21results about How to "Reduce problem size" patented technology

An intelligent compensation method for coupling errors of a dynamic gravimeter comprises the following steps: selecting typical environment interference influencing the dynamic gravimeter in a complex measurement environment, and analyzing an influence mechanism of each typical environment interference on the measurement accuracy of the dynamic gravimeter; according to the influence mechanism under each typical environment interference, analyzing a coupling error influence mechanism between the typical environment interferences; based on the coupling error influence mechanism, constructing a coupling error comprehensive compensation model; adding the coupling effect loss into a data-driven loss function, and defining a model coefficient constrained by physical information; and performing model training by using the training data, and calculating the total loss of the verification set on the obtained model to obtain a final coupling error comprehensive compensation model. According to the invention, the coupling error of the dynamic gravimeter in a complex environment is compensated based on coupling error mechanism analysis and a physical constraint long-short-term memory model, and the measurement precision and environmental adaptability of the dynamic gravimeter under complex multi-factor interference conditions are improved.

The invention belongs to the technical field of equivalent modeling, and discloses a hybrid wind power plant equivalent method and device based on oscillation energy decomposition, and the method comprises the steps: obtaining the operation measurement data of a hybrid wind power plant, carrying out the calculation or frequency scanning of each fan grid-connected branch, and obtaining a broadband equivalent impedance characteristic; constructing a multi-machine oscillation energy equivalent unit comprising a plurality of grid-connected branches according to the broadband equivalent impedance characteristic and the actual grid-connected topology of the wind power plant; establishing an energy function and a dynamic equation according to the multi-machine oscillation energy equivalent unit; performing modal decomposition and branch decomposition on the system oscillation energy, and calculating the oscillation energy contribution degree and the dissipated energy participation degree of each fan grid-connected branch in a dominant oscillation mode to obtain an energy characteristic quantity; performing fan grouping to obtain a fan grouping result of the hybrid wind power plant; and performing differential equivalence and parameter aggregation according to a fan grouping result to obtain a small disturbance equivalence model. The calculation model scale and the calculation complexity of the hybrid wind power plant can be reduced.

This invention provides a non-uniform discrete equivalent circuit modeling method and system for electromigration reliability. It employs an equivalent circuit-based approach to analyze the electromigration stress evolution of general complex interconnect structures, fully considering thermal field effects in the model. This achieves equivalent circuit modeling for stress analysis under multi-physics coupling. Furthermore, by using an adaptive mesh method with non-uniform spatial discretization and a steady-state / quasi-steady-state filtering mechanism, it solves the problem of wasted computational resources in traditional uniform discretization methods when dealing with large-span interconnect structures, and avoids unnecessary transient simulations of immortal interconnect trees, reducing the computational load of full-chip-level analysis. This enables efficient and high-precision transient analysis of the electromigration reliability of large-scale interconnect networks under multi-physics coupling effects.

The application discloses a whole-body-local multi-scale biomechanical model coupling construction method for human body impact simulation, comprising the following steps: constructing a whole-body dynamics model, obtaining the kinematic response and dynamic response of the human body under a preset impact working condition; determining the connection constraint position corresponding to a target local area in the whole-body dynamics model, and extracting the constraint force and / or constraint moment at the connection constraint position; performing equivalent conversion on the constraint force and / or constraint moment to obtain equivalent load for representing the whole-body dynamics effect acting on the target local area; constructing a local simulation model of the target local area, and applying the equivalent load to the local simulation model to realize simulation analysis of the human body impact response. Through the conversion of the whole-body dynamics constraint force to the local simulation model, the application reduces the calculation complexity of the human body impact simulation while retaining the whole-body inertia effect, and improves the simulation efficiency and calculation stability.

