Search

Spherical porous air bearing (SPAB) systems have been extensively used in various mechanical engineering applications. SPABs are promising materials in high-rotational speed, high-precision, and high-stiffness instruments. In SPAB systems, a rotor is supported by gas bearings, which provides higher rotational speed and lower heat generation environment than oil bearings do. Furthermore, SPAB does not cause deformation. Although, the supporting force of gas bearings is less, their stability is better than that of oil films. However, because the pressure distribution in the gas films is nonlinear, they are prone to failure at specific critical speeds, rotor imbalances, or inappropriate operations, which results in nonperiodic or chaotic motion and causes structural fatigue to the system. To understand and control the operating conditions of the SPAB systems during the nonperiodic motion, first, the governing equations of the SPAB system were solved to obtain the dynamic behavior of the rotor center. Then, the performance of the SPAB system were examined under different operating conditions by generating the maximum Lyapunov exponents (MLEs). However, the calculation process of MLE is extremely time consuming and complex. To solve this problem efficiently, a high-precision machine learning (ML)-based MLE prediction model was proposed in this study. The results show that the training process can be finished within few minutes, and the prediction process is able to be completed within few seconds. Meanwhile, the results demonstrate the merit of using the machine learning method for solving the MLE prediction problem and shorten the calculation time significantly. The proposed prediction model achieves excellent prediction outcome and it is more efficient and precise than traditional iteration scheme for the calculation of MLE. The feasibility of the proposed model is validated and the results also are the major contribution of this study.

In speech emotion recognition, the features extracted by handmade design are generally low-level, they may not be enough to distinguish subjective emotions, and speech signals are usually have time sequence and every frame signal has a different role. Therefore, this paper aims at the above problems, a DCNN BiGRU self-attention model is proposed. The model combines the spatial characteristics of convolutional neural networks, the advantages of circulating neural network in learning time series data, and the characteristics of attention mechanisms that can learn feature weights, thereby improving the accuracy of speech emotion recognition. This model achieved an average recognition rate of 89.53% and 91.74% in the EMO-DB and CASIA databases, and through comparison with other literatures, it is proved that this model can obtain more ideal results in speech emotion recognition.

In speech emotion recognition,the extraction of Mel-frequency Cepstral Coefficients (MFCC) will lose many useful spectral feature information,resulting in low recognition accuracy. Therefore,a method for combining MFCC and Mel spectrum for speech emotion recognition is proposed. First,M FCC are extracted from the audio signal,and then the Mel spectrum is extracted from the audio signal. Finally,the fused audio features are used to support vector machine classification. The experimental results in the EMODB data set show that the fusion of multiple features has higher classification accuracy than the classifier using single feature,and the proposed method can effectively improve the recognition accuracy.

In order to improve the accuracy of face recognition, a face recognition algorithm based on sparse representation and feature fusion is proposed. Firstly, the training samples and test samples are pre-processed by gray image conversion, scale scaling, histogram equalization, smooth filtering and so on, and the LBP, Gabor and HOG features of face images are extracted. And then the RSC classification test is carried on the partial samples. A loss function is defined according to the recognition result and the classification residual, then the weight vector is obtained by using the regularized least square method to minimize the loss function. Finally, the final residual is calculated according to the weight vector so as to obtain the final classification result. The experimental results on AR face dataset and LFW face dataset show that the recognition rate of our algorithm is obviously higher than the single feature recognition method, and it is robust to illumination, occlusion and expression.

Gas sensor based on surface acoustic wave (SAW) using piezoelectric zinc oxide (ZnO) thin film was simulated with COMSOL. The gas sensor is designed with an inter-digital transducer (IDT) on a ZnO substrate and covered with a thin polyisobutylene (PIB). The results show that there are the resonant frequency and ant-resonant frequency of SAW in the ZnO substrate arising from the electrode effect of the IDT, and deformed shape of the resonance SAW mode is different with that of anti-resonance SAW mode. The effect of the dichloromethane (DCM) on the gas sensor was also researched. Because the mass of the PIB film increases with PIB selectively adsorbing DCM in air, a shift in resonance to a slightly lower frequency was found.

In this paper, we propose a multi-local means based nearest neighbor classifier (MLMNN). In the MLMNN, k categorical nearest neighbors of a query sample are first found and used to calculate the corresponding k categorical multi-local mean vectors which can represent different local class-specific sample distributions. Then, the query sample is represented by a linear combination of k categorical local mean vectors and the representation coefficient of each local mean vector as the contribution to representing and classifying the query sample is obtained. Finally, the class-specific representation-based distance (i.e. reconstruction residual) between the query sample and k categorical multi-local mean vectors is adopted to determine the class label of the query sample. The experimental results on three popular face databases show that the proposed MLMNN method outperforms the related competitive KNN-based methods.

With the trend of developing intelligent machine tools, there are increasing needs of accelerometers for monitoring the vibration. The accelerometers used for measurement of the vibration signals during milling process require the characteristics of high sensitivity, high resolution and high bandwidth. The common used accelerometer is PZT type, however, is too expensive and difficult to integrate into the intelligent modules. Therefore, the MEMS accelerometer is an alternative with the advantages of lower price and better integration. The present study utilized ADXL001 MEMS accelerometer with the bandwidth of 22 kHz and full scale range of ±70g to develop a low cost dual-Axis accelerometer for measurement of machine tool vibration. In addition, the platform to evaluate the sensitivity of MEMS accelerometer based on ISO-16063-21 back to back calibration method is also developed. Furthermore, we design the package of the MEMS accelerometer and analyze the influence on the temperature and resonance frequency. Finally, the milling experiment is performed to verify the results of comparison between MEMS accelerometer and PZT accelerometer in the time domain and frequency domain. The results demonstrated that the Dual-Axis MEMS accelerometer is suitable for monitoring the vibration of machine tool with low cost.

A fully tunable bandpass filter using tunable transformers and multiple mode resonators (MMRs) is proposed. The proposed tunable MMR makes it possible to realize wide-band tunable filters. Thus the tuning ranges of the center frequency and bandwidth are wide. The novel tunable transformer may help to get the desired external quality factor (Qe) while the center frequency or bandwidth is tuned. Thus a good passband performance may be got.

We compared the effects of young plantations (4 to 5 years old) of two trees (Ulmus pumila and Robinia pseudoacacia (a legume)) and a shrub (Ziziphus jujube var. spinosa) on soil carbon and nutrient concentration, microbial biomass and their accretion during early phase of plantation establishment on mine spoils around a magnesite mine in Northeast China. There was a slight increase in soil total organic C (TOC), total N (TN), and a significant increase in microbial biomass C (MBC) and microbial biomass N (MBN) after restoration of 4 to 5 years for all the three species. Soil TOC, TN, MBC and MBN increased the most in R. pseudoacacia plantation than that in Z. jujube var. spinosa and U. pumila plantation. No significant differences in total P accretions were found among all treatments. These results indicated that soil quality was substantially improved during early phase of plantations establishment for all species. All the three species are promising for ecological restoration of mine spoils around magnesite mine regions, especially R. pseudoacacia has stronger impact on soil improvement.