413 results about "Predictive power" patented technology

A method for multivariate time-series long-term forecasting based on multi-scale temporal feature enhancements, includes a time-series forcasting model TFEformer. The model utilizes a multi-branch structure and a patch-series attention mechanism to extract global and local time-series features at multiple temporal scales, and designs an adaptive feature fusion mechanism to achieve adaptive fusion of multi-scale temporal features. It employs an variate-wise attention mechanism and a redesigned gated feedforward network to perform feature fusion among multivariate variables and within the time-series, respectively. The time-series forcasting model TFEformer proposed by the present invention significantly improves the prediction of long-term trends in time-series and enhances the fitting ability for short-term local fluctuations, comprehensively increasing prediction accuracy across different prediction time lengths in multivariate time-series forcasting tasks.

The invention relates to the technical field of APT killing chain reconstruction and prediction methods, and provides an APT killing chain reconstruction and prediction method and system based on causal reasoning. The method aims at solving the problems that in an existing APT detection and defense technology, multi-source heterogeneous data integration is difficult, space-time dynamic feature modeling is insufficient, killer chain reconstruction is inaccurate, the predictive capacity is limited, interpretability is poor, and real-time performance and accuracy are difficult to consider at the same time. The method comprises the steps of obtaining multi-source heterogeneous data, constructing a causal graph, performing multi-modal feature extraction based on the causal graph, and obtaining node initial feature representation; inputting the node initial feature representation into a time sequence causal graph convolution learning algorithm to obtain an optimized node representation; executing killer chain reconstruction based on the optimized node representation to obtain a possible killer chain sequence; predicting a future attack path according to the possible killing chain sequence to obtain a potential attack path; and generating an interpretability analysis report based on the potential attack path.

The invention relates to the technical field of ship power systems and energy management. The invention provides an energy optimization management method and system for an extended-range hybrid power ship. The method comprises the following steps: collecting ship multi-source heterogeneous data in real time, and constructing a multi-source heterogeneous data set; based on the multi-source heterogeneous data set, constructing a load prediction model based on a time sequence convolutional network and long and short term memory fusion, and dynamically outputting propulsive power demand probability distribution in a future navigation period; a multi-objective optimization model is established, and an improved multi-objective genetic algorithm is adopted to generate a Pareto frontier solution set; and on the basis of a rolling time domain optimization framework, in combination with real-time navigation situation awareness data, carrying out online correction on the Pareto solution set, and generating and executing an optimal energy distribution instruction set. The problems of an existing extended-range hybrid power ship energy management technology in the aspects of working condition adaptability, multi-energy coupling efficiency, real-time state sensing and predicting capacity and emission and economical efficiency balance are solved.

The invention relates to the technical field of electric power meteorological prediction, and aims to solve the problem that a current power grid natural disaster early warning system has defects in feature extraction and cannot fully utilize knowledge in related fields to enhance the prediction capability of a model. The invention discloses a power grid natural disaster early warning method based on a knowledge graph, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: S1, constructing a power meteorological knowledge graph based on meteorological data, geographic information and power equipment data corresponding to a power grid natural disaster; s2, performing feature enhancement on the meteorological data according to semantic features in the knowledge graph to form fusion features; s3, according to the fusion features, constructing a disaster weather prediction model based on a graph convolutional neural network and a self-attention mechanism; and S4, according to the disaster weather prediction model, performing early warning on the natural disasters of the power grid in the target weather. By adopting the method, the accuracy of power grid natural disaster prediction can be improved.

The invention relates to the technical field of power electronics, in particular to a photovoltaic off-grid inverter control method and system, and the method comprises the following steps: applying a test signal to identify an output interaction characteristic parameter, predicting the power generation power through combining with weather and attenuation, calculating the power fluctuation, and adjusting a current control parameter set. And calculating voltage and reactive power compensation parameters to generate an inverter control instruction, and combining a temperature load prediction trend to adjust a power allowable value to obtain a predictive power calibration coefficient. According to the invention, by analyzing the output interaction characteristics, enhancing the perception capability of the inverter to the actual operation environment, combining with the prediction of the photovoltaic power generation amount, realizing the prediction and influence evaluation of the power fluctuation, improving the adaptive adjustment capability of the current control parameters, improving the output voltage and reactive power control precision by utilizing the interaction characteristics, and optimizing the electric energy quality; and through thermal management of multi-point temperature and load prediction, the operation safety is improved, the service life of equipment is prolonged, and the operation performance of the photovoltaic off-grid system in a complex environment is improved.

