21results about How to "Hard to get" patented technology

The invention relates to a transfer learning-based power distribution network measurement data super-resolution sensing method. The method comprises the steps that a high-frequency time sequence is acquired and preprocessed, and training sample data is obtained; a cascade model is constructed based on a graph attention network and a transformer network, the training sample data and an incidence matrix are input into the cascade model, the cascade model is pre-trained, and the incidence matrix represents the topological connection relation of the power distribution network; the cascade model after pre-training is divided into a learning device part and a decision-making device part, transfer learning training is carried out, model parameters reaching a training target are reserved, a final super-resolution perception model is obtained, the learning device part comprises all layers except a transformer linear output layer, and the decision-making device part comprises a transformer linear output layer; and inputting real-time measurement data into the super-resolution sensing model to obtain a super-resolution measurement generation result.

The invention discloses a survey method and system for ocean subsurface vortex and a sub-mesoscale process thereof, and the method comprises the steps: obtaining three-dimensional temperature-salinity-depth flow data of a designated sea area, and drawing a vertical distribution diagram of temperature, salinity, density and a flow field; screening an equal-density surface according to the vertical density abnormal structure and the maximum flow velocity occurrence position; identifying the secondary surface vortex based on the determined potential height contour map and the vertical flow field map of the equal-density surface; acquiring satellite altimeter data of a specified sea area and drawing satellite orbit distribution diagrams of different dates; according to the identified subsurface vortex and the satellite orbit distribution map, carrying out route design and drawing a route; and drawing a site distribution map for hydrological survey according to the route distribution map. The invention aims to provide a scheme for identifying and positioning the subsurface layer vortex, investigating and analyzing the subsurface layer vortex sub-mesoscale process and designing an observation route aiming at the ocean subsurface layer vortex so as to solve the problem that the ocean subsurface layer vortex and the sub-mesoscale process thereof are difficult to position and observe.

The present invention provides a method and system for identifying an impedance model of a new energy power generation system based on a neural network. The training data set and the test data set of the new energy power generation system under different steady-state operating points are obtained by a frequency sweep method and normalized; the neural network is trained by the training data set to obtain a neural network with impedance characteristics of the new energy power generation system; the input data of the test data set is input into the neural network to obtain an impedance identification result; the mean square error is obtained in combination with the output data of the test data set, and the number of hidden layers of the neural network and the number of neurons in each hidden layer are adjusted so that the mean square error (MSE) is less than a set threshold; the input data under any steady-state operating point is input into the obtained neural network to obtain the impedance output under the corresponding operating point. The present invention can obtain the impedance characteristics of the power generation system under the corresponding operating point by only measuring the steady-state operating point data, and has high accuracy.

This invention discloses a method for restoring turbulent degradation images based on complex amplitude detection, belonging to the field of computational optics imaging technology. This method constructs a computational optics imaging system without a hardware correction unit. It utilizes a wavefront sensor and an imaging sensor to simultaneously acquire Hartmann images and target degradation images. A computational unit performs wavefront complex amplitude reconstruction and aberration correction. Based on this, a physically constrained end-to-end neural network architecture is designed. By reconstructing the wavefront complex amplitude, it achieves accurate mapping and inference from the Hartmann image to the point spread function of the imaging system. Finally, a multi-scale deconvolutional network is combined to complete high-quality restoration of the turbulent degradation image. This invention eliminates the need for a hardware corrector, overcomes the physical limitations of traditional adaptive optics, effectively solves the problem of insufficient point spread function accuracy, and enables time-delay-free and highly efficient turbulent image restoration under low-cost conditions.

The invention discloses an artillery barrel inner wall defect image data enhancement method, relates to the technical field of image processing, and adopts the technical scheme that a steel surface defect data set is combined with an artillery barrel inner wall defect-free image through a Poisson fusion method to generate a vivid artillery barrel inner wall defect image. According to the method, the problem that the actual artillery barrel inner wall defect image is difficult to obtain is effectively solved, and rich data resources are provided for training of a deep learning model. A parallel combination of an attention mechanism module and a full-dimensional dynamic convolution (ODconv) module is introduced into an STYLEGAN2 network architecture, and a novel image generation network structure is constructed. According to the structure, the generator can pay attention to local and global features at the same time, and the quality and diversity of generated images are improved.

