328 results about "Prediction interval" patented technology

The invention relates to the technical field of cold-chain logistics monitoring, and discloses a temperature prediction method and system for a multi-modal AIGC cold-chain logistics monitoring platform, and the method comprises the steps: carrying out the multi-modal data collection and data fusion of a cold-chain logistics transportation process based on a cold-chain logistics monitoring platform, and obtaining a multi-modal fusion data matrix; inputting an AIGC model to carry out feature extraction and inter-modal feature weight calculation to obtain a multi-modal feature importance vector; performing cross-modal weighted fusion calculation and associated feature map construction to obtain a temperature state feature sequence; inputting a multi-modal deep belief network to carry out probability distribution modeling to obtain a temperature prediction probability distribution sequence; performing time sequence characteristic analysis to obtain a temperature prediction interval sequence; and performing dynamic threshold judgment to obtain a multi-stage temperature early warning trigger signal. According to the invention, the temperature prediction accuracy of the cold-chain logistics monitoring platform is improved.

The invention relates to the technical field of unmanned aerial vehicles, and discloses an unmanned aerial vehicle battery endurance flight capability prediction system, which comprises a multi-dimensional data acquisition module, a feature mapping module, a prediction module, an optimization module and a feedback optimization module, and can be additionally provided with an early warning module. The multi-dimensional data acquisition module acquires battery data and cleans the battery data to generate standardized data; the feature mapping module maps the data to a feature space, and generates a feature sequence cluster containing a multi-dimensional association relationship by using a time sequence segmentation algorithm; the prediction module divides prediction intervals based on a support vector machine algorithm and extracts prediction indexes; the optimization module generates an endurance prediction strategy by predicting and optimizing the network model; and the feedback optimization module performs multi-source data fusion optimization and outputs a prediction instruction. The early warning module can associate the prediction instruction with the battery health degree, output a grading early warning signal and trigger a response mechanism.

The invention provides a new energy output prediction method and system based on Monte Carlo Dropout, and the method comprises the steps: carrying out the probabilistic prediction of new energy output through a Monte Carlo Dropout technology, generating a dynamic prediction result containing a confidence interval, and quantifying the uncertainty of meteorological sudden change and equipment state; secondly, constructing a multi-stage random dynamic programming model, discretizing a prediction interval into multi-scene input, designing a non-linear objective function based on the discharge depth, and synchronously optimizing the electricity purchase cost and the energy storage aging cost; and finally, realizing rolling optimization of the system in combination with a model prediction control framework, and dynamically adjusting an energy storage aging cost weight by updating prediction data and a scheduling instruction on line and embedding an energy storage health state real-time feedback mechanism. The photovoltaic and wind power consumption rate can be remarkably improved, the full life cycle cost of an energy storage system is reduced, and meanwhile, the robustness of a scheduling strategy in extreme weather is ensured.

The invention relates to the technical field of collaborative optimization, in particular to a virtual power plant power generation-utilization-price collaborative optimization system based on an AI large model, and the system comprises a load confidence matching module, a resource stability mapping module, a source-load capacity coupling module, an electricity price interval adjustment module and a comprehensive regulation and control linkage module. According to the method, the confidence interval prediction of the load demand is realized based on the hybrid neural network modeling of the load behavior data and the equipment temperature control characteristic sequence, and the scheduling matching confidence is measured according to the boundary overlapping condition of the prediction interval and the power generation response characteristic; a stability screening mechanism for adjusting resources is constructed in combination with the output fluctuation ratio and the equipment inertia characteristic, the controllability of load adjustment and the real-time performance of source side response are improved, the price adjustment rhythm is corrected through an electricity price response delay factor, dynamic closed-loop linkage between load adjustment and price guidance is achieved, and the load adjustment efficiency is improved. The execution priority is dynamically updated under the condition that multiple response conditions are matched, and the certainty of resource scheduling and the sensitivity of response are improved.

