55 results about "Fast algorithm" patented technology

The invention provides an efficient transmission algorithm for HPLC carrier communication data, which belongs to the technical field of carrier communication, and comprises the following steps of: step 1, double-link separation processing: respectively processing frame control data and load data by adopting two separated links; 2, a frame control data processing flow: carrying out improved Turbo coding on the frame control data; step 3, a load data processing flow: firstly, carrying out adaptive scrambling on load data, and adopting Turbo coding with different parameter configuration from frame control data; step 4, OFDM symbol generation and transmission: adopting an optimized fast algorithm to reduce the operand and the processing time; and step 5, receiving end processing flow: respectively adjusting frame control and load data by adopting AGC (Automatic Gain Control) and time synchronization. The algorithm is high in anti-interference capability, remarkably reduces the bit error rate, is high in transmission efficiency, can adapt to different power line channel environments and data transmission requirements, and has good universality and stability.

The invention relates to the technical field of automobile electronic intelligent fault diagnosis, and discloses a multi-modal data fusion automobile air conditioner controller abnormal state monitoring method, which comprises the following steps: deploying five types of intelligent sensor nodes at key positions of an automobile air conditioner system to realize hardware-level time synchronization and adaptive sampling; a fast ICA algorithm is combined with sliding window processing to realize signal separation, a self-adaptive wavelet basis selection mechanism is designed to carry out time-frequency analysis, and principal component analysis is used to carry out feature dimension reduction; constructing a lightweight attention fusion network, dynamically adjusting different modal feature weights, and mining feature relevance through a cross-modal interaction module; a hierarchical progressive anomaly detection architecture is designed, a statistical method and a neural network detection result are fused, a time consistency constraint mechanism is established, and multi-time scale anomaly prediction is realized; and fault type identification and severity evaluation are carried out in combination with an expert knowledge base, and a multi-priority protection strategy and an emergency plan are formulated.

The invention discloses a target detection pre-labeling method based on a fast algorithm, and relates to the technical field of computer vision and geometric processing, and the target detection pre-labeling method based on the fast algorithm comprises the following steps: S1, data acquisition and preprocessing; and the collected image data is transmitted to a cloud data center in real time through wireless transmission. According to the target detection pre-labeling method based on the fast algorithm, image information can be locked and recognized more accurately through operation of two times of pre-labeling, so that the collected image information can be recognized and screened more accurately, a certain workload can be reduced for subsequent manual auditing and correction, and the accuracy of target detection and pre-labeling is improved. And meanwhile, it is also ensured that the pre-labeling model can carry out accurate labeling processing on a relatively fuzzy or shielded image, and through operation of a YOLOv5 algorithm, the pre-labeling model can be more efficient and faster when carrying out image recognition.

The invention provides a low-coupling Transform target detection method and system based on FPGA (Field Programmable Gate Array) acceleration. The method comprises the following steps: constructing a Transform target detection network model based on a low-coupling optimization method; and constructing a multi-framework universal FPGA accelerator based on a Winograd fast algorithm, wherein the multi-framework universal FPGA accelerator is used for simultaneously supporting a convolutional neural network used for feature extraction and a Transform codec based on a multi-head attention mechanism in the Transform target detection network model. According to the method, an efficient memory access transformation scheme is provided for the special type of operators, so that the accelerator obtains higher convolution operation performance while having high calculation efficiency for all the operators; a high-precision approximation scheme and corresponding data flow optimization are provided for various special functions contained in an attention mechanism, a difference fitting method is provided for GELU, approximation precision is improved, attention scaling operation and Softmax operator are fused, and scaling operation overhead during operation is saved.

The invention relates to the technical field of wireless communication, in particular to an antenna radiation and mutual coupling collaborative optimization system based on a time domain hybrid fast algorithm. The antenna radiation and mutual coupling collaborative optimization system based on the time domain hybrid fast algorithm is applied to an unmanned aerial vehicle group communication system and comprises an antenna array unit, a data acquisition unit, a time domain hybrid fast algorithm unit, a mutual coupling calculation unit, an optimization control unit and a communication coordination unit. According to the invention, the three-dimensional geographic coordinates, attitude angle data and signal intensity of the unmanned aerial vehicle group are acquired in real time through the data acquisition unit, and real-time mutual coupling coefficient analysis of the mutual coupling calculation unit is combined, so that the system can dynamically sense the formation form change and trigger the optimization instruction, and the optimization control unit performs comparison through a preset threshold value. The node with the maximum deviation degree is adjusted preferentially, it is ensured that the antenna spacing and the radiation pattern are always adaptive to the dynamic environment, the mutual coupling effect is reduced, and the communication stability is remarkably improved.