The invention discloses a motor variable speed fault detection method based on current phase information, and relates to the technical field of motor variable speed. The motor variable speed fault detection method based on the current phase information comprises the following specific steps: step 1, estimating a current signal instantaneous angle through Hilbert transform; 2, dividing the signal into equal-angle (equal-motor rotation angle) segments of (-pi, pi) according to the instantaneous angle of the current signal, and eliminating the first and last incomplete intervals; step 3, resampling the segments through interpolation to enable the segments to have the same number of data points in the equal motor corner segments so as to complete angular domain resampling; 4, reconnecting all the intervals; step 5, extracting a fault order by using order analysis, and overcoming the defect that the rotating speed information is difficult to obtain by replacing the rotating speed signal with the phase characteristic in the current signal; operation steps are tedious and efficiency is low.

The application discloses a ship radiation noise line spectrum estimation method based on multi-snapshot sparse Bayesian learning, and belongs to the field of underwater acoustic signal processing; the method utilizes the sparse characteristics of the ship radiation noise line spectrum, that is, the line spectrum is not uniformly distributed in the whole frequency range, but is discrete and sparse, converts the line spectrum estimation problem into a sparse signal recovery problem, so as to solve the problems in the background technology, under the condition that the data sample length is limited and the sparsity is unknown, the Bayesian framework utilizes probability modeling to adaptively adjust the hyperparameters from the global optimization angle, can effectively reduce the sidelobe, realizes high resolution, reduces the background noise fluctuation variance through multiple observation snapshots statistical average, enhances the signal-to-noise ratio of the line spectrum output, and improves the line spectrum feature detection performance.

This invention belongs to the field of wireless communication, specifically relating to a method for calculating the spatial polarization domain information of received signals based on a three-element circular array receiver. This invention utilizes the similarities and differences between known steering vector configurations to solve the problem of rapidly obtaining parameters when the spatial polarization domain parameters of the received signal are unknown. Using the received signals at each point of the three-element circular array receiver, a model is constructed in pairs. Using the rotation-invariant subspace algorithm and the least squares method, the pairwise differences of the received signals at each point of the three-element circular array receiver are calculated. Given the known steering vector configuration, the elevation and azimuth angles are solved by solving equations. Using the calculated elevation and azimuth angles, and the known steering vector configuration, the phase angle and phase difference in the polarization domain are directly calculated by solving the matrix correlation using the least squares method. Finally, all state estimation results are obtained, and the root mean square error of each angle at different snapshot numbers and signal-to-noise ratios is calculated.

The present invention provides the preparation and application of NK cells. Specifically, the present invention provides a culture medium for culturing NK cells, comprising the following components: IL-2, IL-15, IL-12, IL-21, CD137 antibody, CD3 antibody, vitamins, basal culture medium, serum substitute. NK cells cultured with the culture medium have high activity and have good killing activity against tumor cells.

This invention provides a method for training a multi-task processing model and related products. The method includes acquiring training data for the multi-task processing model and multiple teacher models, wherein the multi-task processing model is configured to perform multiple text analysis tasks in speech synthesis, and each teacher model is configured to perform a single text analysis task; determining a target loss for the multi-task processing model based on the prediction results of the teacher models and the multi-task processing model on the training data; and optimizing the training of the multi-task processing model based on the target loss. Through the technical solution of this invention, the multi-task processing model can be trained to be applied to multiple tasks simultaneously, breaking the traditional pipelined combination analysis method and achieving a balance between quality and performance. In addition, this invention also provides a method for speech synthesis.

The invention discloses a method for determining the morphology of a vertical free outflow gas-liquid interface of a large-diameter pressure water delivery system, and belongs to the technical field of flow calculation of large-diameter pressure water delivery systems. Comprising the following steps: determining characteristic parameters describing vertical free outflow; establishing a coordinate system on the circular section of the vertical free outflow outlet; an appearance extreme value coefficient phi is introduced, and the maximum join height Hmax of a gas-liquid interface of large-diameter vertical free outflow is calculated; an appearance distribution coefficient delta y is introduced, and the distribution rule of the gas-liquid interface water height is determined; calculating the change of the water level rise along with y on the cross section of the vertical free outflow outlet x = 0 by using the appearance extreme value coefficient phi and the appearance distribution coefficient delta y; calculating the change of the water level along with x when y is equal to y0 on the cross section; and a diffusion flow factor a and a water-up flow factor b are introduced to calculate the internal flow characteristic hf of the vertical free outflow. The method can quickly and accurately estimate the appearance and the internal flow characteristics of the water-gas interface of the outlet section, and is simple, convenient and efficient.

