84 results about "Line model" patented technology

The invention relates to the technical field of image recognition, in particular to a building plane element recognition model construction method based on a convolutional neural network, and the method comprises the following steps: obtaining an input APN sample; elements in the APN sample are labeled, and a labeled file is generated; converting the annotation file into a format required by YOLO, and normalizing the annotation file; performing data enhancement on the APN sample, expanding the training data volume, and outputting to obtain an element detection result and a feature vector; installing dependency and carrying out data configuration; setting command line model parameters and customizing training scripts; and carrying out model training and model reasoning. According to the method, the plane layout is abstracted based on the machine learning method, and the vector data type has high flexibility and deformation capability, so that the vector output which keeps the original plane image form unchanged can be converted into various objects according to the purpose of a user. And classifying and identifying the plane elements by using the GNN.

The invention relates to the technical field of mechanical manufacturing and machining, in particular to a dressing wheel and honing wheel design method for improving contact line stability in the gear honing machining process. Based on workpiece gear parameters, a left and right tooth surface mathematical model is established, a honing process space coordinate conversion relation is established, a conjugate meshing relational expression is derived by combining the relative slip speed of meshing contact points and a normal vector, and a tooth surface contact line model is established; and the calculation modeling of the two-dimensional projection of the left and right tooth surface contact lines, the projection intersection point position and the variation under the same rotation angle is realized. And a honing wheel dressing contact line model under dressing wheels with different tooth numbers and a tooth surface contact line intersection projection model in the dressing and honing process under the tooth number difference are established. A honing wheel and variable tooth number dressing wheel design method is provided, and geometric and technological parameters of the honing wheel and the variable tooth number dressing wheel are determined; and a trimming and honing two-degree-of-freedom conjugate meshing relation and a modified tooth surface numerical model are constructed and verified, discretization and point cloud processing are conducted on the model, high-precision tooth surface point cloud is obtained, and accurate machining of a trimming wheel and a honing wheel is guided.

The invention discloses a resonant grounding power distribution network single-phase grounding fault line selection method and system. The method comprises the following steps: when a single-phase earth fault occurs in the power distribution network, acquiring real-time zero-sequence current data of each feeder line; and inputting the real-time zero-sequence current data into the fault line selection model to judge a fault line and output a result. The training process of the fault line selection model comprises the following steps: acquiring a noisy source domain data set from an electromagnetic transient simulation platform and a target domain data set from a real power distribution network; constructing a noise reduction automatic encoder as a feature extractor, and performing unsupervised pre-training by using the source domain data set to minimize a reconstruction error; pre-trained feature extractor parameters are frozen, and a trainable fault classifier is connected to form a transfer learning model; and carrying out supervised training on the model by using the source domain and target domain data sets and taking the fault line selection accuracy as a target to obtain a final model. By implementing the method provided by the invention, the accuracy and generalization ability of the line selection model in a real noise environment can be effectively improved through transfer learning and noise reduction feature extraction.

The invention relates to the technical field of industrial process monitoring and fault diagnosis, and discloses an alarm AI application system based on industrial data, and the system comprises a data access and preprocessing module which is used for collecting and preprocessing the industrial data so as to generate a standardized state vector sequence; the offline model construction module is used for performing offline training and solidifying a physical channel model and a data channel model for subsequent diagnosis; the real-time diagnosis and alarm module is used for loading the model, calculating physical and mode residual errors in parallel and executing root cause diagnosis so as to generate a quantized causal score; and the diagnosis result output module is used for triggering an alarm according to the residual error and the causal score and outputting an explanatory report containing root cause positioning. According to the method, the two-channel residual error model is constructed, and the dynamic Jacobian matrix is utilized to drive the dynamic causal graph network, so that the alarm accuracy and the root cause diagnosis reliability are obviously improved.

The invention relates to the technical field of drawing recognition, and particularly discloses a CAD drawing intelligent recognition method which comprises the steps of drawing preprocessing, primary scheme drawing recognition, character recognition, character content standardization, table structure recognition and image-text alignment and output. Deformation, block shielding and multi-layer table recognition are realized through a multi-vertical-domain special small model, and text information standardization is completed in combination with a vertical domain large model subjected to electrical knowledge fine adjustment and an updatable AI knowledge base. An assembly line framework is adopted, independent deployment and flexible expansion of a multi-product line model are supported, and the newly-added model does not affect an existing system. The whole recognition error rate is reduced, unified specification of multi-source drawings is achieved, seamless joint of recognition results and an enterprise production system is guaranteed, and services can be provided for the outside through an API interface.

