391results about How to "Reduce demand" patented technology

The invention discloses a multi-modal large model automatic evaluation method and system based on Agentic Workflow, relates to the technical field of artificial intelligence model evaluation, and solves the technical problems that an existing large model evaluation method lacks image-text evaluation, is relatively high in subjectivity and is easy to cause illusion. The method comprises the following steps that S1, an input task is analyzed, and a structured task template is generated; s2, generating a multi-round multi-modal task flow chart based on the task template generated in the step S1; s3, based on the flow chart in the step S2, distributing the multi-modal large model component to each task node, executing a task and obtaining an intermediate result; s4, performing multi-dimensional scoring on the intermediate and final results in the step S3 by a plurality of scoring Agents, and fusing to generate an evaluation result; and S5, generating a feedback report based on the process data in the steps S1 to S4, identifying a model short plate, and providing an optimization suggestion. The method has the advantages of comprehensive evaluation of multiple modes, real-time feedback optimization and the like.

The invention discloses a reinforcement learning-based casting cleaning process parameter adaptive control method, which comprises the following steps of: constructing a reinforcement learning model for updating casting cleaning process parameters; the method comprises the steps that in the process that a casting cleaning system cleans a casting according to a set control period, key process signals in each control period are sequentially collected, and observation vectors are generated; inputting the obtained observation vector into a reinforcement learning model, and processing the observation vector by the reinforcement learning model to obtain a process parameter adjusting quantity, and the casting cleaning process parameters of the current control period are updated according to the process parameter adjusting quantity to obtain updated casting cleaning process parameters, and the updated casting cleaning process parameters are adopted to control a casting cleaning system of the next control period. According to the casting cleaning process parameter self-adaptive control method based on reinforcement learning, continuous and self-adaptive adjustment of the process parameters can be achieved, and the problem that fixed parameter control is insufficient in adaptability under complex working conditions is solved.

The invention discloses a thermodynamic system analysis method and system based on a physical information neural network. The method comprises the following steps: constructing a thermodynamic system mathematical model; converting a solving problem of the mathematical model of the thermodynamic system into an optimization problem taking a minimum control equation set residual error as a target function; constructing a physical information neural network; constructing a composite loss function based on the control equation set residual sum of squares, and performing end-to-end training on the physical information neural network to obtain a trained physical information neural network; taking the trained physical information neural network as an approximate solver, and for any given thermodynamic system input working condition, predicting a group of approximate solutions as initial predicted values through one-time forward propagation; and inputting the initial predicted value into a traditional numerical solver for accurate solving to obtain a to-be-solved state variable. According to the method, a feasible technical path is provided for applications with strict requirements on the analysis and calculation speed of the thermodynamic system, such as real-time simulation, online optimization and digital twinning.

This invention belongs to the field of polyurethane elastomer technology and discloses a high-strength, room-temperature self-healing lignin / polyurethane composite elastomer, its preparation method, and its applications. The preparation method includes the following steps: First, a long-chain diol, irregular isocyanate, hydrazine chain extender, and dioxime chain extender are progressively polymerized to obtain a polyurethane matrix with multiple dynamic bonds. Then, high-rigidity lignin and metal ions are introduced to prepare a lignin-reinforced polyurethane composite elastomer. The lignin / polyurethane composite elastomer prepared by this invention exhibits good compatibility, high strength, room-temperature self-healing, and repeatable heat processing characteristics, overcoming the problems of poor compatibility and the inability to simultaneously achieve high strength and low-temperature self-healing in general lignin-based polyurethane elastomers.

The invention discloses an electronic nose gas classification model construction method and system, and relates to the technical field of gas sensing detection and artificial intelligence crossing. The method comprises the steps of performing fine adjustment on a pre-trained Transform encoder based on drift-free source domain data, constructing a source domain reference model and initializing an online updating component; in the real-time classification stage, after preprocessing and feature textualization conversion are carried out on target domain data, the drift mode type of the target domain data is judged by calculating feature distribution similarity, and representative samples are screened from current data; aiming at the new drift mode, training a new LoRA adapter; for a known drift mode, loading an existing LoRA adapter and carrying out lightweight fine tuning; and finally, obtaining a model for gas classification. According to the method, self-adaptive and lightweight online learning of sensor drift is realized, the calculation overhead and the storage requirement are reduced while the classification precision is ensured, and the long-term stability of an electronic nose system in a dynamic environment is improved.

