on the vibration signal by using a local extremum search algorithm and the extracted order; carrying out equal-angle resampling on the vibration signal by utilizing a key phase time scale method according to the fitted instantaneous frequency; performing Hilbert-Huang transformation of CEEMDAN on the resampled isometric domain signal toobtain an order-frequency spectrum of the signal; extracting an impact energy occurrence position in the order-frequency spectrum, and then carrying out binaryzation on the order-frequency spectrum; acquiring an energy weight order sequence capable of reflecting impacts through multi-scale binary spectrum analysis, and carrying out power spectrum analysis on the energy weight order sequence to obtain fault-related impact components. The influence of strong noise and variable rotating speed on vibration signal analysis can be eliminated, and the accuracy of rolling bearing fault diagnosis is improved.

The invention relates to a diagnosis method for impact type failure between a rolling bearing and a gear based on an optimal self-adaptive wavelet filter, which comprises the following steps: firstly establishing an exponentially damped sinusoidal impact type failure signal model, using a mode self-adaptive wavelet generation algorithm publicly disclosed by Hector Mesa to generate a self-adaptive wavelet filter which is matched with the signal model, then utilizing a fast FIR (finite impulse response) filtering algorithm to execute the wavelet filter to obtain the filtering result, then calculating the kurtosis value of the filtering result, and using an evolution differential algorithm to repeat the steps so as to finally obtain an optimized envelope spectrum. By adopting the method, impact failure characteristics in vibration signals can be precisely extracted, and a clearer envelope spectrum can be given out, thereby ensuring that failure symptoms can be displayed more clearly.

An antifriction bearing fault diagnosis method based on an energy operator demodulated depth belief network (DBN) and a particle swarm optimized support vector machine (PSO-SVM); the method comprisesthe following steps: using an energy operator demodulation method to obtain an instantaneous Teager oscillogram and solving a time frequency characteristic statistics parameter thereof; using the DBNto extract secondary characteristics of the time frequency characteristic statistics; finally inputting the extracted characteristic parameter into the PSO-SVM for fault classification. The antifriction bearing fault diagnosis method is higher in accuracy, can greatly shorten the algorithm training time, thus improving the fault diagnosis accuracy and efficiency.

The invention discloses an improved keyless phase fault feature order extraction method comprising the following steps: 1, using a vibration acceleration sensor to obtain equipment state information,and using a time frequency analysis method to preprocess the obtained state information so as to obtain an instant frequency; 2, carrying out normal integral operation for the estimated instant frequency so as to obtain a roughly estimated instant phase, and using a Romberg integral rule to correct the roughly estimated instant phase, thus finally obtaining an accurately estimated instant phase; 3, using the accurately estimated instant phase information to resample an angle domain of the original signal according to a mapping relation between the time domain and the angle domain; 4, using a flexible angle domain synchronous averaging method to denoise the resampled signal of the angle domain, and carrying out order spectral analysis for the denoised angle domain resampled signals, thus extracting the equipment fault feature order. The method can accurately extract the equipment fault feature order under a keyless phase change rotating speed condition.

The invention belongs to the field of mechanical equipment fault diagnosis. The invention discloses a rolling bearing fault diagnosis method based on the Walsh transformation and Teager energy operator. The specific contents are as follows: first, an acceleration sensor is fixed on a bearing seat under test, the acceleration sensor is connected with a data acquisition instrument to obtain a bearing vibration signal, then a bearing fault diagnosis is performed, and subjecting the signal to Walsh transform denoising; fault features are extracted by Teager energy operator demodulation; and the bearing fault type is positioned and identified by comparing with a theoretical calculating value. The invention takes the vibration signal as the research object, combines the Walsh transform denoisingand the Teager energy operator demodulation, and proposes the novel bearing fault diagnosis method. The bearing fault diagnosis method in the invention can effectively diagnose the fault type of a rolling bearing and has the advantages of high diagnostic accuracy.

The invention discloses a gearbox intelligent fault diagnosis method based on a multichannel self-calibration convolutional neural network. The gearbox intelligent fault diagnosis method comprises the following steps: increasing dimension of glassmeter angle field data, converting one-dimensional vibration signals of a plurality of sensors into two-dimensional data, converting the two-dimensional data into gray level images as input, establishing a data set, and dividing the data set into a training set and a test set. And constructing a self-calibration convolutional neural network, and extracting data features. And setting a fusion layer, converting the output of the self-calibration convolutional neural network into one-dimensional data, and fusing feature information. And setting a full connection layer, and mapping the distributed features to a sample marking space. And constructing a Softmax feature classifier to classify the extracted features. And learning the network by using the training set, and testing the trained network by using the test set to realize fault diagnosis of the gearbox. The self-calibration convolutional neural network model provided by the invention is combined with an information fusion method, and can effectively diagnose a single fault of the gearbox under the same rotating speed working condition.

The invention discloses a rolling bearing rolling body weak fault feature extraction method under a transmission path, and the method comprises the steps: decomposing a fault signal into a series of IMF modal components through employing a VMD signal decomposition method, removing an IMF containing a rotating frequency component from the IMF, selecting an IMF containing more fault information from the remaining modal components for signal reconstruction, and therefore, the influence of the frequency conversion component and other interference components on the fault characteristics is limited; for the problem that the fault features of the bearing rolling body signal are weak in the transmission process, the periodic fault features of the bearing rolling body signal are enhanced by adopting a parameter optimization MCKD method; and finally, the 1.5-dimensional spectrum is used as post-processing of the method, and the fault frequency and frequency multiplication of the signal are fully highlighted. According to the method, the periodic fault features of the weak fault signals are fully enhanced by utilizing the method of combining the parameter-optimized MCKD algorithm and the 1.5-dimensional spectrum, and compared with the traditional MCKD algorithm and the 1.5-dimensional spectrum, the combined method has a better feature extraction effect.

