6657 results about "Electric arc" patented technology

A submerged arc furnace intelligent inspection system based on cloud side-end cooperation belongs to the technical field of submerged arc furnace intelligent robot inspection, and comprises an end side device, an edge side platform, a cloud platform and a cross-layer cooperative processing module, the end side device is in communication connection with the edge side platform, and the edge side platform is in communication connection with the cloud platform; and the cross-layer cooperative processing module is used for intelligently deploying global resources of the end-side equipment, the edge-side platform and the cloud platform, predicting a future load trend by adopting a dynamic sparse LSTM model algorithm, inputting a prediction result into a security constraint PPO reinforcement learning algorithm to adaptively adjust a task and a computing power allocation proportion, and carrying out dynamic resource allocation according to an output result. According to the invention, the cross-layer cooperative processing module interacts and communicates with the cloud side, the edge side and the end side, so that a dynamic allocation strategy of resources is realized, and the problems of low resource utilization rate and poor real-time performance caused by a rigid resource allocation strategy during routing inspection are avoided.

The invention discloses a direct-current arc fault diagnosis method and device with adaptive capacity. The method comprises the following steps: 1) collecting a current alternating-current component when a current direct-current system normally works; 2) extracting time-frequency domain characteristics of a current AC component, wherein the time-frequency domain characteristics comprise a current peak-to-peak value, a standard deviation and an amplitude reconstruction harmonic power sum; 3) performing multi-time-frequency-domain feature fusion based on mahalanobis distance to obtain arc detection features; 4) calculating an arc fault judgment threshold value based on a box plot method; 5) collecting a current AC component in real time to calculate a Mahalanobis distance, comparing the Mahalanobis distance with an arc fault determination threshold, and recording an arc event; the method has the advantages that the system adaptability is high, the detection rate is high, the false alarm rate is low, and the method can be widely applied to high-voltage and low-voltage direct-current power distribution systems such as airplanes, vehicles and ships.

The invention discloses a welding robot track self-adaptive correction method based on multi-modal data fusion, and particularly relates to the technical field of industrial robots, which comprises the following steps: synchronously acquiring visual three-dimensional point cloud and arc current and voltage transient waveform data in a welding process in real time; an enhanced working trajectory feature vector fusing the visual trajectory features and the arc deviation compensation amount is constructed; respectively calculating real-time confidence coefficients of the visual sense and the arc sensing mode; inputting the confidence coefficient into a weight distribution strategy model, dynamically resolving vision and arc weight coefficients, and when the confidence coefficient of one sensor is low, lifting the weight of the other sensor to dominate correction; performing weighted fusion on the two pose adjustment amounts by using a weight coefficient to generate a six-degree-of-freedom comprehensive trajectory correction amount; and finally, the robot is driven to execute online adaptive trajectory correction. Through a confidence-driven dynamic fusion strategy, the limitation of a single sensor is overcome, and the welding robot trajectory tracking precision and the adaptive capacity are improved.

The invention belongs to the technical field of intelligent control, and particularly relates to an intelligent regulation and control method for an operating environment in an SF6 ring main unit. The method comprises the following steps: firstly, acquiring temperature, gas pressure and humidity data of a plurality of sampling points in the ring main unit, and constructing a time sequence environment data set through time synchronization and missing interpolation processing; analyzing the temperature data, extracting fluctuation characteristics, determining an environment disturbance factor in combination with an electric arc trigger intensity index, and obtaining a cabinet body temperature control safety factor in combination with a dynamically adjusted temperature rise threshold and a gas insulation state; then constructing an environment equivalent thermal inertia model and a hot-pressure covariant model based on historical data, and calculating a thermal stability index of the cabinet body; and finally, whether adjustment is triggered or not is judged according to the thermal stability index and the temperature control safety factor, if yes, a temperature abnormal area is positioned through a convolutional neural network and a clustering method, the temperature control response level is calculated, a temperature control strategy library is called to be matched with adjustment parameters, and accurate control is achieved.

