Abstract:In order to quantitatively characterize the mixing and segregation phenomenon of municipal solid waste (MSW) layer incinerated on a mechanical grate, an MSW layer model containing three different sizes of par? ticles is established based on the discrete element method, and the mixing index and segregation index are intro? duced to characterize the mixing intensity and segregation intensity. Furthermore, the influence of movable grate pro? cess parameters (movement amplitude and reciprocating frequency) on the mixing intensity and segregation speed is analyzed based on the discrete element method. The results show that the mixing of the MSW layer mainly occurs along the layer’s advancing direction as well as the height of the layer, and a binary linear relationship between mix? ing intensity and process parameters is found. The segregation of the MSW layer is related to the component size. The larger the size of the component, the more concentrated in the top area of the material layer; and a binary nonlinear re? lationship between segregation speed and process parameters is found. The conclusion above provides an effective design reference for the efficient incineration of MSW and the optimization of incineration equipment.
Abstract:The traditional numerical simulation of fluid mechanics does not consider the interaction between the elastic structure of the car body and the external flow field，which makes the results inconsistent with the real situa? tion. Taking a real vehicle model as the research object，the CFD simulation considering the effect of fluid-structure interaction is carried out，and the results of aerodynamic drag，aerodynamic lift，flow field structure and other as? pects are compared with the traditional fluid numerical simulation results. The results show that the fluid-solid cou? pling effect has a greater impact on the aerodynamic lift，and the difference between the two simulation methods can reach 38% with the increase of vehicle speed，which is directly related to vehicle stability and safety. Using fluidstructure coupling CFD numerical simulation to explore the characteristics of wind-induced vibration of the whole ve? hicle，it is proved that the main influencing factors of the actual vehicle vibration amplitude are the wind-induced vi? bration frequency and the magnitude of the force. Moreover，the stiffness of the vehicle elastic structure is optimized to improve the wind-induced vibration of the vehicle，so as to improve the comfort of passengers.
Abstract:In order to reduce the effect of reduction of the magnetron energy dissipation capacity of the magneto? rheological damper under a micro-amplitude，medium and high frequency excitation，this paper proposes a method to increase the working pressure by increasing the initial working pressure of the magnetorheological fluid in order to improve the magnetorheological damper energy dissipation capability under a medium and high frequency excitation. Based on compressible characteristics of magnetorheological fluids，a mechanical model of magnetorheological damper was constructed. The effect of working pressure on the damping force at medium and high frequency excitation was analyzed. A single rod linear liquid spring-type magnetorheological damper without a compensation device was designed and processed，and the accuracy of the mechanical model was verified through experimental tests. The experimental results under sinusoidal excitation showed that：increasing the working pressure of the magnetorheologi? cal damper can make the damping curve much fuller. When the working pressure is higher，the energy consumption capacity of the magnetorheological damper is significantly better than that of the magnetorheological damper when the working pressure is lower. According to the experimental results，it is also found that compared with the initial pressure of standard atmospheric pressure，when the initial pressure is 5 MPa，the maximum damping is increased up to 31.3%. The maximum increase in energy consumption is 78.5%. In addition，by increasing the working pres? sure，the magnetic field in the working area is more likely to saturate. The equivalent stiffness and equivalent damp? ing coefficient are increased with the increase of the working pressure.
Abstract:The existing nine-axis five-linkage CNC slow-feed grinder has many motion links, which may cause many problems during the working process such as insufficient rigidity, deformation of castings like lathe bed and in? accurate machining accuracy. In this work, the maximum deformation of the grinder and the first-order natural fre? quency under the modal analysis are taken as the main optimization goal and the overall quality as the secondary goal. ANSYS Workbench is used to analyze and optimize the grinder, and the original structure is then improved by the sensitivity analysis, finally obtaining the new design scheme for each key part. When the quality of the whole grinder is improved to a certain extent, the maximum static deformation and first-order natural frequency of the grinder is greatly improved. Besides, the extreme learning machine network model (Extreme Learning Machine) opti? mized by genetic algorithm is combined with genetic algorithm to optimize some main parameters of the improved structure of the nine-axis five-linkage grinder. Firstly, the whole machine mass, maximum static deformation and first-order natural frequency of grinding machine are transformed into comprehensive target grey correlation degree via grey correlation analysis. Then, the network model of the extreme learning machine (Genetic Algorithm-Extreme Learning Machine) optimized by the genetic algorithm is used to fit the nonlinear coupling relationship between the main parameters of the grinder and the gray correlation degree of the comprehensive target. Finally, GA′s powerful optimization ability is used to find the optimal process parameters in the trained GA-ELM network model. After opti? mization, the overall quality of the nine-axis five-linkage grinding machine, the reduction of the maximum static de? formation and the first-order natural frequency are optimized when compared with the improved scheme. This method provides a certain theoretical support and reference value for the subsequent technicians to optimize the structure and parameters of the grinder.
