1358 results about "Space model" patented technology

The invention provides an intelligent scheduling and control method and device for an integrated energy system. According to the method, power, gas and heat resource operation data are acquired, multi-scale layered modeling is performed according to a time scale and a space scale, and a power resource state space model, a gas resource flow continuity model, a heat resource heat balance model and a multi-energy coupling characteristic constraint model are established; carrying out feature extraction and dimension reduction representation by adopting a deep auto-encoder network; cooperative training of multiple groups of cognitive models is carried out through a split hierarchical federal learning framework, and a global intelligent model is obtained; constructing a neural architecture search network with a hybrid bionic learning rule, setting a hierarchical scheduling target, and generating a hierarchical intelligent scheduling strategy; and a fault-tolerant control mechanism is constructed, and error detection and correction of operation deviation are realized. According to the invention, multi-time scale collaboration, collaborative learning under multi-device group privacy protection and high-reliability fault-tolerant control are realized, and the operation efficiency and reliability of the integrated energy system are remarkably improved.

The invention relates to the technical field of biomedical engineering, in particular to a multi-mode brain anomaly detection method and system based on machine learning. The method comprises the following steps: acquiring brain medical image data of different modalities, and realizing spatial registration and alignment through a multi-modal registration algorithm based on mutual information; a multi-branch feature extraction model including a convolutional neural network, a converter and a state space model is utilized to perform feature embedding on the original image of each modal; performing frequency decoupling on the features of each mode through adaptive approximate wavelet transform, and decomposing the features into high-frequency detail information and low-frequency global information; a frequency band fusion strategy based on an attention mechanism is implemented on high and low frequency features of different modal images, and fused frequency sub-band features are input into a space-frequency Mama module. Through the adaptive frequency domain decomposition and cross-modal fusion mechanism, the multi-modal brain image information is effectively integrated, and the accuracy and robustness of brain anomaly detection are remarkably improved.

The invention discloses a bad weather image restoration method based on a multi-modal state space model, and belongs to the technical field of computer vision. In order to solve the problem that an existing unified bad weather image restoration method has limitations in the aspects of global receptive field and computational efficiency, the unified restoration of various bad weather images is realized by designing a parallel multi-mode encoder and an adapter to generate comprehensive prompts containing degeneration semantics. Through parallel connection of a state space module and a local context sensing module of a double-attention mechanism, simultaneous capture of global long-range dependence and local detailed features is realized. Experiments show that the method is high in generalization ability, high in processing speed and low in calculation complexity, and can show good adaptive capacity on a plurality of bad weather image removal tasks.

The invention discloses a cargo warehouse-in and warehouse-out management method and management system, and relates to the technical field of information management. Through dynamic storage position optimization, the warehouse-in and warehouse-out frequency, the associated sales relationship and the storage demand characteristics of cargos are combined with a warehouse three-dimensional space model; and generating a cargo storage location allocation scheme by using a multi-objective optimization algorithm. According to the scheme, the picking time of high-frequency warehouse-in and warehouse-out commodities is remarkably shortened, the average time consumption of single-time commodity picking is reduced, meanwhile, the combined picking error rate of associated commodities is reduced, the problem that in the prior art, storage positions are disjointed with warehouse-in and warehouse-out requirements is solved, through predictive process scheduling, cross congestion of operation channels is effectively avoided, and the working efficiency is improved. Therefore, smooth operation of the in-out warehouse process is realized, the occurrence frequency of congestion events of the in-out warehouse channel is reduced, the delay time is shortened, and the problem of congestion of the in-out warehouse process in the prior art is solved.

Neural network architectures for feature extraction from visual input. In at least one embodiment, a neural network architecture for a vision backbone includes hybrid stages with at least one state space model (SSM)-based block preceding at least one transformer block. In at least one embodiment, an SSM-based block includes parallel branches, one including an SSM and one without an SSM, and a concatenation layer for concatenating the output of each branch. In at least one embodiment, the SSM performs a parallel selective scan operation to efficiently map tokens of an input sequence to tokens of an output sequence via GPU acceleration.