The water valve of this application has four ports. The second branch has three connection ports, and the first branch has two ports. The four ports of the water valve connect the four connection ports of the two branches. The two ports of the first branch are connected and connected to one connection port of the second branch. By using a water valve with four ports to connect two branches with five connection ports, the connection of the thermal management system is relatively simple, and the requirement for water valve interfaces is reduced. In addition, only two ports are connected in each working mode of the water valve, which also simplifies the design of the water valve.

The utility model discloses a control system for a wet-method pulverizing production line, which relates to the technical field of mechanical pulverizing equipment and comprises an industrial personal computer and an industrial router connected with the industrial personal computer through Ethernet communication. The jaw crusher control component, the cone crusher control component, the transfer stock bin control component, the double-roller flour mill control component, the vibrating screen control component, the separating tower control component, the magnetic separator control component, the second dewatering screen control component and the first conveying belt control component are connected with the industrial router through Ethernet communication. A first conveying belt control component, a second conveying belt control component, a third conveying belt control component, a fourth conveying belt control component, a first conveying pump control component and a water spraying device control component; the device has the characteristics that the device is convenient to operate and control, one-key start, one-key stop and one-key sudden stop of the whole line can be realized, and the operation data of the whole line can be automatically stored; the device is matched with a wet-method pulverizing production line and is suitable for preparing mineral powder such as quartz sand powder, dolomite powder and calcium carbonate powder.

The application provides a universal vacuum circuit breaker test switching device, a control method, a medium and a product. The device comprises: a portable test box, a switching power supply module for converting alternating current into direct current is arranged in the portable test box, and an automatic energy storage power supply and an unblocking power supply for a high-voltage circuit breaker to be tested are realized; at least one set of test terminals is arranged on a panel of the portable test box, and is used for connecting external test equipment; a switching device socket is used for realizing the connection between an aviation plug of the high-voltage circuit breaker to be tested and the portable test box; and a switching device plug is used for realizing the connection between an aviation socket in a high-voltage switch cabinet and the portable test box. The application has the advantages of safety and reliability, high efficiency and convenience, and can effectively improve the safety and accuracy of the test process, simplify the operation process and improve the work efficiency.

The application provides a three-dimensional reconstruction method based on multi-dimensional information fusion and a convolutional neural network, and comprises the following steps: acquiring polarized sub-images of a measured object under four transmission polarization directions and including multiple detection spectral bands through a multispectral polarization camera; performing image registration and joint denoising on all images; calculating polarization parameters of the measured object, combining spectral information and polarization information to calculate an azimuth angle and a zenith angle of the measured object; and acquiring surface normal vector component three-dimensional data of the measured object through a surface normal reconstruction model constructed based on a convolutional neural network and with the aid of physical prior knowledge. The method provided by the application can effectively restore the texture details of the surface of the measured object, can realize fast three-dimensional information acquisition of a measured object with a surface simultaneously having a diffuse reflection and a specular reflection characteristic in a complex indoor or outdoor scene under illumination conditions, and is helpful to expand the application range of the polarization three-dimensional reconstruction technology.

The present application relates to a gas tracing method, in particular to a kind of environmental identification and tracing method, system and medium based on multi-modal cooperation, method includes: collection contains gas plume visual image and contains gas response signal olfactory signal;Visual image is preprocessed and visual feature is extracted, and visual depth feature vector is obtained;Olfactory signal is preprocessed and olfactory feature is extracted, and olfactory feature vector is obtained;Cross-modal fusion obtains fusion feature vector;Physical constraint loss function is constructed, and physical consistency correction is carried out, and modified fusion feature vector is obtained;Output gas concentration space-time distribution result.Compared with prior art, the present application solves the problem that prior art cannot realize high-precision online monitoring and tracing of multi-component gas.The scheme improves the identification accuracy of gas component, the multi-component distinguishing ability and real-time response performance in complex environment through the integration of visual monitoring and olfactory monitoring, combined with multi-modal fusion algorithm and physical consistency correction.