The invention discloses a unified framework for solving automatic driving track prediction and planning consistency based on a world model. According to the method, through cooperative work of the automatic driving domain controller and the vehicle-mounted sensing system, end-to-end joint optimization of track prediction and planning in a complex traffic scene is realized, time sequence dependence and interaction dynamics among intelligent agents are accurately captured, and the prediction capability and robustness of a model are remarkably improved. The method comprises the following specific steps: firstly, constructing a generative world model, and generating potential future state representation by utilizing a behavior conditional and backtracking expansion technology; secondly, in combination with global modeling and a local convolutional network, multi-scale features are extracted, adaptive fusion is carried out, and a multi-modal prediction trajectory is generated; then, a multi-target planning model is adopted to integrate various driving indexes, and a track with the minimum loss function is generated; finally, path planning parameters are dynamically optimized through real-time environment perception and decision feedback, and the problems of prediction uncertainty and planning consistency of the automatic driving track are effectively solved.

The invention discloses an ablation impedance dynamic prediction method based on multi-modal time sequence fusion and a related device. The method comprises the following steps: acquiring multi-modal time sequence data of an ablation electrode in real time, and constructing a multi-dimensional input vector containing a temperature change rate, an ultrasonic echo gradient and a standardized timestamp; performing parameter fine tuning by using a pre-trained two-channel long-short-term memory neural network, respectively extracting feature vectors of the physical parameters and the image parameters, fusing the feature vectors, and outputting an impedance predicted value in a future preset time window; and dynamically adjusting the ablation output power by adopting a proportional control strategy based on the deviation between the predicted impedance and the target impedance. Through time sequence fusion of the multi-modal data and the prediction capability of the deep learning model, the real-time monitoring and power adjustment precision of the ablation process is improved, and the safety and effect stability of ablation treatment are effectively improved.

The invention relates to the technical field of hydrological forecasting, in particular to a deep learning-based short-term rainfall and flood forecasting method and system, and the method comprises the following steps: constructing a deep learning-based radar echo epitaxial prediction model and a rainfall and flood forecasting model, training and testing the two constructed models by adopting the first data set and the second data set respectively; inputting the rainfall radar echo data of the target watershed in a certain period of time into the radar echo epitaxial prediction model, and predicting the change trend of the rainfall radar echo in a future period of time; and inputting the change trend of the rainfall radar echo in the future period of time into the rainfall and flood forecasting model, and forecasting the rainfall and runoff of the target drainage basin in the future period of time. According to the method, the precipitation and the runoff of the target drainage basin can be predicted more accurately in a short time, and the prediction capability and the decision level of regional water resource management are effectively improved.

The invention provides an intelligent power distribution comprehensive monitoring system integrating the Internet of Things and cloud computing. The system is applied to the technical field of power distribution monitoring and comprises a communication network and a cloud computing platform module, the data acquisition and transmission module is used for acquiring equipment operation parameters, environment parameters and the like of key point locations of a transformer substation or a power distribution substation in real time, and uploading the data to the cloud computing platform after data protocol conversion; the data analysis and processing module is used for carrying out time sequence modeling and fusion by adopting intrinsic mode modeling and warp dynamic linear modeling respectively to obtain a target parameter comprehensive predicted value; the fault prediction module is used for performing fault prediction based on the closed continuous neural network and a Bayesian estimator, fusing prediction results and outputting a final fault prediction result; and the intelligent control module is used for optimizing the equipment operation environment and early warning, and realizing integrated supervision of the transformer substation or the power distribution substation. According to the invention, the operation monitoring and fault prediction capabilities of the substation (distribution substation) equipment are improved.

The invention discloses a short-term power load prediction method in combination with weather parameters. The method comprises the following steps: acquiring power load data and weather parameters of corresponding time; preprocessing the data; using an EEMD method to decompose the power load data into a high-frequency nonlinear component and a low-frequency linear component; in combination with weather parameters, training a CNN-LSTM model which introduces an attention mechanism and is based on multi-scale feature extraction, and outputting a high-frequency nonlinear component predicted value; for the low-frequency linear component, a time sequence model SARIMA is adopted, and a low-frequency linear component predicted value is output; and adding the high-frequency nonlinear component predicted value and the low-frequency linear component predicted value to obtain a complete load prediction result, carrying out error analysis, and carrying out iterative optimization on the model. According to the method, weather data are introduced, high-frequency nonlinear components are extracted through sequence decomposition, weather parameters are used as input to construct a deep learning prediction model, and the prediction capability of complex load fluctuation and the precision and reliability of short-term load prediction are improved.