The invention provides a remote sensing image semantic segmentation method and system fusing a convolutional neural network and a visual state space model, and the method comprises the steps: firstly extracting the multi-scale semantic features of a remote sensing image based on a lightweight ResNet18 encoder; secondly, a decoder based on a visual state space module is used for modeling a long-distance dependency relationship in an image and recovering spatial resolution; the local feature compensation module is used for enhancing the perception capability of fine-grained semantic information and improving the segmentation precision of a small target area; and finally, the multi-scale attention enhancement module is used for fusing deep and shallow layer features and realizing collaborative optimization of spatial details and semantic information. According to the method, the global semantic features and the local detail information of the remote sensing image can be extracted at the same time, and the method has high segmentation precision and a good remote sensing image semantic segmentation effect.

The invention discloses a physical information neural network-driven ocean island typhoon scene generation method and system. The method comprises the steps of collecting multi-source heterogeneous meteorological data and performing space-time alignment preprocessing; constructing a coarse-scale space-time probability prediction model, capturing space correlation of meteorological elements by using a graph topology learning network, efficiently processing long-time-sequence dependence of typhoon evolution by integrating a state space model with linear complexity, and generating a probabilistic typhoon scene with coarse resolution through a multivariable joint distribution probability model; further constructing a physical downscaling model, taking a coarse-scale prediction result as condition input, and performing physical consistency downscaling on a coarse-scale scene by embedding an atmospheric fluid mechanics equation in a loss function as a physical hard constraint; and finally, outputting a high-resolution typhoon scene with probability reliability and physical authenticity.

The invention relates to the crossing field of mechanical engineering and automatic control, in particular to a global path planning and anti-swing control method for a ship unloader, and aims to solve the problems of rigid path planning and poor synergism of out-of-control swing of a lifting appliance in traditional ship unloading operation. The method comprises the following steps: constructing a dynamic three-dimensional operation space model with fusion of a laser radar and binocular vision, and updating obstacle information in real time; an improved fast extended random tree algorithm is adopted to generate a high-smoothness initial path; establishing a six-degree-of-freedom lifting appliance swinging dynamic model and identifying parameters on line; path tracking and anti-swing torque are synchronously optimized through a model prediction controller, and the control period is smaller than or equal to 20 milliseconds; and closed-loop feedback correction is implemented by combining an encoder and an inertial measurement unit. According to the scheme, environment dynamic sensing, path-anti-swing depth cooperation and multi-disturbance self-adaptive compensation are achieved, the operation rhythm is improved by 30% or above, the system still operates stably under the 8-level wind condition, and high precision, high robustness and engineering implementability are achieved.

The invention relates to a robot path planning method based on multi-sensor fusion and free space topology composition, and belongs to the technical field of robot autonomous navigation. The method comprises the following steps: firstly, fusing laser radar and millimeter wave radar data, constructing multi-modal point cloud data, extracting a three-dimensional obstacle boundary by combining depth and normal vector mutation features, and generating a three-dimensional obstacle map and a free region tree; according to the constructed space model, factors such as energy consumption, dynamic obstacles, path smoothness and the like are integrated, target selection weights are dynamically regulated and controlled, and self-adaptive screening of intermediate navigation targets is achieved. And based on the intermediate navigation target, generating a safe trajectory satisfying dynamic constraints by adopting geometric-dynamic dual-mode fusion modeling, and enhancing the feasibility and robustness of the trajectory through trajectory envelope optimization and pre-execution fault-tolerant control. The method can be widely applied to autonomous navigation systems such as mobile robots and unmanned vehicles, and has good environment adaptability and path execution stability.

The invention discloses a costume design assisting method and system based on artificial intelligence, and belongs to the field of costume intelligent design, and the method comprises the steps: extracting multi-dimensional style feature information from a pre-constructed costume image data set and a fashion trend database through employing a convolutional neural network, and constructing a style vector space model; vector similarity matching is carried out, and a costume design sketch is generated by adopting a graph neural network; based on the design sketch, generating a structured process data packet for proofing; the design sketch and the structural data are input into a three-dimensional human body modeling and simulation module, and wearing effect simulation of virtual clothes is realized based on human body motion capture and physical cloth simulation technologies; and constructing a human-computer interaction interface, and quickly responding to a modification request based on the generative adversarial network model. Structured information such as the garment pattern structure, the sewing sequence, the material specification and the process route is automatically generated through the reinforcement learning algorithm, and the garment proofing efficiency and accuracy are remarkably improved.