The invention discloses an optimal scheduling method and system for a wind-light-hydrogen storage micro-grid, and relates to the field of wind-light-hydrogen storage micro-grids, and the method comprises the steps: obtaining multi-source time sequence data, and carrying out the combined denoising and dynamic time alignment to generate a standardized input sequence; adopting a quantile regression model fused with a space-time attention mechanism to output a power prediction interval of wind and light output and load demand in a future time period; constructing a layered multi-objective optimization model; an improved adaptive particle swarm optimization algorithm is adopted to solve the layered multi-objective optimization model, and an equipment scheduling instruction set is generated; and dynamic correction is carried out, and micro-grid instruction distribution is carried out according to the equipment response priority. According to the method, multi-target conflicts such as power balance, equipment loss and energy efficiency are effectively balanced through multi-source data efficient preprocessing, space-time joint prediction interval generation and hierarchical multi-target optimization, and efficient and stable operation of the wind-light-hydrogen storage micro-grid is achieved.

The invention relates to the technical field of land resource monitoring, in particular to a land resource dynamic monitoring and early warning method and system based on multi-source remote sensing data fusion, and the method comprises the steps: employing an unmanned plane to periodically collect optical images, SAR echoes and LiDAR point clouds, constructing a ground three-dimensional digital model, and carrying out the land parcel division; performing fusion to form a multi-dimensional feature vector, establishing an LSTM land parcel feature evolution model, and predicting a change rate interval of each feature in a current period based on a historical sequence; constructing a time sequence difference change detection algorithm, calculating a land parcel change rate, and screening potential abnormal land parcels by taking a prediction interval as an anomaly judgment threshold value; a double-branch convolutional neural network is adopted to identify crop states, growth stages and construction violation behaviors, abnormity is judged and determined, and confidence is given; spatial clustering is carried out on determined abnormal land parcels, accurate boundaries are obtained in combination with a three-dimensional model, multi-level early warning information is generated, and the decision-making efficiency and response speed of land resource monitoring are improved.

The invention provides a short-term power load prediction method and system considering multi-dimensional feature extraction, and the method comprises the steps: obtaining regional loads and historical data of related weather of regional influence loads, and dividing a data set; extracting multi-time-scale time sequence features, including trend features, week features and annual features, in the regional load sequence, and constructing an internal feature set; key dependent meteorology with load correlation higher than a set value in the historical meteorological data is searched by using the maximum mutual information coefficient, and an external feature set is constructed; constructing a prediction model fused with a self-attention mechanism, taking the internal feature set and the external feature set as the input of the model, analyzing a prediction value and an actual value to obtain a conditional probability density function, and calculating load condition prediction error probability distribution by taking the load prediction value of the last part of data as the input of the conditional probability density function, and determining a load prediction interval. The prediction result is high in precision.

The invention provides a photovoltaic output prediction method and a distributed optical storage transformer area active / reactive support capability optimization method, and belongs to the technical field of power system automation. Comprising the following steps: constructing a prediction model based on meteorological information, extracting multi-scale features of meteorological data, reducing dimensions through kernel principal component analysis, predicting photovoltaic power through a long-short-term memory network, quantifying the uncertainty of the photovoltaic power, and forming a prediction interval. Based on a photovoltaic power prediction result, a robust nonlinear programming model is established, the sum of absolute values of active / reactive power transmitted to a main power grid by a light storage area is maximized as a target, and linear and nonlinear constraints such as energy storage operation, power balance, equipment capacity and power factors are considered at the same time. The robustness under the photovoltaic output uncertainty is ensured through a budget uncertainty set; the model is solved by adopting a sequential quadratic programming algorithm. According to the method, the photovoltaic power prediction precision is effectively improved, and organic combination of high-precision prediction and robust nonlinear optimization is realized.