The invention discloses a waterworks double-coagulant accurate prediction method based on automatic machine learning, and the method comprises the steps: firstly obtaining the water quality data of a waterworks, the agent dosage data of two coagulants, and other basic data, and carrying out the preprocessing of the data; an automatic machine learning method is adopted, and a tap water double-coagulant dosage prediction model under the production condition is built; the model is optimized according to an actual effluent target value, and accurate prediction of the dosage of the two coagulants is realized by using the optimal model. According to the invention, accurate calculation and intelligent regulation and control of the dosage of the coagulant in the waterworks are realized, and the predictive capability of the double coagulant is particularly emphasized, so that the innovation point ensures that the effluent quality reaches the drinking water standard, meanwhile, the dependence on manpower is remarkably reduced, the operation cost and the medicament cost of the waterworks are saved, and the economic benefit is increased. And double improvement of economic benefits and environmental benefits is realized.

The invention relates to the technical field of biodiversity assessment, and discloses a dynamic assessment method and system for the influence of engineering construction on biodiversity, and the method comprises the steps: obtaining a basic ecological data set, and carrying out the construction of an ecological patch graph structure and the edge building of hydrological connectivity according to the basic ecological data set, and obtaining a composite ecological graph model; performing plaque function heterogeneity aggregation and disturbance link simulation processing to obtain an ecological propagation path and a sensitivity pedigree of each node; performing spectral domain decomposition and key propagation node identification processing to obtain ecological intermediary nodes and propagation weights; performing reversible modeling and path entropy backtracking analysis processing to obtain a community reconstruction trend index and a function stability attenuation curve; and carrying out ecological response persistence analysis processing to obtain a time sequence biodiversity influence index. According to the method, the problems of static state, splitting, low coupling and the like in the prior art are solved, and the precision, timeliness and predictive capacity of ecological disturbance dynamic response evaluation are remarkably improved.

The invention discloses an error prediction method of inertial navigation equipment, which belongs to the technical field of inertial localization, and comprises the steps of data acquisition and preprocessing, error modeling and preliminary fusion, physical constraint-based LSTM modeling, LSTM optimization, and error correction and feedback optimization, and solves the technical problem of improving the real-time performance and accuracy of a navigation system. According to the method, an IMU drift formula and an environment temperature error are fused for modeling, the prediction precision is enhanced, a physical interpretation force is given to the model, the adaptability of a system to a dynamic environment is improved, a real-time error is used as a reward signal, an LSTM weight and a covariance matrix are dynamically optimized, the prediction capability under long-time work is improved, feedback optimization is performed on corrected navigation data, and the prediction precision is improved. And model self-learning and performance iteration are realized.

The invention relates to a method for constructing a space-time multi-task prediction large model based on data fusion. The method comprises the steps that firstly, a data fusion module is constructed, the data fusion module comprises a time sequence alignment sub-module, a region fusion sub-module and a feature splicing sub-module, and original data sets are fused based on the three dimensions of time, space and features; then, constructing an input embedding module, and extracting time features of the data set while embedding the fused data set into a high-dimensional space; constructing a spatial coding module, and extracting spatial features of the data set; then, a base large model prediction module is constructed, and reasoning prediction is carried out; and finally, designing a multi-task output module, and performing multi-task splitting on a prediction result to obtain a space-time multi-task prediction large model. According to the method, multiple tasks can be predicted at the same time, the multi-task prediction capability of the model is improved, the calculation cost and deployment cost are reduced, and the method has great advantages in an urban traffic comprehensive scene needing quick response.

The invention discloses a multivariate time sequence prediction method based on timestamps and multi-scale modeling, and the method comprises the steps: extracting features through multi-scale feature fusion, carrying out the convolution operation of an input time sequence through employing convolution kernels of different scales, extracting and splicing short, medium and long-term features, and carrying out the fusion through employing a multi-head attention mechanism. Therefore, local details and global trends in the time sequence are captured, and the modeling capability of the model on a complex time mode is improved; timestamps are coded into high-dimensional vectors, the high-dimensional vectors are mapped through an encoder, the complex relation between the timestamps is captured, and the prediction capacity of the model for periodic and tendency changes is enhanced; a self-defined loss function is adopted to carry out model training, so that the model robustness is improved, and meanwhile, relatively high prediction precision is kept; and applying the trained model weight to a test set. Through combination of multi-scale feature fusion, timestamp feature extraction and a self-defined loss function, the precision and robustness of multivariate time sequence prediction are remarkably improved.

