Abstract:

Abstract: The "defining the scales based on water" harmonious coexistence is an emerging water-resources management concept, emphasizes the thorough implementation of the "defining the scales based on water" principles and the concept of "harmonious coexistence" with the rigid constraints of water resources. The concept′s background was systematically analyzed, and based on clarifying the principles of "defining the scales based on water" and the concept of "harmonious coexistence", its definition was established. The deeper connotation was examined from the perspective of interactions between human systems and water systems. The concept was based on three fundamental principles: the regional water balance principle, human-nature symbiotic co-prosperity principle, and human-water relationship harmonious evolution principle. Its main methodological approach included system identification, simulation analysis, metric evaluation, and optimization control. The core concepts were articulated across five dimensions: coordinated development, dynamic regulation, system-level synergy, supply-demand balance, and water-adapted development. Based on these elements, a theoretical framework for the concept of harmonious coexistence of "defining the scales based on water" was established, and its theoretical foundations were systematically elaborated. Furthermore, the concept′s application pathways and prospects were explored from multiple dimensions, including planning, policy, management, strategy, technology and discipline. The study provided a theoretical basis and practical reference for the refined management of water resources and the coordinated development of multiple systems.

Abstract: In response to the unclear mechanism of precipitation variation in the Sanmenxia Reservoir area of the Yellow River Basin and the difficulty of traditional trend analysis methods in capturing internal structural changes in the sequence, innovative trend analysis (ITA), innovative polygon trend analysis (IPTA), and innovative trend pivot analysis method (ITPAM) were comprehensively applied to systematically analyze the multi-scale variation characteristics of the precipitation sequence at the Sanmenxia weather station from 1957 to 2020. And they were compared with traditional trend testing methods such as the Mann-Kendall trend test. The results showed that both the annual precipitation at the Sanmenxia weather station and the high-value category showed a significant downward trend, with a statistic S=-1.031. The overall change in the monthly average precipitation over many years was relatively uniform, but from July to September, a compound pattern of "trend reduction-large variation-high risk level" emerged. The maximum both 1 d and 5 d precipitation showed a significant downward trend, but the precipitation intensity significantly increased (S=0.069), indicating a transformation of the regional precipitation pattern towards "low frequency high intensity". Compared with traditional methods, the innovative trend analysis methods could better identify and detaily analyze different time scales, different value zones, and different risk levels of time series, and had greater flexibility advantages.

Abstract: In response to the significant differences in the spatiotemporal distribution of flash flood disasters and the uncertainty of influencing factors in the evaluation process, which can easily cause distortion of evaluation index values and loss of risk index information, evaluation indices were divided into loss of life, economic loss, environmental impact, and social impact. Their grading standards were subdivided, and quantified. The variable weight theory was used to modify the combination of constant weight, and the attribute measurement comprehensive evaluation vector was used to transform and analyze the uncertainty information. Combined with the attribute set pair coefficient, the severity of disaster consequences was judged, and a risk consequence evaluation model based on attribute-set pair coupling was constructed. The model was applied to the evaluation of flash flood disasters in the 20 a, 50 a and 100 a return rainstorm scenarios in Guxian County, Shanxi Province. The results showed that the model could effectively quantify the risk consequence grades under rainstorm scenarios with different return periods, which was basically consistent with the evaluation grades calculated by other models, demonstrating the good applicability of the proposed method.

Abstract: The insulating gloves worn by power personnel in substations were small target in size and were easily obscured. Aiming at the problem that general feature fusion networks often lost small target information, a multiscale small target feature fusion network named STPFM was constructed. The RT-DETR-R18 model was improved, and the RT-GLV model was designed for detecting whether power personnel were wearing insulating gloves. Firstly, the STPFM network was used to replace the CCFM network. The SSFF module and TFE module of the network were utilized to fuse multi-scale feature information. In addition, a small target detection layer with the SSFF module as the core was added to enhance the model′s ability to learn small target information. Secondly, to address the issue of excessive model parameters after replacing the STPFM network, a lightweight PB_Block module was constructed. Only the modules in the P4 and P5 layers of the Backbone network, which contained less small target information, were replaced. It not only lightened the model but also reduced the loss of small target information. Finally, the PIoUv2 loss function was adopted to enhance the model′s learning ability for both easy and difficult samples. The experimental results showed that the RT-GLV model performed excellently in the detection of whether power personnel were wearing insulating gloves. Compared with the RT-DETR-R18, the mAP@0.5 was increased by 2.1 percentage points, the F1 score was increased by 1.6 percentage points, and the number of model parameters was reduced by 21.5%. In terms of small target detection, the AP@0.5 of wearing insulating gloves was increased by 1.4 percentage points, and the AP@0.5 of not wearing insulating gloves was increased by 6.4 percentage points. Moreover, the model′s detection speed reached 91.3 frame per second, meeting the requirements of accuracy and realtime performance for detecting whether power personnel were wearing insulating gloves.

