43 results about "FastICA" patented technology

The invention provides a rolling bearing fault feature extraction method based on an independent component analysis and cepstrum theory. The rolling bearing fault feature extraction method comprises the steps of acquiring a vibration acceleration testing signal of a rolling bearing by using an acceleration sensor; decoupling and separating the vibration acceleration testing signal by using FastICA based on negentropy maximization; selecting a separated signal capable of representing fault feather information to the maximum extent; carrying out cepstrum analysis on the selected separated signal, and drawing a cepstrum chart; observing whether the cepstrum chart has a fault feature frequency or an obvious peak value at a frequency multiplication position, and furthermore, judging whether the rolling bearing has a fault. By using the rolling bearing fault feature extraction method, the feature information of a fault signal of the rolling bearing can be effectively recognized from a complex sideband signal, a periodical fault component in a sideband can be conveniently extracted, the fault information is remarkably enhanced, the fault diagnosis precision is greatly improved, the fault diagnosis time period is shortened, and the spectral analysis difficulty is simplified; in addition, the rolling bearing fault feature extraction method is easy to realize and good in real-time property.

The invention discloses a detection method of harmonic and inter-harmonic based on single-channel FastICA (Fast Independent Component Analysis). The detection method comprises the following steps of separating the harmonic or the inter-harmonic in an original voltage signal or an original current signal, and respectively carrying out fast Fourier transformation on each harmonic or inter-harmonic so as to obtain parameters, such as the frequency, amplitude and phase. Owing to application of a circulation translation method and a principle component analysis method, the harmonic or the inter-harmonic with prior amplitude is easier to detect, the detection method still keeps good accuracy in the larger noise, the problems that the harmonic or the inter-harmonic with the small amplitude is difficult to detect and is sensitive to the noise in the conventional detection method are solved, and a novel detection method is provided for detection of the harmonic or the inter-harmonic in power quality of a power system.

The invention relates to the technical field of grounding fault identification, in particular to a line selection method based on a FastICA for a low current grounding system in a single-phase grounding fault. According to the method, the FastICA method is utilized to process zero-sequence current of the grounding fault, three observation channels are arranged to separate fault steady-state components, transient-state components and noise signals from observation signals, the line fault characteristics are obvious after processing, and accurate line selection can be conducted on a fault line according to the energy of line separation variables. Because the noise signals are separated at the same time, the method is reliable, high in anti-interference capability and capable of not being affected by interference signals; meanwhile, for different initial phase angles of the fault, satisfactory line selection results can all be obtained by means of the method. By using the method to decompose and analyze the zero-sequence current of the fault, the cancellation or compensation among the components of the signals can be avoided, so that the fault characteristics are obvious to achieve the accurate line selection, and the method is widely applicable to the low current grounding system.

The invention provides a rotating machine fault diagnosis method based on Fast ICA-spectrum kurtosis-envelope spectrum analysis. The method includes the steps that (1), an acceleration sensor is used for obtaining vibration acceleration test signals of a rotating machine; (2), the test signals are separated in a decoupling mode by the adoption of a Fast ICA method with negentropy maximized; (3), the spectrum kurtosis of the separated signals is calculated, and the separated signals capable of representing fault information best are screened out; (4), Hilbert envelope spectrum analysis is performed on the selected separated signals; (5), the frequency corresponding to the peak value of an envelope spectrum is compared with the fault character frequency of a bearing, so that a concrete fault is diagnosed. The effectiveness of the method is verified well by the inner ring fault diagnosis of a rolling bearing of the rotating machine, the fault information is obviously increased, the precision of the fault diagnosis is greatly improved, the fault diagnosis method is easy to achieve and good in real-time performance, and it is illustrated that the rotating machine fault diagnosis method has good application prospects.

The invention relates to a TSP strong industrial electrical interference suppression method based on SVD-ICA. Firstly, a hankel matrix is built for single-channel data interfered by strong industrialpower; SVD decomposition is performed on the built hankel matrix then, and an industrial electrical interference signal is rebuilt then; and, finally, sine and cosine signals as the FastICA input signals are respectively built according to industrial electrical interference frequency, and an effective signal is rebuilt through processing of a FastICA output component by utilizing the cross-correlation technology. It is verified that effective suppression of the fundamental frequency and harmonic interference of the industrial electric power can be achieved; and, compared with the existing industrial electrical interference suppression method, the TSP strong industrial electrical interference suppression method based on the SVD-ICA has the advantages of good noise suppression modification,small damage to the effective signal, no phase shift during processing and no influence on target positioning accuracy. Data processing is fast through the method, the cost of repeated collection dueto too large industrial electrical interference can be reduced, construction efficiency is improved then, and construction cost is saved at the same time; and the data quality is improved without changing collection and construction conditions, and the method is of great significance for improving the accuracy of geological prediction.

The invention discloses a structural modal parameter identification method based on wavelet and ICA improved EEMD. The method comprises steps of eliminating noise residues in the IMFs after classic EEMD decomposition by using a wavelet soft threshold denoising method in wavelets; after FastICA in ICA is used for processing, obtaining new IMFs, and removing modal aliasing; judging a real IMF component by using a correlation coefficient between the new IMFs and the original signal; removing a false mode, then obtaining free vibration response of each real IMFs by using a random decrement technology, and finally, identifying the frequency and damping ratio of the structure by using a method of combining Hilbert transform (HT) and a least square fitting technology. The method has the advantages that the problems of white noise residues after EEMD decomposition, modal aliasing, effective IMFs screening and the like are solved; wavelet soft threshold denoising in wavelet transform and FastICA in ICA are combined to improve the EEMD, and the problems of excessive false modes, noise residues and true IMF judgment in the classic EEMD process are solved.