The invention discloses a wind power prediction method, system and equipment based on multi-model fusion, relates to the technical field of wind power prediction, and realizes non-parametric probability prediction of short-term wind power prediction by combining a Pearson's correlation coefficient, a frost optimization algorithm optimization variational mode decomposition model, a CNN-ASSA-Informer model and quantile regression. Complex time series data are processed through an optimization algorithm, and the complex relation between different frequency components and wind power in the time series data is solved. And a corresponding prediction interval is constructed, so that the precision and confidence of wind power prediction are improved, and a more reliable wind power prediction result can be provided for regional wind power scheduling, renewable energy integration and an electric power market.

The invention relates to the technical field of data processing, in particular to an electron beam welding state detection method and system, and the method comprises the steps: obtaining a voltage data sequence in a welding process; interpolating missing data in the voltage data sequence by using an HTFE algorithm, calculating a correction value of a prediction interval error factor in the HTFE algorithm when each missing data is interpolated, and replacing the corresponding missing data by using the voltage data at the corresponding moment predicted based on the correction value to obtain a complete data sequence; and performing anomaly detection on the complete data sequence, and detecting the welding state based on an anomaly detection result. The welding state of the electron beam can be accurately detected.

The invention discloses a wind power multi-scale decomposition prediction method. At present, single-point prediction is not comprehensive and accurate enough, and cannot adapt to quantitative accurate requirements of a wind power plant and a power grid dispatching mechanism in risk management. The method comprises the following steps of: forming an original wind power sequence from actually acquired wind power data, sequentially performing feature selection and data decomposition processing to form multi-scale modal data, and constructing a depth prediction model according to the multi-scale modal data; a probability prediction interval determination process is completed in the residual error distribution mode depth prediction model through adaptive bandwidth kernel density estimation; after actually obtained wind power data form an original wind power sequence, an initial model is established, feature selection processing is performed on the initial model, that is, weighted marginal contribution is calculated for each feature of the initial model according to all involved feature subsets by using an SHAP algorithm based on a Shapley value in a game theory, and the weighted marginal contribution of each feature of the initial model is calculated; and completing a feature data acquisition process of accurately quantifying interdependence and interaction effect between features.

The invention provides a mining area surrounding soil heavy metal spatial distribution inversion method based on hyperspectral data, and the method comprises the steps: processing multi-source monitoring data of mining area surrounding soil, and obtaining a consistent reflectivity data set; performing soil spectrum purification based on the consistent reflectivity data set to obtain a pure soil signal; spectrum key features are screened out from the pure soil signals; obtaining a modeling data set in combination with the spectrum key features and the heavy metal concentration labels so as to construct a multi-task inversion model, and outputting each metal prediction interval and an over-standard risk probability graph; and outputting a multi-layer package based on the model, performing global and local interpretation and mechanism verification, and generating an interpretation report and traceable evidence. According to the method, full-link unification and purification can be achieved, domain deviation and mixed pollution are remarkably reduced, the accuracy and interpretability of feature screening can be considered, and the model generalization ability, prediction accuracy and space credibility are improved.

The invention discloses a rope-driven mechanical arm motion control method based on model predictive control, which comprises the following steps: taking a joint angle error as a state variable, taking a rope-driven rotation angle as an input variable, and establishing a nonlinear state-space equation of a rope-driven mechanical arm according to the state variable and the input variable; the nonlinear state-space equation is subjected to linearization processing, a linear approximation model is generated, and the linear approximation model comprises a linearization matrix; defining a prediction interval, expanding the linear approximation model into a multi-step prediction form, and generating a prediction model; constructing a cost function, and combining the prediction model into the cost function to solve and obtain an optimal control sequence; and extracting the optimal control increment at the current moment from the optimal control sequence, and updating the state variable, the input variable and the linearization matrix according to the optimal control increment to form closed-loop control. Accurate path tracking control can be carried out on the rope-driven mechanical arm under complex dynamic constraints, and it is ensured that the mechanical arm moves according to the planned trajectory.

