212results about How to "Guaranteed Computational Efficiency" patented technology

The invention relates to an aeroelastic stability fluid-structure interaction prediction method of turbo-machine changed interblade phase angles, which predicts the aeroelastic stability of a turbo-machine through the modal pneumatic damping ratio of different interblade phase angles by adopting an energy method. The method adopts the modal analysis containing prestress on established single-sector finite element models, realizes the vibration displacement transfer on the fluid-structure interaction interface through a data transfer method, adopts the dynamic mesh technology for obtaining mesh files taking the interblade phase angles into consideration, is used for unsteady computation fluid mechanic analysis of Junction Box modules in CFX, further obtains the modal pneumatic damping ratio of each interblade phase angle, and is used for predicting the aeroelastic stability. The method based on the fluid-structure interaction fully utilizes the cutting edge technology of the computational structural mechanics and the computation fluid mechanics, ensures the computation precision in each sub system, and improves the computation efficiency through the division of a fixed region and a movable region. The invention has good practical value and wide application prospects in the technical field of turbo-machine simulation.

The invention discloses a no-grid co-prime array DOA (direction of arrival) estimation method based on virtual array interpolation, and mainly solves a problem of information loss caused by the heterogeneity of a virtual array in the prior art. The method comprises the implementation steps: constructing a co-prime array at a receiving end; receiving an incident signal through the co-prime array, and carrying out the modeling; calculating an equivalent virtual signal corresponding to the receiving signal of the co-prime array; constructing an interpolation virtual array, and carrying out the modeling; constructing a multi-sampling snapshot signal of the interpolation virtual array and a sampling covariance matrix; constructing a projection matrix, and defining projection calculation correlated with the projection matrix; designing an optimization problem based on the minimization of the nuclear norm of the signal covariance matrix of the interpolation virtual array, and solving the optimization problem; and carrying out the estimation of the DOA according to the reconstructed covariance matrix of the interpolation virtual array. The method improves the DOA estimation freedom degree and resolution, and can be used for the passive positioning and target detection.

The embodiment of the invention relates to the image processing field, especially to an image identification method and device. The image identification method and device are configured to identify the types of the image. In the embodiment, the method comprises: obtaining an image to be identified; inputting the image to be identified into a feature extraction model, obtaining the first feature vector of the image to be identified, wherein the feature extraction model is obtained through the sample image and the deep convolutional neural network model; dividing the image to be identified into a plurality of subimages according to a preset rule; inputting the subimages into the feature extraction model, and obtaining the second feature vector of the subimages; merging the first feature vector and the second feature vector, and obtaining a third feature vector; determining the type of the image to be identified according to the similarity of the third feature vector and the a fourth feature vector of each type of sample image in the sample feature library, wherein the sample feature library is obtained for each type of sample image putting into the feature extraction model to determine the type of the image to be identified.

The invention discloses a co-prime array wave arrival direction estimation method based on interpolation virtual array signal atom norm minimization. The objective of the invention is to solve the problem of an information loss caused by non-uniformity of a virtual array in the prior art. The method comprises following steps that a receiving end constructs a co-prime array; based on the co-prime array, an incidence signal is received and modeling is performed; an equivalence virtual signal corresponding to the a co-prime array receiving signal is calculated; an interpolation virtual array is constructed and modeling is performed; a multi-sampling snapshot signal of the interpolation virtual array and a sampling covariance matrix thereof are constructed; a projection matrix is constructed and projection calculation related to the projection matrix is defined; an optimization problem based on the interpolation virtual array signal atom norm minimization is designed and solved; and direction-of-arrival estimation is performed according to the reconstructed interpolation virtual array covariance matrix. According to the invention, the freedom degree and the resolution of the direction-of-arrival estimation are improved and the method can be used for passive positioning and target detection.