The invention provides a fine-grained lightweight DPU performance isolation method. The method comprises the following steps: firstly, entering an offline stage, representing a partial systematic analysis task execution mode through offline analysis resources, and modeling a mapping relationship between task parameters and resource consumption; then entering an online stage, intercepting submission of an application task, evaluating a resource demand according to an offline model, decomposing a large task by applying a self-adaptive task splitting and result recombination part, executing a scheduling process guided by a working load, and realizing dynamic allocation through a global scheduler and an application agent; the global scheduler allocates time slices or subtasks based on workload characteristics; after execution, results are recombined and returned to the application. Therefore, under the condition that hardware or proprietary software is not modified, accurate resource management is achieved by directly constructing on an existing manufacturer SDK, the isolation challenge of the DPU is effectively solved through the technologies, and efficient resource utilization of cloud providers in PaaS and SaaS modes is promoted.

The application discloses a kind of flow control valve fault diagnosis methods, it is related to flow control valve fault diagnosis technical field, and its technical solution points are: establishing fault diagnosis system, the fault diagnosis system includes inference engine, interpreter, data acquisition module, flow prediction module, man-machine interaction module, database and knowledge base.The application provides a kind of new model updating method, avoids the situation that blind model updating leads to the breakdown of diagnostic system.The method can even be used in the case where the training data only includes one working condition, providing an effective model updating scheme;It also provides a fault diagnosis method under the condition of multiple working conditions and incomplete training samples, ensuring the performance of fault predictive alarm, while significantly reducing false alarm;In the offline model identification and online application stage, only DCS data, expert knowledge and user feedback information are used, which greatly saves the online cost of the system.

This invention discloses a production line modeling method based on a multi-layered logical model. The method comprises three levels: equipment-level, unit-level, and production line-level modeling. First, equipment is abstracted as a logical chain of "input-processing-output," describing its geometric, physical, behavioral, and rule information through quadruples. Second, multiple equipment logical models are combined into storage, production, and detection units, with internal collaboration driven by state and event, and inter-unit interaction through standardized interfaces. Finally, a production line-level logical model is constructed through a unified interface, achieving seamless integration and simulation expansion of the multi-layered logical models. This method, combined with a production process simulation module, enables the standardization, modularization, and rapid construction of production line models, improving production line design efficiency and operational flexibility.

The invention relates to the technical field of fault discrimination, and discloses a cosine similarity-based single-phase earth fault line selection method, system and device and a medium, and the method comprises the steps: obtaining real-time signals of zero-sequence current and voltage of a bus; comparing the rated voltage of the bus with the real-time voltage, and judging whether fault line selection is started or not; if so, sampling to obtain zero-sequence current and bus zero-sequence voltage of each line in the first period after the fault; calculating a cosine similarity coefficient of each line; constructing a main criterion, and distinguishing a faulty line and a non-faulty line; and if the cosine similarity coefficient does not conform to the main criterion, constructing an auxiliary criterion, and determining a fault line. The difference between the polarity of the zero-sequence current and the derivative of the zero-sequence voltage is measured through a cosine similarity comparison method, and an equivalent capacitance value does not need to be determined; according to the method, the comprehensive detection criterion based on the synergistic effect of the main criterion and the auxiliary criterion is constructed, the simplification error of the line model is considered, the reliability is high when the high-resistance grounding fault occurs, and the zero-sequence current does not need to be filtered.

The present application relates to a kind of event-oriented power grid-life line-user disaster influence whole chain joint simulation method, to solve the problem that existing technology cannot cover composite extreme event, coupling modeling is not accurate, lack of unified whole chain simulation framework.The method generates extreme event scene such as natural disaster, network attack, constructs multi-granularity life line model, carries out joint simulation in two stages (two-way coupling system iterative simulation, one-way influence system section simulation), to combine time-event double-driven interaction, synchronously model 8 kinds of important users.Can accurately simulate whole chain failure process, output power grid fault evolution, life line function loss and user service availability, provide support for city life line resilience assessment and emergency decision-making.