The invention relates to the technical field of ship energy consumption prediction, in particular to a ship main engine energy consumption prediction method and device based on uncertainty estimation, and the method comprises the steps: receiving historical navigation data of a target ship, dividing the historical navigation data into a training set and a test set, and carrying out the standardization processing; training the base model by using the training set and generating a residual sample; taking a prediction result of the base model as a new input feature, and training a linear regression meta-model to construct a Stacking integrated model; calculating quantiles according to the residual samples and constructing a prediction interval; calling an MAPIE library to automatically realize a CV + conformal prediction process, and automatically optimizing hyper-parameters of the base model by using Optuna to obtain an optimal model parameter combination; and outputting the energy consumption predicted value of each target ship and the corresponding confidence interval. According to the method, the stability and reliability of the ship energy consumption prediction result can be improved, quantitative expression of the confidence coefficient of the prediction result in an energy efficiency management system is realized, and the practicability and decision support capability of the method in scenes such as intelligent shipping scheduling are enhanced.

The invention discloses a property energy consumption abnormity monitoring method and system based on big data analysis, and relates to the technical field of energy consumption abnormity management. According to the property energy consumption abnormity monitoring method based on big data analysis, in a specified energy consumption monitoring time period, an energy consumption power group of a specified property service area is firstly obtained, and the energy consumption power group is compared with an energy consumption prediction interval to screen energy consumption abnormity; according to the comparison result, energy consumption abnormal data in the abnormal energy consumption time period are repeatedly checked, transient normal fluctuation caused by normal working condition switching of equipment or sudden change of the external environment is eliminated, and repeated checking comprises peak value abnormal detection and recheck verification; and finally, after repeated inspection is completed, association verification is carried out in combination with the abnormal energy consumption feature set. According to the method, a scientific basis is provided for property energy consumption management, accurate positioning of energy consumption abnormal points is facilitated, the problem of abnormal identification misjudgment caused by transient change in energy consumption monitoring in the prior art is effectively solved, and energy conservation, consumption reduction and cost optimization are realized.

The invention provides a video abnormal behavior positioning method and device based on a multi-modal large model, and a medium, and relates to the field of computer vision. The method is applied to computer equipment and comprises the following steps: acquiring a to-be-processed video, and determining a plurality of query videos and a plurality of support videos; obtaining a thinking chain text according to the visual language model; obtaining a first robust feature representation and a second robust feature representation according to a semantic time sequence pyramid encoder; obtaining subtitle text fusion features according to a text encoder; and further performing alignment processing to obtain an alignment similarity matrix so as to determine a prediction interval of the query video, determine an abnormal behavior positioning result of the to-be-processed video and determine a dynamic model fine tuning sample. The video abnormal behavior positioning method and device solve the technical problems that in the prior art, visual information is excessively depended on, deep semantic understanding of the action context and the target relation is lacked, misjudgment is easily generated in the alignment process, the positioning precision is insufficient, and the video abnormal behavior positioning efficiency is low.

The invention discloses a power load prediction method, and the method comprises the steps: firstly solving an extreme event data sparsity problem through a generative adversarial network, and constructing an event time sequence library through a time sequence anomaly detection algorithm; then analyzing the causal relationship between the event and the load by applying a causal discovery algorithm, and converting prediction output into probability distribution by adopting a Bayesian neural network to quantify uncertainty; constructing a prediction model triggered by an event, and generating a multi-time scale probability prediction interval; and finally, generating a multi-scene prediction result through Monte Carlo simulation, quantifying the system recovery capability in combination with a toughness index, and integrating the system recovery capability to a decision support system to generate a risk response scheme. According to the method, the accuracy and robustness of load prediction under the extreme climate are remarkably improved, full-chain risk insight from early warning to recovery is realized, and prospective decision support is provided for safe operation of a power system.