The invention relates to a mechanical vibration fault characteristic time domain blind extraction method, and belongs to the technical field of mechanical equipment status monitor and fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method includes: firstly, expanding a vibration observation signal into a high dimension signal subspace; then, obtaining a low dimension signal; afterwards, performing FastICA independent component analysis, calculating normalization kurtosis of all independent components, figuring out a component signal corresponding to the minimum normalization kurtosis, and using an orthogonal matching pursuit algorithm to reconstitute periodic signals; subsequently, removing the reconstituted periodic signal from each independent component, and then using an improved KL distance algorithm to calculate a distance matrix among the independent components after the periodic signals are removed from the independent components, and performing dynamic particle swarm clustering so as to obtain an estimation signal; finally, analyzing an envelope demodulation spectrum of the estimation signal, and performing fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method is suitable for processing a long convolution data problem, can effectively reduce influences from periodic ingredients on a blind separation result, and simultaneously can solve blind separation result order uncertainty problems, and finally achieves bearing fault characteristic extraction.

The invention relates to the technical field of grounding fault identification, in particular to a line selection method based on a FastICA for a low current grounding system in a single-phase grounding fault. According to the method, the FastICA method is utilized to process zero-sequence current of the grounding fault, three observation channels are arranged to separate fault steady-state components, transient-state components and noise signals from observation signals, the line fault characteristics are obvious after processing, and accurate line selection can be conducted on a fault line according to the energy of line separation variables. Because the noise signals are separated at the same time, the method is reliable, high in anti-interference capability and capable of not being affected by interference signals; meanwhile, for different initial phase angles of the fault, satisfactory line selection results can all be obtained by means of the method. By using the method to decompose and analyze the zero-sequence current of the fault, the cancellation or compensation among the components of the signals can be avoided, so that the fault characteristics are obvious to achieve the accurate line selection, and the method is widely applicable to the low current grounding system.

The invention relates to a kurtosis-based FastICA and an STLBO motor bearing fault diagnosis method approximating the solution domain. The method of approximating the solution domain is combined with the initial training of BPNN to determine the upper and lower bounds of network structure parameters and TLBO initialization population; finally, the TLBO correction algorithm of adaptive dynamic learning factor is integrated, and the optimal weights and thresholds are iteratively searched for input into BPNN. The invention can enhance signal characteristics, reduce noise interference, and improve the recognition rate of fault diagnosis; for unknown sources, it does not need too much prior knowledge and theoretical reserves, and has strong generalization; It is necessary to set specific parameters and dynamically change the learning factor according to the current number of iterations, which can improve the convergence speed of the algorithm, and at the same time avoid falling into local convergence, which is convenient for debugging and simple in calculation.

The invention discloses a method for enhancing microphone voice through a small-bore array. According to the method, under the situation that priori knowledge such as a voice signal DOA and an array topological structure is unknown, according to the FastICA basic principle, an objective function with a separation matrix as a variable is built for an output variable, the output variable is maximized, independent isolated components are separated, and then post-processing is conducted. In the process, estimation on a voice source signal direction is removed, and calculated amount is saved.

The invention relates to a single-channel gearbox multi-fault separation dual-core micro-processing system and belongs to the gearbox fault diagnosis technology and signal processing technology field.The system of the invention is the dual-core micro-processing system of DSP and ARM. A DSP core includes a phase space reconstruction parameter estimation module, a phase space reconstruction module,a reconstruction parameter adjustment module, a reconstruction signal separation module, and a Fourier transform module. An ARM core includes a time frequency domain spectrogram drawing module. A time delay reconstruction method is used to expand, which avoids human intervention of a reconstructed signal taking experience as a judgment standard. A principal component analysis method is used to reduce influences of a noise and redundancy in a gearbox on the reconstructed signal. A limited support sample kernel function and FastICA fusion algorithm is used to the estimate a source signal probability density function, and then a nonlinear function according with a statistical characteristic of a source signal is obtained. And finally the signal is effectively separated. In the invention, theinfluence of improper selection of reconstruction parameters and a separated nonlinear function is greatly reduced on the aspect of gearbox multi-fault separation.

The invention relates to a distributed harmonic source identification method based on singe-frequency current transmission characteristics. The method comprises the following steps that (1) a current harmonic source identification model is established; (2) constraint conditions are exported; (3) a current harmonic observation matrix is established and preprocessed; (4) the harmonic source maximum identification frequency T and a non-Gaussianity threshold value F in an electricity network are set, and k is made to be equal to 0; (5) the k is made to be equal to k+1, and a FastICA method is used for solving a mixing matrix Wh; (6) the inverse matrix Wh-1 of the Wh is calculated, the phase position and the row order of the Wh-1 are adjusted according to the constraint conditions, the Wh-1 is denoted as Ph, if the Ph meets the constraint conditions, the step (7) is executed, and otherwise the step (5) is executed again; (7) the reverse matrix of the Ph is calculated, a demixing matrix Wh=Ph-1 is updated, the non-Gaussianity sum phi of identification results is calculated, if phi is no smaller than the threshold value F, the demixing matrix Wh is output, the harmonic source identification result is calculated, the positions of main harmonic sources are analyzed, iteration is ended, and otherwise the step (8) is executed; (8) when k is smaller than or equal to T, the step (5) is executed again, when k is larger than T, the demixing matrix Wh and the non-Gaussianity threshold value are output, and the positions of the main harmonic sources in the network are further analyzed.