HOHAI UNIVERSITY

The present application relates to the technical field of power system fault detection, discloses a high-speed location method for high impedance fault in distribution network based on directional frequency scanning. The high-speed location method includes, intercepting a zero-sequence current signal I 0j collected by each of monitoring points of the distribution network and a bus zero-sequence voltage signal U0 of the distribution network, determining the fault occurrence time tf, extracting a first natural mode function to directly reconstruct a original fault signal, projecting a first natural mode function signal to the reference orthogonal vector group respectively, distinguishing the upstream and downstream of a fault point by a clustering algorithm. The method of constructing virtual reference direction by orthogonal vector group does not need to install voltage transformers at all measuring points, thus reducing hardware investment and maintenance costs, reducing dependence on physical hardware and greatly improving fault detection speed.

A cable tension calculation method simultaneously considering the sag, inclination angle and bending stiffness includes: querying basic parameters of a stay cable according to design data and construction data; considering influences of the sag, the inclination angle θ and the bending stiffness EI, to calculate dimensionless parameters γ, ε and λ 2 ; testing an acceleration response of the stay cable by an acceleration sensor, to identify a frequency ω of the acceleration response of the stay cable, further calculating a dimensionless frequency {circumflex over (ω)} of the stay cable and the dimensionless parameters γ, ε and λ 2 , and substituting the dimensionless frequency {circumflex over (ω)} of the stay cable into a vibration characteristic equation, to establish a function relation between the dimensionless frequency {circumflex over (ω)} and a cable tension H* of the stay cable; and solving a root of the vibration characteristic equation, and identifying the cable tension H* of the stay cable according to the root.

A discrimination method for soft rock deformation based on harmful deformation index is provided, which includes calculating a comprehensive strength stress ratio; determining an activity coefficient of groundwater; calculating a hydraulic coefficient of the a rock mass; determining a shape coefficient of a tunnel based on its cross-sectional shape; determining a long-term strength reduction coefficient; considering the above factors, calculating a harmful deformation index of the soft rock tunnel. Based on obtained harmful deformation index, compare it with a harmful deformation classification table of the soft rock tunnel to determine a deformation category of a surrounding rock. The present disclosure comprehensively considers physical and mechanical properties, geological conditions, hydrological environment, and time effects of the surrounding rock of the soft rock tunnel, overcoming shortcomings of traditional methods that ignore dynamic factors and environmental influences, thereby making deformation identification more comprehensive and accurate, and effectively improving risk prediction capabilities.

Disclosed by the present disclosure is a method for obtaining key parameters for pollutant transportations under coupling effects of temperature and hydrodynamics based on a site scale. The method is to obtain the permeability parameters, thermal conductivity, and dispersion coefficient of the contaminated sites based on multi field coupling numerical simulations of hydrodynamic field-temperature field-concentration field. The transportation parameters for pollutants on the site are key indicators for the investigations and treatments of organic pollution on the site. The method as illustrated in the present disclosure includes establishing a water-thermal-salt coupling model for organic contaminated sites according to site drilling data and site physical test results in the early stage, performing multiple field coupling numerical inversion according to the site model in combination with multi-party data, adjusting the model parameters through manual trial and error method to improve model accuracy and obtaining the optimal model.

An anti-cavitation damping composite metal structure for a flow passage component and a preparation method thereof are provided. The preparation method includes: cladding a gradient functional material layer by layer on a substrate of a flow passage component; forming periodic structures on a surface of each layer of the gradient functional material through etching by an ultrafast laser to absorb a part of an impact load energy caused by cavitation of the flow passage component, where the layers of the gradient functional material form a gradient coating with a toughness increasing layer by layer and a hardness decreasing layer by layer from bottom to top; and forming nano twins on a surface layer by a laser shock peening technique, and implanting a residual compressive stress to further improve anti-cavitation resistance of a surface.

Disclosed is a three-stage modular convolutional neural network (CNN), referred to as Stairnet, for efficient classification of concrete cracks in images. Unlike conventional CNNs that employ repetitive layer structures, Stairnet is organized into three distinct stages—Stair1, Stair2, and Stair3—each with unique architectural characteristics. Stair1 is composed solely of convolution layers; Stair2 introduces a higher layer count and incorporates channel-split and shuffle operations; and Stair3 utilizes larger expansion factors and kernel sizes along with attention mechanisms. Variations in Stair1 and Stair2 enable structural adaptability based on model hyperparameters. Compared to general-purpose CNNs designed for thousands of classes, Stairnet achieves a compact model size, faster training, and high accuracy in concrete crack classification.

A method for calibrating parameters of a hydrology forecasting model based on a deep reinforcement learning includes selecting according to basin characteristics, and the parameters and parameter value ranges to be calibrated by the model are determined. The method includes a calibrating the parameters of the hydrology forecasting model is established, and three elements of the reinforcement learning, that is, a state space, an action space and a reward function are determined. The method includes a deep reinforcement learning method DQN is applied to optimize the parameters to be calibrated by the hydrology forecasting model. In the present disclosure, by setting a stride length of the action value for the deep reinforcement learning model, an accuracy finally optimized by the calibration parameters can be freely controlled, and a DQN algorithm is adopted to search the entire space for the calibration parameters to ensure the optimality for optimizing the calibrated parameters.

The present disclosure provides a method for numerical simulation of reactive transport during CO 2 +O 2 in-situ leaching of uranium at a sandstone-type uranium deposit. Unlike the traditional method for numerical simulation of solute transport during in-situ leaching of uranium with consideration of only convection and diffusion, the method permits establishment of a multi-field coupled reactive solute transport model to simulate the dynamic leaching process of a sandstone-type uranium deposit in Northern China. The method provided in the present disclosure includes: creating a thermodynamic database suitable for CO 2 +O 2 leaching of a sandstone-type uranium deposit in Northern China, and with consideration of the dynamic reaction process of uranium dissolution under combined action of oxygen O 2 (aq) and bicarbonate HCO 3 − , performing numerical simulation of reactive transport during CO 2 +O 2 in-situ leaching of uranium using a TOUGHREACT simulation technology framework.

An intelligent method for efficiently classifying concrete cracks from large amounts of image data is proposed, named inverted residual (IR) 7-Efficient Channel Attention and Convolutional Block Attention Module (EC) network. The IR7-EC network consists of a convolutional layer, seven inverted residual-ECA structures, a CBAM attention mechanism, a pooling layer, and multiple fully connected layers that are sequentially connected. The inverted residual-ECA structure consists of two components: a depthwise separable convolution-based inverted residual structure and an ECA attention mechanism. The new inverted residual structure facilitates the feature extraction of concrete cracks. Compared to conventional network structures like VGG and Resnet, the proposed IR7-EC network excels in both accuracy and efficiency. Once the IR7-EC network is fully trained, it can accurately classify various types of concrete cracks in captured images. This method offers several advantages, including a small number of network parameters, fast training convergence speed, and precise classification results.

The present disclosure provides a method, a system, and a medium for maintaining bridge structures susceptible to damage triggered by ship impact, and relates to the field of bridge health monitoring technology. If ship-bridge collision damage occurs, time-frequency coherence degradation factors at different positions of a time-series matrix, allows determination of an impact damage position and a damage degree based on a convex maximum point in the trajectory curve of the time-frequency coherence degradation factors. This method can effectively separate the influence of environmental factors and provide a fundamental methodological basis for establishing ship-bridge collision damage assessment technology.