Abstract:The detailed chemical reaction kinetics mechanisms of iso-octane and n-heptane were simplified by using the methods of reaction path analysis and sensitivity analysis. With adopting the semi-decoupling methodol? ogy，the small molecule mechanism of C0~C1 was used as the core of primary reference fuel（PRF）for gasoline，a new reduced mechanism of PRF which consists of 41 species and 131 reactions was developed，and then the devel? oped mechanism was verified by the experimental data of shock tube（ST），laminar flame speed（LFS）and jet stirred reactor（JSR）. The results show that the reduced PRF chemical reaction mechanism developed in this paper can not only accurately predict the ignition delay time，but also effectively describe the laminar flame velocity and the evolution of key species.
Abstract:A distributed model predictive control（DMPC）based two-subsystem approach is proposed in this paper to resolve the coupled surge velocity and depth control problem of multiple Autonomous Underwater Vehicles （AUVs）. For the motion coordination of multiple AUVs，each AUV shares information with its neighbors to solve lo? cal predictive optimization problems under the framework of DMPC. For the motion coordination of the surge velocity and depth control subsystems of a single AUV，the predicted surge velocity from the surge velocity control subsystem is approximately lent to the depth control subsystem，of which the predicted pitch angular velocity is sent back to de? couple the dynamics and solve the predictive optimization problem. The simulation results show that compared with the coupling model approach，the proposed two-subsystem approach sacrifices some rate of convergence in depth but reduces the calculation time by 73.3%.
Abstract:In order to solve the problem that the number of classification boundaries of convex hull region formed by polyhedral cone classifier is limited and it is unable to scale in different scales，a rotated and extended polyhedral conic classifier（REPCC）is proposed by adding a rotation factor based on norm vectorization. REPCC in? creases the number of classification boundaries of the convex hull region，and the classification boundaries can be adaptively scaled in each dimension，which can better fit the positive region and improve the classification accuracy. Experimental verification is carried out on two different rolling bearing datasets. The results show that REPCC has better classification accuracy，robustness and generalization ability，and can accurately identify the working state and fault type of rolling bearing. REPCC can be used for intelligent fault diagnosis of rolling bearing.
Abstract:Aiming at the problem of redundant features when using deep learning to extract facial expression im? age features，an improved Xception facial expression recognition network based on multi-layer perceptron（MLP）is proposed. In this model，the features extracted from the Xception network are input into the multi-layer perceptron for weighting，the main features are extracted，and the redundant features are filtered out so that the recognition accu? racy is improved. First，the image is scaled to 48*48，then the data set is enhanced，and these processed images are fed into the network model proposed in this paper. A comparison of ablation experiments show that：The correct rec? ognition rates of this model on the CK + dataset，JAFFE dataset，and MMI dataset are 98.991%，99.02% and 80.339% respectively. The correct recognition rates of Xception model on the CK + dataset，JAFFE dataset and MMI dataset are 97.4829%，90.476%，and 74.0678%，respectively. The correct recognition rates of the Xception + 2lay model on the CK + dataset，JAFFE dataset and MMI dataset are 98.04% and 74.0678%，84.06%，and 75.593%，re? spectively. By comparing the above ablation experiments，the recognition accuracy of this method is significantly bet? ter than the Xception model and the Xception + 2lay model. Compared with other models，the effectiveness of this model is also verified.
