264 results about "Surrogate model" patented technology

This application discloses a method, device, medium, and product for optimizing the structure of a shaped charge shroud, relating to the field of structural design. The method includes: constructing an objective function with structural parameters as design variables and maximizing performance index values ​​as the objective; determining multiple initial sample points using a Latin hypercube sampling method based on the range of structural parameter values; constructing an initial sample library; constructing a surrogate model based on the initial sample library; determining candidate sample points and their corresponding performance index values ​​using a Bayesian optimization loop based on the surrogate model and the range of structural parameter values; updating the initial sample library and the surrogate model; continuing until a termination condition is met; and using the sample point corresponding to the maximum performance index value as the target structural parameter; and optimizing the design of the shaped charge shroud based on the target structural parameter. This application can reduce the computational cost of determining the structural parameters of a shaped charge shroud and improve the efficiency and intelligence of shaped charge shroud structure optimization.

This invention discloses a filter physical perception optimization method based on residual attention and transposed convolution, comprising the following steps: constructing a filter characteristic prediction surrogate model based on residual self-attention mechanism and transposed convolution decoding to establish a nonlinear mapping relationship between filter geometric parameters and electromagnetic response parameters; training the filter characteristic prediction surrogate model using an S-parameter weighted mean square error loss function to obtain a trained filter characteristic prediction surrogate model; and using a physical perception-improved differential evolution algorithm, employing the trained filter characteristic prediction surrogate model as a predictor to find the optimal combination of geometric parameters that satisfies the filter design specifications in the solution space. This method is a filter optimization design method that balances high-precision prediction and high-efficiency optimization. This method can predict the S-parameters of microwave filters more accurately than traditional neural networks, effectively replacing time-consuming full-wave electromagnetic simulation.

This invention provides an active flow control method and system based on intelligent algorithms. The method includes: determining the number of inlet parameter sets for a typical flight state point, constrained by simulation economy and sampling efficiency; performing numerical simulations according to the number of inlet parameter sets to obtain dataset 1, with inlet parameters as independent variables and flow field characteristics as dependent variables; obtaining dataset 2, with flight state as independent variables and flow field characteristics as dependent variables, through numerical simulation for a certain set of inlet parameters; constructing a proxy model of inlet flow field characteristics for a typical flight state point using deep learning based on dataset 1; constructing a proxy model of inlet flow field characteristics for a wide-speed-domain flight state based on the proxy model of inlet flow field characteristics for the flight state point, using a transfer learning algorithm and dataset 2; constructing an environment-agent interaction model based on a deep reinforcement learning algorithm, using the proxy model of inlet flow field characteristics for a wide-speed-domain flight state as the environment model for deep reinforcement learning; setting optimization termination conditions based on the environment-agent interaction model, and obtaining a control sequence with continuously optimal inlet performance under different flight states through deep reinforcement learning. This invention can stably adapt to flight scenarios with a wide speed range and a large airspace, and achieve real-time optimal control of the air intake flow field under a wide speed range, solving the problem of poor generalization ability of traditional methods under different operating conditions.

This invention discloses a PIDNN-SO-based method for optimizing the layout of offshore wind farms, belonging to the field of wind farm optimization design technology. The method first inputs wind farm, turbine, and wind resource parameters, sets constraints on turbine spacing, boundaries, and wake overlap rate, and constructs a high-precision wake flow field physical model based on actuator disk theory, Navier-Stokes equations, and a weighted quadratic wake superposition model. Secondly, it builds a physical information dual neural network (PIDNN) with dual encoders, integrating physical equation loss and data loss to train the model, achieving rapid and accurate prediction of wake wind speed. Then, it uses the snake optimization algorithm (SO) to optimize the environmental evolution factor, completing an efficient search for turbine layout. Finally, it constructs a PIDNN-SO collaborative optimization framework, trains a PIDNN surrogate model through initial sampling, and outputs the optimal layout that maximizes annual power generation. This invention can effectively reduce wake loss and significantly improve the power generation efficiency of offshore wind farms.