The invention discloses an MICP intensity prediction method and system based on machine learning and multi-feature fusion, and relates to the field of solidified data analysis. An accurate quality prediction mapping model and an abnormal compensation parameter mapping model are constructed based on historical data; the system has the capabilities of predicting curing quality from process parameters and intelligently recommending an optimal disposal scheme from an abnormal state; by setting a dual-threshold quality control mechanism and a real-time abnormity monitoring and early warning system, various abnormal conditions in the curing process can be actively identified and timely responded, and complete closed-loop processing from parameter deviation detection to automatic compensation measure execution to hardware standby switching can be realized. The stability of the solidification process and the consistency of the engineering quality are effectively guaranteed, the requirement for manual intervention and the misjudgment risk are greatly reduced, powerful technical support is provided for standardization, normalization and industrialization development of the sand solidification technology, and the engineering efficiency and economic benefits are remarkably improved.

The utility model discloses a rotary adjustable carbon fiber stock solution spinning assembly, and belongs to the technical field of carbon fiber production equipment. Comprising a spinneret assembly mounting frame, a transverse frame is transversely arranged on the middle section of the spinneret assembly mounting frame, an auxiliary frame is arranged below the transverse frame, a plurality of top plates are arranged at the top of the transverse frame, a plurality of long gooseneck pipe frames are arranged at the top of each top plate, and a front placement plate is arranged on the front face of the auxiliary frame; a plurality of short gooseneck pipe frames are arranged on the front placement plate; by arranging the long gooseneck pipe frame and the short gooseneck pipe frame in an up-and-down crossed mode, more structures used for spinning are installed in a small space, meanwhile, the length of the coagulating bath tank is increased in a matched mode, containing of multiple sets of spinning nozzles is achieved, and therefore the whole working efficiency can be greatly improved; meanwhile, the long gooseneck pipe frame and the short gooseneck pipe frame are rotationally installed through a swing seat and a limiting sleeve, and the maximum rotation angle is limited through a limiting sliding groove and a butterfly bolt.

A motor assembly includes an electric motor and an overrunning clutch wherein a shaft of the overrunning clutch is adapted to engage with an external engine, the shaft is connected to an inner member of the overrunning clutch, and the outer member of the overrunning clutch is adapted to engage with a rotor shaft of the electric motor.

The invention provides an SSB time-frequency synchronization method and device, and relates to the technical field of mobile communication. The method comprises the following steps: receiving a synchronization signal block SSB signal sent by a base station; processing the SSB signal to obtain a target time domain sequence generated after processing, and performing sliding correlation on the target time domain sequence to obtain a time domain correlation value sequence; acquiring a peak value of the time domain correlation value sequence, and performing time synchronization of the SSB signal based on time offset corresponding to the peak value; after the time synchronization of the SSB signal is completed, obtaining a cyclic prefix (CP) removal sequence corresponding to the time domain sequence of the SSB signal, and obtaining a pre-frequency offset PSS time domain sequence generated after a primary synchronization signal PSS time domain sequence is subjected to multiple times of frequency offset adjustment; performing correlation calculation based on the CP-removed sequence and the pre-frequency offset PSS time domain sequence to determine a target estimated frequency offset; and completing frequency synchronization of the SSB signal based on the target estimation frequency offset. According to the invention, the PSS can be locked quickly and accurately to realize time-frequency synchronization.

A lithium battery data reconstruction and classification method driven by adversarial learning and feature distribution alignment, relating to the field of lithium battery state monitoring technology, mainly includes the following steps: constructing a feature recognizer to capture the temporal correlation features of lithium battery data throughout its entire life cycle; integrating the feature recognizer into an adversarial training framework, aligning the features of the data generated by the feature extractor with those of the real data through adversarial learning between real lithium battery data and simulated data generated by the feature extractor; constructing a feature transfer-based data reconstruction network based on the feature extractor and combining it with the Unet network to reconstruct battery data consistent with the features of the real data, forming an expanded dataset with the original data; and building a battery state classification network based on the feature recognizer, training it with the expanded dataset to complete the state classification of lithium batteries. This method ensures the effectiveness of the reconstructed data through feature distribution alignment, significantly improving the generalization ability and recognition accuracy of the classification model.