The application discloses a centrifugal pump off-design condition diagnosis method and system based on a vibration signal, and the method comprises the following steps: collecting vibration signals of a centrifugal pump under multiple flow conditions, and constructing a vibration sample sequence with a working condition label; reconstructing the vibration signal by using a variational mode decomposition (VMD); performing a continuous wavelet transform (CWT) on the reconstructed signal, and mapping a one-dimensional vibration signal into a two-dimensional time-frequency graph; constructing a Light-CBAM-CNN diagnosis model, and training the same; inputting the vibration signal collected in real time into the trained diagnosis model after pretreatment, VMD and CWT, outputting a working condition category and a diagnosis confidence, and outputting an "uncertain" label and triggering a retest mechanism or an alarm when the confidence is lower than a threshold value. The application can realize online diagnosis of high-energy consumption off-design conditions of the centrifugal pump only by relying on the shell vibration signal, and has the advantages of small modification workload, low deployment cost, high identification reliability and strong energy-saving scheduling guiding significance.

This application provides a mobile robot task scheduling method and system based on a heterogeneous computing architecture, relating to the field of intelligent scheduling technology. The method includes: constructing a workspace geometric model based on the geometric structure of a planting row; obtaining the motion state parameters of the mobile robot using a first computing unit; determining the look-ahead compensation amount by combining the statistical value of the inference time of the target recognition task by a second computing unit; determining a spatiotemporal pre-aiming slice in the visual field region based on the look-ahead compensation amount and the workspace geometric model, and cropping image data located within the slice from the original image data; the second computing unit performing target recognition on the image data, generating a work instruction corresponding to the recognized target, and controlling the lateral actuator to perform the work according to the work instruction. This can alleviate the spatiotemporal asynchrony problem between perception results and execution position in a moving work scenario, reduce missed and misaligned tasks, and reduce the communication and computational load of the heterogeneous computing system.

This invention proposes a hardware implementation method, apparatus, electronic device, and computer-readable storage medium for the Sigmoid function, relating to the field of data processing technology. The method constructs a function mapping table for the Sigmoid function, containing discrete non-negative floating-point values ​​within the unsaturated interval, based on the symmetry and saturation of the Sigmoid function, thereby substantially reducing the scope of the lookup table. On this basis, the absolute value of the floating-point values ​​belonging to the unsaturated interval is taken to obtain the value to be processed. This value is then converted to the target fixed-point format, and integer operations are used for table lookup, interpolation, and symmetry restoration to obtain the output value of the Sigmoid function based on the target floating-point value, improving computational efficiency. While ensuring the function output capability across the entire range, the size and storage overhead of the function mapping table are effectively reduced, while linear interpolation under the fixed-point format improves output accuracy.

This application provides a deep learning-based method and system for evaporator defect detection, relating to the field of anomaly detection. The method includes: acquiring multi-view image data of the evaporator; extracting multi-scale features, including the number of fins, their arrangement direction, and the spatial location of notches, through the backbone network layer of a deep learning target detection model; nonlinearly enhancing the multi-scale features through the model's feature enhancement network layer to obtain enhanced multi-scale features; and then extracting reconstructed features through a hybrid attention mechanism layer; based on the reconstructed features, obtaining the detection results of evaporator fins and notches through the model's detection layer, and comparing them with preset structural parameters to obtain the defect detection results. This application, used in evaporator defect detection, solves the technical problem of existing technologies being unable to stably identify evaporator defects under changing environmental and evaporator structural parameters.

The embodiment of the invention provides an ultra-low noise tile-type phased array plane based on a silicon-based chip, and the phased array plane comprises an antenna part PCB, and the antenna part PCB comprises an array antenna, an LNA chip, a board-level filter, a BFIC chip based on a silicon-based technology, and a power division network which are connected in sequence; wherein the array antenna is provided with a patch antenna, and a feed point of the patch antenna is directly connected with the LNA chip through a plated-through hole; the LNA chip is connected with the board-level filter through a plated-through hole, the board-level filter is connected with the BFIC chip through a plated-through hole, and the BFIC chip is connected with the power division network through a plated-through hole; the LNA chip is integrated with a dynamic tuning matching circuit, and dynamic impedance matching can be carried out according to the working condition of the patch antenna. Therefore, the tile-type phased array plane provided by the embodiment of the invention realizes the consideration of high integration level, low noise, low loss, dynamic matching and interference resistance, and overcomes the defects of the traditional tile-type phased array plane in the aspects of integration level, noise, loss, matching flexibility, filtering and the like.

