Abstract:

Abstract: In view of the effect of polymer grouting in repairing the dam face disengaging, the ground penetrating radar wave field of the dam face disengaging was studied. A calculation models of the dam with panel disengaging repaired by polymer grouting was established based on the finite-difference time-domain method and the perfectly matched layers boundary conditions. The effects of radar center frequency, degree of panel disengaging repair, size of disengaging area, face thickness and reinforcement on ground penetrating radar (GPR) wave field characteristics of the dam face disengaging repaired by polymer grouting were analyzed. The results showed that the resolution of GPR profile increased gradually with the increase of the excitation source center frequency. The horizontal interfacial reflection wave generated in the GPR profiles increased with the length of the disengaging area. The time interval between horizontal reflectors on the GPR profiles increased with the depth of the disengaging area. The amplitudes of the bypassed and diffracted waves in the disengaging repair area decreased with the increase of the dam face thickness. The electromagnetic waves emitted by the GPR encountered the steel reinforcement and generated a wave field. The reflected waves at the upper and lower interfaces of the disengaging area were divided by the strong bypass waves, which made it difficult to judge the horizontal length of the reflected waves.

Abstract: This study was carried out to understand the current status of forest carbon sequestration and carbon sink value in the middle and lower reaches of the Yellow River, and to contribute to the ecological protection and high-quality development of the Yellow River Basin. Based on the data of national forest inventory and land survey, the principle of accumulation conversion method and the improved continuous function method of biomass conversion factor were used to measure the dynamic changes of forest carbon sequestration and carbon sink value in Henan Province. The results showed that the forest carbon sequestration in Henan Province increased from 1.76×108 t to 2.78×108 t from 2008 to 2018, and the forest type with higher annual carbon sequestration was mixed broadleaf forest and quercus. The distribution of carbon sequestration by age class was the largest in young forests and the smallest in over mature forests. The spatial pattern of forest carbon sequestration was "high in the west and south, low in the east and north", mainly distributed in Nanyang, Luoyang and Sanmenxia. The value of forest carbon sinks increased from 49.247 billion yuan in 2008 to 77.868 billion yuan in 2018, an average annual increase of 2.862 billion yuan. The carbon sink value of mixed broadleaf forest, quercus and populus accounted for 79.93% of the carbon aggregation value in 2018. Finally, suggestions on carbon sequestration capacity and realization of forest carbon sink value were made, for the high-quality development in the Yellow River Basin.

Abstract: In order to solve the problem of accurate identification of arch bridge boom damage, in this study a simplified mechanical model of an under-bearing tied arch bridge was established and the analytic formula of boom strain influence line of under-bearing tied arch bridge was obtained by force method derivation. Based on that, the damage identification method boom was proposed based on quasi-static strain influence line index of under-bearing tied arch bridge. Then the applicability of the method for conventional boom number arch bridge was verified with the help of finite element method. And the influence of test noise, damage location, damage degree, and damage category on the damage assessment results was studied by using finite element model calculations. A scientific implementation plan for vehicle loading was proposed. The results showed that within 10% test noise, the quasi-static strain-influence line difference curvature method could accurately locate the local damage location of arch bridge booms and quantitatively assess their damage degree. The method still had a good recognition effect when other structures, such as tie beams, wind braces, and arch ribs, were damaged.

Abstract: Based on piecewise-linear method, a one-dimensional electroosmotic dewatering model of mud was proposed in order to further explore the influence factors of mud electroosmotic dehydration.In this model, the combined influence of potential gradient and void ratio on the electroosmotic coefficient of mud and the nonlinear stress-strain relationship and large deformation effect of mud were considered. Compared with the results of relevant analytical solutions and laboratory model tests, the error of this model was less than 5%. On this basis, the influences of loading voltage,compression index and electroosmotic coefficient on the process of mud electroosmosis dehydration were analyzed.The results showed that the increase of loading voltage and electroosmotic coefficient could improve the final water removal and shorten the stabilization time. With the increase of mud compressibility, the final dewatering volume of mud increased, but the stabilization time was prolonged.

