615 results about "Physical information" patented technology

The present invention relates to the field of industrial intelligent predictive maintenance, and in particular to an interpretable deep feature fusion network-based industrial intelligent predictive maintenance method, comprising: acquiring gearbox vibration data comprising noise; performing preliminary extraction and noise suppression on features of the acquired data by establishing an interpretable feature extraction module having a physical information constraint; integrating multi-scale features comprising long-distance and local dependencies by means of a dual-branch feature fusion module having global and local feature fusion capabilities; performing dimensionality reduction on a high-dimensional feature and generating an output by means of a classifier to obtain a final fault identification result; and performing interpretability analysis on a diagnosis process of a model. In the present invention, by embedding the signal processing technology having a well-defined physical theory support into a deep neural network, the interpretability and reliability of model inference results are effectively improved while the fault identification accuracy of the model is improved.

The invention discloses an intelligent self-monitoring temperature management system for a box-type substation, and relates to the technical field of intelligent power grid equipment monitoring. The problems that an existing system is large in temperature measurement deviation, low in reliability, delayed in early warning, inaccurate in hot spot positioning and extensive in heat dissipation control are solved. According to the scheme, multi-source signals are acquired in parallel through a data acquisition module, and a temperature time sequence is extracted; a boundary calibration module is adopted to fuse data to generate a three-dimensional boundary condition; the multi-physical field solving module obtains an internal temperature / stress field; the physical information prediction module is fused with a heat transfer physical constraint training graph neural network to predict a hotspot migration trend; the hierarchical scheduling module is used for solving a fan and oil pump collaborative optimization instruction in real time based on model predictive control; according to the invention, the accuracy of internal temperature monitoring of the box transformer substation, the reliability of hot spot prediction and the accuracy of heat dissipation control are remarkably improved, the insulation life of equipment is effectively prolonged, and the operation safety and reliability of the system are improved.

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 a wellbore multiphase flow model numerical solution and gas-liquid distribution state inversion method and system, and belongs to the technical field of petroleum engineering, and the method comprises the steps: 1, constructing and training a physical information neural network for drilling wellbore multiphase flow dynamic simulation and overflow gas distribution state inversion; determining input and output of the physical information neural network; determining a loss function of the physical information neural network; training a physical information neural network; 2, designing an adaptive optimization algorithm, optimizing the final solution precision and convergence speed of the physical information neural network, and obtaining an adaptive physical information neural network; designing an adaptive activation function; designing a self-adaptive sampling mechanism based on residual errors; 3, based on the self-adaptive physical information neural network, numerical solution and gas-liquid distribution state inversion of the shaft multiphase flow model are achieved. According to the method, the problem that a traditional numerical method usually needs high-precision grid division and a large number of computing resources is effectively solved.

The invention discloses a multi-fidelity data fusion and physical constraint-based cold heading process parameter staged optimization method, which comprises the following steps of: 1) performing calculation through Deform finite element simulation and an empirical formula, constructing a multi-fidelity initial data set, and improving data consistency through normalization and deviation calibration; 2) constructing a multi-fidelity physical information neural network (PINN) model, and establishing a mapping relation between process parameters and forming quality indexes by adopting a staged training strategy and an adaptive weight adjustment mechanism; and 3) verifying the generalization ability of the model by dividing a training set and a test set, ensuring that a prediction result accords with a volume conservation criterion and a material forming limit, and realizing optimization of cold heading process parameters. According to the method, through multi-fidelity data fusion and physical information embedding, on the basis of enhancing a physical mechanism and multi-data collaboration, the data acquisition cost is reduced, and the generalization of the model is improved; and through dynamic weight distribution and a staged training strategy, the prediction precision and reliability are improved.

The invention relates to a small sample fault prediction method based on physical information guidance and multi-source adaptive fusion. The method comprises the following steps: collecting real operation data of preprocessing target equipment; according to the physical model or domain knowledge of the target equipment, generating simulation sensor data conforming to a physical rule under various fault modes of different degrees; constructing diversified training samples in combination with real data and simulation data; for different types of sensor data, designing corresponding feature extraction branches, mining potential fault features in the data, and dynamically adjusting the weight of each data source fault feature for fusion based on an output result of a physical model and data-driven feature correlation analysis; inputting the obtained fusion features into a fault prediction model based on a small sample learning framework for training; comprehensively considering the fault prediction result, the real operation data and the analysis result of the physical model, and carrying out quantitative evaluation on the overall health state of the target equipment; and causal diagnosis and visual interpretation are carried out.