The present invention discloses a ship still water shearing force and bending moment calculating method which comprises the following steps of: establishing three-dimensional photoetching format models of a ship housing and all compartments, i.e. STL models; calculating a floating state of a ship in a random loading state; and calculating a buoyancy distribution curve, a weight distribution curve, a load distribution curve, a shearing force distribution curve and a bending moment distribution curve of the ship. According to the present invention, the floating state of the ship under the random loading condition is calculated on the basis of the STL models; transverse inclination and longitudinal inclination of the ship can be simultaneously calculated; the defect of only calculating a longitudinal inclination angle of the ship in a conventional method is overcome; and calculation accuracy is improved. According to the present invention, on the basis of the ship housing STL model, the buoyancy curves of the ship under the random transverse inclination and longitudinal inclination can be calculated. According to the present invention, the concept of a weight distribution table of the compartments is disclosed and in the calculating process, volume distribution among frames only needs to be interpolated according to a loading height, and not only calculation accuracy can be improved, but also calculation efficiency can be ensured.

The invention discloses a user label storage method. The method is executed in calculation equipment. The calculation equipment is respectively connected with a first data storage device and a seconddata storage device. A plurality of user label records are stored in the first data storage device. A plurality of label bitmaps are stored in the second data storage device. The method includes: acquiring a user label record from the first data storage device, wherein the user label record includes a user identifier (ID) and at least one label of a user; determining at least one target label bitmap, which corresponds to the user label record, according to the label; determining a target location of the user label record in each target label bitmap according to the user identifier; and settinga value at the target location in each target label bitmap to 1, and thus storing the user label record from the first data storage device into the second data storage device. The invention disclosesa corresponding user portrait calculation method and the calculation equipment at the same time.

The invention discloses a rapid magnetic resonance imaging method based on an AR2 U-Net neural network. According to the method, an existing R2U-Net convolutional neural network is improved; the method is based on an R2U-Net convolutional neural network. An attention gate module is added, an AG training model is used for implicitly learning and suppressing an unrelated region in an input image, and meanwhile, obvious features useful for a specific task are highlighted, so that the AR2 U-Net convolutional neural network only needs a small amount of training data under the condition of reconstructing an image with the same quality. The problem that the swing amplitude of an optimization loss function is too large in the updating process is solved. According to the method, an Adam optimization algorithm is adopted to replace a conventional SGD optimization algorithm, the convergence rate of the convolutional network can be further increased, the problem that training is ended too early can be effectively prevented, and for learning rate processing, a polynomial attenuation strategy is adopted, so that learning can be stably reduced, and the reduction speed is higher along with increase of the number of turns.

The invention discloses a co-prime array direction or arrival estimation method based on interpolation virtual array covariance matrix subspace rotation invariance, and mainly solves the error of direction or arrival estimation caused by information loss due to the heterogeneity of the virtual array and the base mismatch problem of the predefined space grid points in the prior art. The implementation steps are that a co-prime array is established; the co-prime array receives the signal to perform modeling and calculates the equivalent virtual signal; an interpolation virtual array is constructed and the virtual domain signal statistics is initialized; the optimization problem based on nuclear norm minimization is designed and solved so as to reconstruct the interpolation virtual array covariance matrix; the signal subspace of the interpolation virtual array covariance matrix is formed; two virtual subarray signal subspaces having rotation invariance are constructed; the transformationmatrix of the two virtual subarray signal subspaces is calculated; and the closed-form solution of direction or arrival estimation is formed. The degree of freedom and the accuracy of direction or arrival estimation can be enhanced and the method can be used for passive location and target detection.

The invention discloses a method for calculating the electromagnetic vibration noise of a motor, which comprises the following steps: calculating the distribution of the electromagnetic force densityof the motor according to the operation condition of the motor, and calculating the static deformation and the natural frequency of the motor respectively; according to the distribution of the electromagnetic force density, the electromagnetic resultant force on the stator teeth of the motor is calculated, and the time-space harmonic of the electromagnetic resultant force is calculated. The frequency response function of the motor structure corresponding to each vibration mode is calculated according to the natural frequencies of each vibration mode corresponding to each spatial order of the electromagnetic force, and the transfer function of the motor mechanical structure is calculated according to the static deformation function and the frequency response function. The vibration displacement of the motor surface is calculated according to the time-space harmonics and the transfer function. The modal acoustic radiation efficiency of the motor housing is obtained, and the electromagnetic vibration noise of the motor is calculated according to the modal acoustic radiation efficiency and vibration displacement. In the process of analyzing the vibration and noise of the motor, the invention ensures that the calculation accuracy meets the precision requirement, and reduces the calculation amount to avoid consuming a large amount of calculation resources.