Abstract:In order to achieve the civil aviation safety management goal of‘safety first，prevention first and com? prehensive management’，a deep learning model is established to learn from reports and assess the risk level. Based on the 10-year incident reports available in the Aviation Safety Reporting System，we first establish quantitative indi? cators of incident consequences and classify all incidents into 5 levels according to their severity：high，moderately high，moderate，moderately low and low risk，which helps to eliminate the impact of unbalanced and intricate event consequences. Then，the relationship between the unstructured incident synopsis and the risk level is explored by convolutional neural network（CNN），and the events are classified by the model to determine the risk level. The clas? sification model proves its superiority by comparing it with different quantitative indicators and methods，with an ac? curacy of 96%，which is better than the compared models. Finally，the 2020’s incident reports are predicted by this model，which enables rapid risk assessment of the synopsis of the incident，with an accuracy rate of 80%. The CNNbased civil aviation risk assessment model can fully mine the text-formatted incident synopsis，and quickly assess and actively perceive the risk level，which helps support the early warning of civil aviation safety.
Abstract:Because of the lack of scale calibration on the analysis results of information system business impact analysis as well as the influence of expert evaluation preference on it, an analysis method of system business affecting impact based on crossover probability theory is proposed. First, the experts evaluate the relevance and influence of the business functions of the information system, and the correlation between business functions is represented by a cross-impact matrix. Then, the subjective and objective weight combination method is used to weight the matrix to re? duce the subjective impact of experts scoring on the cross-impact matrix, quantify the correlation between system business functions, and generate a comprehensive cross- impact matrix. Finally, the preference chain generation algo? rithm is used to generate the business preference chain of the information system, and the interaction relationship of each business of the system is associated. On this basis, the influence trend of other businesses after the business in? terruption is obtained by analyzing the position of the interrupted business in the preference chain. Experimental re? sults show that the proposed method can accurately measure the impact and trend of business function interruption on other business functions of information system.
Abstract:Automatically extracting unknown drug-drug interactions from biomedical literature can update the drug database quickly，which is of great importance and medical value in application. Existing neural network mod? els often can only learn a single one-sided feature in a certain aspect from sentence sequences or other external infor? mation，but it is difficult to fully mine the potential long-distance dependency features from sentences to obtain a comprehensive feature representation. This paper proposes a novel drug-drug interaction extraction method combin? ing semantics and dependency. In this method，we not only use the Bi-GRU network to learn the semantic feature representation from the sentence sequence and the shortest dependency path of the target drug entities，but also combine the multi-head self-attention mechanism to further capture the potential dependencies between words. Finally， these multi-source features are fully fused to effectively improve the performance of drug-drug interaction extraction. The experimental results on the DDIExtraction-2013 dataset show that our method outperforms other existing meth? ods and obtains an F1 value of 75.82%.
Abstract:There are some problems with traditional deep reinforcement learning in solving autonomous obstacle avoidance and target tracking tasks for unmanned aerial vehicles（UAV），such as low training efficiency and weak adaptability to variable environments. To overcome these problems，this paper designs an internal and external metaparameter update rule by incorporating Model-Agnostic Meta-Learning（MAML）into Deep Deterministic Policy Gradient（DDPG）algorithm and proposes a Meta-Deep Deterministic Policy Gradient（Meta-DDPG）algorithm inovder to improve the convergence speed and generalization ability of the model. Furthermore，the basic meta-task sets are constructed in the model’s pre-training stage to improve the efficiency of pre-training in practical engineer? ing. Finally，the proposed algorithm is simulated and verified in Various testing environments. The results show that the introduction of the basic meta-task sets can make the model’s pre-training more efficient，Meta-DDPG algo? rithm has better convergence characteristics and environmental adaptability when compared with the DDPG algo? rithm. Furthermore，the meta-learning and the basic meta-task sets are universal to deterministic policy reinforce? ment learning.
Abstract:Functional verification can find defects quickly at a low cost in the early stage of chip design，which is of great significance to ensure design quality. In view of the problems of poor instruction sequence randomnes， cumbersome testcase writing and the verification platform’s poor reusability in core module-level verification，an 8- bit RIS core verification reference model is designed. The model satisfies the verification requirements of the instruc? tion sequences’random combination and reusability through independent modeling of the instruction set and configu? rable parameter design，and the use of automated scripts solves related problem above. The reference model equipped with the UVM platform is applied to verify an 8-bit MCU core with a RISC architecture. The results show that the UVM platform integrated with the designed reference model has good robustness and reusability，the number of defects converges quickly，and the verification cycle is shorter. The code coverage and function coverage of the core module reach 100%.