Abstract: Aiming at the limitation in existing studies focused on the detection of substation local structures, such as lacking methods for rapid discovery and dynamic monitoring over large areas, the capability of identifying potential safety hazards in power grids was enhanced through high-resolution satellite imagery. Firstly, a substation object detection dataset based on high-resolution optical satellite imagery was constructed. Subsequently, an improved YOLOv8 algorithm was proposed, embedding the SimAM lightweight attention module into the backbone network to enhance the ability to focus on detailed features, and replacing the neck with an Efficient-RepGFPN, combined with a DySample dynamic upsampling module to design a novel neck named GDFPN, addressing issues of multilevel feature semantic misalignment. Experimental results demonstrated that the improved method outperformed mainstream detection algorithms, with mAP75 and mAP50-95 increasing to 96.8% and 87.1%, respectively, confirming its superiority in substation detection tasks. The improved YOLOv8 approach proposed could effectively support the rapid discovery and dynamic monitoring of substations over large areas, providing reliable technical support for the safety management of power grids.

Abstract: To address the issue of slow production speed of steel bridge plate units, which directly constrained the construction period of bridge engineering projects, a distributed flexible job-shop group scheduling problem with setup & transportation time (DFJGSPST) model for steel bridge plate unit processing was established to minimize the maximum completion time while considering the processing technology route and production characteristics. A memory-based genetic algorithm with tabu search (MGATS) based on a three-layer encoding strategy was proposed to solve the model. To verify the feasibility of the mathematical model and intelligent algorithm, a DFJGSPST model comprising four types of plate unit groups and fifteen machines was established using a real-world steel bridge plate unit production case. Relevant test instances were selected for experimental validation and comparative analysis with other intelligent algorithms. Experimental results showed that the proposed MGATS algorithm achieved a mean relative percentage difference (RPD) of 2.74%, which was lower than that of the genetic algorithm (GA) at 3.99%, and hybrid genetic tabu search algorithm (GATS) at 3.13%. The success rate (SR) of the MGATS algorithm was 0.15, which was higher than that of the GATS algorithm and the GA algorithm, which verified the stability and robustness of the MGATS algorithm in solving the DFJGSPST model.

Abstract: For the hybrid flexible flowline problem with unrelated parallel machines at each stage, with constraints on deadline and transportation time, an integer programming model was established to minimize the total weighted completion time. A hybrid algorithm of artificial bee colony algorithm and whale optimization algorithm (ABCWOA) was proposed by combining improved genetic algorithm and neighborhood search strategy to obtain near optimal solutions. The algorithm utilized encoding based on job numbers and the NEH heuristic method to generate an initial set of job sequences. In the employed bee phase, an improved genetic algorithm was introduced to produce higher-quality job sequences. In the onlooker bee phase, five neighborhood search strategies were utilized to obtain better neighboring sequences. In the scout bee phase, a whale optimization algorithm based on the worst solution was designed to enhance the search capabilities of the algorithm. Simulation experiments were conducted to test the effectiveness of the improvements within the hybrid ABC-WOA algorithm, as well as to examine instances of varying sizes. The experimental results showed that the proposed hybrid algorithm performed very well.