The invention discloses a production line modeling method based on a multilayer logic model. The method comprises three layers of equipment-level modeling, unit-level modeling and production line-level modeling, wherein firstly, equipment is abstracted into an input-processing-output logic chain, and geometric, physical, behavior and rule information of the equipment is described through a tetrad; secondly, a plurality of equipment logic models are combined into storage, production and detection units, state-event driven cooperation is adopted in the units, and the units interact through standardized interfaces; and finally, constructing a production line level logic model through a uniform interface, and realizing seamless connection and simulation expansion of a multi-layer logic model. The method is matched with a production process analog simulation module, standardization, modularization and rapid construction of a production line model can be achieved, and the design efficiency and operation flexibility of the production line are improved.

The application relates to a behavior intention prediction method and system for a moving target and a readable storage medium. The method comprises the following steps: inputting multi-modal to-be-predicted data into a trained multi-line model to obtain a behavior intention prediction result; a training process of the multi-line model comprises the following steps: constructing a target loss function according to a first loss function of a primary task, a second loss function of a secondary task and a preset alignment loss function between single-line models in an initial multi-line model; the primary task is direction-level intention prediction, and the secondary task is behavior-level intention prediction; and the trained multi-line model is obtained based on multi-modal training data and the target loss function, thereby solving the problem of poor behavior intention prediction performance of single-modal data, processing each modal data through the multi-line model, and strengthening the model performance through a multi-task and multi-modal alignment loss function in the training stage, so that the reliability and adaptability of the model are significantly improved.

The application provides a power distribution network anomaly prediction method, device, equipment, medium and program product. In the method, the line length, line model, power supply capacity and power consumption capacity corresponding to each line level in the target line from the transformer node to the to-be-installed power node are determined according to the acquired power node connection line diagram of the transformer area and the position of the to-be-installed power node. The grid-connected capacity is determined according to the installation purpose and installation capacity of the to-be-installed power node. Then, whether the power distribution network is abnormal is predicted in combination with the current capacity of the transformer and the node type of the to-be-installed power node. The line length, line model, power supply capacity and power consumption capacity corresponding to each line level in the target line determined, the position, installation capacity, node type and installation purpose of the to-be-installed power node are used for power distribution network anomaly prediction, and the prediction accuracy is improved.

The invention relates to a welding type IGBT bonding wire space three-dimensional magnetic field distribution calculation method based on forward modeling, and belongs to the field of power semiconductor devices, and the method comprises the following steps: S1, obtaining welding type IGBT module data, and setting each parameter; s2, carrying out bonding wire geometric preprocessing and wire element discretization; s3, calculating a forward modeling magnetic field; and S4, sorting the magnetic field results at the observation points to generate magnetic field distribution. According to the method, the magnetic field forward modeling model is established based on the three-dimensional space geometric path of the bonding wire, the bending form, the lifting height and the space trend of the bonding wire are directly introduced, errors caused by simplifying the bonding wire into a two-dimensional conductor or an equivalent linear model are avoided, and the space non-uniform distribution characteristics of the magnetic field around the bonding wire are described more accurately; compared with finite element electromagnetic simulation depending on a large number of grid division, the forward modeling calculation method based on the analysis electromagnetic model has the advantages that the calculation process is clearer, and the method is suitable for rapid calculation and comparative analysis of different bonding wire structures.

The invention discloses a power transmission line model drawing method and system based on a standardized model. The method comprises the steps that a maintenance line is divided into segmented lines with different maintenance demand degrees according to the operation state, meteorological environment and fault parameters of the line; pre-defining standardized sub-models corresponding to different maintenance demand degrees; according to the maintenance demand degree of the segmented line, calling a corresponding standardized sub-model, and instantiating to generate a plurality of line sub-models; setting a virtual demarcation point between the adjacent line sub-models, splicing the plurality of line sub-models into a complete power transmission line model according to the connection relation recorded by the virtual demarcation point, and performing global consistency verification; and calling a drawing tool to draw and display the power transmission line model after the global consistency verification is passed. According to the method, on the premise of ensuring graph model drawing standardability and data consistency, simple to complex power transmission line graph models can be accurately constructed, and the influence of conditions such as line conflicts and data barriers is reduced.