The invention discloses a rail profile consistency evaluation method and system based on wheel-rail matching characteristics, and relates to the technical field of railway engineering.The rail profile consistency evaluation method comprises the steps that firstly, wheel-rail contact key indexes are analyzed, the positions of wheel-rail contact points are calculated based on three-dimensional profile data, and a model is established based on the non-Hertz contact theory to solve contact stress; a dynamic equivalent taper is obtained by using an equivalent taper fast algorithm; then, an analytic hierarchy process and an entropy weight method are combined to determine an index weight, weighted summation is performed after normalization processing is performed on original index data to obtain a comprehensive evaluation index CEI, and an evaluation standard is set in combination with a line and a train type; according to the method, multiple dynamic indexes are creatively fused to construct an evaluation system, a subjective and objective combination weighting method is adopted to overcome the defect of single weighting, the constructed quantitative comprehensive evaluation indexes provide accurate support for railway full-life-cycle management, compared with a traditional method, the evaluation accuracy is greatly improved, railway transportation safety and stability are effectively guaranteed, and the method is suitable for popularization and application. And the overall benefit and reliability are improved.

The application relates to a kind of massive SAR data imaging software design methods based on fast BP algorithm, for the existing fast BP imaging algorithm, by accurately estimating the data amount in the operation process, propose to use the available memory of operation server, dynamically correct J operator, carry out sub-scene processing to sub-set, use disk temporary storage intermediate operation result, solve the memory limit problem of SAR full scene full resolution imaging of super large data on the operation server with lower memory, greatly reduce the hardware memory cost of operation server.

The application provides a multi-index dynamic priority unmanned aerial vehicle cooperative path planning method, and the path planning method comprises the following steps: acquiring robot parameter information, path constraint information, task environment information and algorithm parameter information; acquiring obstacle information in a simulation environment, and constructing a simulation environment grid map, so as to uniformly divide a map into grid maps with equal proportions; obtaining an unmanned aerial vehicle cooperative path planning configuration space according to the simulation environment grid map; and searching in the unmanned aerial vehicle cooperative path planning configuration space by using an unmanned aerial vehicle cooperative path planning algorithm based on a WM-MI-CS algorithm, so as to obtain a feasible unmanned aerial vehicle cooperative path. According to an example task requirement, the multi-index dynamic priority unmanned aerial vehicle cooperative path planning method has the advantages of low operation cost, short task completion time and fast algorithm speed.

This invention relates to the field of underground mine positioning and navigation technology, solving the technical problem of difficult underground positioning and mapping in unstructured environments where GPS is ineffective. Specifically, it relates to a SLAM method for unstructured features in underground mines, comprising the following steps: acquiring point cloud information of the current laser frame from a lidar system; calculating the curvature of each point in the point cloud information of the current lidar frame, and extracting corner features and planar point features of each point based on the curvature; extracting information from the current frame using a camera, detecting corner points of the current frame using the FAST algorithm based on the current frame information, and determining whether it is a corner feature point; and tracking feature points of the current sliding window keyframe using the KLT optical flow tracking algorithm based on the current frame information. This invention achieves high-precision positioning in coal mines, balancing accuracy and computational load, and improving the accuracy and real-time performance of positioning and mapping in mine tunnels.

The present application discloses a method for rapid splicing of three-dimensional surface morphology based on an improved ORB, comprising: determining the overlapping area of ​​the micro-nano surface structure to be spliced ​​based on a block matching method; extracting an initial feature point set based on the FAST algorithm, calculating the Harris corner response value of the feature points after non-maximum suppression processing, and screening according to the response value; describing multiple feature points to be matched with the largest response value to obtain feature descriptors; obtaining coarse matching point pairs based on the feature descriptors of the feature points to be matched, performing fine matching on the coarse matching point pairs using a decoupled RANSAC algorithm, obtaining a three-dimensional space rigid body transformation model, and performing splicing and fusion of micro-nano surface structures through transformation interpolation and overlapping area data fusion. The present application also discloses a device for rapid splicing of three-dimensional surface morphology based on an improved ORB, corresponding equipment, and storage medium. The present application can significantly improve splicing efficiency and accuracy, and effectively solve the problem of scale differences in micro-nano surface structure data in different directions.