The invention discloses a high-resolution remote sensing image semantic segmentation method based on multidirectional parallel selective scanning, and belongs to the technical field of high-resolution remote sensing image processing. According to the method, a multidirectional parallel selective scanning model is provided, direction perception modeling is carried out through 8-direction serialization scanning in combination with a state space model SSM, and the multidirectional long-distance dependence capture capability is enhanced while the linear calculation complexity is kept. A pyramid encoder-decoder structure is constructed, multi-level feature extraction is realized through a four-stage OSSBlock module, and local details and global semantics are dynamically fused in cooperation with SE attention jump connection of a decoder. A selective scanning mechanism is adopted to replace self-attention, and linear complexity calculation is realized through a state space parameter matrix; and designing a mixed loss function, and improving the small target segmentation precision in combination with the class balance of Dice Loss and the difficult sample mining capability of Focal Loss.

The invention relates to the technical field of industrial robot control, and particularly discloses an intelligent foreign matter sorting robot control system based on AI recognition, which comprises a dynamic spatial feature extraction module, a manipulator motion state coding module, a collaborative conflict detection module, a dynamic trajectory optimization module and an execution control adjustment module, constructing a three-dimensional dynamic space model through multi-sensor fusion, and extracting spatial topological features by utilizing continuous coherence analysis; manipulator motion parameters are converted into topological space representation, and a track feature coding matrix is established; detecting interaction conflicts among the manipulators in real time by adopting a multi-scale coherence analysis method, and generating graded early warning signals; a collision avoidance track is optimized based on topological constraints and a virtual rejection field technology; precise execution is achieved through inverse kinematics of the Lie group theory and self-adaptive control.

The invention discloses an intelligent building fire identification and simulation early warning method and system, belongs to the technical field of building informatization and disaster prevention and control, and aims to solve the technical problems of how to realize fire early identification, fire intelligent prediction and evacuation path dynamic planning, improve fire identification accuracy and response speed, and improve the safety and reliability of a building. According to the technical scheme, the method comprises the steps of BIM modeling, wherein a high-precision three-dimensional space model is established based on building information modeling, and the high-precision three-dimensional space model is deeply coupled with an FDS fire numerical simulation engine; multi-modal data acquisition: performing data acquisition by adopting a visual, environmental and spatial multi-modal sensor, and performing feature fusion on the acquired data by using a multi-channel deep convolutional network to generate unified space-time fire characterization; a video-sensing-geometric data collaborative sensing network is constructed on the basis of space-time fire characterization, so that the recognition robustness in a complex environment is improved; intelligent identification and decision making; performing early warning control linkage; and post-disaster assessment feedback.

The invention relates to an unmanned aerial vehicle three-dimensional path planning method based on a multi-strategy improved black-wing optimization algorithm, and belongs to the technical field of path planning. A task space model is constructed according to a task scene of a flight task; constructing a flight cost function according to the performance parameters and the task environment of the unmanned aerial vehicle; determining an initial position and a target position of the unmanned aerial vehicle; according to the method, a global path optimization function is established by taking the lowest flight cost as a target through an improved black-wing optimization algorithm, and the optimal flight path of the unmanned aerial vehicle is solved; hybrid chaotic mapping is introduced for population initialization, a self-adaptive dimension strategy is used in a black-wing plinux migration stage, and a simulated FADs behavior and a memory mechanism are added to a black-wing plinux position update tail end; and by adopting a flight cost function, obtaining an optimal black-wing plinary individual from the optimized black-wing plinary population, and obtaining an optimal flight path of the unmanned aerial vehicle. According to the technical scheme, the problem of path planning of the unmanned aerial vehicle in a complex environment can be effectively solved.

The invention provides a laser radar area monitoring method and device, equipment and a storage medium. The method comprises the following steps: acquiring a mapping matrix between a laser radar three-dimensional point cloud coordinate system and an image coordinate system; reversely projecting a preset target monitoring area of the image coordinate system to the three-dimensional point cloud coordinate system through the mapping matrix, and constructing a monitoring area space model; acquiring spatial feature parameters of a moving point cloud cluster of the point cloud data, and judging a target identity corresponding to the moving point cloud cluster according to the feature parameters; if the target identity is an unauthorized target, judging whether the unauthorized target enters a monitoring area space model or not according to the space characteristic parameters; and if yes, generating corresponding intrusion event data. By constructing the mapping matrix between the image coordinate system and the point cloud coordinate system, the image monitoring area is reversely projected to the three-dimensional space, the space model consistent with the point cloud data structure is generated, accurate reduction of the two-dimensional monitoring area in the three-dimensional space is achieved, and the accuracy of intrusion detection judgment is improved.