The invention belongs to the technical field of motor fault diagnosis, and particularly relates to a motor transfer learning fault diagnosis method based on digital-analog driving for a drum blade dampening machine, which comprises the following steps of: preprocessing acquired signals by acquiring key parameters and working condition signals of a drum motor; establishing a mathematical model of the motor; establishing a multi-physical field coupling model of the roller motor, and generating high-precision simulation data under multiple working conditions; constructing a deep convolutional neural network model; performing model training by taking simulation data as a data set for constructing a model, correcting a simulation model error by utilizing actual data, and continuously improving the performance of the model through an interactive iterative optimization algorithm; knowledge migration from a simulation environment to an actual working condition is realized by adopting a migration learning method, a feature extraction layer of a pre-training model is frozen, a classification layer is finely adjusted only for actual data, and training time and data requirements are reduced. By generating and efficiently utilizing simulation data, the method effectively solves the problem that fault data is difficult to obtain under actual working conditions.

This application provides a training method, apparatus, electronic device, and storage medium for a lesion detection model, belonging to the field of image processing technology. It involves acquiring a first medical image and a second medical image, where the first image is unlabeled and the second image is labeled. A preset teacher model is used to detect lesions in the first medical image, and a preset student model is used to detect lesions in both the first and second medical images. Based on the lesion detection results, a loss calculation is performed to obtain target loss data. The first network parameters of the preset student model are adjusted based on the target loss data, and the second network parameters of the preset teacher model are adjusted based on the first network parameters to train the preset student model, thus obtaining a lesion detection model. This lesion detection model is then used to detect lesions in the target medical image, obtaining the location and type of the target lesion, thereby improving the accuracy of lesion detection.

The invention relates to the technical field of unmanned driving, automatic driving and unmanned vehicles, and discloses a laser radar data generation method and device, equipment and a storage medium. The method comprises the steps that a first laser radar installed on a first unmanned vehicle is used for collecting first laser radar point cloud data under the severe weather condition; acquiring second laser radar point cloud data under a common weather condition; and synthesizing the first laser radar point cloud data and the second laser radar point cloud data to obtain synthesized laser radar point cloud data under the simulated severe weather condition. According to the method, on the premise of ensuring the authenticity and diversity of the data, efficient generation of the laser radar data under the severe weather condition is realized, so that sufficient high-quality data support is provided for training and verification of an automatic driving perception algorithm in a complex weather environment.

The invention relates to the technical field of automobile power batteries, in particular to a maximum capacity obtaining method based on PHEV power battery real vehicle data and computer equipment, and the method comprises the following steps: extracting real vehicle SOC data and OCV data according to real vehicle BMS data, and fitting an OCV-SOC curve according to the OCV data and the SOC data; s100, selecting a working section of the battery, calculating OCV data of the working section through real vehicle SOC data and the OCV-SOC curve obtained in the step S100, and further calculating the internal resistance of the battery of the working section; according to the OCV-SOC curve and the internal resistance of the battery obtained in the step S200, the terminal voltage Ut (t) of the working section is expressed as # imgabs0 #, and the initial SOC value SOCt0 and the reciprocal K1 of the maximum available capacity in the formula are obtained in a least square optimization mode through the terminal voltage and current data of the working section. The method can be used for a relatively short charging and discharging working section in the PHEV so as to realize real-time estimation of the maximum available capacity of the power battery of the vehicle and further judge the service life attenuation condition of the power battery of the vehicle.

The application discloses a kind of based on preset performance constraint's main steam temperature model predictive control method and system, it is related to thermal automatic control technical field.The method includes: offline pre-training neural network prediction model using historical operation data, to approximate the dynamic and output mapping of controlled object;Online at each sampling time based on the neural network model is constructed and solved input constraint model predictive control optimization problem, wherein the cost function introduces the barrier term formed by preset performance function, limit the error envelope that converges with time in predicted tracking error;According to the deviation and error margin of actual output and model output, construct compensation control amount, and with predicted control amount superposition form final control rate;Through online parameter update, correct model error to adapt to working condition change.The application does not need accurate mechanism model, can realize main steam temperature fast, small deviation tracking, effectively suppresses overtemperature and frequent manual intervention.