The invention relates to the technical field of power distribution network photovoltaic bearing capacity evaluation, in particular to a power distribution network distributed photovoltaic bearing capacity rapid evaluation method, system and device based on the gravity center theory and a storage medium. An equivalent line model from a root node to a photovoltaic access node is constructed by acquiring a topological structure, line parameters and node load data of a power distribution network, a load active moment, a load reactive moment, a photovoltaic moment and a photovoltaic gravity center are calculated, the voltage and branch current of the photovoltaic access node are calculated on the basis of a moment theory, and the power distribution network is constructed. And determining the maximum photovoltaic access capacity of the node under the condition of satisfying the voltage constraint and the current constraint. A traditional load flow calculation problem needing iteration solution is converted into algebraic calculation based on the moment theory, the complex iteration process is avoided, the evaluation efficiency is greatly improved, and an efficient and practical technical means is provided for distribution network distributed photovoltaic optimal configuration and access planning.

This invention provides a method and system for transmission line inspection and analysis based on digital twins. The method involves: using a digital twin object identifier as the core, uniformly solidifying the transmission line structure and constructing an initial static line structure model; based on continuously collecting multi-source sensing data and forming a time-series operational observation set, constructing a multi-physics coupled evolution model that simultaneously characterizes icing and melting, thermal equilibrium response, and mechanical evolution; identifying hidden risks by jointly discriminating the state evolution trajectory of key state variables and their historical reference distribution; and conducting parameterized perturbations and multi-scenario simulations to assess operational reliability risks. Finally, the risk development trend is reflected back into the spatial line model, generating forward-looking inspection and analysis results. This invention can perform line inspections on transmission lines that are under long-term alternating thermo-mechanical coupling conditions of icing and melting.

The invention relates to the technical field of computers, and discloses a real-time detection and positioning method and device for a subsea pipeline and a related product. The method comprises the following steps: detecting an underwater image to be detected by using a trained deep learning subsea pipeline detection model to obtain a subsea pipeline detection result; detecting an underwater image to be detected by using the trained deep learning operation device detection model to obtain an operation device detection result; generating a subsea pipeline contour point set of the underwater image to be detected based on the subsea pipeline detection result and the operation device detection result; fitting points in the subsea pipeline contour point set to obtain a linear model; and calculating positioning data of the subsea pipeline based on the linear model. According to the method, the underwater subsea pipeline positioning data which is higher in robustness, more accurate and capable of being directly driven and controlled can be generated, it is guaranteed that the stable and reusable positioning data can be obtained in the two stages of searching of the unmanned underwater vehicle and releasing of the operation device, and therefore the short-distance operation efficiency and the success rate are remarkably improved.

The utility model relates to the technical field of living furniture, and particularly discloses a deep line modeling sofa which comprises a deep line modeling component, and the deep line modeling component comprises a sitting part and / or a backrest part; the deep line modeling component comprises at least two modeling modules; the modeling module comprises an internal frame body and an external soft wrapping layer wrapping the internal frame body; every two adjacent modeling modules are tightly attached to each other, so that the visual effect of deep lines is formed between every two corresponding external soft wrapping layers. The deep line modeling sofa provided by the utility model can effectively solve the problem that the existing sofa cannot realize a deep line visual effect.

The application relates to a method, device, computer equipment and storage medium for offline model reasoning. The method comprises the following steps: obtaining to-be-processed data of multiple offline models; performing first parallel processing on the to-be-processed data of the multiple offline models on multiple first pipelines; storing processing results output by the first pipelines into a first cache queue; reading the processing results output by the first pipelines from the first cache queue on multiple second pipelines to perform second parallel processing, so as to complete reasoning of the multiple offline models. The method for offline model reasoning realizes a reasoning calculation method of multiple pipeline parallelization and pipelining operation in an offline model reasoning application scenario. Compared with a method of single-pipeline execution of reasoning calculation, the method greatly improves the speed and efficiency of model reasoning.