The application discloses an automatic replacement system of color photovoltaic adhesive film, and particularly relates to the photovoltaic power station operation and maintenance technical field, comprising a sensing and detecting module, a decision control module, an execution operation module, a material processing and supply module and a safety auxiliary module; the sensing and detecting module is used for acquiring the position, contour and surface state data of the photovoltaic module; the application constructs a three-dimensional grid model of the module through the decision control module, realizes intelligent scoring and global path optimization of the work unit based on a rule base, and generates an optimal work sequence; through adaptive process parameter adjustment in the execution operation module, the hot air power and the winding tension are dynamically adjusted according to the surface color, temperature and defect characteristics of the module, so that the old adhesive film peeling quality and the new adhesive film laying precision are significantly improved; the whole-process automation from detection, decision, execution to quality review is realized, the replacement efficiency and quality consistency are effectively improved, and the work safety is ensured.

This invention belongs to the field of autonomous driving technology and discloses a vehicle lane change control method, device, equipment, and storage medium. This application detects the forward lane change space margin of the target lane corresponding to the high-speed lane change request when a high-speed lane change request is detected. If the forward lane change space margin is greater than a preset lane change space threshold, a reference vehicle is selected from the backward lane change space in the target lane. When the longitudinal distance between the reference vehicle and the main vehicle is less than the original lane change threshold but greater than or equal to the lane change attempt threshold, the main vehicle is controlled to slowly approach the lane line, and the speed of the reference vehicle is obtained. If the speed is less than the speed of the main vehicle, the main vehicle is controlled to change lanes to the target lane. Because it attempts to slowly approach the lane line when the longitudinal distance is between the original lane change threshold and the lane change attempt threshold, and further judges based on the speed before executing the lane change, the requirement for backward lane change space for automatic lane changing is reduced.

This application discloses an insulin pen packaging box, relating to the field of pharmaceutical packaging technology. Its key technical features are: an outer box body composed of a box body and a freely opening and closing lid; a paper tray embedded inside the box body, the projection of the paper tray on the horizontal plane coinciding with the bottom of the box body; the paper tray includes a pad, on which multiple insulin pen placement frames are arranged side-by-side, and each placement frame is fixedly connected to the pad; each end of a placement frame has a slot along its length, and when an insulin pen is placed in a slot, the slot provides support and restraint to the end of the insulin pen. This allows for the simultaneous storage of multiple insulin pens within a single packaging box, thus improving packaging efficiency and patient convenience.

This invention belongs to the field of 3D printing model preprocessing technology, and particularly relates to a method, device, electronic device, and storage medium for generating model edge supports. The method includes: traversing the model; obtaining the minimum model frame; aligning the bottom center point of the minimum model frame to the origin of the zero-plane platform; raising the model by H millimeters; dividing the zero-plane platform into preset squares; obtaining the maximum vertical projection range of the model's bottom plane on the zero-plane platform; obtaining all preset squares within the projection range as candidate squares; obtaining the center points of the boundary squares; sequentially extracting the center points of the boundary squares with a linear interval of L + ΔX millimeters and determining them as extraction points; drawing vertical lines upwards from each extraction point to intersect the bottom of the model as target points; drawing support units downwards from each target point and connecting them between the bottom edge of the model and the zero-plane platform; and storing the 3D data. This method enables the batch automatic generation of support units at the bottom edge of the model.

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

The invention discloses a PCCP pipe bell and spigot joint sealing system manufacturing method and a PCCP pipe bell and spigot joint sealing system manufactured through the PCCP pipe bell and spigot joint sealing system. During manufacturing, a joint is filled with an elastic blocking piece, the elastic blocking piece is divided into a first sub-joint and a second sub-joint, and the elastic blocking piece is arranged in the joint in an abutting mode through a locking mechanism; mortar is poured into the first sub-joint through a reserved first grouting opening, and the locking mechanism is removed after the mortar does not flow; a base layer is arranged on the upper end face of an upper opening of the second sub-joint, anticorrosive paint is applied to the base layer to form an anticorrosive coating, the anticorrosive coating and the base layer form a plugging structure and seal the second sub-joint, and sealant is poured into the second sub-joint through a reserved second grouting opening; the elastic barrier piece is basically non-sticky to mortar and the like; the method is suitable for construction in the low-temperature environment, the obtained sealing system can adapt to different stress interferences and different environmental factors, lasting sealing can be achieved, the construction technology can be simplified after the anti-corrosion coating is formed in a front-arranged mode, and the construction speed is high.