Abstract: The traditional logic threshold-based ABS control method failed to fully utilize the road adhesion coefficient and caused significant slip rate fluctuations during operation. To address this issue, an automotive EMB antilock control strategy based on optimal slip rate estimation was proposed. The proposed strategy initially established a nonlinear model between tire slip rate and road utilization adhesion coefficient, and then employed a segmented estimation algorithm to rapidly and accurately track the optimal slip rate. Subsequently, based on the estimated optimal slip rate, an integral sliding mode controller was designed. By precisely adjusting the EMB braking torque and electric braking torque, the slip rates of the front and rear wheels were maintained at their respective optimal slip rates, ensuring optimal braking distances for automotives under various road conditions. Simulation results indicated that the employed estimation algorithm was capable of identifying the optimal slip rate of the current road surface rapidly and accurately, with the maximum error between the estimated optimal slip rate and the actual optimal slip rate at steady state not exceeding 3%. Furthermore, the integral sliding mode controller could precisely control the slip rate to remain near the optimal slip rate. Compared to the ABS control strategy built into CarSim, the total braking time in a single road condition scenario was shortened by 10.8%, and the total braking distance was reduced by 15.8%. For docking pavement conditions, the total braking time was shortened by 18.0%, and the total braking distance was reduced by 22.2%.

Abstract: To effectively address the issues of noise interference and insufficient multi-scale information within 3D U-Net for protein binding site prediction, a novel model named AMPocket was proposed which incorporated both attention mechanisms and multi-scale information to improve the accuracy of binding site prediction. AMPocket initially employed squeezed attention mechanism that enabled the model to focus on the most critical channels of protein features while diminishing the impact of irrelevant channels, thereby enhancing segmentation accuracy. Additionally, the multi-scale information was introduced to the encoder component, allowing the model to capture feature representations at various levels and thus obtained more comprehensive and robust spatial information. The experimental results demonstrated that AMPocket achieved superior predictive performance across four widely used test sets, in particular, the DCC success rate and DVO metrics on the SC6K dataset outperformed all other competing methods by 93.04% and 55.01% respectively, with a smaller number of parameters. It indicated that the model had better predictive performance.

Abstract: To overcome the high computational complexity of EEG-based emotion recognition methods based on feature extraction or time-frequency representations, a multi-task learning-driven method for emotion recognition based on time series imaging (TSI) was proposed. EEG signals were directly transformed into two-dimensional images using Gramian angular field, Markov transition field, and motif difference field. Built upon the ResNet18 architecture, a multi-task feature fusion framework was designed to integrate features from the three imaging methods to enhance emotional feature representation. Experimental results showed that with the DEAP dataset, the proposed method achieved average classification accuracies of 96.51% and 97.22% for binary classification of Valence and Arousal, respectively, and with the AMIGOS dataset, the accuracies reached 98.59% and 99.64%. When extended to four-class and eight-class classification tasks, the proposed method achieved average accuracies of 91.06% and 87.43% with DEAP, and 97.41% and 89.84% with AMIGOS, respectively. These results demonstrated the robustness of the proposed method in EEG-based emotion recognition.

Abstract: In the auxiliary diagnosis studies of attention deficit hyperactivity disorder (ADHD), many ADHD classification methods suffer from the problem of model integration or lack of biological explanation. To address this, an ADHD classification model based on the binary hypothesis end-to-end deep learning was proposed. Within the binary hypothesis, amplitude of low-frequency fluctuation related to the limbic system was selected as input features. An attention module was incorporated to enable the network to focus on features with high classification contribution. The model adopted an end-to-end architecture, rather than the traditional deep learning and machine learning combined structure, and accomplished the task of detecting biomarkers, thus providing biological explanations. In leave-one-out cross-validation experiments on the ADHD-200 database, the average accuracy across four sub-databases reached 98.1%. Subsequently, statistical and analytical of ADHD biomarkers on the limbic system revealed ADHD biomarkers including the anterior cingulate and paracingulate gyri, right amygdala, olfactory cortex, and left amygdala. These results proved the rationality of the binary hypothesis end-to-end deep learning model.

Abstract: Hydraulic fracturing is the key technology for extracting geothermal energy, which could increase the heat production by improving the permeability of the reservoir rock. The hot dry rock thermal reservoir in Songliao Basin was the research taget, and a two-dimensional numerical model of hydraulic fracturing was established based on ABAQUS software, and the sensitivity analysis of the parameters affecting the characteristics of hydraulic fracture was carried out by combining the orthogonal test method. The results showed that the discrepancy of the numerical simulation results compared with the laboratory test results was 2.6%, indicating that the model was accurate and reliable for studying hydraulic fracturing. The sensitivity analysis of each parameter through the extreme difference analysis method showed that the most influential factor on the fracture width was the elastic modulus of the rock, and the fracturing fluid displacement had the minimal impact. The most influential factor on the fracture initiation pressure was the horizontal stress difference coefficient, and the elastic modulus of the rock had a minimal impact. The results could provide certain guidance for hydraulic fracturing operations of the hot dry rock reservoir in the Songliao Basin.

