16 results about "Level set method" patented technology

The invention relates to the technical field of structure optimization design, and particularly discloses a box-type beam structure lightweight design method based on topological optimization, which comprises the following steps: constructing a multi-physics field coupled three-dimensional parameterized model, and performing cross-scale analysis on integrated structure rigidity, thermal deformation, vibration mode and fatigue life to obtain a three-dimensional parameterized model; a variable density method and level set method mixed topological optimization strategy is adopted, and an initial material distribution scheme meeting the structural function integrity is generated by modeling the relationship between density and stress strain; starting a quantum genetic algorithm and deep reinforcement learning fused intelligent optimization engine; according to the method, through the multi-physics field coupled three-dimensional parameterized model, a variable density method and level set method mixed topological optimization strategy is combined, the relation between the structural lightweight target and the additive manufacturing process is effectively balanced, multi-physics field analysis of thermal response, vibration, fatigue and the like is introduced, and the structural lightweight target is obtained. And it is guaranteed that the box-type beam meets the mechanical property and other functional requirements in actual application.

The radar target scattering center region segmentation method based on local density clustering first carries out Frost filtering to the original radar image to realize global noise suppression of the image; secondly, the image segmentation is carried out to the Frost filtered image based on the level set method, and the suspected target region of interest is preliminarily obtained; then, the area filtering is adopted to extract the maximum connected domain, and the target region of interest is obtained; finally, the scattering center is detected and the region segmentation is carried out by using the local density clustering algorithm. The radar target scattering center region segmentation method based on local density clustering has high accuracy, and can effectively improve the technical problem that the SAR image is difficult to have high enough resolution in the actual situation, so that the traditional image segmentation algorithm is difficult to segment the region containing only one scattering center.

The invention discloses a multi-scale structural stiffness optimization design method based on neural network re-parameterization, which comprises the following steps: microstructure parameterization modeling: carrying out parameterization modeling on a microstructure by adopting a level set method, and controlling the volume fraction and geometric configuration of the microstructure by adjusting a cutting height parameter; proxy model construction: establishing a mapping relation between cutting height parameters and microstructure equivalent mechanical properties by using a neural network to form a high-precision proxy model; neural network re-parameterization design: constructing a neural network optimizer, taking coordinates of each discrete unit in a design domain as input, and outputting corresponding cutting height parameters; neural network parameters are updated through automatic differentiation and back propagation, and continuous optimization of design variables is achieved; and a multi-scale topological optimization process: calculating a global stiffness matrix and structural flexibility in combination with an agent model and finite element analysis, and obtaining an optimal multi-scale structural topology meeting volume constraints through loss function minimization iterative optimization.

The invention discloses a structural response analysis and robustness topological optimization design method based on a convolutional neural network, and belongs to the technical field of structural response analysis and topological design. Comprising the steps that sensitivity analysis and substructure unit modulus clustering of a substructure numerical value shape function about substructure unit moduli are carried out, and substructure sampling and sample transformation under a level set method framework are carried out; constructing a convolutional neural network between the substructure numerical value shape function and the substructure unit modulus and a continuum structure topological optimization model, and defining input of the continuum structure topological optimization model; based on the structure statics response analysis of the extended multi-scale finite element, the displacement response solution of the whole structure is realized; and carrying out topological optimization model sensitivity calculation, level set function updating and convergence determination. According to the method, a network structure with fewer layers and lighter weight is adopted, the neural network model is simplified, and the difficulty of neural network training is reduced.

The application discloses a multi-constraint geometric nonlinear level set topology optimization method, device and medium, and aims to solve the problem that the current linear condition level set method is difficult to well optimize the large deformation of a to-be-optimized component. The method divides a design domain corresponding to the to-be-optimized component into a plurality of finite element units, and establishes corresponding level set radial basis functions based on volume constraints; according to the relationship between the element strain increment and the node displacement increment of the finite element unit, displacement constraint terms and frequency constraint terms are introduced to extend the Lagrange function; according to the nonlinear element unbalanced force, the element thermal load and the overall thermal load are calculated; based on the extended Lagrange function and the overall thermal load of the to-be-optimized component under the set temperature, the design domain is topologically optimized. The application considers the displacement constraint and the frequency constraint under the nonlinearity, is suitable for the to-be-optimized component under the large deformation condition, and can achieve an optimal topology result.