The invention discloses a mine slope risk prediction method based on multi-modal deep learning, and relates to the technical field of mine slope risk prediction, and the method comprises the steps: Stepl: inputting multi-modal data, Step2: preprocessing effectiveness data, Step3: constructing a sample set, and Step4: realizing mine slope stability prediction based on a gating cycle unit GRU of a mutual information mechanism. According to the invention, a multi-modal data integration GRU network is adopted to carry out sequence modeling on fused data, more complex sequence features are extracted, slope safety event prediction analysis is completed, high-precision monitoring of a mine slope is realized, and the accuracy and comprehensiveness of a prediction result are improved; according to the method, the mutual information mechanism is applied to the GRU network to maximize the mutual information between the sample and the feature, the dependency between the feature and the label is improved, the prediction capability is guaranteed, and the generalization capability on new data is also guaranteed.

The invention relates to the technical field of hydraulic systems, in particular to a hydraulic and thermal fluid dynamic lubrication calculation model and an optimization method, comprising the following steps: S1, modeling a multi-scale heterogeneous system; s2, carrying out heat-friction-viscosity synergistic effect modeling; and S3, dynamic boundary lubrication and lubricant behavior prediction. According to the method, by integrating nanoscale, mesoscale and macroscopic level lubrication system modeling, physical phenomena on different scales and interaction of the physical phenomena are comprehensively captured, and the step covers multiple aspects of multi-scale heterogeneous system modeling, nanoscale lubrication behavior modeling, macroscopic fluid field and structure field coupling and the like; high-precision predictive capacity and systematic thermal management are achieved by combining multiple physical fields such as heat, friction, viscosity and fluid dynamics, through collaborative modeling of multiple scales and multiple physical fields, the accuracy and stability of a lubricating system are improved, a solid theoretical basis is provided for optimization design, and the method is suitable for popularization and application. And stable operation in high-load and high-temperature environments is ensured.

The invention belongs to the technical field of power grids, and particularly relates to a power grid equipment safety analysis model optimization method based on an intelligent test tool. According to the method, historical data and real-time data are collected and analyzed, existing information resources are fully utilized, the accuracy of fault and risk identification is improved, a Bayesian optimization method is utilized to construct and optimize an equipment safety analysis model, the fault and abnormity prediction capability is improved, and through risk assessment and an intelligent test tool, the fault and abnormity prediction efficiency is improved. Key equipment needing to be optimized is accurately positioned, and the utilization rate of maintenance resources is increased; and through a Bayesian optimization method, the model is continuously updated and perfected based on new data, and it is ensured that the prediction capability of the model is continuously improved along with time lapse.

The invention provides a biodiversity influence comprehensive evaluation method and system based on multi-source ecological factors, and relates to the technical field of biodiversity analysis, and the method comprises the steps: obtaining multi-source ecological factor data of a geological disaster control influence region, and carrying out the maximum likelihood estimation and kernel function learning processing of the multi-source ecological factor data to obtain a space-time intensity field; performing adaptive time-frequency decomposition processing based on the time-space intensity field and the time sequence of the community functional traits to obtain a multi-scale modal set; performing low-rank dynamic recognition processing on the multi-scale modal set to obtain an approximate operator and control gain pair; performing sparse regression modeling processing to obtain a nonlinear state equation set; and carrying out distribution displacement measurement processing based on the nonlinear state equation set and preset quadrat baseline distribution to obtain a comprehensive evaluation result of the biodiversity influence. According to the method, the problems of single index, rough modeling and lack of multi-scene prediction capability in the prior art are solved.

The invention provides a filling process prediction method and device based on a discrete element method and data driving, and relates to the technical field of bulk material forming, and the method combines the physical modeling advantage of the discrete element method and the powerful prediction capability of a data driving method. The filling process of the granular material under different process parameters is simulated through a discrete element method, key quality indexes are obtained to serve as input of a data driving model, and the process parameters serve as output for training. The model not only can accurately predict the filling quality, but also can adjust the process parameters in real time according to the target quality index, so that the intelligent optimization of the filling process is realized. According to the method, the calculation efficiency is greatly improved, the generalization ability of the model is enhanced, the dependence on empirical data is reduced, a more efficient and accurate solution is provided for the filling process in the industries of building materials, pharmacy, powder metallurgy and the like, the product quality is improved, the production cost is reduced, and the research, development and application of novel materials and equipment are accelerated.

