Abstract:Addressing the limitations of LOF anomaly detection algorithm, such as with high time and space complexity and insensitivity to cross anomalies and outliers around low-density clusters, this paper proposes isolation-based data extracting LOF (iDELOF) anomaly detection algorithm, which puts the isolation-based K-nearest-neighbor search space extraction (iKSSE) in front of LOF, to efficiently cut out a large amount of useless and interfering data and obtain a more accurate search space. Based on this, the theoretical and four groups of experimental analysis are completed, and in each group of experiments, iDELOF is compared with many typical algorithms such as LOF, iForest and iNNE. The results show that iDELOF improves the detection capabilities of LOF by widening the gap between the local outlier factor of normal and abnormal points, and enhancing the ability to identify cross anomalies and abnormal points around low-density clusters.Additionally, iDELOF has the same obvious superiority as LOF in identifying axis-parallel anomalies. The data subset obtained by iDELOF through iKSSE is significantly smaller than the original dataset and the data volume of most subsets is less than 1% of the original dataset. Therefore, the time and space complexity of iDELOF is significantly reduced, and the larger the amount of data in the original dataset, the more obvious the superiority is. When the amount of data is large enough, the running time of iDELOF will be lower than that of the IF algorithm.

Abstract:Continuous detection of space precession cone targets using broadband radars can generate high-resolution range profile (HRRP) sequences. The HRRP sequences contain information such as spatial geometric information and precession laws of space cone targets, which are applicable to target association, tracking and classification. Therefore, it is of great significance to perform HRRP sequence prediction for spatial precession cone targets. The ConvLSTM network effectively combines the characteristics of CNN and LSTM, which can fully mine spatial and temporal information of HRRP sequences to achieve prediction of HRRP. This article establishes a HRRP sequence dataset based on the precession model of spatial cone targets, which incorporates different parameters such as size, motion speed and motion direction, and uses this dataset to design and implement a ConvLSTM network model suitable for HRRP prediction for spatial precession cone targets according to HRRP characteristics. In order to validate the predictions by the ConvLSTM network designed in this paper, the ConvLSTM network is compared with the two-dimensional convolutional neural network mode. Simulation and experimental results show that ConvLSTM network is in good agreement with HRRP calculated using physical optics method, and is more accurate than predictions of 2D convolutionneural network. The Pearson correlation coefficient is as high as 0.973 1, and average absolute error reaches 0.033 4. The ConvLSTM network model can effectively extract temporal and spatial features of HRRP sequences to achieve high-precision prediction of HRRP sequences.

Abstract:While the application of UAV images in target detection and tracking has been widely studied amid the popularization of UAV aerial photography technology, there has been relatively few researches on the matching between UAV images and satellite images. Due to their different imaging mechanisms and viewpoints, there are large-scale differences between UAV images and satellite images. Effective matching between UAV images and satellite images are difficult to realize using existing image matching methods. In order to solve this problem, this paper proposes a method for matching UAV images with large-scale differences from satellite images. This method registers the direction and scale of UAV images using the longitude and latitude coordinate of satellite images and the camera pose information of UAV images. Using the position of the UAV images, the satellite images are then roughly matched to obtain the satellite sub-images including the UAV images matching area. The CNN features of the registered UAV images and satellite sub-images are then extracted using neural networks, and the precise matching between UAV images and satellite images is realized based on CNN features. Simulation results show that the proposed method can effectively match large-scale UAV images with satellite images. The comparison in terms of matching accuracy between this matching method and existing image matching algorithms show the effectiveness and superiority of this matching algorithm.