The invention relates to the technical field of industrial equipment state monitoring and fault diagnosis, in particular to a trend fault prediction method based on cooperation of a dynamic mode and a threshold value. According to the method, a theoretical prediction interval dynamically changing along with a load is generated in real time by establishing nonlinear mapping between working conditions and key parameters, parameter drift interference caused by working condition fluctuation is effectively eliminated, a real-time health baseline of equipment is quantified in combination with maintenance records, and the width of an early warning threshold value is cooperatively adjusted according to feature similarity and the health level. Self-adaptive monitoring of different aging stages of the whole life cycle is realized, mode matching is performed by utilizing multi-dimensional feature vectors and fusing physical field information, deviation severity and form similarity are comprehensively evaluated through fuzzy reasoning, abnormity is locked in advance according to high feature goodness of fit when a numerical value does not seriously exceed a limit, and a real-time monitoring result is obtained. Early weak symptoms are accurately captured, abnormal sources are output, and the diagnosis precision under variable working conditions is remarkably improved.

The invention relates to the technical field of civil engineering, and discloses a road subgrade settlement prediction method and system based on multi-source data. The method comprises the following steps: acquiring and correcting a precipitation sequence and roadbed soil moisture distribution data; analyzing rainfall accumulation characteristics and carrying out penetration risk classification, thereby extracting a penetration depth estimation value, and utilizing finite element analysis to simulate a soil body saturation state change trend; quantifying a soil body supporting capacity reduction range, constructing and calibrating a settlement initiation probability calculation model, and obtaining settlement probability distribution; identifying a high-risk evolution area, integrating path weights and accumulated influence factors, and generating a settlement prediction interval subjected to multi-dimensional verification; and fusing the prediction interval and actual roadbed structure data through a geographic information system to generate a comprehensive evaluation report. According to the method, the whole-process accurate evaluation of the roadbed settlement risk from multi-source perception, mechanism simulation to space prediction is realized, and the accuracy and timeliness of the roadbed settlement risk prediction and the pertinence of engineering maintenance are improved.

The invention relates to a power fee accounting method based on intelligent management, and the method comprises the following steps: S1, obtaining multi-source heterogeneous data, carrying out the preprocessing, and carrying out the collection, data standardization, time calibration and automatic tagging of the multi-source heterogeneous data, and obtaining a standard structured data stream; s2, constructing a dynamic accounting rule data set based on the standard structured data stream and transaction data, market rules, contract terms, regions and type price standards; s3, constructing an intelligent prediction model based on LSTM and ProphetAI models, automatically detecting reasonability and correcting a prediction interval according to the dynamic accounting rule data set, and obtaining a pre-accounting result set; s4, according to the pre-checking result set, in combination with the price standard and the amount fee strategy, configuring subsidy and refunding policies and parameter processing, automatically matching a settlement unit and a settlement rule, and flexibly applying different checking templates according to distributed, direct-purchase and non-direct-purchase power plants to obtain a checking intermediate table; and S5, generating a final settlement packet according to the accounting intermediate table and the user configuration parameters.

The invention discloses a current transformer state analysis monitoring method based on interval modeling, and the method comprises the steps: collecting the ratio error data of an electronic current transformer, carrying out the preprocessing, and dividing the preprocessed data into a training set, a test set, and a verification set; the method comprises the following steps of: constructing a hierarchical deep learning model BITCN-BiLSTM-MHA; the preprocessed data are input into the constructed hierarchical deep learning model BITCN-BiLSTM-MHA, and model parameters after training are obtained; according to the trained model, measuring errors of the current transformer at different moments in the future are predicted, and deterministic prediction results of different time periods are generated; and according to the prediction result, carrying out probability distribution modeling on the prediction result by using an adaptive window width kernel density function, and generating prediction intervals of different confidence intervals. According to the method, in a dynamic complex signal environment, the variation trend of the measurement ratio error interval of the electronic current transformer in the future can be well predicted.

The invention relates to the technical field of power system automation, and discloses a power distribution network operation risk assessment regulation and control method based on a graph attention network and reinforcement learning, and the method comprises the steps: mapping a topological structure of a power distribution network into a graph structure; constructing a node feature matrix and an adjacent matrix based on the node features and the edge features; inputting the node characteristic matrix and the adjacent matrix into a short-circuit current prediction model based on a graph attention network (GAT), and outputting short-circuit current prediction values of a bus and a line in the power distribution network and corresponding short-circuit current prediction intervals after the distributed power supply is connected to the grid; constructing a reward function aiming at reducing the short-circuit current risk and guaranteeing the load power supply, and training a strategy network and a value function through a SoftActor-Critic algorithm so as to obtain a reinforcement learning model; and collecting the state of the power distribution network in real time in a preset scheduling period to realize dynamic regulation and control of the short-circuit current risk of the power distribution network. According to the method, topology and electrical association is accurately captured, and the prediction precision is improved.