The invention discloses an inspection robot dog scheduling method considering charging planning in an electric energy meter detection scene, and belongs to the technical field of inspection scheduling. The technical scheme of the invention comprises the following steps: deeply fusing inspection task allocation and charging planning decision, accurately calculating the shortest feasible path of a robot dog between equipment points and between the equipment points and charging piles, and calculating the optimal path of the robot dog; by means of the technical scheme, the inspection task execution efficiency of the inspection robot dog is improved, the practicability and reliability of the inspection robot dog scheduling scheme are improved, the algorithm response speed is high, dynamic events such as newly-added inspection tasks and robot dog state changes can be flexibly processed, and the service life of the inspection robot dog is prolonged. Rapid replanning of the inspection task sequence is realized, and efficient and stable execution of the inspection task is ensured.

The invention discloses a 4D wireless spectrum map reconstruction method in a dynamic environment, and the method comprises the steps: carrying out the modeling of missing spectrum data and mixed noise into a four-dimensional tensor model according to the four dimensions of two-dimensional space, frequency and time; in order to reduce the degree of freedom of the tensor model, the model is decoupled into an incomplete spectrum map corresponding to a single radiation source through a non-negative matrix factorization algorithm. Further considering the problem of mixed noise influence and calculation complexity, proposing an optimization function based on a dynamic threshold value and a fast algorithm combining random singular value decomposition and Nesterov accelerated gradient, and solving under a non-negative matrix factorization algorithm framework; and finally, reconstructing the frequency spectrum map corresponding to each radiation source by adopting parallel computing to obtain a complete 4D frequency spectrum map. According to the method, an accurate 4D spectrum map can be obtained based on spatial sparse sampling in a dynamic time-varying environment, and the high-precision spectrum situation sensing requirement in a complex noise environment is met.

The present application discloses a tire polishing curve fitting method, including the following specific steps: scanning data is subjected to mean linear filtering, and coordinate system conversion and processing, the scanning data is converted into polar coordinate representation, the pattern groove data is extracted, and further converted back to rectangular coordinates for maximum filtering to obtain approximate data of the tire surface without the pattern groove; on the basis of data preprocessing, the present application uses logical analysis method to summarize the law of traditional genetic algorithm, proposes an optimal state genetic algorithm, and applies it to tire laser scanning data processing, the optimal state genetic algorithm has the advantages of fast algorithm convergence speed, small calculation amount and the like in processing tire scanning data and fitting tire pattern groove curve, effectively improves the algorithm search speed and accuracy, and makes up for the deficiency of tire scanning algorithm theoretical research and practical application.

The invention provides an interface interaction method and device, and the method comprises the steps: determining a first virtual article, rendering the first virtual article to a UI interface, and then setting a trigger point and a trigger point lighting effect; according to the method, key points which are uniformly distributed are found on the surface of a complex three-dimensional object through a fast algorithm in a simple mode, different trigger points are generated for different virtual objects in an adaptive mode, and trigger interaction information corresponding to the virtual objects is prompted through the light effect of the trigger points, so that a user is guided to perform corresponding trigger activity operation, and the user experience is improved. The method is suitable for running in a mobile client; further, when the user gradually clicks the plurality of trigger points, gradually revealing a plurality of attribute values of the first virtual article at the target position of the UI until the overall score of the whole first virtual article is revealed; through the interactive design, the user obtains comprehensive experience, and the participation sense and exploration experience of interaction between the user and the virtual object are improved.

This invention discloses a lossless data compression and encrypted transmission method based on federated middleware, belonging to the field of computer application technology. The method includes: performing run-length encoding on the data to be processed based on federated middleware to obtain first compressed data; processing the first compressed data using a compression algorithm based on a raw-middleware dual dictionary to obtain second compressed data. This compression algorithm uses both the raw dictionary and the middleware dual dictionary, adding only values ​​that appear at a set frequency to the dictionary; sequentially performing Huffman coding and DES processing on the second compressed data to obtain encrypted data and a key; performing RSA processing on the key to obtain an encrypted key; and transmitting the encrypted data and the encrypted key to the target end using an encrypted compression transmission plugin of an experiment-and-train-enabled architecture. This method achieves lossless data compression and encryption based on federated middleware, with fast algorithm speed and high security.