Abstract:In order to address the issues of low efficiency and large scale of test sequence suite on the extended finite state machine (EFSM) model, a method of small-scale EFSM test sequence suite generation for all transitions was proposed based on the improved adaptive multi-population genetic algorithm (IAMGA). Firstly, the fitness function was designed by the transition coverage gain so that each generated feasible transition path can improve transition coverage. Next, the individual was selected within the subpopulations divided by feasible transition paths, which can overcome the premature convergence and improve the success rate of all transitions coverage. Then, the probabilities of crossover and mutation were adaptively adjusted based on the average path pass rate of the population so as to speed up convergence. Finally, by traversing the test sequence suite in reverse order, redundant sequences were removed, which further reduced the size of the test sequence suite. The experimental results show that compared with the generation test sequence method for single transition, this method which is designed for all transitions, can generate a test sequence suite that meets all transitions coverage with a success rate of more than 90% at a time. Compared with the traditional genetic algorithm, the average size of test sequence suite generated by IAMGA is reduced by 50%, while the average number of iterations is reduced by 20%. The method proposed in this paper can effectively improve the efficiency and quality of the generating EFSM test sequence suite.

Abstract:The offshore wind turbine tower has been used in the marine corrosive environment for a long time. The tower structure in marine corrosive environment is prone to local buckling and collapse under the impact of wind, wave, earthquake and other reciprocating loads. In this paper, carbon fiber reinforced polymer (CFRP) is proposed for local reinforcement of the corroded wind turbine tower. The single shear test and numerical simulation study of CFRP steel composite structure are carried out. On the basis of simple shear test, the hysteretic performance of corroded wind turbine tower strengthened with CFRP is studied using ABAQUS software and python. Structural damages and energy consumption mechanisms under four corrosion conditions and six reinforcement conditions are compared. Results show that the high temperature environment softens the contact between CFRP and steel, which reduces the bearing capacity of CFRP steel composite structures. The fiber fracture of CFRP steel composite structure occurs earlier than the damage of adhesive layer under tensile force. The contact relationship of CFRP steel can be simplified as the “Tie” in numerical simulation. Under the cyclic load, the corroded tower structure shows early buckling, the proportion of material plastic energy consumption decreases and the proportion of “buckling hinge” energy consumption increases. After CFRP strengthening, the buckling of the structure can be delayed, the plastic region of the material can be expanded and the overall energy dissipation capacity of the structure can be improved. In order to prevent the sudden damage of the structure when the number of CFRP strengthening layers is small, attention should be paid to the structural ductility change during strengthening.

Abstract:This paper introduces a prefabricated concrete-filled double-skin steel tubular (CFDST) joint designed for easy repair following a disaster. To investigate the influence of endplate thickness, axial compression ratio, bolt diameter, concrete filling degree and shape of inner steel tube on the seismic performance of the joints, quasi-static tests were performed on six scaled 1∶2 joint specimens. The bearing capacity, ductility, stiffness degradation, strength degradation, and energy dissipation capacity of the joints were analyzed. The failure modes of joints included endplate bending, flanges buckling, weld cracking between endplate and flange, bolt warpage, and bolt fracture. The load-displacement hysteretic curves were plump, indicating the joint has good energy dissipation capacity. The displacement ductility coefficients of joints were greater than 4.89, indicating that the joint has good plastic deformation ability and ductility. The strength degradation coefficients were basically maintained at 0.9-1.0, showing good bearing capacity stability. Increasing the endplate thickness significantly improved the seismic performance of the joint. Increasing the axial compression ratio, the concrete filling degree, and replacing the inner circular steel tube with a square steel tube increased the bearing capacity of the joint, and changing the bolt diameter had little effect on the bearing capacity. Changing the axial compression ratio and the bolt diameter had little effect on the energy dissipation of the joint. Increasing the concrete filling degree and replacing the inner circular steel tube with a square steel tube noticeably decreased the energy dissipation of the joint. The failure modes and bearing capacity of the joint obtained by the established nonlinear finite element model were in good agreement with the test results.

