Microseismic source location is the essential factor in microseismic monitoring technology, and its precision has a large impact on performance of technique. Here, we discuss problem low-precision identification for events mine, as may be obtained using conventional methods that are based arrival time. In this paper, characteristics mining analyzed according to mine's wavefield. We review research progress mine-related recent years, including combination Geiger method with linear method,...

VSC-HVDC is a kind of HVDC technology which based on voltage source converter and turn-off devices. Its poor over voltage/current capacity are prone to failure. Based the established system simulation model, DC waveforms under various fault conditions achieved, then character decided according amplitude fluctuation range voltage. Moreover, in full consideration influence transmission power, wavelet analysis method adopted extract feature faulty signal, combined with artificial neural network...

Anchor bolt corrosion is a complex and dynamic system, the prediction identification of its degree are significant importance for engineering safety. Currently, non-destructive testing using ultrasonic guided waves can be employed detection. Building upon analysis anchor mechanisms, this paper proposes method evaluating bolts based on multi-scale convolutional neural networks (MS-CNNs) that address multi-mode propagation dispersion effects wave signals in testing. Electrochemical experiments...

Probabilistic Neural Network (PNN) overcame the shortcomings of entrapment in local optimum, slow convergence rate which was BP algorithm. With enough training samples, PNN obtained optimal result Bayesian rules. Because fast rate, samples can be added into at any time. So, is fit to diagnose fault power transformer and has auto-adaptability. In order improve classification accuracy, conception combination introduced PNN. The diagnosis consisted 4 Probability neural networks this paper. PNN1...

Feedforward neural network (FNN) is an information processing system that simulates human brain function to a certain extent by referring the structure of biological and working mechanism neurons. As most important part, architecture FNN essentially influences its application performance. This paper proposed self-adaptive algorithm for (SSAFNN) based on mechanism. First, neuron growing pruning indexes are idea grow factor competition, respectively. The dynamically adjusted according hidden...

Grout is an important part of the bolt anchorage system; once having damage, cohesive force will not reach requirement, which may affect support effect. Moreover, as buried in rock, defects that happened grout have unknown characteristics. In order to identify defect, a multiscale convolutional generative adversarial network (MSCGAN) method proposed this article. The generator based on supervised learning designed MSCGAN framework, can generate data for enriching diversity system Together...

Fisher discriminant is a classical linear technique widely used in pattern classification and feature extraction. When fisher of power transformer fault based on dissolved gasses analysis (DGA), it transformed to its nonlinear version high dimensional space by means the kernel trick. The (KFDA) presented diagnose faults oil-immersed transformer. In order improve accuracy, conception combination introduced. diagnosis consisted 4 KFDA. KFDA1 classify normal fault. KFDA2 thermal discharge KFDA3...

Based on learning principle of support vector machine (SVM), damage detection bolt anchorage is studied in this paper. Characteristic matrix composed 17 time domain or frequency characteristic properties including mean value, peak value the denoised signal, etc. Principle component analysis(PCA) used for feature extraction and several main components with higher cumulative contribution rate are chosen SVM classifier training testing. Particle swarm optimization(PSO) method to optimize...

To improve the accuracy of terrain classification during mobile robot operation, an adaptive online method based on vibration signals is proposed. First, time domain and combined features time, frequency, time–frequency domains in original signal are extracted. These adopted as input random forest algorithm to generate models with different dimensions. Then, by judging relationship between current speed its critical speed, model dimensions adaptively selected for classification. Offline...

Transformers fault diagnosis plays a vital role in running security and reliability. The detected information is collected from the disperse sensor, which lack of data fusion analysis easily lead to decision error leak. A model oil-immersed transformer based on collaborative method Kernel C-Means Clustering (KCM) multi-source presented. basic idea that trained samples are clustered first by using KCM, then Dempster-Shafer theory evidence Fusion used train chosed sample, decide fault. result...

The parameters such as the length of rock bolts, anchorage length, free or construction defects (e.g., position less density in body) are important to evaluate quality bolts. Because harsh environment, measured signal includes many kinds noise, which have a great effect signal. It is difficult detect effectively by traditional information processing technology, and therefore new technology based on non-destructive detecting provided. steps follows: first, excite bolts using pseudo-random...

Deep learning is a hot topic in the field of machine learning, which provides new method for nondestructive testing bolt anchorage. Aiming at importance identifying type defects, this paper proposes DS-DBN-SVM(Differential Search-Deep belief network-Support vector machine) model anchorage defects. The DS algorithm used to optimize weights and thresholds DBN network. original acceleration signal anchor as input DS-DBN extract high-level features signal. Finally, SVM classifier defect...

In nondestructive testing of anchor bolt quality, it is important to identify the state accurately. this paper, a combination method spectral kurtosis and K-means clustering algorithm proposed different types models: maximum nonzero minimum values are extracted as eigenvalues from distribution model signals which was calculated through fast kurtogram algorithm, then automatically classified by realize identification model. Through verification test, proved be fast, effective high accuracy...

Superconducting magnet (SM) plays a crucial role in the success or failure of superconducting generator. The SM 50 kW synchronous generator (SSG) was studied this paper. Magnesium diboride wires are employed to wind SM. First, initial structure is designed by finite element modeling. After which an approach based on response surface methodology and particle swarm optimization used optimize Finally, order verify whether conforms requirements SSG, made tested optimized size. results indicate...

The voltage source converter is an important part of VSC-HVDC (Voltage based high direct current) system which has low over-voltage and over current bearing ability thus prone to various faults. In this paper, on the PSCAD/EMTDC simulation model, mathematical model DC fluctuation component caused by AC single-phase break fault established. Also, both-side converters built. Based research mentioned above, two VSC harmonic transfer patterns are studied. It proved that n-th positive sequence...