The present application relates to the technical field of aero-engine aerodynamic optimization and computer-aided design, and discloses a numerical optimization design method and structure for aero-engine low-pressure turbine blade leading edge modeling based on response surface modeling. The method realizes precise matching of Klebanoff stripe and large-scale vortex generation timing by adjusting flow or rotor speed to change wake injection angle, and promotes stripe impact and accelerates large-scale vortex breakup. Meanwhile, by constructing a blade leading edge modeling parameterization model and establishing a response surface proxy model for the combination of modeling design variable parameters and boundary layer velocity mean square deviation, the optimal design variable parameter combination is determined in the constraint domain with the maximum velocity mean square deviation in the boundary layer as the target. The present application effectively reduces the boundary layer momentum thickness and total pressure loss under the premise of ensuring constant blade load, improves the design efficiency and engineering applicability under low Reynolds number conditions, and does not need to introduce additional flow control structures, thus being simple in structure and easy to implement.

This invention discloses a method for assessing the vulnerability of bridge piers under ship impact, comprising: obtaining the tensile stress threshold at the pier base for different failure states of the bridge pier under ship impact; extracting multiple parameter combinations from the core parameter space affecting the tensile stress at the pier base using the optimal Latin hypercube sampling method; calculating the tensile stress at the pier base corresponding to each parameter combination using a finite element model of the bridge pier under ship impact, and fitting an initial surrogate model for the tensile stress at the pier base; based on the prediction accuracy of the initial surrogate model, supplementing the extraction of multiple parameter combinations in the parameter space region where the prediction error or the degree of nonlinearity of the function is greater than the corresponding threshold; and using the supplemented extracted parameter combinations to return to the operation of calculating using the finite element model to obtain a new surrogate model; repeating this process until the prediction accuracy of the final surrogate model meets the requirements; and fitting the pier vulnerability curve for each failure state. This embodiment is more consistent with the damage mechanism of concrete cracking and structural failure.

This application provides a method for predicting the fatigue life of turbine blades based on adaptive gene expression programming. The method includes: establishing a finite element model of the turbine blade and determining key parameters affecting its fatigue life; extracting sample points, calculating fatigue life using the finite element model to form an initial dataset, which is then divided into a training set and a validation set; performing adaptive gene expression initialization programming, mutating and crossovering the population, and selecting and iterating based on fitness to obtain a surrogate model; evaluating the accuracy of the surrogate model; if the accuracy does not meet the requirements, adding the worst-performing points from the validation set to the training set for retraining until the accuracy meets the requirements; and using the obtained surrogate model to predict the expected value and standard deviation of the upper and lower bounds of the fatigue life. This method aims to handle complex, uncertain, coupled problems and improve the efficiency of fatigue life prediction while ensuring computational accuracy.

This invention relates to the field of wireless communication network optimization technology, specifically to an antenna layout and relay configuration optimization method based on a channel prediction surrogate model. The method includes: acquiring a digital twin environment model of the area to be optimized; constructing an original sample set by calculating channel parameters between candidate antennas and relay nodes through ray tracing; training and constructing a channel prediction surrogate model using improved sparse Gaussian process regression; and using location parameters as input and output channel quality indicators. A joint decision vector for antenna, relay location, and transmit power is defined, and an optimization objective function for total system capacity and deployment cost is defined. The surrogate model is embedded in Bayesian optimization as a fast evaluator, and the joint decision vector is iteratively sampled and evaluated. After convergence, the optimal layout configuration scheme is output. This method replaces real-time simulation with a dedicated surrogate model, achieving multi-parameter joint optimization and improving the efficiency of channel evaluation and optimization iteration.

The application provides a kind of collaborative jet airfoil optimization design method based on combined parameterization, belongs to the field of aircraft aerodynamic design and flow control.This method firstly carries out CST parameterization to the original basic airfoil for constructing original collaborative jet airfoil;With blowing / air suction port position, size as design variable, the curve of subsidence section is fitted by original basic airfoil linear interpolation, realizes the adaptive association of subsidence section and original basic airfoil.Subsequently, a joint design space containing CST parameterization variable, collaborative jet airfoil design variable and jet control parameter is constructed, the initial proxy model is constructed and updated to convergence by using SBMO framework.The application carries out global collaborative optimization to the subsidence section channel geometry and the overall profile of original basic airfoil, solves the parameter coupling problem of the basic geometry shape of collaborative jet airfoil and subsidence section, and can significantly improve the aerodynamic performance of collaborative jet airfoil.

