14results about How to "Achieve high-precision forecasting" patented technology

This invention discloses a method for predicting the remaining life of brake friction pads, belonging to the field of brake life prediction. The method includes the following steps: S1, obtaining an initial effective feature set based on Pearson correlation coefficient and strong wear correlation features; S2, constructing a multi-physics coupled nonlinear wear degradation model for the friction pads and outputting multi-physics degradation parameters; S3, obtaining the optimal low-dimensional feature set through dual-source fusion; S4, optimizing the hyperparameters of the hybrid basis learner group using a dynamically perturbed improved gray wolf algorithm; S5, constructing a weighted fusion remaining life prediction model and completing accuracy evaluation and error correction; S6, online adaptive updating to achieve real-time remaining life prediction and life threshold warning. Using the above-mentioned method for predicting the remaining life of brake friction pads, high-precision prediction of the entire life cycle of brake friction pads through online dynamic adaptive updating is achieved, significantly improving prediction reliability and engineering application value.

This invention discloses an optimization method and system for a predictive model of diabetic complications. It involves collecting routine laboratory test data from diabetic patients, performing preprocessing such as missing value removal, label merging, and SMOTE resampling to construct a balanced dataset. The average importance of indicators is evaluated using multiple machine learning models to select a subset of key features. Random forest, XGBoost, support vector machine, and multilayer perceptron are selected as base classifiers, and after hyperparameter optimization, a stacked ensemble learning model is constructed to further improve predictive performance. Finally, the model is packaged as an API service and embedded into a hospital information system to achieve high-precision, low-cost, and real-time risk prediction of diabetic complications. Experimental results show that this invention achieves an accuracy of 98.5% and an AUC of 99.76% in predicting diabetic nephropathy complications, outperforming single models.

The method for predicting the service life of the coiled tubing under the cooperation of plasticity and damage is characterized in that the fatigue service life and the strain amplitude under the specific curvature radius are obtained through a full-size bending fatigue test, and the fatigue service life and the strain amplitude are substituted into a Manson-Coffin model to invert an unknown coefficient; the method comprises the following steps: determining mechanical parameters of a material in combination with a uniaxial tensile test, constructing a finite element mechanical model, carrying out finite element simulation on bending working conditions of a damaged coiled tubing on a roller and a guider, extracting strain amplitudes, and further establishing strain amplitude calculation models under different working conditions by adopting a Levenberg-Marquardt algorithm; coupling the model with a Manson-Coffin equation, and respectively constructing fatigue life prediction models of the roller and the guider; and on the basis of the Miner theory, three times of bending cycles with different curvatures in one trip are regarded as complete load cycles, an accumulated damage value is calculated, a correction coefficient is determined through low-cycle fatigue finite element simulation, a calculation model of the number of remaining trip times is established, and quantitative prediction of the remaining life is achieved. The method is suitable for the technical field of petroleum and natural gas drilling engineering.

The present application relates to artificial intelligence, deep learning, financial technology, time series prediction technology field, specifically refers to a kind of financial time series prediction method and system of frequency domain enhancement and morphology embedding, comprising: first, obtain stock market daily line level opening price, highest price, lowest price, closing price, transaction volume multidimensional K line data and do moving average smoothing;Again, reorder column dimension, extract K line morphology features by multilayer one-dimensional convolution and project into high-dimensional vector;In the training phase, the data is enhanced in the frequency domain, and only the high-frequency component is disturbed;The features are input into the iTransformer encoder to extract the time series features;Finally, the real prediction value is restored by adaptive inverse normalization.The present application realizes high-precision prediction of stock price by morphology embedding, frequency domain fine enhancement, adaptive inverse normalization and iTransformer modeling, eliminates the numerical cliff effect, improves the trend following and inflection point capturing ability, and enhances the overall robustness of the algorithm.

This invention discloses a multi-process optimization method for CNC grinding machines that balances accuracy and energy consumption, relating to the field of CNC grinding machine process optimization. The method establishes a first relational model characterizing the relationship between process parameters and axial error. This first relational model reflects the relationship between process parameters and the axial error of the CNC grinding machine during actual machining under those process parameters, enabling high-precision prediction of axial error in the actual machining process. A second relational model is also established to reflect the relationship between process parameters and grinding machine energy consumption. Furthermore, multi-objective optimization is performed with axial error and grinding machine energy consumption as objectives, achieving the goal of suppressing spindle axial error while improving resource utilization efficiency and reducing production costs.