The invention discloses a false star filtering and star map recognition method based on multi-sky-area parallelism, which is applied to autonomous attitude determination of a star sensor and comprises the following steps: constructing a multi-sky-area parallelism index database of a whole celestial sphere, dividing a whole celestial sphere space represented by a unit spherical surface, and generating a plurality of candidate sky-area sets which cover the whole celestial sphere and are allowed to be overlapped with one another; the method for extracting star pattern features based on the discretized radial angular distance comprises the following steps: extracting position parameter information of observation star points based on star map data acquired by a star sensor, and constructing a radial feature vector for representing a local spatial distribution relationship of fixed stars for each observation star point in a star map; performing feature matching of the multi-day-area parallel star mode feature vectors, and calculating a matching evaluation index based on the feature matching; and a final star map recognition result is obtained based on iterative screening recognition and false star elimination. And outputting a final star map recognition result for subsequent attitude calculation. The invention further discloses a corresponding system, electronic equipment and a computer readable storage medium.

The invention discloses a water-rich rock mass fracture precursor identification method based on a lightweight neural network, and belongs to the technical field of deep rock mass engineering safety monitoring, and the method comprises the steps: collecting pulse waves and interference waves of deep rock mass fracture to form a data pair, building a waveform library based on the data pair, building a lightweight model for predicting rock fracture, and carrying out the prediction of rock fracture. The method comprises the following steps of: extracting and splicing data pairs to obtain a fused feature map, extracting local features, fusing to obtain a time sequence feature map, compressing by utilizing global pooling, converting into a nonlinear channel association vector, normalizing to obtain a weight vector, combining with the time sequence feature map, and compressing to obtain a channel statistical vector; and performing nonlinear feature transformation and numerical regularization by using a full connection layer, and obtaining ternary probability distribution through a Softmax function. According to the invention, key technical support is provided for real-time monitoring, and urgent demands of water-rich rock mass engineering on intelligent, real-time, high-reliability and safe monitoring are precisely met.

The utility model discloses a brake cylinder cover press-fitting device, and belongs to the technical field of brake cylinders. The brake cylinder cover press-fitting device comprises a support mechanism which comprises a support and a supporting column, and the supporting column is arranged on the support; the clamping mechanism comprises a first clamping piece, a second clamping piece and a clamping driving cylinder, at least one of the first clamping piece and the second clamping piece is rotatably arranged on the supporting column in a sleeving mode, the first clamping piece is provided with a first clamping part, the second clamping piece is provided with a second clamping part, the first clamping part and the second clamping part are oppositely arranged, and the clamping driving cylinder is arranged on the supporting column; a cylinder bottom and a cylinder rod of the clamping driving cylinder are rotationally connected with the first clamping piece and the second clamping piece respectively; and the press-fitting mechanism comprises a press-fitting driving cylinder and a jacking piece, the press-fitting driving cylinder is connected with the support, the jacking piece is connected with a cylinder rod of the press-fitting driving cylinder, and a positioning hole matched with the piston rod is formed in the jacking piece. The brake cylinder cover press-fitting device is easy and convenient to operate, safe, efficient and low in cost.