The application relates to a parallel multi-objective optimization method and system for engineering product CAE simulation, and belongs to the technical field of simulation optimization. The method solves the problems that the simulation-driven design cycle is long and it is difficult to obtain a high-quality multi-objective trade-off optimization scheme. The method comprises the following steps: based on multiple objective functions and constraint conditions, initial sample data is obtained through two-stage collaborative optimization and is put into a sample library; a joint surrogate model is constructed and trained, and a hierarchical error compensation mechanism is initialized; multiple rounds of iterative optimization are performed until a preset optimization termination condition is met; each round of iterative optimization comprises the following steps: based on the current joint surrogate model and the hierarchical error compensation mechanism, multiple candidate design points are generated in parallel, and then the multiple candidate design points are distributed to multiple computing nodes for parallel simulation calculation to obtain new sample data which is put into the sample library; the joint surrogate model and the hierarchical error compensation mechanism are updated; and a multi-objective optimization design scheme set of a product to be optimized is obtained from the final sample library. The optimization efficiency is improved.

This invention discloses a small-sample Bayesian optimization sampling method and system based on cloud droplet data augmentation, belonging to the field of transportation engineering material design. Addressing the problems of limited initial samples and poor fit of the Bayesian optimization surrogate model in modified asphalt formulation design, this invention acquires initial small-sample data; constructs a clustering cloud model to obtain expectation, entropy estimates, and hyperentropy estimates; generates cloud droplet virtual samples using a forward cloud generator and merges them with the original samples to expand the dataset; trains a Gaussian process regression surrogate model based on the expanded data; constructs an expectation-improved acquisition function to optimize and solve candidate formulations; updates the data after physical testing verification; and iterates repeatedly until the termination condition is met to output the optimal formulation. This invention mines the distribution information of small-sample data through cloud droplet data augmentation, improves the accuracy and sampling efficiency of the surrogate model, significantly reduces the number of expensive physical tests, and lowers the cost of formulation design. It can be widely applied to the formulation optimization of modified asphalt and similar high-cost experimental materials.

This invention discloses a heat sink structure optimization design method based on a non-dominated sorting genetic algorithm, belonging to the field of heat sink optimization design technology. The method first determines the structural parameters to be optimized (heat sink length, width, height, fin thickness, fin spacing, substrate thickness) and the optimization objective function (maximum junction temperature of power devices, heat sink mass, heat sink entropy productivity). Then, Latin hypercube sampling is used to sample parameters and establish a geometric model. Sample data is constructed through thermal simulation, and a surrogate model is established using response surface methodology. Finally, multi-objective optimization is performed based on the non-dominated sorting genetic algorithm to obtain the Pareto optimal solution set, and the best solution is selected through comprehensive performance evaluation indicators. This invention effectively reduces heat sink mass and cost while ensuring heat dissipation performance, shortens the R&D cycle, and is applicable to the heat dissipation optimization design of power devices in power electronic systems.

The application discloses a hydrogen network optimization method and system based on a Gaussian process regression agent model, and belongs to the field of petroleum chemical process system engineering and technology.The method first constructs Gaussian process regression agent models of desulfurization, denitrogenation and dearomatics reaction kinetics for different hydrogenation devices, respectively, to accurately predict hydrogen consumption and product impurity content with operating temperature and pressure as input;then embeds the agent models of the devices into a hydrogen network optimization model to establish an integrated optimization problem of minimum total annual cost;finally, solves the optimization problem by using an improved differential evolution algorithm combined with a three-stage search strategy of global exploration-local development-convergence verification to obtain optimal operating parameters.The application shortens the optimization calculation time to about 90 seconds, solves the solving difficulty of traditional methods coupled with complex kinetics, realizes collaborative optimization of hydrogenation operation and hydrogen network operation while ensuring product quality, reduces hydrogen consumption and total annual cost of the system, and significantly improves the economic benefit and operation efficiency of the hydrogen system of a refinery.