Abstract:In order to present a restoring force model and a determination method of the parameters for steel reinforced concrete self-centering coupled wall panels, the study investigated the failure modes and working mechanism of the coupled wall panels under lateral loads utilizing cyclic loading tests. The failure process of the wall panels included the rocking of wall panels and the shear yielding of the dampers. The failure of wall panels was localized in the dampers, which demonstrated excellent resiliency. On basis of hysteretic characteristics of the wall panels, the restoring force models were proposed, which consisted of skeleton rules, unloading rules and reloading rules. Ten parameters affecting restoring force model were then proposed. The skeleton rules were defined by three stiffness parameters and three strength parameters. The unloading rules and reloading rules were defined by four displacement parameters. Methods for the determination of stiffness parameters, strength parameters and displacement parameters affecting the restoring force model were proposed in light of the cooperative working mechanism among coupled wall panels, dampers, prestressed tendons and framed beams. The comparison between test results and simulation results shows that the test hysteresis curves are in good agreement with those of the numerical simulation, and the test energy dissipation is basically the same as that of the simulation. It is then concluded that the proposed restoring force models do well in describing the hysteretic features of the steel reinforced concrete coupled wall panels and that the methods for determining the parameters have satisfactory analytical accuracy.

Abstract:To address defects of the traditional method for predicting the strip outlet crown of hot tandem rolling, which suffers from low accuracy and poor interpretability, a model for combined prediction of hot strip crowns based on mechanism and data driving is proposed. The strip crown reference value is obtained using the strip crown mechanism prediction model. The deviation between the reference value and the actual value is used as the prediction variable of machine learning models, and then the deviation prediction value and the reference value are summed to obtain the strip crown prediction value of the combined prediction model. This combined prediction strategy is verified using multiple neural networks. It is found that the proposed strip crown combined prediction model has better prediction performance than the traditional model, with over 97% of the predicted data having an absolute error of less than 0.02 mm and more than 82% of the predicted data showing an absolute error of less than 0.01 mm. Additionally, the model is both satisfactorily feasible and widely applicable. The proposed model integrates the relative strengths of the mechanism model and the data-driven model, resulting in a representation that is more closely aligned with the actual physical phenomena. The combined model not only alleviates the problems of poor interpretation and low reliability with the results from the black-box neural network prediction, but also compensates for the defects of the mechanism model, which often produces results that deviate from the production conditions and cannot be adjusted in real time. This proposed model makes a significant contribution to the shape control of hot strip and the improvement of hot strip product quality.

Abstract:Seawater sea sand concrete (SSC) has a wide range of applications in the construction of islands and coastal projects. In marine environments, concrete is prone to cracking, which can significantly affect the durability of structures. In order to ensure the safe service of this new type of concrete in the marine environment, it is crucial to study the fracture mechanical properties of SSC and to determine the fracture parameters reasonably. However, the size effect is inevitably present due to the neglect of material inhomogeneity when using the determined fracture parameters based on traditional linear elastic fracture model. To address this issue, this article aims to determine the size-independent fracture parameters of SSC, utilizing a non-linear fracture theory based on boundary effect model, taking into account the material discontinuity and heterogeneity. In this paper, SSC with the maximum aggregate sizes of 10 and 20 mm were prepared. Three-point bending tests of with heights of 100 and 200 mm were carried out respectively, and the initial crack length-to-beam depth ratios were set from 0.1 to 0.7 in each group. Additionally, fresh water and river sand were used to replace seawater and sea sand in order to prepare ordinary Portland concrete (OPC) as the control group for the experiments. Based on the boundary effect model and by incorporating the average aggregate size of concrete, the size-dependent tensile strength f_t, fracture toughness K_IC of SSC can be obtained analytically from small and medium-sized specimens. Furthermore, the means, upper and lower limits of two fracture parameters with 95% reliability were determined based on the normal distribution analysis. The ultimate load of SSC specimens under any size condition was successfully predicted by using obtained tensile strength. Moreover, the results show that under the same aggregate gradation, compared with ordinary Portland concrete, SSC exhibits a higher proportion of aggregate fracture on the fracture surface and demonstrates higher tensile strength and fracture toughness. With the increase in the maximum aggregate size, the proportion of aggregate fracture in both SSC and OPC decreased. However, the fracture toughness K_IC increased, while f_t decreased. The above model and related results can provide references for the practical engineering design of seawater sea sand sea sand concrete.