Abstract: Studies on the long-term aerosol optical depth ( AOD) based on the large-scale Yellow River basin was limited, and most of them only focused on meteorological conditions. In this study, the MODIS aerosol optical depth ( AOD) products was collected, and then the temporal and spatial pattern of AOD and comprehensively quantified the impact of geographical environment, natural weather and social economy on AOD in the Yellow River Basin were analyzed based on the geographically weighted regression (GWR) . The results showed that the AOD exhibited a downward trend in the Yellow River Basin. The AOD value decreased from 0. 38 in 2001 to 0. 22 in 2022. Moreover, the distribution of AOD also showed obvious seasonal differences that AOD values in spring and summer were higher than in autumn and winter. This wight be the result of a combination of factors such as temperature, atmospheric diffusion conditions, and vegetation cover. From the perspective of spatial distribution, the AOD in the study area gradually increased from west to east. This trend was opposite to the distribution of DEM in the Yellow River Basin, indicating a close correlation between terrain and aerosols. The analysis of influencing factors based on GWR model showed that, for the entire Yellow River Basin, terrain and vegetation had the greatest impact on AOD in the Yellow River Basin, followed by socio-economic factors and natural meteorology. Prominent cities in the Yellow River Basin were also analyzed in the study, and the results showed that the inter-annual variation of AOD in different cities in the study area was quite different. The AOD values of Xining, Yinchuan and Baotou in the upper reaches of the Yellow River Basin showed a low level, with the highest value appearing in winter and the lowest value appearing in summer, while the AOD values of cities in the middle and lower reaches were the highest in summer and the lowest in winter.

Abstract: To tackle the problem of low accuracy of detection and recognition for object in complex scenes, YOLOv5 object detection and recognition algorithm based on attention and multistage feature fusion(AMFF) was proposed in this study. The main ideas included adding the proposed dual space directions pyramid split attention (DSD-PSA) mechanism to the backbone network of the traditional YOLOv5s model to enhance the learning of the feature map space and channel information, adopting multistage feature fusion(MFF) structure in the bottleneck network to fuse the features of different branches, increasing richness of the feature and improving the ability to cope with complex scenes. In addition, C3Ghost module and depthwise separable convolution were used to replace C3 module and common convolution to reduce the number of parameters and the complexity of network. Compared with the traditional YOLOv5s algorithm, the mean average accuracy of the proposed algorithm in the VOC2007+2012 data set reached 85%, and the mean average accuracy of the smart retail cabinet commodity identification data set reached 97.2%, which verified the effectiveness and feasibility of the proposed algorithm.

Abstract: Previous surveys of urban resident travel were hindered by prolonged durations and insufficient granularity in traffic zone divisions, which impeded the timely and accurate acquisition of travel data. To address this issue, this study proposed a method for extracting the dynamic origin-destination ( OD) matrix of urban residents based on mobile phone signaling data. Firstly, methods to address two complex types of noise inherent in the signaling data: ping-pong switching data and drifting data were proposed. Specifically, a window thresholdbased detection and equivalent location replacement method for ping-pong switching data was proposed, as well as a complex drift point detection and marking method for drifting data. Secondly, an enhanced ST-DBSCAN clustering algorithm was proposed, which incorporated a temporal isochronization method to integrate temporal and spatial information, enabling the identification of dwell points during travel. Finally, a road network with key nodes was established using geographic information system ( GIS) , aligning resident travel OD with the network nodes to effectively derive the dynamic OD matrix of urban residents. Experimental results showed that the enhanced STDBSCAN clustering algorithm outperformed the traditional ST-DBSCAN, improving clustering efficiency by 6. 10% and identification speed by 5. 26%. Furthermore, the dynamic OD matrix extraction method based on the enhanced ST-DBSCAN clustering algorithm achieved approximately 16. 98% and 21. 55% reductions in mean squared error compared to the conventional statistical methods and the second-order statistical methods, respectively. By applying the proposed dynamic OD matrix extraction method to the case of Beijing, this study was able to conduct timely and effective analyses of daily and peak travel patterns of urban residents.