The invention is suitable for the field of atomic layer deposition, and discloses an atomic layer deposition process management method and system based on digital twinning, and the method comprises the steps: building a digital twinning model of an atomic layer deposition production system based on a multi-scale physical information neural network model; acquiring real-time reaction chamber process parameters and real-time state variables, and inputting the real-time reaction chamber process parameters and the real-time state variables into the digital twin model to obtain physical field distribution in the reaction chamber and a predicted growth state of a film on the surface of the substrate; based on the physical field distribution in the reaction cavity and the substrate surface film prediction growth state, decision analysis is carried out through a pre-trained reinforcement learning agent, and correction parameters used for optimizing the film growth quality are obtained; and generating a correction instruction based on the correction parameter, and updating the process parameter of the reaction cavity based on the correction instruction, so that the film on the surface of the substrate in the reaction cavity reaches the target film thickness and uniformity, and thus real-time intelligent closed-loop control of perception-prediction-decision-execution can be formed.

The invention relates to the technical field of geophysical exploration, in particular to a multichannel deep learning magnetotelluric inversion method based on physical information constraint. The method comprises the following steps: generating a synthetic data set containing a geoelectric model and forward modeling response thereof, and adding a noise simulation actual observation condition; constructing a hybrid network architecture combining Transform and U-Net, taking apparent resistivity and impedance phase as dual-channel input, extracting global features by using an encoder, gradually recovering spatial resolution through a decoder, and outputting an underground resistivity model; network training adopts a composite loss function fusing model loss and data loss, and an inversion process is constrained by introducing a magnetotelluric forward modeling physical rule, so that a result is ensured to fit observation data and conform to a physical mechanism; after training is completed, preprocessed actual measurement data are input into the model, and a resistivity image can be directly obtained. The method is used for geological structure identification and reservoir interpretation, and the inversion precision and reliability are effectively improved.

The invention discloses an underground water exploitable quantity evaluation method and system based on coupling of a physical mechanism and a neural network, and relates to the technical field of water resource management, meteorological data, surface water hydrological monitoring data, geological survey data, underground water resource development and utilization data and hydrogeological parameter data are integrated and preprocessed, and a basic data set is obtained; a three-dimensional geological model of a heterogeneous structure is constructed by utilizing a digital twin physical engine and combining drilling geological data. By introducing a digital twin technology and a numerical model, the underground water flowing process can be simulated more accurately, a digital twin physical engine is utilized to construct a heterogeneous aquifer digital model, a dynamic interaction process of surface water and underground water is embedded, and the model is enhanced through a graph neural network and a physical information neural network, so that the underground water flowing process is simulated more accurately. Errors caused by hypothesis of aquifer homogeneity, neglecting of surface water-underground water coupling process and the like of a traditional model are effectively reduced, and the accuracy of underground water exploitable quantity evaluation is improved.

The invention discloses a geothermal abnormal point identification system based on remote sensing detection. The system comprises a data layer which collects thermal infrared remote sensing data, multispectral / hyperspectral remote sensing data, radar data, geological and topographic data, meteorological data and geophysical prospecting data; the processing layer is used for extracting key features, capturing dynamic changes of geothermal anomalies, distinguishing continuous geothermal anomalies and transient events, combining hyperspectral data and magnetotelluric sounding data, and constructing'earth surface-deep part 'associated features; the analysis layer is used for learning a diffusion rule of geothermal anomalies along a fault zone through a graph propagation model, taking a geothermal field heat conduction equation as a loss function constraint by utilizing a physical information neural network, jointly training a temperature field prediction model, constructing a deep three-dimensional geological structure based on gravity and geophysical prospecting data, and constructing a geologic structure model; constructing a geothermal anomaly detection model through the features provided by the processing layer to recognize geothermal anomaly points; and the output layer is used for generating a geothermal anomaly probability graph.