Abstract:Aiming at the conflict of the stationary and rapidness of the terminal mode switching when the variable speed control moment gyro is applied as the actuator on the agile remote sensing satellite for attitude maneuver, on the basis of considering the frame speed error, the attitude error parameter is designed as the switching index, the transition rule within the error parameter switching area is formulated, and the command torque is assigned to CMG and flywheel in real time and solved respectively. A control torque gyroscope/reaction flywheel fuzzy smooth switching control law is proposed. In order to make the attitude of the satellite at the end of attitude maneuver reach the requirements of attitude stability and pointing accuracy in a shorter time, the clustering variation improved particle swarm optimization algorithm is proposed to optimize the parameters of the control law and determine the best switching region and switching parameters. Finally, the simulation results show that the improved particle swarm optimization algorithm always shows better fitness than the traditional particle swarm optimization algorithm in the same iteration times, with faster convergence speed and higher convergence accuracy. Compared with the existing control law, the fuzzy smooth switching control law after parameter optimization can complete the smooth switching of dual modes in a shorter time. At the end of attitude maneuver, the requirements of attitude stability and pointing accuracy can be reached more quickly, and the control performance of agile maneuver and high stable pointing of remote sensing satellites can be improved, which raises the quality of imaging mission.

Abstract:In order to verify the applicability of the stochastic finite fault method in simulating ground motions caused by subduction slab earthquakes, the 2021 Chiba, Japan M_j 6.1 subduction slab earthquake was taken as an example, and a total of 25 sets of surface and borehole station records within a range of 100 km around the epicenter were obtained from KiK-net and simulated by the stochastic finite fault ground motion method. The ground motion characteristics such as spectrum, duration, peak value, and spatial distribution of simulated and observed records were analyzed. Results show that the simulated and observed response spectra of pseudo-spectral acceleration (A_ps) with 5% damping ratio were well matched in the band range of 0.1–10 Hz. The simulated records of duration model based on 70% energy duration were consistent with the observed records in the strong motion section. The simulated and observed peak ground acceleration (A_pg) from the surface stations were in good agreement, and the A_pg attenuation characteristics were basically the same. The A_pg contours based on the simulated records were very similar to the observed A_pg contours. In addition, the simulated results and observed records were compared with the results of commonly used ground motion prediction equations (Zhao16) for subduction slab earthquakes in Japan. Results show that the A_pg prediction of Zhao16 was generally overestimated, and the A_ps prediction was underestimated and overestimated to some extent at low and high frequencies respectively, which may be caused by the basin effect and soft soil layer in the study area. The research results can provide basis for the applicability of the stochastic finite fault method in ground motions caused by subduction slab earthquakes, and will offer further reference for exploring the application of the method to areas with similar tectonics in China.

Abstract:The monitoring data of permanent magnet synchronous motor (PMSM) exhibit complexities such as non-smoothness, non-linearity, multi-source heterogeneity and low value density. These characteristics make it challenging to accurately model the type and extent of motor faults using simulation data. The serious imbalance between normal and faulty data samples leads to problems such as overfitting and low accuracy in the training of fault diagnosis models. In this paper, an improved auxiliary classification generation adversarial network (ACGAN) is proposed to study the expansion of real fault data for PMSM by learning the distribution characteristics of the original samples, while the generated fault dataset provides a data base for the next fault diagnosis and health assessment. Firstly, to address the problems of poor convergence and the tendency for gradients to disappear or explode in ACGAN networks, the Wasserstein distance is used to constrain the reconstruction loss of the generated data, and the gradient penalty is used instead of weight clipping to optimize the model and mitigate model training instability. Secondly, to analyze the change relationship between data and the historical change pattern, recurrent neural network is introduced in the generator to improve the quality of the generated data. Finally, the effectiveness of four data expansion methods, ROS, SMOTE, ADASYN and improved ACGAN, is compared and analyzed in improving the performance of fault diagnosis models using fault data from PMSM inter-turn short circuits. Results show that the model trained using the improved ACGAN method is more stable, converges faster and produces expanded data of superior quality than those adopting other data expansion methods.