The invention relates to the technical field of circuit control, in particular to a control method and system of a bidirectional charging and discharging circuit. The method comprises the following steps: collecting current monitoring data and voltage monitoring data of a bidirectional charging and discharging circuit in real time; performing arc characteristic analysis on the current monitoring data and the voltage monitoring data to obtain arc characteristic data; analyzing a high-frequency harmonic component and a current ripple fluctuation amplitude of the bidirectional charging and discharging circuit based on the arc characteristic data; performing direct-current arc detection on the bidirectional charging and discharging circuit according to the high-frequency harmonic component and the current ripple fluctuation amplitude to generate a direct-current arc fault judgment result; and calculating the instantaneous current change rate of the current monitoring data, and carrying out short-circuit detection on the bidirectional charging and discharging circuit through the instantaneous current change rate. By monitoring current and voltage data in real time, arc and short circuit characteristics are accurately analyzed, relay control is optimized, an accurate control instruction is generated, and the safety and reliability of the two-way charging and discharging circuit are improved.

The invention discloses a gradient welding strength control method and system for a welded galvanized steel pipe for fire fighting, particularly relates to the technical field of welding automation control, and aims to solve the problem of abnormal welding seam strength gradient caused by instantaneous unbalance of an electric field and a thermal field during dynamic parameter switching in the prior art. Electromagnetic distortion characteristic quantity is extracted through frequency domain energy analysis, and thermodynamic offset state quantity is deduced in combination with a thermal diffusion trend; on the basis of frequency band correlation mapping of electromagnetic and thermodynamic parameters, thermoelectric cooperative imbalance levels are divided through mutual information entropy evolution, and a self-adaptive harmonic attenuation weight tuning instruction is generated; analyzing a mismatching relation between transient disturbance of an environment magnetic field and arc voltage modulation distortion in real time, dynamically generating a magnetic field compensation factor and reversely superposing the magnetic field compensation factor to a control instruction; and in combination with the matching degree of molten pool oscillation energy and a harmonic frequency spectrum, correcting wire feeding rate calibration, reconstructing a multi-band harmonic energy proportion spectrum, outputting an anti-interference welding control signal, and realizing dynamic balance between arc stability and molten pool heat input.

The invention is suitable for the technical field of performance detection, and provides a method for detecting the ablation resistance of a burning-resistant nozzle and an arc contact, and the method comprises the steps: collecting a temperature field matrix, an arc image sequence, an electrical parameter time sequence and a surface displacement curve of an ablation region, and generating a multi-modal data set; based on the data set, extracting spatial features of the ablation core region through a cavity convolution layer and a channel attention mechanism; inputting a preset bidirectional LSTM network, fusing thermal stress data of the COMSOL simulation model to drive gating weight update, and extracting time sequence evolution characteristics; generating a space-time fusion feature vector containing the temperature field gradient, the arc form change rate and the surface roughness evolution feature; inputting the space-time fusion feature vector, the thermal stress distribution data and the material phase change threshold parameter into a constructed ablation prediction model for training; the ablation prediction model after training is completed is optimized; and a performance prediction result is generated in real time by using the optimized ablation prediction model, so that the detection accuracy is effectively improved.

The invention provides a self-adaptive teaching-free welding robot intelligent gas saving monitoring system, which relates to the technical field of industrial robots, and comprises a three-dimensional sensing module used for identifying the workpiece material category through a multispectral sensor integrated with a welding gun of a welding robot, meanwhile, a stereoscopic vision camera and a line structure light scanning camera are used for obtaining three-dimensional point cloud data of the welding seam in real time; the parameter adjusting module is used for calling a heat conductivity coefficient database based on the three-dimensional point cloud data, and correcting the initial flow of the welding current, the arc voltage and the protective gas in real time; and the matching control module is used for calculating the arc length variable quantity based on the arc voltage, controlling the opening degree of the gas flow valve through a dynamic regulator, realizing dynamic matching of the gas flow and the changed arc length, and generating gas dynamic matching control parameters. Through cooperative operation of self-adaptive attitude compensation and real-time gas consumption monitoring, teaching-free accurate operation in the welding process is achieved, and intelligent control over gas consumption is improved.