The invention discloses an ice condition prediction method based on multi-feature fusion and physical constraint, and belongs to the technical field of hydrological forecasting. Firstly, various types of data of a target area are collected and preprocessed to serve as a data set, features of the various types of data are extracted, feature fusion is conducted on obtained image features, time sequence features and environment features through a multi-head attention mechanism, and a fusion feature vector is generated. Secondly, adopting a physical information neural network model, taking the fusion feature vector as input, taking ice thickness and ice stress as output layers, carrying out constraint by using a composite loss function, carrying out model optimization by using a verification set, carrying out processing through a full connection layer in the network, and carrying out end-to-end training and regularization of a prediction model; and finally, evaluating the final prediction model obtained by training through the test set, and verifying the effectiveness and generalization ability of the final prediction model. The method combines multi-source and multi-mode observation data, can effectively capture complex relations between ice surface changes and various factors, can improve prediction precision, stability and reliability, and can improve model training efficiency and generalization ability.

The invention discloses a self-adaptive full-band efficiency correction method for an LIBS (laser-induced breakdown spectroscopy). The method comprises the following steps: for each laser pulse, inputting collected spectral data into a trained physical information deep network to obtain a current plasma temperature; when the current plasma temperature is not in a preset ideal temperature interval, predicting by using a physical information deep network to obtain an optimal laser energy adjustment sequence; wherein energy adjustment is performed according to the optimal laser energy adjustment sequence, so that a future temperature sequence can be converged to a preset ideal temperature interval; controlling an energy controller of the laser to emit a next laser pulse according to the optimal laser energy adjustment sequence to obtain a stable plasma; and after the stable plasma is obtained, collecting target spectral data, and outputting a full-band efficiency calibration curve according to the target spectral data. And high-precision and high-robustness full-band efficiency correction can be realized in a complex and changeable real environment.

The invention discloses a state evaluation method, system and device and a medium, and belongs to the field of machine learning, and the method comprises the steps: obtaining multi-modal sensor data, carrying out the preprocessing and fusion, and obtaining a fusion feature vector; detecting the fusion feature vector based on a mixed architecture model constructed by a spline enhanced multi-layer sensing network, a convolutional neural network and a Transform attention mechanism to obtain an anomaly type and confidence; when the detection result is abnormal, generating an abnormal event according to an event driving mechanism; responding and acquiring time sequence data from historical sensor data, and performing simulation prediction according to the physical information neural network model to obtain a prediction result; the prediction result is fed back to the detection model to update the confidence degree, the target detection result is obtained, the state evaluation result is obtained in combination with the prediction result, and therefore the state evaluation accuracy and reliability can be improved.

The invention relates to the technical field of power equipment fault diagnosis, in particular to a transformer overheating fault intelligent diagnosis system based on a physical information neural network. According to the invention, physical constraint filtering is carried out on signals through the adaptive noise suppression module; a loss function fusing a heat conduction equation, a gas decomposition kinetic equation and an insulation aging kinetic equation is constructed through a physical information neural network, and coupling modeling of temperature and gas and collaborative prediction of the aging state of insulation paper are achieved; spatial topological features and time sequence features are extracted through a multi-modal feature fusion module, and cross-modal association is established by adopting an attention mechanism; predicting a future feature trajectory through a fault evolution prediction module and outputting hierarchical early warning; and finally, outputting fault types, severity and maintenance suggestions. According to the method, a physical mechanism and data intelligence are deeply fused, advanced early warning and accurate positioning of an overheat fault are realized, and meanwhile, the blank that insulation failure cannot be pre-judged in the prior art is filled.

The invention discloses a physical field solving method based on a Bayesian physical information extreme learning machine, and the method comprises the steps: constructing a single-layer full-connection neural network, carrying out the random initialization, and fixing the weight of an input layer; based on a partial differential equation of a physical system and boundary conditions thereof, defining a training loss item containing physical information; a physical system solving problem is converted into a linear least square problem, and a linear equation set is constructed; supposing that an output layer weight parameter obeys Gaussian prior distribution with the mean value being zero, and controlling a covariance matrix by an adjustable hyper-parameter; constructing a Gaussian likelihood function based on the observation data, and calculating posterior distribution of the output weight in combination with the prior distribution; carrying out iterative optimization on the hyper-parameter by adopting an evidence maximization method to obtain a mean value and a covariance of posterior distribution; based on posterior distribution, adopting a Monte Carlo integral method to generate prediction output of the physical system; and performing uncertainty quantization based on the variance of prediction output, and outputting a prediction value and a confidence interval thereof.