Abstract: To prepare high-performance epoxy resin cementitious repair materials (ECRM) for effectively addressing dam crack rehabilitation, the influence of the content of epoxy resin, nano-SiO2, and steel-PVA hybrid fiber on the bonding properties of cementitious repair materials was analyzed by the interface flexural bonding strength test. The strengthening mechanism of the bonding properties of cementitious repair materials was revealed by the scanning electron microscope test. The results showed that the interface flexural bonding strength of cementitious repair materials increased first, and then decreased with the increase of epoxy resin, steel fiber, and PVA fiber content. When the content of epoxy resin was 9% (mass fraction, the same below), the volume content of PVA fiber was 0.9% (volume fraction, the same below), and the volume content of steel fiber was 1.2%, the interface flexural bonding strength of cementitious repair materials reached the maximum, which was an increase of 68.2% compared to the control group (without epoxy resin, nano-SiO2, PVA fibers, and steel fibers). As the content of nano-SiO2 increased from 0% to 2.0%, the interface flexural bonding strength of cementitious repair materials gradually rosed with an increase of 14.7%. Compared to nano-SiO2, steel fiber, or PVA fiber, the addition of epoxy resin had a more significant effect on the improvement of the bonding properties of cementitious repair materials. The microscopic strengthening mechanism of cementitious repair materials could be concluded as follows, the addition of epoxy resin and steel-PVA hybrid fiber could inhibit the formation and expansion of cracks in the matrix and improve the integrity of the matrix. The addition of nano-SiO2 could reduce the hole defects in the matrix and improve the compactness of the matrix.

Abstract: In a certain subway construction project in Suzhou, the cement soil mixed piles in the open-cut excavation area of the river channel experienced excessive deformation during the initial excavation, resulting in the failure of the waterproof curtain. Even after treatment with double-liquid grouting combined with MJS reinforcement, it remained ineffective in stopping the flow of sand and water. By using liquid nitrogen artificial ground freezing to form an effective waterproof curtain, the repair construction of the failed waterproof curtain on the condition of large seepage was successfully realized. An analysis of the failure of the foundation pit′s waterproof curtain was conducted, and a construction plan for liquid nitrogen freezing repair was proposed. Using on-site measurements, the study statistically evaluated the temperature of the frozen wall, the consumption of liquid nitrogen, and the growth rate of the freezing wall during the repair construction process. The results showed that the growth rate of the freezing wall at the leakage site of the cement-based improved soil was 67.3 mm/d, which was 58.6% of the average expansion rate of the frozen wall at the non-seepage area, and the temperature difference between the frozen soil at the seepage point and the non-seepage area exceeded 40 ℃ , which showed a clearly inhibitory effect on the expansion of the frozen wall. During the active freezing period, each cubic meter of soil required 1.671×103 kg of liquid nitrogen, and each set of freezing pipes consumed 3.49×103 kg of liquid nitrogen daily during the maintenance freezing period, which was 48.5% of the empirical estimate.

Abstract: To accurately consider the effect of particle breakage on the establishment of constitutive models for sand, one approach directly introduced the particle breakage ratio into a nonlinear critical state line (CSL) in the e-ln p plane, while another approach indirectly reconstructed the critical state line in a different scale space, using a linear or other simple functional relationship. The implementation and performance of the two kinds of CSL were compared by incorporating the modified SANISAND model. Using the return mapping algorithm based on cutting plane, model simulations of the drained and undrained tests on Toyoura sand were carried out. It was found that the CSL directly incorporating particle breakage ratio was only suitable for modelling sand with partial initial states, i.e., experimental simulations with small changes in initial perimeter pressure or pore ratio, while for sand with large changes in initial states, the critical state line with indirect consideration of particle crushing was more suitable.