The invention relates to a light-weight pest image detection method based on dynamic adaptive scanning and attention mechanism joint optimization, and solves the problems that a light-weight model sacrifices a feature modeling capability during parameter compression, so that missing detection is more, small-scale pest information is difficult to extract, and the detection precision is low. And although small-scale features can be extracted by a high-parameter-quantity scheme, the calculation complexity is high, the high-parameter-quantity scheme is difficult to deploy to edge equipment, and the unification of high detection precision and low calculation power requirements cannot be realized. The method comprises the following steps: constructing a multi-category crop pest data set; constructing a lightweight pest image detection network; training a lightweight pest image detection network; acquiring a pest image to be detected; and obtaining a pest image detection result. Target features are extracted through the dynamic self-adaptive scanning module, long-distance dependency relationships in different directions are captured by utilizing the features of a state space model, and meanwhile, the calculation efficiency is kept; meanwhile, the detection network is optimized based on the attention mechanism, the detection precision and robustness of the pest target are remarkably improved, and the real-time performance of pest detection is achieved through light-weight design.

The invention relates to the technical field of digital twinning, in particular to a digital twinning task scheduling system and a digital twinning task scheduling method oriented to multi-machine cooperation, which realize intelligent mapping between resources and tasks by introducing a topological resource space model. Comprising a topological resource space builder, a task analysis and modeling device, a task resource mapping decision-making device, a global monitoring and optimizing unit and a data fusion processor, the topological resource space builder builds a topological space model of a resource and task mapping relation, and the task analysis and modeling device extracts feature information and a dependency relation of a digital twin task; a task resource mapping decision maker generates a task scheduling decision based on a topological space model and feature information, a global monitoring and optimizing unit dynamically adjusts connectivity parameters of the topological space model according to execution information, a scheduling strategy is optimized, the resource utilization rate is increased, and a data fusion processor merges execution result data and improves the task scheduling efficiency. And through multi-dimensional task feature analysis and intelligent decomposition combination, complex tasks are flexibly dealt with, and the execution efficiency is improved.

The invention discloses an intelligent power prediction method considering dynamic load change, and relates to the technical field of power grid load prediction, and the method comprises the steps: collecting original load data, carrying out the preprocessing, constructing a VMD constraint optimization model, carrying out the four-stage improvement of optimization parameters through employing an improved dung beetle optimization algorithm, so as to generate IMF components, reconstructing the IMF component by calculating a sample entropy to obtain a low-frequency component and a high-frequency component; establishing a Kalman filtering state space model based on the low-frequency component, and decomposing the low-frequency component into a residual component and a pseudo trend component through a Kalman filtering recursive algorithm; external features are obtained, the high-frequency component, the residual component and the pseudo trend component are aligned and spliced with the external features, multi-component collaborative prediction is carried out through a local-global interactive attention mechanism, and a final load prediction result is obtained; and generating a power demand visualization chart based on the final load prediction result. And reliable decision support is provided for power dispatching and energy management.

The invention discloses a nonlinear aerodynamic damping estimation method based on LSTM, and the method comprises the following steps: 1, building a nonlinear state space model of a structure based on structural response, including a state equation, an observation equation and a relation between nonlinear aerodynamic damping and structural vibration amplitude; 2, performing updating and covariance prediction on response data by using unscented Kalman filtering; 3, correcting the Kalman gain in real time by using a long short-term memory network; 4, performing state updating and covariance updating based on the corrected Kalman gain; 5, training the long-short-term memory network through an unsupervised learning mode, optimizing the filtering performance, and defining a mean square error of a posterior observation predicted value and a real observation value as a loss function; and 6, calculating the nonlinear aerodynamic damping according to the estimated nonlinear aerodynamic damping parameters. The invention further discloses a nonlinear aerodynamic damping estimation system based on the LSTM and a storage medium.