The invention provides a geological disaster intelligent early warning method based on fusion of a physical information neural network and space-air-ground monitoring, and the method comprises the following steps: S10, obtaining geological structure features of different depths under the ground of a city, and recognizing the spatial distribution and thickness of an underground abnormal body; acquiring a surface deformation time sequence, crack and landform information, a three-dimensional point cloud and a continuous vibration signal of a surface-shallow stratum; s20, constructing a three-dimensional twin substrate, performing space-time alignment and resampling on multi-source heterogeneous data, mapping the data to a three-dimensional grid of a unified coordinate system, and extracting and fusing geological disaster precursor features; s30, taking the generated fusion feature field as an input training neural network model, carrying out geological disaster forward prediction and parameter inversion, and outputting future stability probability distribution, a potential slip plane and key parameter evolution; and S40, adaptively adjusting a risk threshold and an early warning rule, constructing an incremental data set, and carrying out periodic incremental training and parameter optimization on the physical information neural network in the step S30.

The invention discloses a pumped storage power station building foundation dangerous point management and control system based on BIM, and the system comprises a data layer which enables linear engineering monitoring data to be organized into a high-order tensor of pile number * time * sensor type * spatial position through four-dimensional tensor modeling, and carries out the Tucker decomposition, strain rate abrupt change points are dynamically recognized in the compression process by combining sliding window anomaly detection, original data are directly stored, and then a hierarchical storage strategy is implemented on non-abrupt change data through a data value evaluation model driven by deep reinforcement learning; the model layer is used for fusing a physical information neural network, embedding a rock-soil mechanics constitutive equation into an LSTM network to construct a mixed loss function, pre-training a general geologic model through transfer learning, and then carrying out few-sample fine tuning in combination with project data; and the application layer is used for dynamically generating a hazard source list through a BIM model and triggering instant early warning based on the compressed and reconstructed data of the data layer and the output of the model layer, and simulating a disaster situation development path in combination with a digital twinborn technology.

The invention belongs to the field of industrial thermal engineering and intelligent modeling, and discloses a heat exchanger heat exchange coefficient prediction method and system based on a physical information neural network. The method comprises the following steps: acquiring multi-dimensional operation data through a signal acquisition system, cleaning abnormal and blank values, standardizing, and segmenting into time sequence samples by adopting a sliding window method; a double-layer physical information long-short-term memory network is constructed, and a time sequence feature and a physical equation residual error are combined to generate a space-time fusion feature matrix. And a composite loss function including data loss, physical equation loss and physical consistency loss is designed, physical and data driving influences are balanced through hyper-parameter tuning, and accurate prediction of the heat exchange coefficient is achieved based on a gradient descent optimization model. The method combines field physical laws and data features, improves the reliability and physical interpretability of prediction, and is suitable for operation optimization of the heat exchanger of the desulfurization wastewater treatment system of the thermal power plant.

The invention discloses a power system time domain simulation method based on a physical information depth operator network model. The method comprises the following steps: initializing an MATLAB environment; calling actual example data; setting a fault type and a scene; performing time domain simulation by using a PST tool box; recording the data of the simulation curve and the state variable; preprocessing the data; generating and storing a training sample; initializing parameters of the depth operator network model; calling a training set and a test set; training parameters are set; determining key parameters; calculating a loss function; storing model parameters; and setting the feasibility and the actual effect of the fault scene verification model. By applying a DeepONet innovative algorithm, the calculation time of time domain simulation of the power system is greatly shortened, the efficiency of the method is far higher than that of a traditional numerical integration method, and the method can meet application scenes with extremely high real-time requirements; good generalization ability is achieved, and stable performance can be kept in different fault scenes; by embedding the physical constraint, the physical interpretability and reliability of the simulation result are improved.

The invention discloses an open-pit mine slope risk prediction method and system based on multi-source data, and relates to the technical field of intelligent geological engineering, and the method comprises the steps: collecting geological data, displacement monitoring data and environmental activity data of an open-pit mine slope, carrying out the preprocessing, and constructing a three-dimensional digital twinborn body; carrying out fusion calculation on the preprocessed displacement monitoring data and environmental activity data through a physical information neural network, predicting a displacement value and a stress value of the strip mine slope, and obtaining full-field physical quantity prediction data; according to the strip mine slope slip crack surface damage evolution path, the stability risk levels of the whole strip mine slope and different areas in the strip mine slope are judged, high-risk areas are positioned, and a strip mine slope risk prediction report is generated. According to the method, the damage evolution path of the slip crack surface of the strip mine slope is dynamically deduced, so that space-time continuous modeling of the damage evolution process, early recognition of the slope instability precursor and accurate insight of the risk evolution trend are realized.