The application discloses a multi-mode travel O-D demand estimation method based on multi-source data check. The method comprises: S1, selecting basic elements in a transportation network and establishing an urban multi-mode transportation network model based on travel needs among the basic elements; and S2, carrying out pre-treatment on multi-source historical traffic data for checking an O-D demand estimator and carrying out statistic calculation to obtain needed basic data; and estimating the multi-mode O-D demand by using the model from the S1 and the calculation result from the S2 to obtain a novel total O-D demand. According to the method disclosed by the application, statistics is carried out based on multiple data sources; and he restriction on the previous single-mode O-D demand estimation is eliminated, the traffic characteristics of various modes of travel in the road network are considered comprehensively, the mutual feedback mechanism between O-D demands and section flowsis reflected, the problem of consistency between the O-D estimation value and the actual observation value is solved, and the obtained urban multi-mode O-D demand distribution result meets the user balance criteria.

The invention provides an alliance chain data fragmentation method and system, equipment and a storage medium, and the method comprises the steps of allowing a first slave fragmentation node to receive a plurality of transactions distributed by a main fragmentation node, and wherein a sender account and/or a receiver account of each transaction correspond/corresponds to the first slave fragmentation node; executing each transaction in the fragment, sending an execution result to a main fragment node so that the main fragment node generates a block according to the execution result after receiving the execution result, and storing block data corresponding to the block, wherein the execution result comprises updated account state data of each account. According to the invention, the fragmentation architecture is designed, so that the execution problem of a single node is solved, and the possibility of transverse expansion is provided for an execution module from the architecture level; and on the other hand, for the whole transaction batch, in the transaction distribution step, the transaction is distributed to each fragment according to form and to, so that unrelated transaction parallelism to a certain extent is realized, and in addition, execution of cross-fragment transactions is realized through a distributed shared memory.

The invention discloses a lightweight missing spatio-temporal data reconstruction method. The overall steps are as follows: 1, performing spatio-temporal data representation: Abstracting the point data and the mesh data of the static reference into a unified space-time state matrix for representation; 2, performing time dimension interpolation: Introducing an average correlation coefficient to automatically select a time window so as to improve the modeling time dependence capability of the SES algorithm; 3, performing spatial dimension interpolation: respectively adopting constant equal distance and a correlation distance based on a Gaussian function to assign a weight to each spatial neighbor so as to improve the capacity of modeling spatial dependence of the IDW algorithm; 4, performingspace-time integration: introducing an extreme learning machine as a learning algorithm of the neural network model, and integrating an estimation result of a space-time dimension to obtain a final predicted value of missing data. According to the method, a plurality of improved lightweight models are integrated, so that the reconstruction precision of mass missing spatio-temporal data is furtherimproved on the premise that the calculation efficiency of the reconstruction algorithm is ensured.

The invention provides a method for eliminating a control rod tine effect in reactor core calculation. Specific steps are: step 1: dividing a whole solving region into several node regions, and writing a to-be-solved neutron diffusion equation; step 2: by using a variable principal, writing each node functional style; step 3: expanding neutron flux density and a neutron source space basis function and substituting the neutron flux density and the neutron source space basis function into the node functional style; and step 4: by using the node functional style, obtaining a relation between an emission neutron flux and an incident neutron flux of a node surface and a relation between a node inner flux and an emission flux and an incident flux of the node surface. The method provided by the present invention is based on a variational node method; a cross section of a node functional is written into a space function; a hypothesis that a macro cross section in the node is a constant is not made during a deduction process; and finally non-uniformity in the node is displayed in a response matrix of the node, so as to realize computation of a non-uniform node and further eliminate the control rod tine effect.

The invention relates to a power system cascading failure analysis model and probability calculation method based on a Markov process. The method comprises the following steps of: differentiating operation characteristics of different components of a power system, and, in combination with state information of the power system, where system components are, analyzing the shutdown probability of the components; considering the correlation between failure events in a cascading failure evolutionary process, describing a cascading failure transfer process through the Markov process, searching a cascading failure event of the system under different initial failure conditions, establishing a cascading failure event model, and generating a power system cascading failure state space in the Markov process; and, analyzing the system transfer probability in a cascading failure state space, and calculating the different-scale cascading failure probability. According to the method disclosed by the invention, while the calculation efficiency is ensured, the cascading failure event of the system under different initial failures is discovered; furthermore, the cascading failure probability is calculated; a corresponding system improvement measure is provided; the cascading failure is prevented; the system risk is reduced; and the power supply capability of the power system is increased.