The invention discloses a 3D printing path planning method based on electric arc additive anisotropy and a stress field, and the method comprises the steps: constructing a CAD three-dimensional model of a part, and obtaining the stress field of the part through finite element simulation; determining a slice plane; mapping the stress field to a slice plane to form a force flow line; according to the obtained force flow line, rotation transformation regeneration is carried out according to the anisotropy of the used electric arc additive, and a reference trajectory considering the anisotropy of the material is obtained; under the principle of alternate arc starting and extinguishing, the corresponding printing sequence in the layers and between the layers is further planned, and the reference trajectory lines are connected end to end to be planned into a continuous path; and generating a code file, and printing according to the printing sequence. The comprehensive mechanical property of a printed piece is improved.

The invention relates to the technical field of intelligent fire fighting, in particular to a building electrical fire identification method and system based on multi-source information fusion, and the method comprises the steps: firstly collecting the multi-source monitoring data of an electrical system in real time, then carrying out the preprocessing of the multi-source data, including noise filtering, time alignment and feature extraction, and constructing a dynamic weight fusion model; the weight of each data source is adaptively allocated based on the feature entropy, then fusion features are input to the fire risk grading model, a fire probability index and an early warning level are output, and when the early warning level exceeds a threshold value, a linkage control instruction is triggered and an alarm is given; according to the method, through dynamic fusion of multi-source information, the electrical fire identification precision is remarkably improved, multi-dimensional parameters such as current, temperature, smoke and arc are adaptively integrated by adopting an entropy weight assignment mechanism, the problems of false alarm and missing alarm of traditional single-threshold detection are solved, the missing alarm rate and the false alarm rate are remarkably reduced, and the reliability of fire early warning is ensured.

The invention discloses an intelligent response method for an arc fault of a high-altitude switch cabinet, relates to the technical field of safety protection of power equipment in a high-altitude area, and aims to solve the problems that the arc time of the arc fault is prolonged, the pressure in the cabinet is sharply increased, a traditional disposable pressure relief device cannot be reset and is lack of insulation state evaluation, and the reliability is poor. The invention aims to provide a response method which combines optical signal and pressure double-criterion detection, resettable pressure relief, gas component analysis and intelligent decision so as to improve the safety and reliability of the switch cabinet.

A method protects against fault arcs in an electric power distribution system in which electric power from an electric feed point is distributed via a common superordinate main distribution line and from the common main distribution line via multiple subordinate output lines. The main distribution line can be interrupted by a main switch, and each of the output lines can be interrupted by a paired output switch. A fault arc current threshold which, when exceeded by a fault arc current of a fault arc, is a precondition for triggering the main switch is matched to the maximum forward current and / or the output trigger duration of the output switches such that a fault arc is quenched selectively either by one of the output switches or by the main switch before the energy released by the fault arc exceeds a specified energy threshold.

The invention features a replaceable cartridge for a plasma arc torch. The cartridge includes a cartridge body having a first section and a second section. The first and second sections are joined at an interface to form a substantially hollow chamber. The interface provides a coupling force that secures the first and second sections together. The cartridge also includes an arc constricting member located in the second section; an electrode included within the substantially hollow chamber; and a contact start spring element affixed to the electrode. The spring element imparts a separating force that biases the electrode toward at least one of the first section or the second section of the body. The separating force has a magnitude that is less than a magnitude of the coupling force.