Abstract: Aiming at the problems of long paths, low accuracy and prone to local optima of the artificial fish swarm algorithm in robot path planning, an improved artificial fish swarm algorithm was proposed, which aimed to improve the efficiency and accuracy of the algorithm. An improved artificial fish swarm algorithm aimed at improving algorithm efficiency and accuracy was proposed in this study. Firstly, an optimization cycle was added to the algorithm′s foraging behavior to reduce the randomness of the algorithm′s selection of location points in path planning, enabling the robot to move towards the target point faster. Then, the tabu search algorithm was integrated, and the tabu table was introduced to record the path where the algorithm might fall into the local optimum, so that the algorithm can avoid the local optimum region when selecting new location points, and could avoid the algorithm′s local excessive cycle. At the same time, it could optimize the planned path, delete the paths between duplicate grid points, and ensure that there would be no duplicate grid points in the path. When the improved artificial fish swarm algorithm was applied to a new type of 3D raster map, simulation experiments showed that compared to other comparative algorithms, the average path length obtained by improving the artificial fish swarm algorithm in maps 1, 2 and 3 was reduced by 10%, 15% and 30%, respectively, and the success rate of path planning in complex maps was increased by 75%.

Abstract: A complex road traffic object detection algorithm was proposed to address the issue of traffic target detection algorithms′ inability to resist complex background interference and insufficient detection performance in the current autonomous driving scenario. At first, the multi-head self-attention residual module (MHSARM) was used to improve the feature information of the target to be inspected while decreasing the complex background interference. Secondly, in the feature fusion area, CoordConv was used instead of traditional Conv, so that the network could perceive spatial information and improve network detection accuracy. The improved YOLOv5s algorithm had stronger feature extraction ability and good generalisation ability in complex roads, and mAP_0.5 reached 93.3% and 47.4%, respectively, which was higher than that of YOLOv5s 0.9% and 1.4%. In addition, compared with the latest target detection algorithms YOLOv7 and YOLOv8, the mAP_0.5 of improved YOLOv5s improved by 1.3% and 2.2%, respectively. Compared with the latest research results of Sim-YOLOv4 algorithm on Kitti dataset, mAP_0.5 improved 2.2%.

Abstract: To address to the low feature information, low detection rates, and high false rate and missing rate in the target detection task, a Tr-SSD algorithm based on multiscale feature fusion and a hybrid attention mechanism was proposed. Firstly, a Resnet50 residual network was utilized as the backbone network for the SSD algorithm to enhance its feature extraction capabilities. Secondly, a hybrid attention mechanism was designed and applied to the mid-scale feature maps of the network to enhance effective information within the feature maps and establish longrange dependencies between pieces of information. Finally, a FPN (feature pyramid network) structure was formed by using network layers centered around the Transformer instead of the original backbone network in the SSD algorithm, which fused feature information of different scales to more accurately locate small targets. Experimental results showed that the Tr-SSD algorithm achieved mAP values of 81. 9%, 87. 5%, and 88. 4% on the PASCAL VOC dataset, HRSID dataset, and RSOD remote sensing dataset, respectively. This represented an improvement of 4. 7 percentage points, 6. 8 percentage points, and 9. 2 percentage points compared to the original SSD algorithm. Moreover, the detection speed could meet the requirements for real-time detection.

Abstract: In the traditional K-means and many improved algorithms, the inability to explicitly handle outliers, resulted in their poor clustering performance. To solve this problem, in this paper, an efficient K-means with region segment and outlier detection was proposed. Firstly, to obtain better clustering results, an unified clustering model to form an interactive collaboration between outlier detection and clustering was constructed. Secondly, to improve algorithm efficiency, clusters were adaptively segmented through near neighbor clusters search to reduce redundant calculations. Finally, on synthetic datasets and real datasets were tested to verify the effectiveness of the proposed method. The experimental results showed that EK-means algorithm outperformed other algorithms in terms of clustering performance and execution efficiency. The ACC could reach 0. 911 in the Wine dataset.