The invention provides a cerebral cortex thickness measurement method and system based on a level set image segmentation algorithm, and belongs to the technical field of neural image analysis, and the method comprises the steps: obtaining a brain magnetic resonance imaging image, and carrying out the skull stripping processing; performing three-dimensional visualization on the stripped image through volume rendering, and acquiring three-dimensional coordinates of points of interest and a region of interest on the surface of the cortex by using a preset interaction means; using a level set method to carry out brain grey matter and brain white matter segmentation on the region of interest to obtain a brain grey matter segmentation layer and a brain white matter segmentation layer, and storing three-dimensional coordinate data of points in the segmentation layers; calculating the minimum Euclidean distance according to the three-dimensional coordinate data of the points in the segmentation layer and the three-dimensional coordinates of the points of interest to obtain the cortex thickness of the points of interest; the method solves the problems that the existing cerebral cortex thickness measurement lacks a visual means, is not supported by a mathematical model, cannot be accurate to the thickness of an interest point, and cannot display and retain comprehensive information.

The invention relates to a concrete skewback segment crack simulation analysis method, and belongs to the field of concrete construction. According to the technical scheme, the concrete skewback segment crack simulation analysis method comprises the following steps that S1, simulation analysis basic conditions are determined, specifically, concrete damage parameters are set, and finite element software is selected; s2, constructing a concrete skewback segment simulation model by adopting an extended finite element method, defining a level set function to characterize a crack interface through a level set method, and constructing a displacement function by adopting a discontinuous shape function; s3, setting a concrete crack initiation criterion and a concrete crack propagation direction criterion; and S4, anti-cracking simulation analysis is conducted on the construction stage and the operation stage of the skewback section, and the cracking state and stress field distribution of skewback section concrete are obtained. According to the method, the anti-cracking performance under the load effect in the operation stage is evaluated, the whole-stage cracking state and the stress field distribution rule are output, a whole-cycle data support is provided for engineering practice, and the problem that single-stage analysis and evaluation are one-sided in the traditional technology is solved.

The application provides a deep neural network image segmentation method based on smoothness constraint, adopts a smoothness loss calculation method based on curve evolution as a loss function of a neural network segmentation model, and the loss function of the application firstly converts a segmentation label into a signed distance function, proposes a content loss measurement based on distance penalty, the loss size of a pixel point with segmentation error is positively correlated with the distance from the label boundary, and small regions with missegmentation are effectively avoided; then, an unsupervised length loss measurement is proposed, the measurement minimizes the boundary length of a predicted mask based on a level set method, and the smoothness of a segmentation result is ensured; finally, in order to solve the situation that a segmentation target disappears in a predicted image due to excessive constraint of the length loss function, an area compensation loss measurement is proposed, so as to provide an outward "compensation" loss for the process of segmenting an image.

Disclosed is a Gabor wavelet-fused multi-scale local level set ultrasonic image segmentation method. In the method, non-uniformity of the grayscale of an ultrasonic image is taken as a texture having cluttered directions, the multi-directional property of Gabor wavelets is used to process the image, and intermediate images in different filtering directions are fused by taking maximum values, so as to obtain an intermediate image having a weakened texture effect and an enhanced difference between a foreground and a background. For the feature of a weak edge of an ultrasonic image, a concept of multi-scale is used to improve the conventional LIC method, Gaussian convolution kernels having different variances are set, and a final edge is obtained by means of average fusion.

The invention relates to the technical field of batteries, and discloses a thermal ablation simulation method for a laser irradiation photoelectric unit. The method comprises the following steps: establishing a photoelectric unit multilayer structure geometric model; tracking fusion, vaporization and ablation interface evolution by adopting a level set method; adopting a heat conduction and steam cavity coupling model in stages; introducing a gasification cavity heat effect and transient pressure model; coupling upper and lower heat effects through an equivalent steam heat flow boundary; and analyzing the thermophysical property influence of the substrate material by adopting a dimensionless thermal diffusion number. The method can accurately track the interface evolution, truly reproduces the morphology of the ablation pit, and is used for the anti-laser protection design of the spacecraft solar cell.

The application relates to the technical field of structural optimization design, and specifically discloses a box beam structure lightweight design method based on topology optimization, which comprises the following steps: a three-dimensional parameterized model of multi-physical field coupling is constructed; cross-scale analysis of structural stiffness, thermal deformation, vibration mode and fatigue life is integrated; a variable density method and a level set method mixed topology optimization strategy is adopted; the relationship between density and stress and strain is modeled; and an initial material distribution scheme meeting the structural functional integrity is generated; an intelligent optimization engine of quantum genetic algorithm and deep reinforcement learning fusion is started; through the three-dimensional parameterized model of multi-physical field coupling, the variable density method and the level set method mixed topology optimization strategy are combined, the relationship between the structural lightweight target and the additive manufacturing process is effectively balanced, multi-physical field analysis such as thermal response, vibration and fatigue is introduced, and the box beam can meet the mechanical performance and other functional requirements in actual application.