The invention provides an arc discharge detection method and device for a photovoltaic inverter and electronic equipment, and the method comprises the steps: collecting sample current data corresponding to the photovoltaic inverter, and classifying and marking the sample current data as arc data and arc-free data; taking the sample current data as a data sample, taking the arc data and the arc-free data as sample labels corresponding to the data sample, and training a preset neural network model; converting the trained neural network model into an algorithm program, deploying the algorithm program to a preset micro-control unit, inputting real-time current data collected by the micro-control unit into the neural network model, and determining a data type to which the real-time current data belongs; and when the data type is the electric arc data, determining that the photovoltaic inverter has a direct current arc discharge phenomenon. And low-cost and high-precision intelligent arc discharge detection can be realized.

The invention discloses an electric arc additive manufacturing temperature field and stress field real-time reconstruction method, which comprises the following steps of: 1, constructing a sequential coupling finite element model in which heat conduction analysis and thermoelastic-plastic mechanical analysis are combined on the basis of process parameters of an electric arc additive manufacturing process and three-dimensional geometry of parts, generating a multi-physical field coupling data set among heat flux density, temperature and stress; step 2, constructing and training a neural network model based on a multi-physics field coupling data set, and realizing mapping from heat flux density to temperature and stress; 3, dynamically optimizing heat source model parameters in combination with real-time monitoring data and process parameters in the electric arc additive manufacturing process; and 4, based on the optimized heat source model parameters and the trained neural network model, real-time reconstruction of a temperature field and a stress field in the electric arc additive manufacturing process is achieved. According to the invention, the requirements of rapid sensing and intelligent regulation and control of temperature and stress states on an additive manufacturing site are met.

The invention provides a water turbine runner crack electric arc additive repairing track planning method based on an irregular model, and relates to the technical field of hydroelectric power generation equipment crack repairing. The method comprises the steps that crack defects are removed through carbon arc air gouging, and a model of a blade area to be repaired is established; based on a reverse reconstruction technology and a Boolean operation technology, generating a three-dimensional grid model of the irregular repair area; a mixed trajectory planning strategy is optimized, and the repairing area is sliced to generate an electric arc additive repairing trajectory; and welding gun posture optimization based on collision detection is carried out, and the contour track is adjusted according to a given rule. A geometric model of a to-be-repaired area of a blade is established, and a three-dimensional grid model of an irregular repaired area is quickly and accurately established based on a reverse reconstruction technology and Boolean operation.

The invention discloses a new energy medium-voltage switch intelligent control method based on multi-data interaction, and particularly relates to the field of switch intelligent control. According to the method, electrical, mechanical and temperature multi-dimensional operation data are collected in real time through a multi-source sensor, fault diagnosis is performed by adopting a hybrid reasoning architecture, features are extracted in combination with Hilbert-Huang transform and Mel cepstrum coefficient algorithms, and a three-level reasoning mechanism is constructed to realize accurate classification; based on a diagnosis result, a system dynamically optimizes a control strategy, the control strategy optimizes a switch-on and switch-off time sequence based on phase-space reconstruction and LSTM prediction, plasma impedance closed-loop adjustment and a quantum genetic algorithm are combined to realize arc suppression and mechanical compensation, and a six-dimensional compensation vector is introduced to correct mechanical operation deviation; the self-learning unit is based on a depth deterministic strategy gradient and an online knowledge distillation technology, verifies the validity of parameter update by using digital twin simulation, and forms a fault feature-control strategy-maintenance scheme triple knowledge base.

The invention discloses a direct-current hollow electrode hydrogen plasma furnace mineral powder smelting system and method and application, and belongs to the technical field of hydrogen metallurgy. The system comprises a hollow electrode gas powder injection hydrogen plasma gun system, a direct current bottom electrode ore-smelting electric furnace system, a fine ore and hydrogen supply system and the like. The hollow electrode and the furnace bottom electrode form a direct-current arc, hydrogen carrying mineral powder is sprayed into an arc area to generate high-temperature hydrogen plasma, mineral powder reduction is achieved, and metal and slag are separated. According to the method, hydrogen serves as a reducing agent, carbon emission is reduced, the reaction efficiency is high, energy consumption is low, and the method can be used for steelmaking, smelting of micro-carbon ferromanganese and titanium slag and recovery of valuable metal in nonferrous smelting waste slag and has good green metallurgy application prospects.