The invention provides a face recognition method and system, computer equipment and a storage medium, and belongs to the field of face recognition, and the method comprises the steps: collecting a plurality of face images of a target person under a non-ideal condition, and extracting a high-dimensional semantic feature set of the face images; through an iterative optimization process, a customized dictionary exclusive to the target person is obtained based on the feature set learning, and sparse coding and dictionary atom updating are alternately carried out in the iterative process; finally, forming a dictionary database by the customized dictionaries of all the target persons; and extracting to-be-recognized face features, calculating reconstruction errors between the to-be-recognized face features and each dictionary in the dictionary database, and judging the identity according to a minimum error principle. According to the method, the recognition accuracy and robustness under small samples are remarkably improved, the real-time performance is ensured through a fast algorithm, the dependence on a large number of training samples is reduced, and the method is suitable for actual application scenes such as access control and security and protection.

The application discloses a unit ascending and descending load working condition optimization method based on historical working conditions, and collects historical ascending and descending load working conditions by fitting calculation on screening parameters and screening according to average fitting deviation and maximum deviation limit value. In actual matching process, the starting working condition load and the ending target load are set to ensure that the operation requirements of load ascending and descending operation are met. Different business priority needs in different business scenes can be met through different optimization methods, all working conditions under the same load ascending and descending condition are screened through setting a filtering time range, the optimized working condition is ensured to be accurate and reliable according to the optimization standard, the operation parameters under the specified load condition are adjusted according to the ascending and descending load working condition after optimization and the returned index data, the operation state of the unit reaches the optimum, the general unit ascending and descending load working condition optimization method shields business differences, provides an efficient and fast algorithm, greatly saves manpower, improves the calculation efficiency and improves the accuracy of business processing data.

The present application relates to the technical field of pipeline engineering safety monitoring and state evaluation, and discloses a kind of liquid-solid two-phase flow based on tee pipeline erosion defect fast algorithm, comprising: S1, along the inner wall of the tee pipeline of metal material in liquid-solid two-phase conveying system axial and circumferential, potential erosion area is discretely divided, form several independent erosion unit, define equivalent topographic feature parameter vector for each erosion unit, for corresponding pipe wall area erosion defect geometry state is characterized;S2, sensor is arranged at the preset monitoring position of tee pipeline, and multiple-source physical response signals and the like in the running process of tee pipeline are collected;The method realizes accurate inversion and dynamic evolution prediction of erosion defect under limited information by erosion defect parameterization modeling, running response mapping construction, online rapid inversion and consistency correction, solves the problems of low calculation efficiency, poor engineering applicability and insufficient evaluation accuracy of prior art.

The application discloses a satellite formation relative attitude estimation method and system based on monocular vision, and the method comprises the following steps: acquiring a satellite image in a satellite formation to be calculated; wherein the satellite image refers to an image of a measured satellite located in a field of view of a monocular measurement camera arranged on a detection satellite; extracting feature points of the satellite image based on a FAST algorithm and a BEBLID algorithm; completing feature matching of the satellite image based on the extracted feature points; and completing relative pose calculation by using a combination of EPnP and orthogonal iteration based on a feature matching result, so as to obtain a pose of the measured satellite relative to the monocular measurement camera as a pose of the measured satellite relative to the detection satellite. The application can effectively improve the pose estimation precision.

The invention relates to the technical field of electromagnetic spectrum monitoring and signal processing, and discloses a low-altitude unmanned aerial vehicle electromagnetic interference complexity early warning method and system based on a fast FFT algorithm. According to the method, space electromagnetic signals are collected in real time, FFT frequency domain analysis is carried out to obtain spectrum data, unmanned aerial vehicle signal frequency and power and left and right interference signal frequencies and interference energy are extracted based on the spectrum data, the frequency coincidence ratio and the energy coincidence ratio are calculated in an unmanned aerial vehicle system working frequency band, and objective complexity is obtained through the product of the frequency coincidence ratio and the energy coincidence ratio. And determining an electromagnetic interference complexity grade by combining a preset grade division standard look-up table, and further outputting an alarm signal and / or prompt information. The system comprises an electromagnetic signal acquisition unit, a signal processing unit and an early warning output unit, wherein the signal processing unit completes parameter extraction, overlap ratio calculation, complexity calculation and grading output. According to the scheme, complexity grading evaluation with low calculation amount is realized, and the method is suitable for real-time early warning of a low-altitude unmanned aerial vehicle scene.