The application discloses a kind of non-cooperative target centroid intelligent positioning method and system fusing multi-dimensional physical constraint.The method is directed to sparse, noisy point cloud data, and constructs a comprehensive evaluation function containing four-dimensional information of geometry, orbital dynamics, time sequence continuity and surface physical properties;Using a two-stage solution framework of surrogate model and hybrid optimization, first, a lightweight neural network surrogate model is used with a differential evolution algorithm for global coarse search, and then switching to a high-fidelity orbit model combined with an adaptive particle swarm optimization algorithm for local fine search;Finally, the optimal state estimation and its covariance matrix, evaluation decomposition and other decision support information are output.The application improves the accuracy, robustness and computational efficiency of centroid positioning, and enhances the interpretability of the results, suitable for on-orbit servicing, space debris removal and other high-risk space missions.

The general aspects of the present disclosure relate to a method for determining a surrogate model of a state space model. The method includes receiving a time-continuous state space model describing a system to be closed-loop controlled. The method further includes forming a time-discrete representation of the state space model describing the system to be closed-loop controlled, wherein at least the time-discrete representation of the state space model has a time dependence based on a stochastic distribution. The method further includes determining a surrogate model of the time-discrete representation of the state space model.

The application discloses a MOFs synthesis route multi-objective optimization method based on machine learning. The method is to collect the synthesis conditions of prepared Ce-UiO-66, and evaluate the defect content and thermal stability thereof through a thermogravimetric analysis curve, as initial data set; the data set is randomly divided into a training set and a test set, eight algorithms are adopted to model each performance of Ce-UiO-66 and select the proxy model for performance prediction; the target achievement probability (PA) value of each performance in the synthesis space is calculated and expanded into a multi-objective evaluation factor; the Ce-UiO-66 material is prepared; the obtained material is subjected to characterization test, if the test data does not meet the requirement, the data set and the proxy model are updated. The application has the advantages of low cost, short cycle and the like in optimizing the catalytic and stable performances of MOFs based on reliable experimental data and machine learning, and can be popularized to the design of synthesis routes of other materials.

This invention relates to the field of computer-aided design technology, and more particularly to a method and system for constructing wind dams for concentrated wind energy capture. The method includes the following steps: acquiring multi-source wind field data; processing the multi-source wind field data to obtain wind field structure data; generating a guiding structure from the wind field structure data to obtain wind dam geometric structure data; performing fluid-structure interaction (FSI) simulation based on the wind field structure data and the wind dam geometric structure data to obtain FSI data; and performing multi-constraint dynamic optimization on the FSI data to obtain construction control strategy data. This invention achieves proactive guidance and efficient matching of the wind dam structure to airflow convergence behavior, improving the efficiency of concentrated wind energy capture and the reliability of structural design. By using a surrogate model and multi-constraint dynamic optimization, the highly complex simulation process is transformed into a rapidly reasoning-based decision-making process, effectively reducing computational costs.

A visual navigation method for robot swarms based on differentiable physics and agent models is proposed. A high-level navigation policy network processes raw image information acquired from a depth camera, using a deep learning model to understand the environmental geometry, identify obstacles and dynamic agents, and output velocity commands for obstacle avoidance, collision avoidance, and target navigation. A low-level motion policy network translates these velocity commands into joint position and torque commands for the quadruped robot's legs, generating stable and flexible body movements to execute velocity tracking strategies. After training using a forward-backward asymmetric optimization framework, swarm control is achieved in the online phase. This invention improves training sample efficiency and policy transferability in scenarios relying solely on local visual perception, while overcoming the curse of dimensionality in multi-agent learning. It can be flexibly extended to large-scale robot swarms and achieves zero-sample transfer from simulation to real-world scenarios.