Abstract:In the expert recommendation algorithm of the Q & A community，the graph neural network mainly uses the interactive relationship between users and questions to build a model，and its model performance depends on the density of interactive data. So it is difficult to effectively represent and learn users and questions without inter? active information. This paper proposes an attention graph neural network expert recommendation method based on memory. Firstly，a multi-dimensional feature-oriented subnetwork is designed，and then a memory network is con? structed to store the similar questions answered by users for each question. At the same time，an attention mechanism is introduced between user representation and similar question representation，and the vector representation of new questions is constructed by fusing similar questions from different users′ perspectives，Finally，experts are recom? mended based on the representation of users and questions，which effectively improves the accuracy of expert recom? mendation. The proposed method is validated on the Q & A community data set and public data set，and its perfor? mance is improved compared with other similar models.
Abstract:To solve the problem of edge blurring and color distortion of license plate images taken in foggy weather，an end-to-end depth multilevel wavelet U-Net based algorithm for license plate fog image removal is pre? sented. Taking MWCNN as the main frame work of the defogging network，the feature information in the wavelet do? main is integrated using the“SOS”enhancement strategy and the cross-layer connection between the codec. The pixel-channel joint attention block of the discrete wavelet transform is used to reduce the fog residue in the defrosted license plate image.In addition，the cross-scale aggregation enhancement blocks are used to supplement the missing spatial domain image information in the wavelet domain image，which further improves the quality of the defogging li? cense plate image.The simulation results show that the network has obvious advantages in structural similarity and peak signal-to-noise ratio，and it performs well in dealing with the composite plate fog image and the actual photo? graphed plate fog image.
Abstract:This paper proposes a global uniform and local continuity repair method for mural image inpainting. It uses the relationship between linear system and image repair to construct the similarity-preserving overcomplete dictionary with global weighted feature. Meanwhile, a novel sparse repair model with elastic net regularization based on similarity-preserving overcomplete dictionary is formulated to enhance the global feature consistency, and then an estimated method of neighborhood similarity is presented to guarantee local feature consistency, finally, a global fea? ture patch and local feature patch weighted method are applied to obtain the target patch. Experimental results on damaged murals demonstrate the proposed method outperforms state-of-the-art inpainting methods.
Abstract:The anti-tumor peptide M1-21 monomer was used to create nanoparticles（Nano-M1-21，about 50 nm）by the desolution method. The characteristics of Nano-M1-21 were determined by the dynamic optical disper? sion（DLS）and the low pressure transmission electron microscopy（TEM）. The stability of Nano-M1-21 particles was measured in vitro at 37 degrees centigrade. To verify the anti-tumor effects of Nano-M1-21，we used breast can? cer MDA-MB-231，MCF-7 and 4T1 cells to show that Nano-M1-21 was easily absorbed by the cells and its tumor inhibitory concentration was lower than that of monomer peptide. With a mouse breast cancer 4T1 cell-grafting tumor model，we found that Nano-M1-21 showed a good suppressive effect on the tumors. In conclusion，compared to the monomer peptide M1-21，the optimized Nano-M1-21 nanoparticles provide benefits to improve the cellular uptake. and anti-tumor effects
Abstract:Phosphorus doped PdNi nanoparticles were prepared by one-pot synthesis method，which was care? fully characterized by transmission electron microscope，scanning electron microscopy and X-ray photoelectron spectroscopy methods，showing an ultrafine particle size with a diameter of（4.4 ±1.0）nm. A cyclic voltammetry study demonstrated this material exhibited an excellent electrochemical property with an electrochemically active sur? face area（ECSA）of Pd at 42.11 m2 /g，which was 1.5 times higher than its commercial Pd/C counterparts. Pd/NiP nanoparticles showed a remarkable electrocatalytic dechlorination activity，in which direct electron transfer and indi? rect atomic hydrogen mediated electron transfer two pathways simultaneously occurred during the dehalogenation pro? cess. Meanwhile，voltammetric analyses showed that the electron transfer coefficient（k）was much smaller than 0.5， suggesting the dissociative electron transfer（DET）to the C-Cl bond of 4-CP following the concerted DET mecha? nism. Results indicated that the synergistic effect of trimetallic alloy and the compressive strain caused by the doping of Ni and P atoms markedly promoted the generation and retainment capacities of the adsorbed atomic hydrogen （H*ads）and the absorbed atomic hydrogen（H*abs），and effectively inhibited the hydrogen evolution reaction， thereby enhancing the electrocatalytic dechlorination efficiency. That was advantageous to improving the electrocata? lytic reduction dechlorination efficiency. Electrolyses of 2 mM 4-CP showed that for the material catalyzed by the Pd/ NiP nanoparticles（1.02 mg），87.5% of 4-CP could be completely dechlorinated within 360 min with an apparent rate constant of 0.0057 min-1，demonstrating an excellent dechlorination performance and high stability of the devel? oped material. This work provides a facile strategy for the synthesis of phosphorus-doped bimetallic nanoparticles， which provides an excellent catalyst and an important reference for the future electrocatalytic dechlorination works.