The invention relates to the technical field of intersecting line track calibration, in particular to a circular pipe intersecting line welding line coordinate system calibration method which comprises the steps that an intersecting line track mathematical model is established, and the coordinate relation of intersecting lines in a main pipe coordinate system and a branch pipe coordinate system is determined; based on the transformation relation between a robot base coordinate system and a master coordinate system, an anti-symmetric matrix is adopted to express an attitude matrix, and a calibration constraint equation for eliminating a translation vector is established; and randomly selecting three teaching points on the intersecting line track, substituting the coordinates of the three teaching points in the two coordinate systems into the constraint equation, solving to obtain an attitude matrix and a translation vector, completing the solving of a conversion matrix, and realizing the calibration of the main coordinate system. According to the method, calibration can be completed only through any three teaching points, auxiliary equipment is not needed, the method is adaptive to all types of intersecting line models, the method has the advantages of being easy and convenient to operate, high in efficiency and precision and wide in application range, and an accurate coordinate system conversion basis can be provided for track planning and interpolation calculation of scenes such as robot intersecting line welding and cutting.

The invention discloses a weld deformation tracking compensation method based on a convolutional neural network, and aims to solve the problem that continuous and stable tracking of a weld is difficult under complex working conditions such as arc light, smoke dust, splashing and welding gun shielding. Outputting a shielding probability graph by using the shielding sensing branch, and generating modulation parameters to perform channel and space modulation on the features; a weld joint probability graph is output through the weld joint position estimation branch, weld joint candidate points are determined, and meanwhile, a position variance graph is output through the uncertainty estimation branch to represent the positioning uncertainty of the candidate points; and determining a fitting weight by combining the shielding probability and the position variance, performing weighted geometric fitting on the candidate points to obtain a weld center line model, performing continuous gating on the compensation gain according to the fitting residual error and the tracking credibility, and smoothly outputting a compensation instruction in cooperation with amplitude limiting and time. Robust tracking and stable compensation control over the center line of the welding seam under shielding interference are achieved, and the tracking reliability and the welding track compensation stability are improved.

The invention provides a lane line tracking method, device and equipment. The method comprises the following steps: detecting a lane line position in a target image to obtain a lane line detection result; according to a lane line detection result, coordinates of a corresponding point chain are determined, and the point chain is a discrete point set which corresponds to the lane line and is arranged according to a spatial sequence and is used for representing a continuous geometrical shape of the lane line; based on the coordinates of the point chain and a preset highest polynomial order, constructing a plurality of lane line models for fitting the point chain; the plurality of lane line models respectively correspond to different polynomial orders and do not exceed a preset highest polynomial order; fitting errors corresponding to all the lane line models are calculated, and the lane line model with the minimum fitting error serves as a final lane line model; and performing lane line tracking based on the final lane line model to obtain a lane line track. By verifying the applicability of different models, the problem of high-order term overfitting caused by noise is avoided, the accuracy of trajectory prediction is ensured, and the lane line fitting stability is improved.

The application discloses a kind of general line shape calculation methods of railway line, it is related to railway design technical field, including the following steps: constructing line model management link table container;Through the curvature expression of constructing straight line, circular curve, easement curve, create general line shape data structure;Using curvature expression, combined with numerical integration and nonlinear equation solution, calculate line parameter;Interconversion is carried out to two kinds of data organization models of intersection method and unit method;Build mileage conversion system, carry out the conversion of line model mileage and broken chain;Through using model splitting, merging, editing means, realize the interactive modification of line model;According to the design specification and design speed used by line model, carry out line model compliance check;Line model coordinates, line spacing are calculated and output results;The calculation method is novel and unique, can greatly improve the precision and calculation efficiency of railway line line shape calculation, with strong versatility and expandability.

The invention relates to the technical field of semiconductor AI analysis, in particular to a method and system for constructing a model pool of a machine group, and the method comprises the steps: dividing the machine group into at least one sub-group according to a preset classification condition, and enabling the sub-group to comprise at least two wafer machines; creating a primary model for the wafer machine; the method comprises the following steps: screening operation data according to a node business rule and creating a personalized primary model; and creating a secondary model for the sub-group, namely screening the operation data according to a group business rule, and creating a universal secondary model. The invention provides a double-route model interactive updating method, which provides two sets of parallel routes of a personalized primary model and a mastered secondary model, and the primary model and the secondary model can influence each other in an updating process, namely, interactive updating is realized. Therefore, smooth operation of the current production line model can be maintained, and the secondary model can quickly provide a reliable model scheme for starting a new production line.