The application discloses a four-direction overlapping combined sparse total variation Cauchy noise image restoration method, relates to the image restoration technical field, and solves the technical problem that the existing image restoration method has a poor Cauchy noise restoration effect and ladder effect after restoration, obtains a degraded image polluted by Cauchy noise; an image degradation model is established based on an overlapping combined sparse non-convex mixed regular term; an improved alternating direction multiplier fast algorithm is proposed to solve the model; and stability analysis is performed on model parameters to obtain a final restoration image; the application can effectively remove Cauchy noise in the image, relieve the ladder effect of the image edge area, and well protect information in a complex texture area of the image, and has great advantages in visual effect and numerical analysis.

The invention relates to the field of steam turbine vibration monitoring, in particular to a bending interference fringe correction method based on angular point detection and a phase correlation function. The method comprises the following steps: determining angular points in an interference fringe image based on a FAST algorithm; angular point scale and direction enhancement is carried out, and the feature discrimination degree of angular points is improved; fourier transform is carried out on the reference column and the offset column to obtain a phase correlation diagram, and the position of the maximum response value is found in the phase correlation diagram; taking the position of the maximum response value as the center, and intercepting the response value of the m * m neighborhood; fitting neighborhood data by using an interpolation function; searching a real pole value of the difference function in a peak value region, wherein the offset corresponding to the pole value is the calculated sub-pixel offset; and according to the offset, carrying out translation transformation on the image information of the offset wavelength channel, or carrying out equivalent adjustor position information. The accuracy of vibration monitoring is improved.

The application relates to a real-time splicing method for core drilling images, and belongs to the field of image processing.The method comprises the following steps: for any two adjacent core drilling images, respectively adopting a FAST algorithm to quickly detect corner points, and adopting a Harris algorithm to remove weak corner points and reserve true corner points with high response values to obtain two true corner point images; adopting a quadtree division method to realize uniform distribution of the true corner points in each true corner point image, and adopting a non-maximum suppression method to screen high-quality corner points; adopting a fast approximate nearest neighbor matcher to perform initial matching on the high-quality corner points in the two images to obtain a plurality of initial matching pairs; adopting a homography geometric consistency inlier purification algorithm based on sparse feature points to perform accurate matching on the plurality of initial matching pairs to obtain an optimized homography matrix; and splicing the two adjacent core drilling images according to the optimized homography matrix.The application guarantees the real-time performance and accuracy of the splicing of the core drilling development images.

This invention discloses a fast forward modeling optimization method for three-dimensional gravity and gravity tensor, and a further optimization method for the system. The fast forward modeling method divides the underground three-dimensional space into multiple horizontal layered media and constructs a discrete model of the underground space and observation points. Based on the model division, the forward modeling coefficient matrix and coefficient spectrum matrix of gravity and gravity tensor are calculated. Based on the coefficient spectrum matrix, a fast algorithm is used to calculate the gravity value and gravity tensor value of each depth layer corresponding to the density model. The forward modeling values ​​of each layer are superimposed to obtain the gravity forward modeling anomaly of the density model. Furthermore, based on the parity characteristics of the fast Fourier transform of real number sequences, the spatial complexity of the forward modeling coefficient matrix is ​​reduced while ensuring fast calculation efficiency, providing support for further increasing the practicality of three-dimensional gravity inversion.

This invention provides a rapid prediction method for lunar orbit rendezvous guidance, achieving high-precision and rapid prediction to meet mission requirements. The prediction method of this invention applies the gravitational field acceleration calculation method used for GPS autonomous orbit determination of Earth orbit satellites to the field of lunar orbit rendezvous. Addressing the problems that the algorithm cannot be directly applied to the design of lunar orbit rendezvous guidance strategies, improvements and upgrades are made one by one, resulting in a high-precision and rapid algorithm suitable for lunar orbit rendezvous flight control missions to meet mission requirements. Specifically, this invention is the first to transplant this algorithm to the field of lunar orbit applications, which is unprecedented. Considering the special characteristics of the lunar gravitational field, where the acceleration measures of various gravitational perturbations are of similar order and cannot be approximated by the J2 term alone, the J22 term perturbation acceleration is added to the algorithm, thereby significantly improving the computational accuracy and numerical stability of the algorithm.