The invention provides a training method of an energy consumption probability prediction model, an energy consumption probability prediction method, computer equipment and a readable storage medium, and relates to the technical field of building energy management and intelligent prediction. The method comprises the following steps: performing feature extraction on historical building data of a target building in a preset historical time period to obtain sample energy consumption related features; constructing a sample data set according to the sample energy consumption related characteristics and the real energy consumption probability; dividing the sample data set into a plurality of sub-data sets; respectively training the plurality of initial energy consumption prediction models according to the plurality of sub-data sets to obtain a plurality of trained energy consumption prediction sub-models; and according to the plurality of trained energy consumption prediction sub-models, constructing a target energy consumption prediction model. According to the method, the target energy consumption prediction model is obtained according to the multi-sample energy consumption related features, so that the prediction interval coverage degree is high, the method adapts to dynamic load changes, and then the accurate prediction requirement for building energy consumption is met.

The invention provides a charging remaining time prediction method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring an initial temperature, an initial state of charge and an initial charging rate of a target battery at an initial moment; the initial temperature, the initial charging rate and the sum value of the initial charge state and the preset charge state step length are used as input, and the temperature of the target battery at the end point of each prediction interval is predicted through a temperature rise prediction network; the preset state-of-charge step length is used for indicating the state-of-charge variation of the adjacent prediction intervals; determining the charging rate of the target battery in each prediction interval from the mapping relation between the temperature and the charge state of the target battery and the charging rate; the preset charge state step length is compared with the charging rate of the target battery in the prediction interval, and the charging time of the target battery in each prediction interval is determined; and determining the sum value of the charging durations of all the prediction intervals as the charging remaining time of the target battery.

The invention discloses an automatic driving vehicle risk assessment method and device, electronic equipment, a readable storage medium and a computer program product. The method comprises the following steps: collecting historical motion state information and road structure information of a vehicle and surrounding vehicles; fusing the historical motion state information to obtain a fused context feature vector; performing trajectory prediction according to a pre-constructed vehicle trajectory prediction module and the context feature vector to obtain a surrounding vehicle prediction trajectory and a self-vehicle prediction trajectory; according to the surrounding vehicle prediction trajectory, the own vehicle prediction trajectory and the road structure information, calculating a quantitative perception risk time sequence of preset look-ahead time of the own vehicle in the prediction interval; and performing fusion perception risk calculation according to the quantitative perception risk time sequence to obtain a potential risk assessment result of the vehicle in the driving scene. According to the method, the problem of inaccurate evaluation result caused by difficult parameter adjustment in a multi-vehicle interaction scene of an existing risk evaluation method can be solved.

The invention discloses an intelligent fusion terminal power quality monitoring evaluation data processing method and system. The method comprises the following steps: dividing a plurality of intelligent fusion terminals into consensus evaluation groups according to a power grid topological structure, and allocating reputation values to the terminals; and each terminal periodically measures the electric energy quality parameter and generates a prediction interval of the group consensus value in the current evaluation period. And then, according to the actual measurement value and the reputation value of each terminal, a group consensus value is determined by calculating a weighted cumulative weight. And dynamically adjusting the reputation value of each terminal by comparing the group consensus value with the prediction interval of each terminal. And finally, outputting the group consensus value and the updated reputation value, and generating an electric energy quality comprehensive evaluation result based on the group consensus value. Meanwhile, each terminal adaptively updates a measurement error compensation model in the terminal according to the deviation between a self-measurement value and a group consensus value. According to the invention, the credibility of the power quality monitoring data and the accuracy of the evaluation result are effectively improved.