The invention belongs to the technical field of computer-aided manufacturing, and relates to a method and system for predicting the propagation rate of fatigue cracks across welding zones in electric arc additive remanufacturing, and the method comprises the following steps: S1, manufacturing a sample; s2, performing in-situ SEM fatigue test on the sample; s3, obtaining a corresponding relation between the crack growth rate and the stress intensity factor range of the sample; s4, fitting to obtain a fatigue crack propagation rate model of the sample; s5, establishing a CZH sample fatigue crack growth rate prediction model; and S6, predicting the fatigue crack propagation rate of the sample across the welding zone. The system further comprises an electric hydraulic fatigue test system, a data acquisition module, a database, a fitting module and a prediction module. According to the method, by using the correction coefficient, the fatigue crack growth rate is predicted by using the fatigue crack growth rate parameter of the matrix material area.

The invention relates to the technical field of welding parameter correction, and provides a welding parameter correction method for real-time molten pool morphology feedback, which comprises the following steps of: capturing a molten pool image, an arc spectrum and an infrared radiation signal in real time, accurately extracting geometrical characteristics and contour vibration frequency of a molten pool, and dynamically evaluating arc stability and a thermal attenuation state. And when defects are detected, welding parameter correction is generated based on a process knowledge graph and Bayesian reasoning, collaborative correction of parameters such as welding current, welding voltage and welding speed is achieved through reinforcement learning optimization weight distribution, and the change trend of morphology characteristics of a molten pool is continuously monitored after correction. According to the method, the depth belief network is judged and optimized by combining the descending rate of the morphology defect probability, and the corrected successful case is supplemented to the process knowledge graph, so that the welding quality stability is remarkably improved, the method is particularly suitable for automatic production of a high-precision welding scene, the defect rate can be effectively reduced, and the production efficiency can be effectively improved.

The invention discloses an arc fault detection method and device based on voltage time difference variation, and belongs to the technical field of low-voltage power distribution network fault protection, and the method comprises the steps: enabling an arc fault point to be equivalent to an arc voltage source, and employing an arc voltage change rate and an arc voltage effective value, establishing detection feature description of a fault waveform generated by the arc voltage source, wherein the fault waveform presents a square-wave-like waveform; sampling adjacent cyclic wave voltages of a load side monitoring point, performing feature identification on the tail end voltage difference between adjacent cyclic waves based on detection features, and identifying a first fault waveform; and after the first fault waveform occurs, performing continuous identification counting on the arc faults, realizing identification of the continuous arc faults by using the difference between the non-fault waveform before the fault occurs and the waveform to be identified, and confirming the occurrence of the arc faults. The detection method is good in reliability, clear in physical significance, high in detection sensitivity, not affected by line parameters, capable of achieving detection only through a single monitoring point and lower in cost.

The invention relates to the technical field of fire control, in particular to an electrical fire control method and system based on multiple sensors, and the method comprises the steps: obtaining related data of electrical fire influence factors, and carrying out the preprocessing; extracting a current peak value by using an AMPD peak value detection algorithm according to the preprocessed current data, and constructing an arc fault strength index; combining the arc fault strength index and temperature data, analyzing the relationship between arc fault abnormal heat release and temperature change, constructing an electrical heat dissipation goodness, integrating the electrical heat dissipation goodness and the change rule of carbon monoxide and carbon dioxide concentration, constructing a fire risk index, and quantifying the risk degree of electrical fire occurrence; and correcting the SOS anomaly detection algorithm based on the fire risk index to obtain a corrected significant anomaly score, and preventing and controlling the electrical fire according to the significant anomaly score. According to the invention, the fire risk index is constructed in combination with various sensor data, and the electrical fire early warning accuracy and prevention and control efficiency are improved.