The application discloses a CPU double-path acceleration method for large-scale LLM sparse matrix multiplication, and relates to the technical field of high-performance computing. The method first constructs a fast matrix multiplication algorithm lookup table, generates an optimized algorithm through random walk and performs Hensel promotion to compensate for numerical errors; then performs two-stage division and zero block identification on the weight matrix, classifies it into complete zero blocks, near-zero blocks and dense blocks; then aggregates non-zero elements for the near-zero blocks and merges dense blocks into large blocks using dynamic programming; uses multi-core CPU to concurrently perform sparse path and dense path calculation, respectively calls sparse kernels and matching fast algorithms, and finally fuses the results based on position mapping using vector instructions. The application solves the problem of low utilization of unstructured sparse matrices through double-path shunting strategy and software and hardware co-optimization, significantly improves the calculation throughput and hardware resource utilization while ensuring the model accuracy.

The application relates to the technical field of unmanned aerial vehicle electromagnetic environment evaluation, and discloses an unmanned aerial vehicle electromagnetic interference complexity evaluation method based on S transform characteristic parameters and complexity interval weight coefficients. The method collects 2400-2476 MHz frequency band electromagnetic interference signals and screens effective signals, extracts time occupation degree, frequency spectrum occupation degree, energy occupation degree, space coverage and power spectrum density by using an S transform fast algorithm, matches actual values of various characteristic parameters in combination with a pre-constructed characteristic parameter complexity interval database and a characteristic parameter weight database, generates a characteristic parameter evaluation matrix, and outputs a complexity level after obtaining a comprehensive evaluation vector through fuzzy transformation. The method can improve the pertinence of 2.4G frequency band unmanned aerial vehicle electromagnetic interference feature extraction and the automation degree of evaluation.

A kind of multi-source shortest path method based on bidirectional SPFA, comprising: according to the traffic road network map of obtained construction directed weighted graph;Obtain multi-source shortest path query request;Select specific node as pre-processing starting point, calculate the shortest distance of pre-processing starting point to all other nodes using shortest path fast algorithm, construct and initialize the evaluation function array;For each shortest path query request, empty the node scheduling queue for storing pending node and reset each auxiliary array;Based on the evaluation function array dynamic optimization node scheduling queue, calculate the shortest distance in combination with shortest path fast algorithm;Iterative processing all shortest path query requests and return query result.The present application can effectively process the directed graph containing negative weight edge but not containing negative weight ring, and is suitable for large-scale network environment under traffic road network map efficient shortest path calculation.

The application discloses a photovoltaic array maximum power point tracking method and system based on a quantum marine predator-disturbance observation fusion algorithm, and specifically is as follows: on the basis of marine predator optimization, the duty cycle of a front-stage DC-DC converter is taken as a to-be-optimized control variable, candidate duty cycles are normalized and mapped into rotation angle parameters of a parameterized quantum circuit, quantum state coding, quantum gate operation and measurement sampling are performed under a quantum computing framework, and position updating of the marine predator algorithm is corrected according to measurement results, so that global optimization of quantum computing is realized; when a search solution enters a maximum power point neighborhood, a disturbance observation method is switched to for local fine tracking; and when environmental mutation or output power deviation exceeding a threshold is detected, quantum global search is restarted. The application improves the search ability for a global maximum power point under the condition of local shadow and multi-peak power curve, has fast algorithm convergence speed, strong steady-state operation performance and a wide application range.

The invention discloses a partition energy storage configuration method and device for a high-voltage power grid, and the method comprises the steps: firstly, building an electric coupling strength matrix, and carrying out the optimal partitioning of the power grid through an improved Newman fast algorithm based on the electric coupling strength matrix; then, taking economy and power grid vulnerability as optimization targets, comprehensively considering constraint conditions such as energy storage constraint and power flow constraint, and establishing an urban power grid energy storage configuration model; and then, a Pareto frontier is obtained by adopting a multi-target particle swarm optimization algorithm, and the optimal site selection and capacity of energy storage of each partition are determined through a rank sum ratio comprehensive evaluation method. According to the method, the grid-side energy storage power station can be reasonably configured, and the autonomy and economical efficiency of urban power grid partition operation are effectively improved.