This invention discloses a method, apparatus, device, medium, and product for generating guide line effects. The method involves: determining the target axis scaling information of the guide line model; determining the final scaling information of the guide line model based on the target axis scaling information and a preset standard length for a single guide line effect; obtaining the tiled texture coordinates of the guide line model based on its base texture coordinates and the final scaling information; determining the scaled texture coordinates of the single guide line effect based on the tiled texture coordinates; determining the sequence frame texture coordinates based on the tiled texture coordinates and the scaled texture coordinates of the single guide line effect; and sampling the sequence frame texture map based on the sequence frame texture coordinates to generate the guide line effect. This invention enables consistent lengths of guide line effects across guide lines of varying lengths.

The embodiment of the application provides a project whole life cycle integrated management method and system based on a digital main line technology, relates to the technical field of project management, and first acquires an original project planning document set in a project starting stage, generates a project task decomposition structure tree according to this, then constructs a task time sequence dependent relationship network, clearly defines the logical relationship and an estimated duration parameter between task node units, then carries out resource demand matching processing to obtain a resource allocation quota set, and finally generates a project initial digital main line model by fusing the task time sequence dependent relationship network and the resource allocation quota set, so that the application realizes integrated management of project whole life cycle information, and improves the scientificity and accuracy of project management.

The invention provides a newspaper layout character block detection and recognition system based on scanning, and relates to the field of newspaper layout character detection and recognition. According to the method, four models of newspaper layout analysis, character block detection, text line detection and character recognition are trained through a deep neural network algorithm, a newspaper image based on scanning is divided into a plurality of sub-modules through the layout analysis model, and character block detection processing is carried out on sub-module images of all versions; therefore, the sub-layout titles and the distribution information of the multiple columns of character blocks are obtained, and image noise elimination processing is carried out on the sub-layout titles and the distribution information. The text line detection model carries out single-line text detection on the character blocks and calls the character recognition model to carry out character recognition, and a language model is introduced to carry out secondary correction on a recognition result. And finally, recombining and typesetting the identification result through position coordinate information output by the layout analysis model, the text block model, the text line model and the like, and restoring the original sample to a newspaper layout typesetting state before identification.

The invention discloses a multi-dimensional intelligent risk control early warning system based on stream batch fusion and dynamic graph learning, and the system comprises a five-layer three-engine architecture system which comprises a data access layer, a data storage layer, a core calculation layer, a function service layer, and a unified interaction portal. The three engines are a real-time calculation engine, an offline calculation engine and a graph learning engine. By setting a'five-layer three-engine 'architecture, the cooperation of stream batch fusion and dynamic graph learning is realized, and a complete risk control closed loop of'real-time feature extraction-dynamic anomaly reasoning-mapping root cause tracing-offline model iteration' is constructed. Actual measurement shows that the real-time processing delay of the system is less than or equal to 1.5 seconds, the abnormal identification F1 value is greater than or equal to 0.93, and the root cause positioning time consumption is less than or equal to 10 seconds, so that the problems of insufficient real-time performance, low precision and difficult traceability of a traditional risk control system are effectively solved, and the intelligent risk control requirement in the field of network security audit can be adapted.

This application provides a telephone fraud risk assessment method based on the fusion of multi-agent and machine learning, belonging to the field of telecommunications fraud risk assessment technology. It involves collecting and processing historical call data, constructing a call behavior feature system based on the processed data, and training a machine learning model using this feature system. A fraud case knowledge base is constructed based on fraud case data. The method receives the number to be assessed and obtains its real-time call data. Using a multi-agent approach, the machine learning model and the fraud case knowledge base are used to perform risk analysis on the real-time call data, resulting in multiple independent sub-scores. These independent sub-scores are then adaptively weighted and fused to obtain a final risk score. The final risk score is then mapped to a risk level. Thus, through offline model training and case knowledge base construction, online multi-agent parallel analysis, and adaptive weighted fusion of multi-dimensional scores, a high-precision, interpretable, and real-time telephone fraud risk assessment is achieved.