The invention discloses an active learning and physical loss-based fatigue crack growth rate prediction method, which comprises the following steps of: splicing and fusing a preprocessed stress intensity factor, a stress ratio, pre-strain and stress amplitude, and taking the fused characteristics as the input of a model; embedding the Jones model into a loss function of the neural network, constructing a physical information time sequence model based on the Jones model, setting a parameter set, selecting a most valuable training sample from a training set by using active learning, retraining the model by using the selected most valuable training sample, selecting an optimal model, and evaluating the optimal model by using a test set; fusing the actually measured stress intensity factor, stress ratio, pre-strain and stress amplitude of the to-be-predicted material, and inputting into the obtained prediction model to obtain a predicted value of the fatigue crack growth rate. By adopting the technical scheme of the invention, the fatigue crack growth rate of the material in the whole life period can be efficiently and accurately predicted at low cost.

The invention belongs to the technical field of underwater sound, and discloses an underwater target sound scattering modeling method and system based on a physical information neural network. The method comprises the following steps: establishing a loss function of a physical information neural network by establishing an underwater target acoustic-structure coupling mathematical model; a decoupling parallel dual-network architecture is established, a fluid domain neural network is used for learning and outputting a scattering sound pressure field of a fluid domain, an elastomer domain neural network is used for learning and outputting a displacement field of an elastomer domain, and the two networks realize physical coupling by sharing a sampling point and a corresponding boundary condition loss function at an acoustic-structure coupling boundary; and generating training data, performing iterative training on the decoupling parallel physical information neural network, and calculating a scattering sound pressure field outside the target and a displacement field inside the elastomer by using a trained model. According to the method, the physical information neural network is innovatively applied to the underwater target sound scattering field, and a new thought and a new way are provided for accurately forecasting the underwater target sound scattering field.

The invention provides a three-dimensional wind field prediction method and system based on multi-modal complementary fusion learning. The method comprises the following steps: S1, acquiring remote sensing observation data and numerical simulation data; s2, obtaining standardized remote sensing features and simulation features; s3, obtaining a unified scene representation; s4, splicing the unified scene representation with the to-be-predicted space-time coordinates, inputting the spliced scene representation and the to-be-predicted space-time coordinates into a physical enhancement decoder, and outputting three-dimensional wind speed vectors at the corresponding space-time coordinates; and S5, iteratively optimizing parameters of the bimodal encoder, the cross-modal attention fusion module and the physical enhancement decoder to form a closed-loop prediction model. According to the method, multi-modal data complementation and physical information deep fusion are realized, through innovating a neural network architecture and a constraint mechanism, the prediction precision under a sparse data condition is remarkably improved, the physical credibility of a result is enhanced, and a technical support is provided for intelligent development of the wind power industry.

The invention relates to the technical field of wind power generation, artificial intelligence and fluid mechanics, and discloses a prediction method and device for a multi-working-condition flow field of a wind driven generator based on a neural network, and the prediction method comprises the steps: obtaining a sample data set which comprises multiple groups of multi-working-condition data; and inputting the sample data set into a physical information field adversarial neural network model for training to obtain the total loss of forward propagation of the sample data in the field adversarial neural network model. According to the total loss, determining whether training of the wake flow field prediction model is completed; and under the condition that a new round of training is carried out on the wake flow field prediction model, back propagation is carried out on the total loss, and neural network parameters are optimized. In this way, a dual-branch loss collaborative optimization mechanism is formed. A physical information neural network and a domain adversarial neural network are combined, and complementary advantages of the two are fully exerted. When the data is limited or the distribution difference is large, high-precision and physically consistent wake flow field prediction can be realized.

The invention belongs to the technical field of virtual power plants, and particularly relates to a multi-physics field coupled virtual power plant energy storage health state monitoring and predicting method, which comprises the following steps: acquiring electrochemical parameters and thermodynamic parameters of an energy storage facility in real time, preprocessing the acquired data, and storing the preprocessed data into a database; calculating the current SOH of the energy storage facility, constructing a physical information neural network (PINN), and predicting the SOH of the energy storage facility based on the current SOH and historical data in the database; displaying a curve graph of data acquired in real time, the current SOH, an SOH prediction chart and alarm information in a visual mode; and with maximization of SOH and minimization of energy loss and thermal risk as targets, a Pareto optimization problem is constructed, and an optimal charging and discharging strategy is solved. According to the method, multi-parameter perception, physical mechanism and deep learning are fused, and high-precision real-time monitoring and prediction of the energy storage health state of the virtual power plant can be realized.