The application relates to the technical field of artificial intelligence and intelligent decision-making, and provides a business process optimization and intelligent decision-making method and device.The method comprises the following steps: determining multi-modal fusion features according to multi-modal business data and a business scenario; inputting the multi-modal fusion features and timestamp data into a target time sequence fusion transformer model to obtain a business demand prediction result output by the target time sequence fusion transformer model; determining a target decision threshold according to the business demand prediction result and a preset threshold adjustment strategy; and determining a target business process decision scheme according to the multi-modal fusion features, the business demand prediction result and the target decision threshold.The business process optimization and intelligent decision-making method and device provided by the application can realize business process optimization and high-efficiency, high-precision intelligent decision-making.

This invention provides a method and apparatus for optimizing a filtered antenna. The method includes: obtaining the value range of the antenna structural parameters to be optimized in the target filtered antenna; iteratively optimizing the full-band electromagnetic response sequence of the target filtered antenna using a multi-objective optimization algorithm based on a pre-constructed multi-objective evaluation function and the value range, to obtain the optimal full-band electromagnetic response sequence of the target filtered antenna; and optimizing the target filtered antenna based on the target values ​​of the antenna structural parameters corresponding to the optimal full-band electromagnetic response sequence. The multi-objective evaluation function is determined based on a trained target multi-resolution Transformer surrogate model, which is trained based on sample value data corresponding to the antenna structural parameters and simulated values ​​of the full-band electromagnetic response sequence. This invention can achieve collaborative optimization of multiple performance indicators, with high optimization accuracy, short optimization time, and low optimization cost.

The invention relates to the technical field of new energy, and discloses an optical storage and charging integrated device energy management system, which comprises a data acquisition module used for acquiring a multi-source heterogeneous original data stream and carrying out preprocessing and state feature extraction; the prediction module is used for intelligently predicting photovoltaic output and charging load; the optimization scheduling module is used for carrying out multi-target energy optimization scheduling; the real-time control module is used for carrying out equipment power control and coordination operation; the health management module is used for performing energy storage health state evaluation and system performance analysis operation; the strategy learning module is used for acquiring operation scene features and carrying out control strategy learning optimization and operation effect verification; according to the invention, by constructing the multi-module collaborative intelligent energy management system, the photovoltaic consumption rate is improved, the operation cost is reduced, the adaptability and reliability are enhanced, and powerful technical support is provided for large-scale application of the photovoltaic storage and charging integrated energy station.

This invention discloses a method for constructing an intelligent hyperspectral inversion model for water environment prediction using unmanned aerial vehicles (UAVs), specifically relating to the field of water environment remote sensing monitoring technology. First, a spectral-spatial joint data matrix is ​​constructed, and an initial spectral feature matrix is ​​extracted. Based on spatial coordinate relationships, a water body neighborhood spectral coupling network is constructed, and the spectral perturbation propagation weights are calculated. The initial spectral feature matrix is ​​then trajectory-corrected, and a water environment parameter response function is constructed. A nonlinear mapping relationship between spectral features and water environment parameters is established, forming an inversion model. The spatial distribution results of water environment parameters are generated based on the inversion model, and local anomalous spectral indices are extracted through prediction residual neighborhood propagation analysis. When the anomalous index exceeds a threshold, the weights of the inversion model are adaptively updated, thereby outputting the final water environment parameter prediction results. This invention can effectively suppress the influence of spectral perturbations under complex water environment conditions, improving the accuracy and prediction stability of water environment parameter inversion.