Abstract:In the field of UAV smoke detection, due to the significant variations in different detection scenes, existing smoke detection algorithms often suffer from issues such as low detection accuracy and slow speed. To address these issues, in this paper we constructed a new UAV smoke dataset (USD) in multiple scenes, and proposed an improved YOLOx UAV smoke detection algorithm in multiple scenes. Firstly, we introduced an improved attention module into the YOLOx network to improve the extraction process of channel features and spatial features respectively, which can extract more representational smoke features. Then, we presented a two-way fusion network to enhance the fusion ability of multi-scale feature fusion module for small smoke target features. Finally, we utilized a Focal-EIOU loss function to address the issues such as the imbalance of positive and negative samples in the training process, and the distance and coincidence degree of two frames cannot be reflected when the prediction frame and real frame do not intersect. Experimental results show that the proposed algorithm has good robustness when applied to UAV smoke detection tasks in multiple scenarios. Compared with several classical smoke detection algorithms, the accuracy of the proposed smoke detection method on different data sets has been improved respectively. For instance, compared with the original YOLOx-s model, the accuracy was improved by 2.7%, the recall rate was improved by 3%, and the speed reached 73.6 frames per second.

Abstract:To better understand the mechanism of impact of windbreak wall wake on the galloping of catenary positive feeder, this study establishes an analysis model for wind-induced vibration response of positive feeder in light of the aerodynamic theory, focusing on the positive feeder of LanzhouUrumqi high speed railway. The fluid-structure interaction method is used to analyze the time history of two-dimensional models with different natural freguency rations and degrees of freedom. The results show that both the degree of freedom and the frequency ratio significantly impact the galloping amplitude of the positive feeder. A lower frequency ratio amplifies the impact of wind speed on the amplitude of the positive feeder and extends the range of wind speeds triggering the galloping of the positive feeder. The galloping amplitude of the positive feeder in the vertical single-degree-of-freedom system is greater than that in the vertical-horizontal two-degree-of-freedom system, indicating that the horizontal vibration of the positive feeder has a certain limiting effect on the vertical vibration. When the positive feeder vibrates in the wake of windbreak wall, the windward angle of the positive feeder is constantly changing. A larger windward angle is more likely to induce substantial galloping of the positive feeder. The galloping mode of the positive feeder of the catenary in the strong wind section is determined as Den Hartog galloping without icing. The research results offer a deeper understanding of the galloping mechanism of the positive feeder under the condition without icing in the strong wind section, providing theoretical insights into mitigating and controlling the galloping of the positive feeder of the catenary.

Abstract:To realize nondestructive testing of ancient wood components, we proposed linear prediction formulas for wood density and compressive strength based on combined testing technology. The ultrasonic wave velocity and compressive strength of 12 kinds of wood in different grain directions, as well as the impedance ratio in radial and oblique directions, were measured using the ultrasonic-needle resistance instrument combined detection technology. The regression coefficients in the prediction formulas were fitted, and then the prediction formulas were applied to the prediction of the density and compressive strength of two kinds of new wood and a kind of ancient wood. Results show that the ultrasonic wave velocity and impedance ratio of wood were positively correlated with wood density and compressive strength, and the goodness of fit of the linear relationship obtained by multiple regression method was significantly improved when the ultrasonic-needle resistance instrument combined detection technology was adopted. When predicting the density of new and ancient wood, only needle resistance instrument technology achieved accurate prediction. The prediction error of new wood was less than 5%, and that of ancient wood was less than 1%. When predicting the compressive strength of new and ancient wood, the prediction error of single detection technology was large, while the prediction effect of combined detection method was better. The prediction error of new wood was less than 5%, and that of ancient wood was less than 9%. The tests verify the feasibility of using ultrasonic needle resistance instrument combined detection technology to predict wood density and compressive strength. The results can provide important technical support for the health monitoring of ancient building wood components.