This application discloses a method, method, and apparatus for training an information extraction model. The main technical solution includes: acquiring a third dataset, which includes multiple unlabeled third text samples; using a pre-trained first information extraction model to extract target information of a preset type from the third text samples in the third dataset, as a second target sample; inputting the second target sample into a pre-trained text generation model to obtain text generated by the text generation model using the second target sample as a second text sample; forming a second sample pair by combining the second target sample and its corresponding second text sample; obtaining a second dataset using multiple second sample pairs; and training a second information extraction model using the second dataset. This application can improve the information extraction effect of the information extraction model while reducing manual costs and time consumption.

This application belongs to the field of marine observation and monitoring technology, and relates to a long-term mooring test device and method for marine sensors. The method includes: an energy management step: formulating and implementing a graded energy management strategy based on energy supply and demand balance, marine environmental parameters, and test mission requirements, and initiating data acquisition; an anomaly response step: continuously analyzing the changing trends of marine environmental parameters, automatically switching to emergency operating mode and sending alarm information when an anomaly is detected; real-time monitoring of the optical window transmittance of each sensor, and assessing the degree of sensor fouling based on the optical window transmittance; automatically performing anti-fouling treatment when the degree of sensor fouling reaches a preset anti-fouling trigger threshold; otherwise, determining that the sensor is at an acceptable level of fouling, and continuing to monitor transmittance and assess the degree of fouling.

The application relates to the technical field of integrated circuits, and discloses a digital auxiliary adaptive zero-crossing detection circuit. The circuit comprises a digital auxiliary comparator module, a digital logic module, a detection discrimination module and a counter module. The digital auxiliary comparator module compares a switch node voltage signal with a power ground signal to generate a first output signal; the digital logic module processes the first output signal to generate a zero-crossing detection signal and control the turn-off of a power tube; the detection discrimination module receives a second output signal of the digital logic module and the zero-crossing detection signal to generate a digital tuning signal; and the counter module receives the digital tuning signal to generate a six-digit digital signal, dynamically adjusts a resistance array in the digital auxiliary comparator module, adaptively adjusts the flip offset of the comparator by changing the resistance value of the resistance array, and thus changes the turn-off timing of the power tube. The application can accurately control the turn-off timing of a synchronous rectification power tube, reduce the reverse current and reduce the dynamic power consumption.

The present application belongs to the field of microbial preparation, and discloses a kind of microbial stabilizer, including non-bacillus or its fermentation bacteria liquid, chemical stabilizer, wherein, the non-bacillus includes at least one of Serratia entomophila, Acinetobacter calcoaceticus.The present application also discloses the preparation method of the microbial stabilizer and its application in prolonging the shelf life of microbial agent.The present application studies the microbial stabilizer containing non-bacillus, finds that the microbial stabilizer formed by the suitable collocation of non-bacillus and chemical stabilizer can maintain the stability of viable bacterial concentration for a long time, can prolong the shelf life of microbial product, reduces the need for frequent production and storage, can ensure the supply of product in market more stable, thereby directly reduces the production cost, improves economic benefit, can be used in the field of microbial preparation technology.

The utility model relates to the technical field of agricultural unmanned aerial vehicles, and discloses an agricultural plant protection unmanned aerial vehicle which comprises an unmanned aerial vehicle body, a pesticide box is fixedly connected to the bottom of the unmanned aerial vehicle body, a plurality of fixing rods are fixedly connected to the peripheral side of the unmanned aerial vehicle body, the end of each fixing rod is connected with a vehicle arm through a folding assembly, and each folding assembly comprises a sleeve fixedly connected with the corresponding fixing rod. The side wall of one side of the sleeve is provided with a through groove, the side wall, located on one side of the through groove, of the sleeve is rotationally connected with a rotating shaft, the rotating shaft is fixedly connected with a connecting block, the machine arm is fixedly connected to the side of the connecting block, the connecting block is provided with a pressing plate, the pressing plate is connected with a locking pin capable of transversely sliding, and the side wall of the sleeve is provided with a locking groove. The locking pin can be embedded into the locking groove to limit rotation of the machine arm, and the locking pin is sleeved with a first reset spring used for driving the locking pin to be embedded into the locking groove. According to the agricultural unmanned aerial vehicle, a folding type vehicle arm structure is adopted, rapid folding of the vehicle arms is achieved, the transportation and storage space requirements are remarkably reduced, and the convenience of the agricultural unmanned aerial vehicle is improved.