The invention provides a resource index prediction method and device based on a Transform framework and a medium, and the method comprises the steps: carrying out the pre-training of a PatchTST model through employing the historical data of resource indexes, such as a CPU, a memory, and the like, improving the future values of the prediction indexes of a conventional PatchTST model into the future distribution of the prediction indexes, obtaining a prediction mean value and a prediction variance, and carrying out the prediction of the prediction mean value and the prediction variance. And further converting into prediction value output, prediction interval upper bound output and prediction interval lower bound output. Compared with a traditional time sequence prediction method based on a Transform architecture, the method has the advantages that the prediction accuracy is high, the training time is short, the upper and lower bounds of a prediction interval can be output, and the predicted upper and lower bounds can be used for representing a confidence interval of a predicted value. According to the invention, richer and more accurate guidance can be provided for intelligent operation and maintenance scenes such as dynamic capacity expansion and contraction, capacity management, capacity planning and the like.

The invention discloses a fuel cell performance degradation prediction method based on deep learning, and the method comprises the steps: collecting the historical operation data of a fuel cell, carrying out the preprocessing of the data through a simulated annealing isolation forest algorithm and a Gaussian filtering algorithm, carrying out the feature selection through employing SHAP-TCN, extracting key information, and carrying out the prediction of the performance degradation of the fuel cell. According to the method, the prediction interval of performance degradation is constructed in combination with a DeformamleTST deep learning model and a quantile regression method, so that uncertainty in the prediction process is comprehensively captured, a partially enhanced optimizer algorithm is further improved through an adaptive regularization weighting method, and an improved IPRO is obtained to perform hyper-parameter tuning on the model. According to the method, the residual life prediction precision is improved, the problem of uncertainty in a traditional method is effectively solved, the method has high engineering application value, and theoretical basis and technical support are provided for predictive maintenance and management of the fuel cell.

The invention provides a sea attitude motion self-compensation method based on deep learning and model prediction, and the method comprises the steps: constructing a sea ship attitude prediction model based on deep learning, and predicting the attitude change of a ship in a future time period; designing a six-degree-of-freedom platform driving instruction resolving algorithm, and resolving a driving instruction of the six-degree-of-freedom platform according to the posture forecast value of the ship at the future moment; and based on a model prediction technology, designing an attitude compensation control model, and controlling the six-degree-of-freedom platform to realize attitude compensation according to the driving instruction. The prediction attitude compensation error in the prediction interval is used as a model prediction control instruction, a six-degree-of-freedom platform driving instruction most suitable for current instruction compensation is solved by adopting an optimization control algorithm, and accurate prediction of ship attitude change is realized through a deep learning technology. And the six-degree-of-freedom self-stabilization platform is efficiently controlled in combination with a model prediction control algorithm, so that the problems of response lag and insufficient control precision of an offshore self-stabilization platform in the prior art are effectively solved.

According to the data-driven electricity-to-water coefficient composite model and the prediction method provided by the invention, the spatial interpolation model, the neural network model and the regionalization prediction model are operated in parallel and are subjected to comparative analysis to form the composite model, so that the composite model can adapt to different characteristics of data, the prediction precision and generalization ability are remarkably improved, and the prediction efficiency is improved. The method is suitable for estimating the water conversion coefficient by electricity in different scenes. Wherein the EPC-STIN model (neural network class) and the regional prediction model are innovative architectures of converting electricity into water prediction models. Besides, a plurality of evaluation indexes are adopted, and the prediction result of the composite model is comprehensively evaluated through a comprehensive evaluation system, so that the prediction model with the optimal accuracy is output for different data sets, and it is ensured that a comprehensive score can reflect the balanced performance of the model in multiple aspects such as error accuracy and prediction interval reliability. And the scientificity of model selection is improved. Therefore, compared with the prior art, a more comprehensive evaluation basis is provided for electricity-to-water coefficient prediction, a user is helped to scientifically quantify the uncertainty of a prediction result, and the credibility of decision making is improved.