The invention discloses a regional electricity consumption prediction method based on a hybrid model and a terminal. The method comprises the following steps: acquiring historical electricity consumption data of a target region; performing data preprocessing according to the historical power consumption data to obtain an original data set; performing data processing on the original data set through an empirical mode decomposition model to obtain a combined data set; constructing an electric quantity prediction model based on a convolutional neural network model and a dual echo state network model, and training the electric quantity prediction model according to the combined data set; and predicting the future power consumption of the target area through the trained power consumption prediction model. According to the method, the spatial feature extraction capability of the convolutional neural network and the time sequence feature prediction capability of the echo state network are effectively fused, so that the electric quantity prediction model can effectively learn spatial features and time features from historical power consumption data; according to the electric quantity prediction model, the prediction precision is effectively improved.

The invention discloses a chromium-molybdenum-vanadium steel creep life prediction method based on data driving, which comprises the following steps: S1, data acquisition: converting the acquired data into a structured digital format, and manually marking the data; s2, performing sample generation by adopting a hyperbolic space embedding-based generative adversarial network algorithm, and further expanding the data; s3, adopting a neural network algorithm based on dynamic group evolution optimization as a feature extraction model, and performing feature extraction model training on the expanded data; s4, adopting a self-encoding neural network based on feature refinement as a feature dimension reduction model, and performing feature dimension reduction model training on the data after feature extraction; s5, adopting a fractional order neural network based on quantum state transfer as a classifier model, and performing classifier model training on the data after dimension reduction; and S6, processing and predicting a new sample by using the completely trained model. According to the method, the prediction capability of the creep life of the material is improved.

The invention relates to the technical field of intelligent operation and maintenance of ship power equipment, in particular to a DLinear-Transform-based anomaly detection method for a marine power generation steam turbine system, and the method comprises the steps: S1, carrying out the collection and processing of non-stationary time series data of the marine power generation steam turbine system; s2, constructing a DLinear-Transform model framework, and carrying out the construction of the DLinear-Transform model framework; s3, carrying out anomaly detection on the marine power generation steam turbine system on the basis of DLinear-Transform; and S4, verifying the performance of the model public data set. According to the method, a DLinear-Transform fusion framework is constructed, and anomaly detection of the key parameters of the ship power generation steam turbine system under the non-stable working condition is realized based on the DLinear linear characteristic and the Transform context modeling capability. Through SMD, SWaT and actual marine power generation steam turbine data set verification, the method is superior to existing optimal models such as TransAD and the like. Compared with an existing ship intelligent operation and maintenance technology, the system equipment anomaly detection and fault prediction capability is improved, and a new technical architecture is provided for predictive maintenance of ship power equipment.

The invention relates to the technical field of concrete quality control, in particular to a concrete strength prediction system and method based on SVR. According to the invention, multi-dimensional data of rebound test, compressive strength test and carbonization depth value measurement are collected through the data acquisition module, rich input characteristics are provided for the model, and the prediction capability of the model is enhanced; and abnormal data is eliminated by adopting a Grubbs criterion, so that the quality and consistency of input data are ensured, and the influence of noise on model training is reduced. According to the method, a concrete rebound curved surface regression equation and a support vector regression model are established respectively and serve as prediction criteria, data fusion is carried out through a Kalman filter, and the prediction precision and reliability are further improved. Through automatic and standardized data processing procedures including data acquisition, preprocessing, regression prediction, SVR prediction, data fusion and the like, rapid prediction of the concrete strength is realized, and the prediction efficiency is improved.

The invention discloses an asset use analysis method and system based on an intelligent prediction and optimization algorithm, and aims to solve the problems of low efficiency, insufficient prediction capability and unscientific purchase decision in existing asset management. According to the invention, asset use records, maintenance information and purchase history are obtained through the data acquisition module; constructing an asset use prediction model by using time sequence prediction, regression analysis and machine learning algorithms, and predicting future asset demand quantity and use trend; generating an optimized purchase plan and an asset allocation scheme in combination with an inventory state, an asset life cycle and a budget constraint; and the prediction result and the optimization suggestion are displayed through a visual tool, and support is provided for decision making. According to the invention, efficient planning and purchase optimization of asset use are realized, and the method can be widely applied to the field of enterprise asset management and supply chain optimization.