The present application relates to the field of coking, synthetic ammonia and the like containing ammonia wastewater treatment, and discloses a novel ammonia evaporation energy-saving process system, a primary ammonia evaporation tower, an intermediate pressurizing pump set, a water flow regulating device, a secondary ammonia evaporation device, a flash tank, a steam ejector, a pressurizing pump set and a control unit. After the ammonia-containing wastewater is preliminarily treated by the primary ammonia evaporation tower, it is delivered to the secondary ammonia evaporation device by the pressurizing pump, and introduced into new steam for secondary ammonia evaporation. After the secondary ammonia evaporation, the wastewater flows into the flash tank to produce flash steam by cooling and pressure reduction, and the flash steam is pressurized by mixing with the new steam through the steam ejector, and then returns to the primary ammonia evaporation tower or the secondary ammonia evaporation device, so that the steam is recycled multiple times. The system can effectively reduce the new steam consumption by 35-45% under the premise of ensuring the stability of the ammonia nitrogen index of the effluent, improve the energy utilization rate, reduce carbon emissions, and is suitable for direct ammonia evaporation and indirect ammonia evaporation systems.

The application relates to the technical field of PCIe devices, and discloses a parameter configuration method and device, a computer device and a storage medium, the method comprising the following steps: setting a plurality of first preset value combinations; configuring preset values of a plurality of PCIe devices according to the first preset value combinations, and performing first testing on the plurality of PCIe devices to obtain first testing results; determining an effective preset value range that can be configured by each PCIe device according to the first testing results; selecting an effective preset value from the effective preset value range corresponding to each PCIe device to form a second preset value combination; configuring preset values of the plurality of PCIe devices according to the second preset value combination, and performing second testing on the plurality of PCIe devices to obtain second testing results; and determining an optimal second preset value combination according to the second testing results corresponding to the plurality of second preset value combinations. The application can determine optimal preset values corresponding to each PCIe device respectively, and can simultaneously ensure the stability of PCIe links of each PCIe device.

A kind of 900MPa grade magnetic pole steel for pumped storage hydroelectric generating set, its component and wt% are as follows: C: 0.25~0.30%, Si: 0.10~0.20%, Mn: 1.1~1.5%, Cr: 0.3~0.4%, Ti: 0.2~0.3%, B: 0.0025~0.0030%, P: ≤0.015%, S: ≤0.010%, Als: 0.02~0.10%; production method: after conventional smelting, cast into blank; the cast blank is heated; adopt 7 passes hot rolling to product thickness; laminar flow cooling is carried out; coiling; annealing; straightening is carried out.The application does not contain valuable alloy elements such as Nb, Mo, etc., and its yield strength is greater than or equal to 900MPa, elongation A is greater than or equal to 12%, magnetic induction performance B 50 ≥1.60T, can fully meet the demand of 900MPa grade magnetic pole steel for pumped storage hydroelectric generator rotor suitable for rated speed of 600r / min, product thickness is not more than 2mm, and cost can be reduced by 600~900 yuan / ton of steel compared with prior art.

The application discloses a power distribution network topology fast adaptive inference method and device based on meta learning, and belongs to the technical field of smart grids. The method comprises the following steps: acquiring a first data set; the first data set comprises a plurality of groups of first topology data and first measurement data of the power distribution network; the first topology data represents the connection relationship between nodes in the power distribution network, and the first measurement data represents the operating state of each node; the first topology data is taken as the label of the first measurement data, and a meta initial model is trained; second measurement data and topology label data of a target power distribution network are acquired; the topology label data represents the connection relationship between part of nodes in the target power distribution network; the topology label data is taken as the label of the second measurement data, the meta initial model is trained, and a target topology inference model is obtained; real-time measurement data of the target power distribution network is acquired, the real-time measurement data is input into the target topology inference model, and target topology data is obtained. The application can improve the efficiency of power distribution network topology inference.