ALBERT

Description: Structural health monitoring (SHM) technology is a monitoring process and early warning method for the health status or damage of special workpiece structures by deploying sensors. In recent years, there have been many studies on SHM, such as ultrasonic, pulsed eddy current, optical fiber, magnetic powder, and other nondestructive testing technologies. Due to their sensor deployment, testing environment, power supply, and transmission line wiring mechanism, they bring problems such as detection efficiency, long-term monitoring, and unreliable systems. The combination of wireless sensing technology and intelligent detection technology is used to solve the above problems. Therefore, this paper studies the tag antenna smart sensor, which is used to characterize the extension of metal defects in SHM. Then, it presents a wireless passive three-dimensional sensing antenna, and simulations verify the feasibility of the antenna. The simulation results show that the antenna can characterize the two extension directions of depth and width of the metal surface structure smooth defect. At the same time, the antenna can characterize the position of smooth defects on the surface of metal structures relative to the antenna and then realize the smooth defect positioning.

Description: Aiming at the shortcomings of single network classification model, this paper applies CNN-LSTM (convolutional neural networks-long short-term memory) combined network in the field of music emotion classification and proposes a multifeature combined network classifier based on CNN-LSTM which combines 2D (two-dimensional) feature input through CNN-LSTM and 1D (single-dimensional) feature input through DNN (deep neural networks) to make up for the deficiencies of original single feature models. The model uses multiple convolution kernels in CNN for 2D feature extraction, BiLSTM (bidirectional LSTM) for serialization processing and is used, respectively, for audio and lyrics single-modal emotion classification output. In the audio feature extraction, music audio is finely divided and the human voice is separated to obtain pure background sound clips; the spectrogram and LLDs (Low Level Descriptors) are extracted therefrom. In the lyrics feature extraction, the chi-squared test vector and word embedding extracted by Word2vec are, respectively, used as the feature representation of the lyrics. Combining the two types of heterogeneous features selected by audio and lyrics through the classification model can improve the classification performance. In order to fuse the emotional information of the two modals of music audio and lyrics, this paper proposes a multimodal ensemble learning method based on stacking, which is different from existing feature-level and decision-level fusion methods, the method avoids information loss caused by direct dimensionality reduction, and the original features are converted into label results for fusion, effectively solving the problem of feature heterogeneity. Experiments on million song dataset show that the audio classification accuracy of the multifeature combined network classifier in this paper reaches 68%, and the lyrics classification accuracy reaches 74%. The average classification accuracy of the multimodal reaches 78%, which is significantly improved compared with the single-modal.

Description: The microseismic signals in the coal minefield are very complex because of its special environment with a large number of blast vibration signals, and how to effectively identify the microseismic signals is still a big problem. S transform (ST) and Manifold Learning (ML) methods are introduced to extract the characteristics of the microseismic signals, and Gaussian Mixture Model based on the improved Bee Colony optimization algorithm (IBC-GMM) is established to identify the microseismic signals accurately. Firstly, the time-frequency characteristics of microseismic signals in coal mine are extracted by ST analysis. It is found that there are obvious time-frequency differences between rock-fracturing signals and blast vibration signals. Blast vibration signals have short duration, high frequency, and complex frequency spectrum, and their dominant frequencies are mainly over 100 Hz. However, rock-fracturing signals are relatively slow, with low frequency and stable spectrum change, and their dominant frequencies are generally below 100 Hz. Then, combining with the microseismic data of Xiashinjie coal mine in Tongchuan, China, the feature dimension reduction is carried out by Manifold Learning (ML) method, and the processed feature vectors are automatically recognized by IBC-GMM. Field test results show that the method summarizes the characteristics of the microseismic wave which are difficult to emerge as the learning vector, and the features reflect the key features of microseismic signals well. The identification accuracy is as high as 94%, and its recognition effect is superior to other recognition models (such as traditional Gaussian Mixture Model based on Expectation-Maximum (EM-GMM), Backpropagation (BP) neural network, Random Forests (RF), Bayes (Bayes) methods, and Logistic Regression (LR) method). Therefore, IBC-GMM could be used to mine engineering microseismic monitoring waveform recognition to provide the reference.

Description: The automated mechanical transmission (AMT) based on the electromagnetic linear driving device (EMLDD) has good potential for shift performance. However, the direct-drive shifting mechanism based on the displacement sensor is difficult to meet the compactness of the structure and control robustness in complex environment. Through analyzing the working principle of the electromagnetic linear driving device and features of sensorless control strategy, a new displacement prediction method based on the improved GA-BP neural network is proposed to replace the displacement sensor. With current, voltage, and input shaft speed of the electromagnetic linear driving device as input, displacement prediction is obtained by the GA-BP neural network with improved selection factor. Finally, the experiment validated the effectiveness of displacement prediction based on the improved GA-BP neural network of shift control. The results showed that prediction accuracy of the improved GA-BP neural network was greater than 96% under all shift working conditions. The average RMSE was reduced by 21.8%, absolute error of displacement prediction was controlled within ±0.5 mm, and average shift time was less than 0.18 s. In this paper, the BP neural network is applied to complex linear displacement prediction, which has important application and popularization value.

Description: Study on permeability evolution of an aquifer coal seam in Western China is of great significance for preventing water inrush disaster and realizing water-conserving coal mining. The permeability evolution of an aquifer coal seam is related to a loading path closely under plastic flow. In this work, permeability variations of the Xiaojihan water-bearing coal seam and Longde nonwater coal seam are researched using a transient method under plastic flow. The experiment results indicated the following: (1) Under the same axial strain, the permeability, relative residual strain, and confining pressure influence coefficient of Xiaojihan coal specimens all decrease in plastic flow with the increase of loading-unloading times and confining pressure, while the permeability recovery coefficient increases during this process. (2) The permeability of Xiaojihan water-bearing coal specimens decreases with the growth of axial strain in plastic flow, resulting in the increase of relative residual strain and reinforcement of plasticity. Besides, the confining pressure influence coefficient decreases and the permeability recovery coefficient decreases slightly with the axial strain. (3) Finally, the permeability of Xiaojihan coal specimens is greater than that of Longde coal specimens, while the confining pressure influence coefficient and permeability recovery coefficient of Longde coal specimens are greater than those of Xiaojihan coal specimens. The closure rate of internal cracks of the water-bearing coal specimen is lower than that of the nonwater coal specimen, which is beneficial for water storage and transport.