The invention relates to a BIM model mobile terminal and mixed reality cooperation method and system and a medium, and the method comprises the steps: carrying out the space segmentation and space grid division processing of a BIM model according to a pre-created index structure, obtaining the index information of a plurality of space models and a plurality of space grids, and obtaining a plurality of space models and a plurality of space grids based on the index information; and constructing a unique identification code of the spatial model. Identifying the identification code through the mobile terminal, loading the spatial model corresponding to the identification code, obtaining spatial coordinates in the spatial model according to the gesture operation, converting the spatial coordinates into spatial grid index information through a preset conversion rule, and generating a loading instruction of the mixed reality equipment. And the mixed reality equipment loads a space model corresponding to the space grid index information according to the loading instruction, and superposes the space model to a real building scene through a positioning strategy. The problem of low cooperation efficiency of a BIM model mobile terminal and mixed reality in related technologies is solved.

The invention belongs to the technical field of unmanned aerial vehicle path planning, and particularly discloses an unmanned aerial vehicle path planning method based on a multi-strategy fusion information acquisition optimization algorithm, and the method comprises the steps: obtaining the digital elevation model map data of a target region, and constructing a three-dimensional environment space model based on a task environment; constructing a path cost function and initializing an optimization algorithm population, wherein the initialization adopts three strategy fusion mechanisms of hypercube initialization, chaos initialization and random initialization to enhance the diversity of the population; a crisscross strategy is introduced, the adaptability and stability of the algorithm in the dynamic search process are enhanced, and a sine disturbance mechanism is adopted, so that the global optimization capacity and local development precision of the algorithm are enhanced; the flight path of the unmanned aerial vehicle is constructed by optimizing the path point sequence, and the path is smoothed by introducing B spline interpolation, so that the path is ensured to have continuity and feasibility, the flight requirement of the unmanned aerial vehicle is met, and the robustness and environmental adaptability of path generation are remarkably improved.

The invention relates to the technical field of substation operation and maintenance, in particular to a robot inspection path optimization method and system applied to an extra-high voltage substation, and the method comprises the steps: triggering the dynamic optimization of an inspection path in response to the detected equipment operation state change or inspection environment change of the extra-high voltage substation; based on a preset transformer substation environment space model and real-time equipment states and environment parameters collected in real time, a first target inspection path is generated through path cost analysis, and simulation verification and fine adjustment are carried out; when interference risks, positioning deviation overrun and / or multi-machine conflicts exist in execution of the first target inspection path, path planning is updated based on local environment information collected by the robot in real time, and a second target inspection path is generated; and executing the second target inspection path. According to the invention, the equipment state and the environment change can be responded in real time, the inspection path can be automatically adjusted, key inspection tasks are prevented from being omitted, and the continuity of the inspection process is ensured.

The invention discloses a power electronic networking equipment stability analysis method, system, equipment and medium, and the method comprises the following steps: building a full-system state space model, carrying out the cross validation of the full-system state space model through employing a modal analysis method, an impedance analysis method and a complex torque analysis method, and obtaining the stability of the full-system state space model; obtaining a multi-mode coupling stability analysis result; according to the multi-mode coupling stability analysis result, constructing a power grid s domain node admittance matrix; establishing a virtual power angle dynamic equation of the network-forming converter, and optimizing control parameters according to an evaluation result; an inertia center reference system is adopted, a dynamic interaction model of the synchronous machine and the network-forming converter is established, and dynamic coupling characteristics between devices are analyzed through a relative motion equation. According to the method, the comprehensiveness, the precision and the engineering applicability of dynamic stability analysis of the novel power system are improved through fusion of a multi-dimensional collaborative evaluation framework and an innovative technical system.

The invention relates to a machine learning technology, and discloses a wire and cable wiring optimization method and system based on artificial intelligence, and the method comprises the steps: obtaining a three-dimensional model of a target wiring space, obtaining a three-dimensional space model, recognizing an evasion region and a power utilization node topological graph in the three-dimensional space model based on a pre-trained target recognition model, and carrying out the recognition of the evasion region and the power utilization node topological graph. And generating a shortest wiring path according to the power utilization node topological graph and the three-dimensional space model, calculating a multi-dimensional constraint condition according to pre-acquired cable parameters, and optimizing the shortest wiring path according to the avoidance area and the multi-dimensional constraint condition based on intelligent feedback iteration to obtain an optimized wiring path. According to the invention, the accuracy and efficiency of cable wiring optimization can be improved.