The invention provides a cross-domain equipment fault diagnosis method and system based on large and small model cooperation, and relates to the technical field of equipment fault diagnosis. According to the method, the causal field generalization structure is introduced into the small model, explicit decomposition is carried out on the stable causal law and the field specific difference, and meanwhile, the causal field generalization structure is corrected by using the large model, so that the small model can automatically identify and retain the causal relationship which is universally applicable to each device and each field; therefore, the influence of inter-domain distribution difference is effectively eliminated. Theoretical analysis shows that the generalization error of the model mainly depends on the accuracy of the stable causal item, and the structure can minimize error drift caused by distribution drift. Therefore, the robustness of health state evaluation and fault prediction can be remarkably improved in a cross-domain scene, and the fault diagnosis model can still keep the prediction capability close to the training domain level under the condition of no target domain annotation data.

The invention belongs to the technical field of power equipment monitoring, and particularly relates to a transformer load remote monitoring and detecting system. Comprising a data acquisition module, an edge calculation module, a load adjustment module, a load prediction module and a load analysis module. The data acquisition module acquires transformer secondary side load power, current, vibration signals and winding temperature data; the edge calculation module processes load early warning including data denoising and the like; the load adjusting module calculates and adjusts the primary side load power; the load prediction module constructs a model to predict future primary side load power; the load analysis module analyzes the future load condition according to the prediction result. The system realizes remote real-time monitoring, improves data processing precision, load adjustment accuracy and prediction capability, and ensures safe and efficient operation of the transformer.

The invention relates to the technical field of data processing, in particular to a high-pressure water pump pressure compensation method based on real-time data analysis, which establishes a prediction model of input water pressure change based on historical water pressure data, and can effectively capture a long-term dependency relationship in time sequence data. Therefore, the prediction precision of the future water pressure change trend is improved, the high-precision prediction capacity enables a high-pressure water pump control system to make adjustment in advance, more accurate pressure compensation is achieved, and the prediction precision of the water pressure change trend can be improved according to an error signal between real-time input water pressure data and a preset pressure target value. The rotating speed of the driving motor and the opening degree of the mechanical valve are dynamically adjusted to ensure that the actual output water pressure is always close to a target value, so that efficient closed-loop control is realized.

The invention discloses a city weather simulation prediction method based on an improved mesoscale WRF model, and the method comprises the steps: obtaining geographic space information data, carrying out the preprocessing, coupling local climate partitioning and population density grading, carrying out the detailed division of a built earth surface, carrying out the classification conversion of a natural earth surface according to a butt-joint mapping relation, and carrying out the prediction of the natural earth surface. The method comprises the following steps: obtaining a land cover classification data set and calculating to obtain corresponding urban canopy parameters, then improving default underlying surface data built in a mesoscale WRF model, and performing simulation prediction on urban meteorology by using the improved mesoscale WRF model. According to the method, through detailed division of the built-up surface and classification conversion of the natural surface, a surface coverage fine classification data set conforming to the real condition and considering the complex non-uniformity of the built-up area and corresponding city canopy parameters are constructed, then default underlying surface data built in the mesoscale WRF model are improved, and the city meteorological simulation prediction capability is remarkably improved.

The invention discloses an electric power meteorological prediction method and device based on adaptive enhancement, and relates to the technical field of electric power meteorological prediction, and the method comprises the following steps: collecting historical meteorological data and electric power data, and cleaning the collected data; according to the method, a model reasoning path tracing module is constructed based on Chain-of-Though logic, a prediction result is decomposed into subtasks such as meteorological factor contribution degree and power system response logic, interpretability of the prediction process is achieved, reliability of a generated result is evaluated by adopting sequence probability confidence, a visual interpretation report is dynamically generated, and the prediction efficiency is improved. According to the method, operation and maintenance personnel can visually acquire key information, sensitivity of the model to key meteorological events is optimized through adversarial training, prediction capability of the model to extreme meteorological conditions is improved, the trust problem of the black box model in power decision is finally effectively solved, a visual and reliable decision basis is provided for the operation and maintenance personnel, and the operation and maintenance efficiency is improved. And the operation safety and stability of the power system under the complex meteorological condition are improved.