This invention discloses an InSAR interferometric network optimization method based on a multi-factor coherence surrogate model. The method includes: constructing a multi-factor feature dataset containing SAR imagery, baseline data, and soil moisture data of the study area; constructing a candidate interferometric pair set, and randomly selecting a subset of sample interferometric pairs from the candidate set; constructing a coherence surrogate prediction model, which uses the multi-factor feature data corresponding to the sample interferometric pair subset and the ground truth values ​​of sample coherence to train the model parameters; inputting the multi-factor feature data corresponding to each interferometric pair in the candidate set into the trained coherence surrogate prediction model and outputting the predicted coherence of each interferometric pair and the optimized interferometric pair network. This invention improves the stability and accuracy of unwrapping and time-series inversion, while also reducing the difficulty of interferometric processing and coherence calculation under large-scale data, and can support the needs of large-scale, long-term, and near-real-time monitoring.

The application discloses a helicopter tension-torsion strip multi-physical field coupling parameter inversion method and system based on a physical information neural network, first, parameterized geometric modeling is carried out on the tension-torsion strip, a material parameter set is defined, and a task spectrum containing four flight stages of taking off, cruising, hovering and landing is constructed; simulation results are obtained by solving each stage in the task spectrum respectively, input and output of each group in the above simulation process are defined as a sample point, and a PINN is constructed; the geometric, material and task spectrum parameters are taken as input, the maximum Mises equivalent stress of each stage is taken as output, a fatigue damage evolution equation is embedded in a loss function as a physical constraint, the PINN is trained again by using finite element sample points, the trained PINN proxy model is combined with a target fatigue life, and the geometric parameters or material parameters meeting the life requirement are inverted. The problem that the prior art cannot efficiently and accurately inversely deduce the key geometric parameters or material parameters of the tension-torsion strip according to the target fatigue life is solved.

This paper presents a sparse grid polynomial-based orbital game situation analysis method, belonging to the field of aerospace orbital adversarial and intelligent decision analysis technology. Addressing the problems of lack of systematic representation of multi-source uncertainties, low computational efficiency of traditional quantification methods, and difficulty in achieving multi-objective equilibrium optimization in existing high-orbit multi-to-multi dynamic adversarial scenarios, this paper constructs a POMDP multi-to-multi dynamic game model incorporating uncertainties in high-orbit dynamics and control execution, strategy behavior, and communication delay. A surrogate model is established using sparse grid sampling and polynomial chaotic expansion, and global sensitivity analysis is performed. Based on this, a multi-objective optimization model is constructed, and the Pareto optimal policy set is solved. This achieves game strategy generation that balances performance, robustness, and economy, and is applicable to spacecraft strategy optimization and space situation analysis in high-orbit multi-to-multi dynamic adversarial missions.

This invention discloses a method for optimizing the design of complex equipment based on a hybrid adaptive sampling surrogate model. This method divides the design space into several Voronoi polygons controlled by sample points. Three indices—leave-one-out error, local nonlinearity, and polygon region size—are used to evaluate the importance of each sample point region. An entropy-weighted distance method is used to select the most suitable sensitive region for adding points. A learning function based on error estimation is then combined to determine new sampling points. This iterative update constructs a high-precision surrogate model, establishing a mapping relationship between design parameters and optimization objectives, thereby solving the design optimization problem. This invention innovatively proposes a multi-index adaptive weight allocation method to select sensitive sample regions and proposes a novel learning function to determine the location of added points. This method can construct a high-precision surrogate model using fewer sample points and can be used for the optimization design of complex equipment, including but not limited to tunnel boring machines.

This disclosure relates to a battery intelligent design system and method, electronic device, storage medium, and program product. The method includes: a requirement input module configured to acquire user requirement information of the battery to be designed; a hybrid surrogate model determination module configured to determine a historical hybrid surrogate model matching the historical chemical system type from a historical project database; a multi-objective optimization module configured to optimize a specified initial set of schemes using the historical hybrid surrogate model with the indicator requirements of multiple design indicators as optimization objectives, and obtain the indicator prediction results of each design scheme in the target scheme set; and a result output module configured to select at least one design scheme as the target design scheme based on the weights of each design indicator and the indicator prediction results of each design scheme in the target scheme set, and output the target design scheme and the indicator prediction results. This enables an efficient, adaptive, and automated battery design process.