Abstract: The commonly used deep learning methods based on BERT pre-trained model in geological named entity recognition were character-based approaches, and could not utilize word-level information. Additionally, the dropout mechanism in neural networks might cause inconsistency between the training and inference stage. To address this issue, a geological named entity recognition model MBCR based on MacBERT and R-Drop was proposed. Firstly, MacBERT was used to learn text feature representations, which could fully utilize character and word information. Then, BiGRU was employed to encode context features, effectively extracting complete semantic information. Subsequently, CRF was adopted to capture dependencies between labels and generate the optimal label sequence. Moreover, R-Drop was introduced during the training process to further enhance the model′s generalization capabilities. Compared with BiLSTM-CRF, BERT-BiLSTM-CRF, and other models, the proposed MBCR model improved the F1-score on the NERdata dataset by 2. 08-4. 62 percentage points and on the Boson dataset by 1. 26-17. 54 percentage points.

Abstract: In order to improve the modulation speed of seven segment two-level SVPWM algorithm and reduce the use of logic resources, a hardware architecture of SVPWM based on FPGA was proposed. After inputting the reference voltage, the hardware architecture first carried out the coordinate transformation based on the inverse Clarke transform, constructed three groups of intermediate variables containing three-phase duty cycle through a series of addition operations, and obtained the simplified 2 bit sector judgment conditions from the above hardware wiring through two XOR operations. Then, according to the simplified 2 bit sector judgment conditions, the three-phase duty cycle was selected from the above three groups of intermediate variables, and clamp protection was carried out, and PWM was output according to the natural sampling method. The above process formed a whole. The whole process from reference voltage input to three-phase PWM output had been completed in two clock cycles with only three triggers in FPGA, which effectively improved the calculation speed. In addition, the resource usage of the hardware architecture with different FPGA platforms was also given. Compared with other methods, the LUT usage was reduced from at least 500 to about 300, and the logical resource usage was reduced. The effectiveness of the proposed hardware architecture was verified by simulation and physical test.

Abstract: In order to further enhance boiling heat transfer,multiple low thermal conductive material plates were embedded near the upper wall of the solid heater,alternating temperature variations with space on the heating surface for boiling heat transfer were obtained in this study. The single-component multiphase lattice Boltzmann method was used to investigate the effects of the number and gap spacing of low thermal conductive material plates on boiling heat transfer performance and bubble dynamics, the mechanism of enhancing boiling heat transfer by adding low thermal conductive material plates was revealed from a microscopic perspective. The results indicated that the dynamic behavior of bubbles changed with the increase of gap spacing, leading to bubble merging, independent growth, and other bubble detachment processes. Based on the analysis of bubble dynamic behavior, temperature field, and flow field, it was found that bubbles first nucleated and grew at the gap heated surface,the vortex separated from bubbles could promote the lateral migration and merging process of growing bubbles on the heated surface. Morever, within a certain gap spacing range, bubbles would fuse with each other to form a liquid bridge,which could promote the evaporation of the micro layer around the root of the bubbles on the heated surface, and pushed the cold fluid to wet the gap heated surface again. The combined effects of adding multiple low thermal conductive material plates, heat accumulation at gaps, bubble merging to forma liquid bridge, re-wetting of gap heated surfaces, lateral migration of bubbles, and rapid merging of multiple bubbles could enhance boiling heat transfer performance.

Abstract: To solve the flexible job shop scheduling problem(FJSP), a hybrid Levy flight, reverse search, and parameter adaptive adjustment strategy improved beetle swarm optimization (LRA-BSO) was proposed based on the beetle antennae search algorithm which could simulate the foraging behavior of beetles in nature and the swarm intelligence optimization theory. Firstly, a FJSP model was established. Secondly, the initial population was generated based on the Tent chaotic mapping, which would improve the quality of the initial population. Then, the Levy flight strategy and reverse search strategy were used to improve the global search ability of the LRA-BSO algorithm, and the search step size and the search distance of the beetle swarm were adjusted through fitness feedback to avoid falling into local optimum. Finally, the optimization ability of the algorithm was validated through 6 multi-dimensional standard test functions. In addition, the applicability of the LRA-BSO algorithm in FJSP was verified by 10 standard test cases and 1 practical case. The test results showed that the algorithm performed better or equal to other intelligent optimization algorithms in eight standard test cases and demonstrated good optimization ability. In the practical cases, the improved algorithm had a 48% improvement in convergence speed compared to the original beetle swarm optimization algorithm.