The application discloses a kind of based on the joint imaging method and system of seismic wave full waveform and gravity data, belong to geological exploration technical field.First, preliminary modeling is carried out to research area to obtain prior information, then data is collected to construct total energy loss function, density and wave velocity model are constructed based on level set method, interface structure is linked to different physical data corresponding inversion parameter, prior information is initialized model parameter, calculated prediction data, determine data proportion coefficient, judge energy loss, calculate update gradient, use asynchronous iteration Adam algorithm to update parameter, complete joint inversion.The application effectively fuses multiple physical data, reasonably sets data proportion, improves imaging effect by the aid of structure similarity, fully utilizes prior information and optimizes regularization term, has good scalability, has strong ability to capture complex interface, can improve geological structure imaging resolution, reduce interference, reduce calculation cost, improve the precision, reliability and overall quality of inversion result.

This invention discloses a vibration analysis method and system for irregularly shaped plates with cracks. The invention employs non-uniform rational B-splines to uniformly describe the geometry and displacement field of irregularly shaped plates, directly inheriting high-precision representations from CAD models and avoiding geometric errors caused by mesh generation in traditional finite element methods. To address displacement discontinuities and stress singularities caused by cracks, enrichment functions are introduced for local enhancement, and the crack location is independently characterized using the level set method, completely decoupling the crack description from the uniformly geometric mesh. This method eliminates the need to reconstruct the mesh due to crack changes, significantly reducing the computational cost of crack dynamics modeling under complex boundaries, improving the computational accuracy of crack regions, and simultaneously solving the problem of stress oscillation at the crack tip of irregular plates. It provides an efficient and robust numerical tool for vibration analysis of geometry-damage coupled structures.

This invention proposes a critical through-impact vibration prediction system for GCVFs based on vortex flow patterns. The system includes a flow field modeling module, which provides data to a fluid-structure interaction (FSI) analysis module. Simultaneously, the flow field modeling module optimizes the mesh through a dynamic mesh optimization module before outputting the data to the FSI analysis module. A signal processing and feature extraction module analyzes and processes the vibration characteristic signal data output from the FSI analysis module and transmits it to a data fusion and prediction module. The data fusion and prediction module predicts the impact vibration intensity, frequency, and evolution trend of the GCVF using its internal prediction model and transmits the prediction results to a display and alarm module. A level set method is used to establish the GCVF flow field model, and the dynamic evolution of the gas-liquid interface is tracked through a zero-value surface. A re-initialization algorithm is introduced to address the numerical diffusion problem, accurately describing the gas-liquid interaction and fluid transport laws, thus effectively solving the complexity problem of gas-liquid coupling flow field modeling in existing technologies.

The invention provides a coal mine goaf directional drilling grouting slurry diffusion distance prediction method, which comprises the steps of defining a global coordinate system for an operation scene of a goaf, establishing a goaf three-dimensional numerical model, and performing grid division on the goaf and a grouting drilling area by adopting a hybrid grid division strategy; based on an O-shaped ring theory and a key layer theory, establishing a three-dimensional space distribution model of porosity and permeability of the goaf; introducing a free medium and porous medium flow coupled slurry-water two-phase flow theory and a level set method, and establishing a slurry diffusion distance prediction model; the grouting pressure, the grout dynamic viscosity and the porous medium permeability are used as influence factors, the grout diffusion volume is used as a response variable, a three-factor and three-level response surface test is designed, and the optimal parameter combination of all the influence factors and the corresponding prediction result of the goaf grouting grout volume fraction are obtained. The method provides theoretical basis and engineering support for goaf grouting process parameter optimization and goaf water disaster prevention and control.

The application provides a medical image segmentation method, comprising: acquiring PET indirect imaging original image data, pre-processing the image, including denoising, contrast enhancement and gray level normalization, to obtain a pre-processed PET image data set; designing a segmentation middleware, integrating multiple segmentation algorithms, including threshold segmentation, region growing and level set method, selecting appropriate segmentation algorithms according to feature vector analysis results for different types of lesion regions; generating labeled PET image data using simulation technology, combining real case data to construct a training data set, using data enhancement technology to expand sample size, generating a large-scale data set for algorithm training; using the segmentation middleware to refine the segmentation result of U-Net, focusing on the edge of the lesion area, accurately positioning the lesion boundary through edge detection algorithm, to obtain a fine segmentation result.