The invention discloses a nuclear reactor neutron transport calculation method and device, and the method comprises the steps: carrying out the space merging of a fine network multi-group neutron transport model in each time of source iteration in a process of calculating the fine network multi-group neutron transport model through a source iteration mode, and obtaining a coarse network multi-group neutron transport model; correcting the first coarse mesh multi-group neutron standard flux through the coarse mesh multi-group correction factor to obtain a second coarse mesh multi-group neutron standard flux, and solving the coarse mesh multi-group neutron transport model to obtain a third coarse mesh multi-group neutron standard flux; performing iterative solution on the coarse mesh multi-group neutron transport model to obtain a target coarse mesh multi-group neutron standard flux so as to determine a fine mesh multi-group correction factor; and correcting the initial fine-net multi-group neutron standard flux through the fine-net multi-group correction factor to obtain a target fine-net multi-group neutron standard flux, and calculating the fine-net multi-group neutron transport model through the target fine-net multi-group neutron standard flux to improve the efficiency of nuclear reactor neutron transport calculation.

The invention relates to a shale condensate gas reservoir circulation gas injection process optimization method and device and a storage medium. The method comprises the steps that geological characteristic parameters and fracturing construction microseismic monitoring data corresponding to a target shale condensate gas reservoir are obtained; establishing a geological numerical model and a crack model based on the data, establishing a component mathematical model, performing state calculation based on a state equation, performing condensation flash calculation, performing full implicit numerical solution, performing relative permeability and capillary pressure model, and optimizing a circular gas injection process. According to the method, microseism inversion, phase equilibrium calculation and numerical simulation are integrated, the minimum total volume of stratum gas condensate and the maximum cumulative gas production rate serve as optimization targets, the circulation gas injection rate is quantitatively determined, and the shale gas condensate reservoir circulation gas injection process is quantitatively optimized; the problems that in the prior art, due to the fact that optimal gas injection technological parameters need to be found through a large number of field tests, mining time is long, efficiency is low, and cost is high are solved.

The invention relates to the field of unmanned aerial vehicle image processing, in particular to a corn phenological period monitoring method based on unmanned aerial vehicle images and deep learning, and adopts the technical scheme that a corn planting test point is tested by taking a year as a time unit, and N planting densities are set for M hybrid varieties of the test point, so that the diversity of a data set is improved; features in different scenes are considered; the training subset is used for training a MaizePhenoNet model, the verification subset is used for monitoring model training, and the phenomenon of overfitting of model training is prevented; compressing the hidden dimension of the classification head from a 1280 channel to a 512 channel so as to reduce parameters of a full connection layer; scaling expansion channels in the FUSDIB module and the UIB module and carrying out 8-times alignment, so that main parameter overhead caused by point convolution is reduced; under the condition that a backbone topological structure and a down-sampling strategy are not changed, the method effectively reduces the model parameter scale and storage requirements, and is suitable for an end-to-end corn phenological period identification task.

The invention discloses an alarm information processing method and device, equipment, a storage medium and a program product. Comprising the following steps: acquiring an alarm information data set comprising a plurality of alarm information; projecting each piece of alarm information to a plurality of preset dimensions to obtain a weight vector of each piece of alarm information; inputting the weight vector of each piece of alarm information into a preset alarm labeling model, and outputting to obtain an initial classification label of each piece of alarm information, the alarm labeling model being obtained by training a plurality of samples and labels of each sample; the multiple pieces of alarm information, the corresponding weight vectors and the initial classification labels serve as a training data set, a preset large language model is trained in combination with a preset first low-rank matrix and a preset second low-rank matrix, and an alarm noise reduction large language model is obtained; the ranks of the first low-rank matrix and the second low-rank matrix are far smaller than the dimension of a weight matrix of the large language model, the row number of the first low-rank matrix is the row number of the weight matrix, and the column number of the second low-rank matrix is the column number of the weight matrix.