This invention relates to the fields of automated robot technology and coal yard safety management technology in thermal power plants, and particularly to a coal yard debris cleaning device based on an automated robot. The device includes a sensing module that collects and fuses data to output a comprehensive data stream. A twin module dynamically constructs and updates a dynamic digital twin through coupled discrete element method and computational fluid dynamics simulation. A risk prediction module uses a graph neural network surrogate model and Monte Carlo simulation to generate a spontaneous combustion probability spatial distribution map as dynamic risk information. A decision-making module integrates dynamic risk information and robot state to generate a set of candidate joint action strategies. A control module verifies the strategies through parallel simulation and performs multi-objective arbitration based on dynamic risk information to select the optimal control law and decouple it into low-level control commands. A feedback module executes commands and collects actual data, feeding it back to the twin module to correct the dynamic digital twin and to the control module to optimize decision weights.

This invention discloses a method and system for solving the performance of a gas turbine engine by integrating differentiable operators and static computation graphs. The method includes: constructing differentiable component operators: reconstructing the input-output mapping relationship of the core components of the gas turbine engine into a globally differentiable surrogate model, which possesses analytical gradient solving capability; constructing the overall engine static computation graph: utilizing a programming framework supporting automatic differentiation, establishing a general graph construction mechanism decoupled from the engine topology, which automatically instantiates the corresponding static computation graph based on the input engine topology description; and reverse-mode automatic differentiation solving: calculating the residual vector through forward propagation, and using reverse-mode automatic differentiation technology to propagate the gradient flow backward along the static computation graph, analytically obtaining the Jacobian matrix required for Newton iteration, and completing the solution of the engine's overall performance residual equation. This invention achieves efficient, accurate, and stable simulation of gas turbine engine performance.

This invention relates to the interdisciplinary field of carbon emission reduction and industrial intelligent optimization, specifically a parameter optimization method and system based on a direct air carbon capture (DAC) system. The method includes: acquiring a physical simulation model of the DAC system, controllable parameter types, controllable parameter ranges, environmental parameter ranges, and current environmental parameters; setting constraints based on the controllable parameter types; constructing a proxy model for evaluation indices using Gaussian process regression based on the controllable parameter ranges, environmental parameter ranges, constraints, and the physical simulation model; and iteratively obtaining the optimal parameter combination using a multi-objective optimization algorithm based on the controllable parameter ranges, constraints, current environmental parameters, and the proxy model. This invention solves the problem in existing technologies where the controllable parameters of a DAC system cannot be adaptively adjusted according to environmental parameters, resulting in insufficient operating efficiency, through the use of a proxy model and optimization algorithm.

This invention belongs to the field of underwater acoustic communication technology and provides a performance simulation verification method, system, and storage medium for underwater acoustic communication systems. The method includes constructing a multi-source heterogeneous input feature space, constructing and training a deep neural network surrogate model in machine learning, generative learning and online inversion of the spatiotemporal dynamic error field, embedded online compensation and closed-loop verification iteration, and uncertainty quantification and robust decision-making. Through the end-to-end collaborative design of multi-source data fusion, machine learning surrogate modeling, spatiotemporal dynamic error field generation, embedded closed-loop simulation, and robust decision-making, the dynamic error field can adapt to the spatiotemporal nonstationarity of the marine environment and communication scenario in real time, accurately generating fine-grained error compensation amounts. The embedded closed-loop mechanism realizes automated iteration of error learning, compensation, and model optimization, continuously improving simulation accuracy without manual intervention and significantly improving verification efficiency.

This application discloses a method and system for developing a closed-loop laser stealth cutting process, relating to the field of laser cutting technology. The method includes: acquiring historical processing parameters from past cutting processes and using these parameters as initial sample data; iteratively training a preset surrogate model using the initial sample data to obtain a trained first surrogate model; determining a combination of required parameters based on the first surrogate model and a multi-objective optimization strategy, and iteratively updating the first surrogate model multiple times using this combination until the first surrogate model outputs optimal process parameters that satisfy the preset optimization objective; and performing laser cutting on the target material using the optimal process parameters. This application achieves the technical effect of adaptively adapting to various laser cutting conditions for different cutting materials, thereby improving processing quality and efficiency.