Abstract:In this paper，four kinds of MOFs catalysts were synthesized by hydrothermal method. The composi? tion and microstructure of the catalysts were characterized by XRD，SEM and IR，respectively. Then，the NRR ac? tivity of the catalysts was studied. The results showed that Fe-BTC had the highest ammonia yield and Faraday effi? ciency at -0.376 V（vs.RHE）and 80oC，with values of 3.63×10-10 mol s-1 cm-2 and 0.31%，respectively. Under constant voltage，when the alternating magnetic field intensity is 4.355 mT，frequency is 50 kHz，the catalyst ammonia production rate and efficiency of Faraday′s largest，were 3.61×10-9 mol s-1 cm-2 and 5.67%，respectively.This phe? nomenon is mainly due to the increase of the adsorption amount of N2 on the Fe-BTC surface by the alternating mag? netic field，and the superposition of the induced electromotive force generated by the alternating magnetic field and the electric potential of the electric field itself，which provides extra energy for the NRR reaction.
Abstract:In the present work，CoFe-B-P nanoparticles were prepared by solvothermal method and investi? gated their oxygen evolution reaction（OER）performance were investigated. The morphologic structures and composi? tion of the resultant samples were characterized by SEM，TEM，XRD，XPS，and ICP-OES. It was found when the mo? lar ratio of Co and Fe is 4，and the addition amount of NaH2PO2·H2O is 2.0 mmol，the resultant Co4Fe1-B-P catalyst exhibited the best OER performance with a low over-potential of 285 mV at a current density of 10 mA cm?2 and a small Tafel slope of 52.70 mV/dec，as well as excellent long-term stability in 1.0 M KOH after continue testing 20 h. The improved OER activity benefits from the electronic interaction between transition metal and nonmetal.
Abstract:The 2D conductive material MXene（Ti3C2Tx ），electroactive tannic acid and robust aramid nanofibers are composited and vacuum-filtered to prepare self-supported flexible films with layer structure. The microstruc? ture，morphology，mechanical and electrochemical performance of Ti3C2Tx /TA/ANF films are systematically investi? gated. The results show that Ti3C2Tx /TA/ANF film exhibits a high tensile strength of 36.2 MPa and excellent flexibil? ity，which can be arbitrarily bent，folded and twisted. The Ti3C2Tx /TA/ANF film assembled flexible solid-state super? capacitor possesses a volumetric specific capacitance of 826.56 F cm-3，a superior volumetric energy density of 28.7 Wh L-3 and electrochemical stability under various bending angles.
Abstract:The introduction of deformation before aging can change the aging precipitation process of the precipitation-strengthened Al-Cu-Mg（-Si）alloy，and this change is directly related to the amount of deformation. In order to study the deformation strengthening mechanism of multi-element alloys，it is necessary to use electron backscatter diffraction（EBSD），X-ray diffraction（XRD），transmission electron microscopy（TEM）and other methods to test the Al-3.0Cu-1.8 Mg-0.5Si（wt.%）alloy treated by different deformation and aging process，quanti? tatively or semi-quantitatively，so that the contribution of different strengthening methods to the strength is listed， which provides theoretical support for the application of thermo-mechanical treatment in aluminum alloys. It is found that the increase in deformation can increase the peak hardness and strength of the alloy due to three reasons. One is the increase of the grain-boundary area，the second is the increase of the dislocation density，and the third is that the size of the precipitates is smaller and the distribution is more uniform. Furthermore，the contribution of pre? cipitates to the strength is getting higher with a larger deformation prior to the aging process，when the amount of de? formation is greater than 6%.