The invention discloses a welding robot self-adaptive wire feeding control method, a platform and a medium, relates to the technical field of welding automation process control, and is suitable for welding process management under complex working conditions. According to the method, an arc voltage sequence and a molten pool characteristic parameter sequence in the welding process are collected, a welding state characteristic value is obtained through calculation, a reference database containing welding process parameters, welding working condition parameters and the welding state characteristic value is constructed based on historical experimental data, and corresponding preset reference indexes are generated through a mapping relation function; and in real-time welding, the welding state characteristic value of the current wire feeding control period is obtained and compared with a preset reference index, the deviation value is calculated, the welding state is judged according to the grading threshold value interval, the wire feeding speed is dynamically adjusted, and self-adaptive matching of the wire feeding speed and the welding state is achieved. According to the method, instability caused by traditional adjustment depending on experience can be overcome, dynamic control over the welding process is achieved, and the stability of a welding pool and the welding seam forming quality are improved.

The invention discloses an energy consumption control method and system for metallurgical equipment, and relates to the technical field of metallurgical equipment energy consumption control, and the method comprises the following steps: in the actual operation process of an electric arc furnace, various sensors are installed around an electrode, and the operation state of an electric arc is monitored in real time; key features reflecting the arc drift state are extracted, and complex original data are converted into quantitative indexes reflecting arc drift and the severity of the arc drift; and taking the key quantitative index as a feature vector, training the feature vector by adopting a deep learning method, constructing an arc drift prediction model, and after the construction is completed, outputting an arc state prediction result based on the current input feature vector, and identifying arc drift. According to the method, arc drift is recognized in advance by fusing multiple sensors and deep learning, through intelligent electrode height and response speed self-adaptive regulation and control, the electric energy utilization rate is increased, the smelting time and equipment loss are reduced, and low-energy-consumption, high-efficiency, automatic and green smelting is achieved.

The invention discloses narrow-gap swing laser-MIG composite filler wire welding equipment and method for a thick-wall component, belongs to the field of thick-wall component welding, and aims to solve the problems that the side wall of a narrow-gap groove of a thick plate of 50-250 mm is not fused and the welding efficiency is low. The swinging narrow-gap MIG welding gun and the auxiliary wire feeding pipe are arranged on the two sides of the laser, dynamic cooperation of the main / auxiliary double wire feeding systems is adopted, the dynamic characteristics of a molten pool are improved in combination with laser spot swinging, and the deposition efficiency is improved. The heating range of the composite heat source is expanded through the coupling effect of swing laser and electric arc, sufficient fusion of the side wall is ensured, meanwhile, bubble discharging is promoted through molten pool stirring, grains are refined, and the welding seam quality is improved. The laser and the MIG electric arc form a composite energy field, the auxiliary welding wire is directly melted, the energy utilization rate is remarkably improved, the deposition amount in unit time is increased by 30%-50%, the single-layer single-pass welding thickness can reach 3 mm-5 mm, the welding passes are reduced compared with a traditional MIG fuse wire mode, and the method is suitable for efficient and high-quality thick plate welding.

The invention provides a reduction melting separation furnace and a reduction melting separation method for preparing DRI from low-grade ore, and belongs to the technical field of metallurgy. The reduction melt separation furnace comprises a melt separation furnace body, a material bearing pore plate, a gas injection device and an electrode, the melt separation furnace body is hollow and forms a furnace cavity, the material bearing pore plate is arranged in the furnace cavity and divides the furnace cavity into a reduction chamber and a molten pool which are arranged up and down, the material bearing pore plate is formed by splicing at least two pore plate bodies, and the at least two pore plate bodies can be opened and closed; when the at least two pore plate bodies are closed, the material bearing pore plate bears the DRI, when the at least two pore plates are opened, the DRI falls into the molten pool, the gas injection device is arranged in the furnace cavity and injects reducing gas to the DRI on the material bearing pore plate, and the electrode is inserted into the molten pool and used for being electrified to generate an electric arc so as to melt the DRI. According to the method, further efficient reduction of FeO in the low-grade DRI is achieved, the metallization ratio and the yield of the DRI are effectively increased, and low-grade ore resources are utilized to the maximum extent.