The invention discloses a carbon capture device and method, which are used for integrating CO2 absorption and desorption processes of carbon capture in the same device, the device at least comprises: an absorption unit arranged at the upper part of the device; in the absorption unit, the absorption slurry is in contact with carbon-containing gas from top to bottom so as to carry out gas-liquid mass transfer under a high-pressure condition; the solid phase part of the absorption slurry is powder of a supported ionic liquid, and the ionic liquid is an amino-containing ionic liquid; the desorption unit is arranged below the absorption unit and is isolated by a liquid seal; under the vacuum condition, the absorbed slurry is subjected to CO2 desorption in a desorption unit, the desorbed slurry is recycled after being subjected to gas-liquid separation, and CO2 is discharged. According to the invention, the gas-liquid mass transfer efficiency can be effectively improved and the absorption rate of CO2 can be increased through the nano-micro interface strengthening effect; absorption and desorption are integrated into the whole device, different areas are isolated in a liquid sealing mode, and the functions of high-pressure absorption and vacuum desorption can be achieved in the same device.

The invention discloses a process-based scene annotation judgment method, and relates to the technical field of scene annotation, the process-based scene annotation judgment method comprises the steps of feature index construction, scene annotation analysis, scene annotation judgment and early warning prompt, indexes are constructed by constructing VPR features of a sub-map, and then geographic priori corresponding to a target scene is executed, so that the scene annotation judgment accuracy is improved. Retrieving to obtain candidate reference images matched with the scene image corresponding to the target scene, executing analysis and judgment between the matching degree of the scene image and each candidate reference image, and executing geographical reordering of the candidate reference images corresponding to the target scene based on a judgment result; according to the method, the final positioning result of the target scene is determined, and the flow optimization of the labeling corresponding to the target scene is analyzed, so that the precise and efficient labeling of the flow corresponding to the scene is realized, the problem of mismatching of different scenes caused by lack of semantic constraint in the current labeling is intelligently solved, and the scale and resource consumption of online retrieval are effectively reduced.

This disclosure relates to the field of power system technology, and provides a method, apparatus, medium, and equipment for calculating the operating envelope of distribution network nodes. The method includes: acquiring basic data including network topology, line parameters, transformer parameters, transmission capacity limits, node voltage limits, and distributed energy output limits, and constructing a feasible region model containing linearized power flow equations and various security constraints based on this data; defining a fully symmetric multicell as its independent operating envelope for each node connected to distributed energy, and ensuring that the Cartesian product of all node envelopes is included in the feasible region model by constraining each node to adjust its power independently within its envelope without violating grid security constraints; and solving for the operating envelope of each node with the objective of maximizing the size of all node envelopes. This embodiment achieves accurate and efficient definition of the safe operating range of distribution network nodes under high-proportion distributed energy access.

The utility model discloses an overline spliced bridge based on a prefabricated UHPC-T beam structure. The overline spliced bridge comprises T beams, steel cross beam components, shear key components and transverse prestressed tendons. The bottom of the T-beam is arranged on an abutment through a bridge support and comprises a web plate and a wing plate; the steel cross beam component is used for connecting and fixing the adjacent T beams, so that the adjacent T beams are connected into a whole to improve the structural strength and stability, the steel cross beam component comprises a pre-buried connecting seat pre-buried in a web and a steel cross beam, and the two ends of the steel cross beam are fixed to the pre-buried connecting seat of the adjacent T beams through high-strength bolts; the shear key component comprises key teeth and key grooves which are arranged on the two sides of the wing plate, and the key teeth of the adjacent T beams are embedded in the key grooves. The transverse prestressed tendons penetrate out of the wing plates to be tensioned, and the multiple T-beams are connected into a whole. According to the overline spliced bridge, the modularized UHPC-T beams are adopted, the structural height is reduced by means of the ultrahigh mechanical property of UHPC, and the overline spliced bridge adapts to the condition that the clearance under the bridge is limited; the shear keys are matched with the steel cross beams to replace a traditional cast-in-place wet joint and cross beam technology, old bridges do not need to be dismantled, the longitudinal section of a road does not need to be lifted, and the construction period is remarkably shortened.