The invention discloses a defect-containing material life prediction method based on an M integral physical information neural network, and the method comprises the steps: constructing a damage parameter, namely an equivalent damage area, of a defect-containing metal material according to an energy equivalent method based on a classical fracture mechanical parameter M integral; establishing an M integral fatigue model according to a power-law relationship between the equivalent damage area and the M integral, and constructing a physical loss function; constructing a mixed loss function by adopting the physical loss function and the data related loss function, and constructing a physical information neural network based on the mixed loss function; optimizing hyper-parameters of the physical information neural network to obtain an optimal life prediction model; and evaluating the fatigue life of the defect-containing metal material by using the prediction model. According to the method, the influence of multi-defect features on fatigue failure is considered, the life prediction result meets the correlation between the fatigue data and conforms to the physical law, the dependence on the fatigue data is reduced, and the prediction precision and the analysis efficiency are remarkably improved.

The invention provides a transformer magnetic field prediction method and system based on a physical information neural network, and the method comprises the steps: inputting a magnetic field distribution diagram sample of a transformer into a U-Net model, and obtaining a predicted magnetic field image, outputted by the U-Net model, of the transformer; determining a data driving loss function according to the pixel difference between the predicted magnetic field image of the transformer and the target magnetic field image, determining the magnetic field energy density of the transformer according to the predicted magnetic field image of the transformer, and determining a physical loss function according to the magnetic field energy density of the transformer; determining a comprehensive loss function according to the data-driven loss function and the physical loss function, and training the U-Net model by using the comprehensive loss function; and inputting the current magnetic field distribution diagram of the transformer into the trained U-Net model to obtain a current predicted magnetic field image of the transformer. According to the method, the deep learning model and the physical loss constraint are fused, and the precision, the physical consistency and the generalization ability of transformer magnetic field prediction are remarkably improved.

The invention discloses a rolling bearing generalization fault diagnosis method based on a quantum physical information neural network, and the method comprises the following steps: S1, obtaining vibration signals and working condition parameters in the operation process of a rolling bearing, and carrying out the preprocessing and domain division to obtain a meta-training set and a meta-test set; s2, constructing a quantum-physical information neural network model which comprises a quantum path and a physical path; s3, feature fusion and classification; and S4, based on physical perception element learning training, repeating inner ring-outer ring iteration until the model is converged. Compared with the prior art, the rolling bearing fault diagnosis field generalization method based on the quantum physical information neural network has the advantage that the field offset problem existing in a traditional fault diagnosis method is solved.

The invention discloses an underground medium multi-scale forward and reverse modeling method and system based on a decoupling neural network, and belongs to the technical field of underground multi-physics field coupling. The underground medium multi-scale forward and reverse modeling method based on the decoupling neural network comprises the following steps: generating a spatio-temporal evolution data set of multiple physical field variables; a physical field decoupling physical information neural network architecture is constructed, the architecture comprises three special sub-networks, each sub-network takes time-space coordinates as basic input and dynamically receives output of other sub-networks as auxiliary input features, and strong coupling modeling between physical fields is realized through feature sharing and a joint loss function; executing a two-stage training strategy; and executing forward intelligent prediction or key physical property parameter inversion based on the trained neural network architecture. According to the method, a unified neural network framework composed of a plurality of special sub-networks is constructed, physical consistency and data-efficient cross-scale modeling from a rock core scale to a site scale are realized, and forward simulation and parameter inversion are synchronously supported.

The invention discloses a power system probability load prediction method, system and device, and a storage medium, relates to the technical field of hydrogen energy ship power system load prediction, and aims to solve the technical problems that multi-source data and physical information rules are not integrated and multi-scale chaos in an HPV complex operation environment is not considered in the prior art. The method specifically comprises the steps of collecting and preprocessing multi-source data; screening strong correlation data by using the maximum information coefficient, and reconstructing multi-source data into a unified spatial-temporal characteristic matrix; constructing an electrical and environmental parameter chaos model based on a Lorentz dynamic equation as a physical constraint term of a loss function; a BERT-PINN framework with an attention mechanism is established, and multi-step probability prediction is realized through two-stage training. According to the method, the data driving model and the physical information rule of the multi-source data are effectively integrated, the feature fusion problem of the multi-scale chaotic features is solved, and the model prediction performance is improved.