The invention relates to the technical field of content distribution networks, and discloses a routing planning method, device and equipment for a content distribution network, a medium and a product. The method comprises the following steps: acquiring real-time data information of each acceleration node in the content distribution network, the real-time data information comprising a domain name; performing data analysis on the real-time data information, and determining popularity information of the domain name; determining a route planning type based on the popularity information of the domain name, wherein the route planning type comprises static route planning and dynamic route planning; performing corresponding path planning on the domain name based on the routing planning type to obtain a routing result; and pushing the routing result to the acceleration node, so that the acceleration node updates the forwarding path according to the routing result. According to the technical scheme of the invention, the method achieves the dynamic planning of the route in the content distribution network, achieves the calculation of the real-time back-to-source route of the content distribution network, and meets the high-quality requirements for the dynamic change of a service.

The application discloses a time-delay-based IPv6 address credit risk detection method and device, which comprises the following steps: aggregating an IPv6 address set to obtain a candidate first prefix set; performing time-delay measurement and aggregation based on a live address first prefix set obtained by detecting the candidate first prefix set, and extracting time-delay set features after the time-delay measurement and aggregation; inputting the time-delay set features into a trained IPv6 mobile network address identification model for feature identification and classification, so as to obtain a first prefix set of IPv6 mobile network addresses according to an IPv6 mobile network identification and classification result; calculating the score of each first prefix in the first prefix set based on a risk index of the credit of the first prefix of the IPv6 mobile network address, so as to obtain different credit scores, and obtaining a risk detection result of the credit of the first prefix of the IPv6 mobile network according to the credit scores. The application solves the problems of data deviation and difficulty in obtaining input data, and reduces the cost of active detection.

The present invention belongs to the field of image processing technology, and discloses an image denoising method, device and terminal based on a partial differential equation learning network, including: image denoising data set acquisition and input, output of predicted denoised image time series through a partial differential equation learning network, comparison with the real denoised image time series to obtain loss, minimizing loss to train parameters, and then introducing random integral theory to construct an improved numerical solver for calculating the solution of the partial differential equation, and finally the network outputs the partial differential equation corresponding to the image denoising data set and the solution of the equation, and then the trained network is combined with a denoising post-processing module for image denoising to obtain a denoising result. The present invention can solve the problem of difficulty in adjusting parameters of the image denoising algorithm based on partial differential equations, and improve the error problem of the existing numerical solver, improve the accuracy of the solution, and finally use the image deconstruction operation of the denoising post-processing module to obtain a more competitive denoising result.

The invention discloses a dry-type reactor turn-to-turn insulation fault detection method. The method comprises the following steps: S1, setting a sensor; s2, signal acquisition and transmission; and S3, signal processing and analysis. Compared with the prior art, the method is a partial discharge joint detection method based on a high-frequency method and an optical detection method, an optical sensor and a high-frequency current sensor are arranged, optical pulse voltage signals and electromagnetic pulse voltage signals are combined for processing and analysis, and a data set reflecting fault position information is obtained; accurate positioning of an insulation fault is realized; the problem that the partial discharge optical signal of the dry-type air-core reactor is difficult to obtain is solved; and accurate positioning of an insulation fault is realized. The invention also discloses a dry-type reactor turn-to-turn insulation fault detection device, the signal acquisition assembly and the analysis processing assembly in the device can carry out time sequence analysis of electromagnetic and optical fault signals at the same time, and the accuracy of fault signal acquisition, processing and analysis is improved.

The invention belongs to the technical field of helicopter fault diagnosis, and relates to an intelligent diagnosis method and system for potential faults of a helicopter rotor based on sound signals. The method comprises the steps that typical fault features are implanted into feature parameters, and rotor models in various fault states are generated; flow field parameters are solved, and the dynamic pressure load on the surface of the rotor wing blade is obtained; a sequence signal of sound pressure changing along with time is collected through the virtual acoustic sensor array; extracting fault features; constructing a fault classification and discrimination model based on machine learning; and outputting the fault state and the fault type of the rotor wing. According to the method, the influence of the rotor wing fault on the sound radiation characteristic can be comprehensively reflected, and the problem that a helicopter rotor wing fault sample is difficult to obtain is effectively solved through high-fidelity geometric modeling, fault implantation and a CFD and CSD coupling strategy; the multi-dimensional fault features are extracted, the fault diagnosis model is constructed in combination with the machine learning algorithm, the fault diagnosis accuracy is high, and automatic recognition and type positioning of faults can be achieved.