Abstract: In order to solve the problem of poor estimation accuracy and even divergence coused by the covariance matrix of state prediction error in iterative computation of forecasting-aided state estimators, in this study, a robust forecasting-aided state estimation for power systems based on SRUPF (square root unscented particle filter) was proposed. Two mathematical methods, matrix QR decomposition and matrix Cholesky factor update were adopted, and square root technology were introduced to dynamically update the state covariance matrix, thereby maintaining the positive definiteness of the state prediction error covariance matrix. The results of testing using MATLAB showed that in the non Gaussian noise testing of IEEE 30 systems, the average root mean square error of the SRUPF voltage phase angle was 0. 09% of the corresponding test value of UPF, and the average root mean square error of the SRUPF voltage amplitude was 0. 14% of the corresponding test value of UPF. In the IEEE 57 system non Gaussian noise test, the average root mean square error of the SRUPF voltage phase angle was 0. 67% of the corresponding test value of the UPF, and the average root mean square error of the SRUPF voltage amplitude was 0. 57% of the corresponding test value of the UPF. The SRUPF proposed in this paper had a good effect on solving the problem of non positive of the covariance matrix of state prediction errors in auxiliary predictive state estimation, with high estimation accuracy and robustness.

Abstract: Aiming at the problems of poor dynamic performance of traditional VSG technology and difficulty to determine the optimal values of important parameters J and D, a VSG control and parameter optimization strategy based on droop control and neural network prediction was proposed to realize dynamic adjustment of key parameters J and D in VSG technology. The proposed strategy applied the active power-frequency droop control to the control algorithm of VSG. Then, simulated the rotor motion equation and the voltage and reactive power control characteristics of synchronous generator, the small signal analysis model of VSG was established, and the initial setting of key parameters rotational inertia and damping coefficient were completed. Finally, an artificial neural network was established for analysis learning and network training, and the weight was adjusted to change the VSG moment of inertia and damping coefficient. The error between the output and the input was compared by the error function, and the parameter reached the expected value after multiple learning and training. The neural network optimization algorithm was combined with the droop control strategy to optimize the VSG control strategy. Traditional VSG control, constant parameter droop control and adaptive parameter droop control based on neural network optimization were used to simulate a numerical example, and the results showed that, compared with traditional VSG control, the proposed adaptive parameter droop control based on neural network optimization reduced the maximum frequency variation by 26.7%, and the frequency stabilization time by 0.25 s. The strategy was effective.

Abstract: The activated carbon derived from furfural residue using steam activation was investigated for the adsorption 4,4′-thiodiphenol(TDP) and bisphenol F(BPF) from aqueous solution. Adsorption conditions including adsorption time, FRAC dosage, pH value, temperature and initial concentration were discussed. The results showed that adsorption equilibrium data of TDP and BPF onto FRAC were well described by the Sips and Koble-Corrigan isotherm models. Thermodynamic parameters revealed that the adsorption process of TDP and BPF on FRAC was spontaneous and exothermic process. The adsorption kinetics process of TDP and BPF conformed to the pseudo-second-order kinetic model. Besides, the adsorption of TDP and BPF on FRAC were mainly influenced by the hydrogen bonding, hydrophobic effect, electrostatic interaction and π-π interaction. At 298 K, the maximum adsorption capacities of FRAC for TDP and BPF were 5.408 3 mmol/g and 3.695 5 mmol/g, respectively, implying that the FRAC had a good application in endocrine disruptors wastewater treatment.