The invention discloses a quasi-direct current optimal power flow method considering temperature influence for an electrical power system. According to the quasi-direct current optimal power flow method considering the temperature influence, an active power balance equation in a traditional direct current model is corrected through the mathematical connections between an active power transmission equation and a reactive power transmission equation of a circuit so as to consider the influences of inactive power and circuit resistance in the model to build a quasi-direct current model of the electrical power system to complement and perfect the direct-current model. According to the electric heating coupling relationship between the circuit temperature and circuit resistance, the circuit temperature influence is considered in the quasi-direct current model of the electrical power system. The quasi-direct current optimal power flow method considering the temperature influence uses a simplified interior point method to solve the quasi-direct current optimal power flow model considering the temperature influence; the computational efficiency of the algorithm can be well guaranteed to meet the online calculation requirements. The simulation result indicates that compared with a traditional direct current optimal power flow, the algorithm is capable of greatly improving the computational precision of the algorithm based on guaranteeing the computational efficiency of the algorithm, and the algorithm is significant for the online dispatching of the electrical power system.

The invention provides a decision-making method for maintaining middle and small span concrete bridges. The decision-making method provided by the invention comprises the following steps of: (2) determining a decision-making object and a predicated age limit, wherein the decision-making object is a bridge network OD (Outside Dimension) model; (2) indentifying the bridge network OD model: converting an image-formed OD model into a numerical-value-formed OD model and identifying a minimum path set matrix P which can be used for a decision-making calculation; (3) establishing a cost model of a maintenance policy and an effect model which can be used for processing an event tree analysis on the OD model to obtain the maintenance policy; (4) carrying out an improved non-dominated sorted genetic algorithm: defining a coding method of the maintenance policy and taking a connecting probability Rnw and a maintenance cost Cnw of the bridge connection as two targets to be participated in a genetic evolution; carrying out a Pareto optimality and searching for the globally optimal solution. The method provided by the invention has the advantages of simple principle, convenience in using, wide application range, good reliability and the like.

The invention discloses a harmonic noise suppression method based on waveform morphology sparse modeling. The harmonic noise suppression method includes the steps of 1) constructing Chirplet transform according to waveform morphological characteristics of harmonic noise in seismic records acquired via vibroseis slip sweep, and constituting an over-complete dictionary with continuous wavelet transform; 2) quickly implementing Chirplet positive and inverse transforms; 3) determining Chirplet transform parameters based on time-frequency distribution characteristics of correlated data; and 4) according to the starting frequency of a reference scanning signal, determining the filter cutoff frequency of the harmonic noise to achieve fidelity separation of an effective signal from the harmonic noise. The invention solves the problem of harmonic noise interference in the seismic data acquired via vibroseis slip sweep, and achieves the purpose of improving the signal-to-noise ratio of the seismic data. The method of the invention determines the Chirplet transform according to the time-frequency distribution characteristics of the harmonic noise, and ensures the sparseness; the fast implementation of the Chirplet transform guarantees the efficiency of transform, and the correction coefficient guarantees the accuracy of inverse transform reconstruction; and the Chirplet transform parameters are automatically determined according to the data drive, a strong adaptability is gained and single-channel calculation facilitates parallel processing.

The invention discloses an assimilation method for multi-scale data of a ground surface temperature field on the basis of photovoltaic power station measurement data. The method comprises steps as follows: firstly, photovoltaic power station measurement points are clustered with a DBSCAN (density-based spatial clustering of applications with noise) algorithm, a multi-scale grid of the temperaturefield is established according to a clustering result, finally, the multi-scale grid is subjected to data assimilation, and an assimilation result of the multi-scale data of the temperature field is obtained. According to the method, the problem that calculation accuracy and calculation efficiency cannot be realized simultaneously due to fixed grid resolution of a common data assimilation method can be solved, and accuracy of automatic numerical weather prediction calculation of the electric power industry is improved; the multi-scale grid is established according to photovoltaic weather station data easily obtained by the electric power industry as well as the distribution condition of the photovoltaic power stations, a high-resolution data assimilation result is obtained in the area nearby the photovoltaic power stations, and the accuracy of automatic numerical weather prediction calculation of the electric power industry is improved.