Abstract: The complex topographic features of the reservoir area, especially the common narrow sections, have a significant impact on the flood evolution process. To reveal the influence mechanism of narrow sections and other natural special terrains on the evolution of dam failure floods, a study was conducted on terrain refinement and dam failure calculation methods. Taking a reservoir project as an example, Civil 3D and HEC-RAS software were used to refine the narrow sections of the upstream reservoir area and the downstream main channel in the DEM data. An improved calculation method was proposed, defining the reservoir area as a two-dimensional flow region, compensating for the limitations of conventional methods in reflecting the real terrain of the upstream reservoir. It allowed for a more accurate simulation of the impact of actual terrain on the evolution of dam failure floods. The calculation results showed that the flood arrival time was delayed by an average of 1 hour, and the maximum inundation depth of dam failure flood peak was reduced by 49.52%. It was evident that narrow sections could play a significant role in peak shifting and peak cutting, increasing the emergency evacuation time for downstream residents and effectively reducing inundation risk. The promotion and application of the improved method could help optimize the design of evacuation schemes and improve the economic and scientific basis for flood control decision-making.

Abstract: A short-term power load forecasting model based on secondary decomposition and temporal convolutional networks was proposed in response to the high complexity and strong fluctuation of transformer load data in the station area. Firstly, the maximum information coefficient method was used to extract features from the high-dimensional load dataset. Secondly, complete ensemble empirical mode decomposition with adaptive noise and optimized variational mode decomposition were employed to perform secondary decomposition on the transformer load data. Then, the sub-sequences obtained from the two decompositions were input into the temporal convolutional network model for prediction. Finally, the prediction results of each sub-sequence were superimposed to obtain the final load forecasting result. Simulation analysis was conducted on the load data of a distribution transformer in a certain district of Zhengzhou City. Compared with the traditional time convolutional network model, the proposed model reduced MAE, MAPE, and RMSE by 64.29%, 9.66 percentage points, and 59.00% respectively. The experimental results showed that the proposed combined forecasting model had better forecasting effects and higher prediction accuracy.

Abstract: To address the key challenges in control strategy and capacity configuration design for flywheel energy storage (FES) coupled with thermal power unit (TPU) in primary frequency regulation, a coordinated optimization of primary frequency regulation control strategy and capacity configuration for TPU-FES coupled system considered state of charge (SOC) management was proposed. Firstly, a dynamic control strategy was developed for the TPUFES coupled system, which considered the SOC management of FES. Secondly, an economic evaluation model was established, which comprehensively considered primary frequency regulation performance, life-cycle cost, and economic benefit. On this basis, a coordinated optimization method for control strategy and capacity configuration was constructed and solved using the particle swarm optimization algorithm. Finally, a case study based on the typical daily frequency regulation data from a real 350 MW double-reheat TPU were conducted to validate the simulation results, along with the sensitivity analysis of key parameters. The results demonstrated that the proposed method significantly enhanced both frequency regulation performance and economic benefits. Specifically, the life-cycle economic benefit of the FES increased by 15.09%; the investment payback period was reduced by 13.14%; and the effectiveness of the proposed coordinated optimization method was proved.

Abstract: To address the challenge of automated surface defect detection (e.g., damage, stains, and defects from human violations) in the power IoT, a lightweight SSD detection algorithm for edge computing devices was proposed. The proposed algorithm aimed to achieved efficient detection through three key innovations. Firstly, a dense connection mechanism was introduced into the bottleneck structure of MobileNetV2 to enhance image feature representation dynamically. Secondly, a cross layer attention mechanism implicit feature pyramid network (CL-IFPN) based on No-Local attention mechanism was constructed, and its deep integration with MobileNetV2-SSD significantly improved small-defect detection. Finally, a feature fusion module was added to the convolutional layer, and the QFL function was used to boost prediction accuracy of defects at different sizes and the balance of positive and negative sample training. Experimental results showed that on the public dataset VOC2007, the proposed algorithm achieved a detection accuracy of 79.62% and a speed of 36 frames per second, outperforming similar algorithms. On the self-built power device defect dataset, the detection accuracy reached 95.19% and a speed of 24 frames per second, demonstrating the algorithm′s practicality in power device defect detection. The proposed algorithm offered an effective technical solution for intelligent operation and maintenance of power IoT devices in edge computing environments.