The application provides a dynamic emotion recognition method and device based on symbolic music structure, and relates to the technical field of music information processing. The method comprises the following steps: extracting a chord sequence from input symbolic music data, and dividing the symbolic music data into multiple continuous structural segments; performing multi-channel feature extraction on each structural segment, inputting the obtained multi-dimensional structural feature sequence into an emotion recognition model, obtaining an emotion prediction value corresponding to each structural segment output by the emotion recognition model, arranging the emotion prediction value corresponding to each structural segment in sequence, and generating a dynamic emotion curve corresponding to the symbolic music data. The dynamic emotion recognition method and device based on symbolic music structure provided by the application take a slice mode based on structure perception and a continuous emotion regression modeling mechanism as cores, realize high-precision prediction of the evolution process of music emotion values over time, and enhance the structure alignment capability and time sequence modeling capability of the model for emotion trends.

This application provides a method for precise monitoring of water and salt dynamics in coastal saline-alkali land based on the HYDRUS model, belonging to the field of coastal saline-alkali land monitoring technology. This application first determines the soil salinity profile type and characteristic ions in different regions; then, it sets the soil profile stratification and initial water and salt transport parameters of the HYDRUS model in conjunction with surface vegetation information. Meteorological and groundwater data are used as boundary conditions to run the model, simulating the conductivity and characteristic ion concentration at different depths of each point over time. This data is input into a two-dimensional convolutional neural network to extract spatiotemporal state features; an adjacency graph is constructed based on geographic location, and adjacent node information is fused through a graph neural network to optimize the spatiotemporal features into global state features reflecting spatial interaction. Finally, by combining the correlation between characteristic ions and total salt, salinization prediction information is generated, achieving precise monitoring. This method integrates water and salt transport simulation with spatiotemporal deep learning to achieve high-precision prediction of water and salt dynamics in coastal saline-alkali land.

The application provides a charging pile cluster dynamic scheduling method and system based on edge agent cooperation, and relates to the technical field of intelligent management of charging facilities. The problems of dependence of the existing system on centralized decision and stable communication, poor robustness in extreme environment, lack of accurate battery aging evaluation and multi-target cooperative optimization capability are solved. The method comprises the following steps: obtaining multi-dimensional monitoring data; dividing the cluster into monitoring and scheduling units, and calculating the port health index; screening high-priority units according to the health index, emergency discharge capacity and battery aging confidence interval; switching the scheduling mode according to the communication state and weather warning, and optimizing the scheduling by using a multi-agent cooperative algorithm; and generating and executing the scheduling instruction. The application improves the scheduling robustness, equipment life and overall efficiency in extreme scenarios.

This invention discloses a method for predicting the melting settlement of ice-rich permafrost, relating to the fields of geotechnical engineering and disaster prevention and mitigation in cold regions. The method includes: constructing a layered parameter model of the permafrost; vectorizing the temperature field based on Fourier's law and the sensible heat capacity method; and automatically identifying the soil type based on an initial void ratio threshold. For ordinary permafrost, a large-strain consolidation model is used to calculate settlement; for ice-rich layers, a volumetric collapse model is used to directly convert the melted ice volume into settlement. The method combines Lagrange dynamic mesh technology to update node coordinates and physical parameters in real time, achieving bidirectional coupling of thermodynamic parameters, and introduces adaptive time step control to improve computational efficiency. This invention, through the coupling of consolidation and collapse mechanisms, solves the problems of inaccurate settlement prediction and low computational efficiency in traditional methods for high-ice-content permafrost. It can accurately simulate the settlement deformation of complex strata such as ice wedges and ice lenses under long-term thermal action, and is suitable for stability assessment and disaster prevention in cold region engineering.

This invention discloses a CFD-DEM coupled modeling and verification method for paddy field soil based on JKR surface energy calibration. The steps include: initially estimating the range of JKR surface energy parameters between soil particles through soil tilt angle tests, and then accurately determining the optimal JKR surface energy value from the range of JKR surface energy parameters by combining soil firmness tests; establishing a two-way coupled soil CFD-DEM model based on EDEM discrete element software and Fluent fluid dynamics software; simulating the weeding wheel cutting process and the direct shearing process of the soil using the two-way coupled soil CFD-DEM model; and evaluating the accuracy of the model based on the average relative error and static shear stress error. This CFD-DEM coupled modeling and verification method for paddy field soil solves the key bottleneck in modeling high-moisture-content cohesive paddy field soil, achieves high-precision prediction, and provides a highly reliable technical means for the accurate simulation of agricultural machinery-soil interaction processes.