The invention discloses a retina image analysis method applied to calculation equipment. The method comprises steps that a to-be-analyzed retina image is acquired, and the to-be-analyzed retina image is pre-processed to acquire an enhanced image; the enhanced image is filtered to extract a background image; a bright region image and a dark region image are determined according to RGB color values of pixels of the enhanced image and the background image; a videodisc region and a bright noise region are determined, and the videodisc region and the bright noise region are removed from the bright region image to acquire an exudation region; a blood vessel region and a dark noise region are determined, and the blood vessel region and the dark noise region are removed from the dark region image to acquire a hemorrhage region. The invention further discloses a retina image analysis device of the retina image analysis method, and the calculation equipment equipped with the retina image analysis device. The method is advantaged in that the exudation region and the hemorrhage region of the retina image can be rapidly acquired through segmentation.

The invention discloses a method for simulating the break of a windscreen by using EFGM in a car crash. The method specifically comprises the steps of presetting background grids of the windscreen by using an N-surface body unit; adopting an integral modeling, endowing the windscreen with a dispersed constitutive model with disabled plastic damage, and judging that how much force is suffered between discrete points by using a cohesive fracture method when separation occurs; calibrating the windscreen constitutive parameter of the corresponding car model and EFGM algorithm parameter by a calibration test; assembling the windscreen in a whole car model, performing simulated analysis on the whole car crash, and if the simulation process and the result of the whole car crash meet expectant truth, evaluating the damages of parts; and if the simulation process and the result of the whole car crash do not meet the expectant truth, adjusting the model. According to the method, the cohesive fracture rule is introduced into the EFGM, so that the defect that the traditional FEM depends on the grids is avoided, thus ensuring the random character of the crack path, and more truly expressing material performance and restoring the crash event.

The invention provides a whole-process voltage stability analysis fault screening method and system. The method comprises the steps of obtaining a static state voltage stability severe fault set according to an obtained power system static state power flow data; performing evaluation on the static state voltage stability severe fault set according to the obtained current operation mode to obtain atransient state voltage stability severe fault set; and performing evaluation on the transient state voltage stability severe fault set to obtain a medium and long term voltage stability severe faultset. By virtue of the method, voltage stability severe fault screening and sorting methods for the static state, transient state and medium and long term three analysis stages are provided, so that medium and long term dynamic simulation plays an important role in a fault scanning process, thereby avoiding scanning of a large number of fault sets in the medium and long term analysis for ensuringcalculation efficiency, and effectively screening out faults which cause medium and long term voltage instability.

The present invention relates to a figure simulation method for an unsteady-state heat transfer process of a large-sized crude oil floating-roof tank, which solves the problem that existing figure simulation researches cannot comprehensively reflect an influential rule of related factors on a large-sized floating-roof oil tank temperature field. A mathematical model, which is established by particularly adopting a grid generation technology, of an oil product heat transfer process in a large-sized floating-roof crude oil tank can ensure that a figure simulation process coincides with an actual woking condition to the largest extent; and the model is solved by using an iteration method. The method has the characteristics of simplicity, flexibility and high versatility, and is easy to be implemented on the computer; and not only can computing efficiency be ensured, but also time is saved.

The invention discloses a CO2 enhanced shale gas reservoir development dynamic simulation method based on a cross-scale multi-flow space gas transmission mechanism. The method comprises the steps: establishing a double-pore/double-seepage numerical model through a multi-component multi-phase three-dimensional numerical simulator; carrying out logarithm interval encryption on the hydraulic fractureby adopting a local grid encryption technology; establishing three flowing spaces of a matrix, a natural fracture and a hydraulic fracture in the shale reservoir; respectively calculating the density, viscosity and molecular weight of the single-component and multi-component natural gas on the basis of a density, viscosity and molecular weight calculation formula; calculating the apparent permeability of the corresponding single-component and multi-component natural gas; giving an average pore radius and a certain pressure interval; and enabling the multi-component multi-phase three-dimensional numerical simulator to call the corresponding natural gas permeability correction factor at each time step based on the calculated pressure, so the purpose of correcting the Darcy constitutive is achieved.