The invention relates to the technical field of electrical variable monitoring, and further relates to a pole-mounted circuit breaker electrical variable monitoring method. The method comprises the steps of 1, collecting a plurality of real-time electrical variables of the pole-mounted circuit breaker, performing normalization processing on each real-time electrical variable, and performing nonlinear fusion to form a single-dimensional electrical variable observation sequence; 2, acquiring historical sampling values of multiple electrical variable observation sequences within a fixed time delay interval so as to avoid embedding dimension explosion caused by vector operation and explicitly reflect a low-dimensional manifold hypothesis; and step 3, performing combined mapping on the intrinsic state quantity and the difference value of the insulation leakage current and the arc voltage, and predicting the fracture phase voltage in the future time. According to the invention, high-precision, high-robustness and prospective localized fault early warning and protection control can be realized, and the intelligent monitoring and safe operation capability of the pole-mounted circuit breaker in a complex power distribution environment can be obviously improved.

High-frequency irreversible electroporation (H-FIRE) is a tissue ablation modality employing bursts of electrical pulses in a positive phase-interphase delay-negative phase-interpulse delay pattern. Despite accumulating evidence suggesting the significance of these delays, their effects on therapeutic outcomes from clinically-relevant H-FIRE waveforms have not been studied extensively. The present invention provides methods of pulse delivery, including delays, that mitigate bubble formation and / or minimize the risk of arcing, such as due to the presence of bubbles, and / or minimize muscle stimulation are described herein.

The invention discloses an alternating current power grid fault current waveform recording method, and relates to the technical field of power grid fault detection, when power grid waveform distortion or a high-resistance fault symptom is monitored, denoising and wavelet enhancement are performed on multi-source wave recording and simulation data through a data acquisition and preprocessing module, and an initial data set with multi-label high-order features is generated; the multi-task learning module corrects data distribution deviation in combination with domain adaptation and adaptive weighted loss, integrates physical priori such as arc discharge, and recognizes composite faults such as harmonic distortion, short circuit and grounding; through cross validation and noise simulation evaluation, the model verifies robustness through high-order calculus measurement, characteristic attention distribution can be verified in a visual mode, incremental training is automatically triggered when real-time recording reaches a threshold value, a comparison memory mechanism is applied, new fault information is adaptively absorbed, and a multi-label diagnosis closed loop is formed. And the accuracy and continuous optimization capability of multi-scene fault identification of the smart power grid are effectively improved.

The invention discloses a photovoltaic arc detection method and system based on a neural network. The method comprises the following steps: generating a data set; preprocessing the data set to obtain a training data set; a hybrid architecture detection model combining a CNN model and a Transform model is constructed, the CNN model processes the voltage time sequence data to extract feature information, the Transform model fuses the feature information at each moment to extract fused feature information, and a predicted value is calculated to adjust parameters of the hybrid architecture detection model to obtain an arc detection model; sampling voltage time sequence data of the photovoltaic system are collected and imported into the arc detection model, and the arc detection model detects whether an arc occurs or not and outputs detection information. According to the method, the accuracy of arc fault detection in the photovoltaic system is greatly improved by fusing the convolutional neural network CNN and the Transform technology, and the safety and stability of operation of the photovoltaic system are improved. The key features of the arc fault can be accurately captured from the complex electric signal, so that the detection is more efficient and stable.