The invention provides a rolling bearing fault diagnosis method and device, a medium and computer equipment, wherein the method comprises the steps: intercepting a vibration signal of a rolling bearing, and obtaining an initial signal; creating an initial Hankel matrix based on the initial signal, and reconstructing the initial Hankel matrix by using a singular value decomposition algorithm to obtain a reconstructed signal; demodulating the reconstructed signal by using a 1.5-dimensional symmetric differential analysis energy operator demodulation algorithm to obtain a demodulated signal; determining a 1.5-dimensional energy spectrum of the fault characteristic signal based on the demodulation signal; carrying out fault diagnosis on the bearing based on a 1.5-dimensional energy spectrum; therefore, reconstructing the original vibration signal of the rolling bearing by using a self-adaptive singular value decomposition algorithm, removing background noise in the vibration signal, and obtaining fault features; and processing the fault features through a 1.5-dimensional symmetric difference analysis energy operator demodulation algorithm. Residual noise is suppressed, the fault features are improved, and the precision of a fault diagnosis result is ensured.

The invention discloses a self-diagnosis method for faults of a rolling bearing retainer. The method comprises steps of rapidly filtering vibration signals through a high-precision rapid filter bank;selecting an optimal signal according to the spectral kurtosis of the mean square envelope autocorrelation signal of each filtered signal, and calculating an envelope spectrum of the optimal signal; then, automatically selecting the first M-order actual fault characteristic frequency according to the theoretical fault characteristic frequency of the retainer; automatically finding out a thresholdvalue meeting the probability requirement according to the statistical characteristic of the frequency spectrum; and finally, calculating a global test index and a fault occurrence rate of cyclic stability, and realizing self-diagnosis of the fault of the rolling bearing retainer. According to the invention, an effective self-diagnosis tool is provided for an intelligent bearing to analyze the fault of the rolling bearing retainer.

The invention discloses a method and a system for positioning a single-pole grounding fault section of a direct-current power distribution network, and the method comprises the steps: cooperatively inputting a resonant branch to inject a current detection signal through an MMC additional control strategy, then setting an action sequence of a load switch/circuit breaker based on a delay level difference principle, and completing the positioning and isolation of a fault section based on single-end information. The section positioning method avoids the installation of additional devices, and has the characteristics of high sensitivity and no need of double-end communication.

The invention relates to an autonomous underwater robot propeller fault feature extraction method based on empirical mode decomposition, fractal dimension and an SHFC positioning algorithm, and belongs to the technical field of underwater robot fault diagnosis. The method comprises the following steps: performing data preprocessing by adopting empirical mode decomposition to replace a common filtering method in fractal dimension; in a high-frequency part after modal decomposition, introducing a rolling time window through extraction of fractal dimension fault features of a small sample in eachtime window, capturing sudden change of a moment fractal dimension feature value from a fault to a fractal dimension, extracting a maximum value of the sudden change of the fractal dimension, and then enhancing an extraction effect of the fault features. Fault characteristics can be enhanced, whether the underwater robot thruster breaks down or not can be conveniently detected, and the method isparticularly suitable for state monitoring of the autonomous underwater robot thruster and wide in application prospect.

The invention discloses a DC power distribution network grounding fault line selection method and system. The method comprises the steps of: starting MMC additional control by using a bus voltage imbalance criterion, and injecting a detection signal; collecting a positive electrode current and a negative electrode current of the head end of each feeder line by delaying delta t; filtering the positive electrode current and the negative electrode current of the head end of each feeder line, and setting the center frequency fmp as the characteristic frequency of the detection signal to obtain the filtered positive electrode current and negative electrode current of the head end of each feeder line; solving a zero-mode current of each feeder line for the filtered positive and negative electrode currents of the head end of each feeder line, carrying out normalization processing, and summing to obtain a reference current; and calculating waveform correlation between the normalized feeder zero-mode current and the reference current one by one, carrying out fault judgment according to a Pearson correlation coefficient to determine a fault line and a sound line, and completing grounding fault line selection of a DC power distribution network. The DC power distribution network grounding fault line selection method effectively enhances fault features, avoids installation of additional devices, and has the advantages of high sensitivity and no need of double-end communication.

The invention provides an autonomous underwater robot propeller fault feature extraction method, which is specifically based on empirical mode decomposition, fractal dimension and an SHFC positioningalgorithm, and belongs to the technical field of underwater robot fault diagnosis. According to the invention, empirical mode decomposition is adopted to carry out data preprocessing to replace a common filtering method in fractal dimensions; in a high-frequency part after modal decomposition, a rolling time window is introduced, through extraction of fractal dimension fault features of a small sample in each time window, sudden change of a moment fractal dimension feature value from a fault to a fractal dimension is captured, a maximum value of the sudden change of the fractal dimension is extracted, and then an extraction effect of the fault features is enhanced. According to the invention, fault characteristics can be enhanced, whether the underwater robot propeller has faults can be detected conveniently, and the method is especially suitable for state monitoring of autonomous underwater robot propellers.