The invention discloses a dual-branch attention and boundary enhanced remote sensing image change detection method and system based on a Mama framework, and belongs to the technical field of computer vision and remote sensing image processing. The method comprises the following steps: carrying out cutting, data set division and normalization preprocessing on a dual-tense remote sensing image; multi-scale features are extracted through a Mamba encoder, and global spatial dependence is modeled by using a state space model; constructing a boundary enhancement branch, and combining channel compression and up-sampling to generate a boundary probability graph; in the decoding stage, a space-time state space (STSS) module is introduced to fuse space-time features, and a dynamic channel attention mechanism is embedded to strengthen a change sensitive area; using deformable convolution to optimize edge offset, and performing weighted fusion on main branch and boundary branch output to generate a final change graph; and training the model through a composite loss function and optimizing parameters. According to the invention, the detection precision of a change area in a complex scene is significantly improved, and the problems of boundary blur, small target leak detection and background interference are solved.

The invention belongs to the technical field of underwater positioning navigation, and relates to an integrated navigation system and method based on robust adaptive Kalman filtering, and the system comprises a strapdown inertial navigation system, a Doppler velocimeter, a depth sensor and a data processing module; the method comprises the following steps: constructing a state space model based on an integrated navigation system, and estimating system noise parameters on line in real time by introducing a Sage-Husa adaptive filtering mechanism; performing sequential processing on the multi-dimensional measurement information by adopting a sequential measurement and variance limitation method, and setting upper and lower limits of scalar measurement noise variance; an IGGIII criterion is introduced, and gross errors of measurement information are detected through a residual filter; according to the method, noise and outlier interference can be suppressed in a complex underwater environment, and the positioning precision, robustness and stability of a navigation system are improved.

The invention discloses a cooperative parameter optimizing and setting method and system for a network construction type reactive power compensation device, terminal equipment and a storage medium, and belongs to the field of power system control. The method comprises the steps of obtaining operation condition data and a power flow initial value of a power grid region where a reactive power compensation device to be set is located, constructing a discretization state space model, calculating a damping ratio according to a calculated system characteristic root, and determining a dominant oscillation mode; respectively changing the installed capacity, configuration points and control parameters of the reactive power compensation device to be set in a preset step length, and generating a characteristic root moving track; the installed capacity, the configuration points and the control parameters are continuously optimized based on the feature root moving trajectory until the preset stabilization capability index is larger than 0 and the optimized feature root real part is smaller than 0, the optimal installed capacity, the optimal configuration points and the optimal control parameters of the to-be-set reactive power compensation device are obtained, the to-be-set reactive power compensation device is adjusted, and the to-be-set reactive power compensation device is adjusted. According to the invention, the problem of lack of cross-band parameter setting in the prior art is solved.

The invention discloses an end-to-end defogging network fusing Mama and multi-scale context enhancement, and belongs to the technical field of computer vision and image restoration. The existing defogging technology has the problems of complex global modeling calculation, poor local detail recovery and insufficient feature fusion self-adaption. The problem is solved through the design of four core structures: 1, a Mamba state space model is introduced, and global fog long-range dependence modeling is realized with linear complexity; 2, designing a multi-scale cavity convolution module and a local context enhancement unit, and covering multi-scale local features; 3, constructing an enhanced adaptive fusion module to realize dynamic adaptation of global and local features; and 4, gradually recovering details by adopting a multi-scale input-cross-scale fusion-multi-scale output framework. The network can efficiently process high-resolution fog-containing images, and is suitable for scenes needing real-time high-precision defogging, such as automatic driving, security and protection monitoring, remote sensing imaging and the like.

The invention discloses a dynamic simulation prediction model construction method based on data driving. The dynamic simulation prediction model construction method comprises the following steps: S1, collecting multi-source heterogeneous multi-modal original data and preprocessing to obtain a standardized data set; s2, extracting a long-term dependent feature sequence by adopting a Mamba depth state space model; s3, establishing a parameter diffusion space by using the self-adaptive multi-mode diffusion model, and obtaining diffusion characteristic parameters; s4, performing reverse denoising convergence processing on the diffusion characteristic parameters to generate real-time adaptive model parameters; s5, performing dynamic weighted fusion of the feature sequence based on real-time adaptive model parameters; s6, inputting the fusion sequence into a dynamic simulation prediction model framework for training, and obtaining trained model parameters; and S7, solidifying model parameters to complete model construction. According to the invention, the dynamic prediction precision and generalization performance are improved.