235results about How to "Accurate removal" patented technology

The invention discloses a wafer defect atlas intelligent identification and classification method and system, and relates to the technical field of semiconductor integrated circuit manufacturing and detection, and the method comprises the steps: obtaining a detection point set, an original image and standard reference attribute data; constructing a space weight matrix based on reference data; calculating a generalized connection cost based on a path weighted cumulative value, constructing an initial graph model, and performing probability sparsification; determining verification points based on the weight matrix and extracting local gradient features of the original image; correcting the connection probability by using a Bayesian inference model, and updating a graph topological structure; and finally, outputting a classification result according to the connected sub-graph features. According to the method, layout physical attributes and image microscopic features are combined, and accurate reconstruction and classification of defect forms under a complex background are realized.

A river and lake shoreline monthly monitoring remote sensing image automatic interpretation and result management system comprises a data resource management and access module, a remote sensing image intelligent preprocessing module, an intelligent interpretation and change identification module, a man-machine collaborative interaction and correction module and a database and report generation module. The data resource management and access module is used for accessing image data, the remote sensing image intelligent preprocessing module is used for data preprocessing, the intelligent interpretation and change recognition module is used for element automatic extraction and change discovery, and the man-machine collaborative interaction and correction module is used for man-machine interaction and result correction. And the database and report generation module is used for data storage and monitoring result generation. The invention provides an improved self-adaptive gating network sub-pixel shoreline interpretation algorithm for automatic interpretation of remote sensing images, and provides an improved cross-temporal attention and uncertainty collaborative change detection algorithm for monthly dynamic analysis of river and lake shorelines. A better scheme is provided for a river and lake shoreline monthly monitoring remote sensing image automatic interpretation and result management system.

The invention relates to an inspection algorithm for a composite machine, and the method comprises the steps: obtaining the operation parameters of a function cooperation assembly of the composite machine, eliminating abnormal data through a density clustering model, eliminating false anomalies through the combination of assembly cooperation logic, and generating a standardized parameter set through an interval scaling method; constructing a double-branch feature extraction model, mining single-component local features and cross-component coupling features, and fusing through a feature interaction layer to obtain a multi-dimensional feature set; inputting the feature set into a dual-module detection model, optimizing a random forest to output an initial anomaly probability, performing correction in combination with a working condition and a historical fault rule, performing secondary verification through a component collaboration rule base, and outputting an anomaly parameter type, a time sequence node and a severity degree; abnormal core parameters and grades are positioned, an abnormal conduction path map is generated, a component physical connection relation and a historical maintenance case are matched, and an abnormal basic cause is locked; according to the invention, data reliability and identification accuracy are improved, efficient fault tracing is realized, and maintenance shutdown loss is reduced.

This application provides a method, system, medium, and equipment for detecting thermal hazards in tunnels, relating to the field of tunnel safety detection. The method includes: acquiring geological feature data of the target tunnel; identifying a group of thermal hazard parameters with a correlation greater than a set threshold; setting a quantitative scoring range based on this; using the analytic hierarchy process (AHP) to assign weights and calculate hazard scores; then, combining relevant data to establish a dataset and training a lightweight gradient boosting model; finally, inputting data from an unexcavated tunnel section into the model to output the thermal hazard level. This application, by acquiring geological feature data, can accurately understand the environment, screen key parameters to eliminate redundancy, quantify parameters, and rationally allocate weights to ensure objective and accurate scores. It enables tunnel thermal hazard detection and early warning, providing guidance for prevention and control, reducing impact, and is of great significance to tunnel engineering.

The application discloses a kind of unmanned vehicle multispectral remote sensing water fine classification extraction method, it is related to remote sensing image processing technical field, the joint feature matrix including spectral feature, gray texture feature and elevation difference feature is creatively constructed in the present application, effectively eliminates redundant interference information, provides high-quality input with very distinguishing degree for multilayer perception machine model.Simultaneously, illumination geometry and digital elevation model are deeply coupled in the feature construction process, not only can accurately identify and reverse inhibit the shadow pseudo water body feature projected by complex terrain or building, break through the failure limitation of traditional single elevation filtering in flat shadow area, but also cooperate with the standardization closed operation process based on morphological dilation and corrosion filtering, eliminate the small holes and jagged artifacts in prediction result, realize the high-fidelity restoration of water body boundary, so as to improve the classification accuracy of complex water body such as narrow river, fragmented pit pond and urban waterlogging.

The invention belongs to the technical field of power equipment state monitoring, and discloses a transformer discharge detection method and system with an anti-interference function, and the method comprises the steps: obtaining a multi-mode discharge original signal of a transformer, carrying out the synchronization of a coefficient through a channel, anchoring a synchronous interference signal, extracting initial multi-dimensional features, and carrying out the detection of the multi-mode discharge original signal. Performing weighted average fusion on the initial multi-dimensional features according to a channel synchronization coefficient to obtain an anti-interference feature vector; constructing a feature screening-interference immune lightweight neural network model, inputting anti-interference features, taking a channel synchronization coefficient as an interference immune factor, and outputting a discharge type identification result; and obtaining an evaluation index by combining the peak amplitude of the multi-mode discharge original signal of the transformer after the de-noising preprocessing, and obtaining a discharge risk level based on the evaluation index. According to the invention, accurate identification and risk grade evaluation of transformer discharge are realized, and the problems of poor anti-interference performance and the like are solved.

The invention relates to the technical field of magnetic suspension compressor assembly, and particularly discloses a magnetic pollution source prevention and control optimization method in a magnetic suspension compressor assembly process, which comprises the following steps: in a pretreatment stage of magnetic suspension compressor assembly, dynamically optimizing thermal cycle parameters in a low-temperature environment by adopting a self-adaptive genetic algorithm; ferromagnetic impurities hidden in microdefects of the compressor component are accurately induced to be released in advance; carrying out feature extraction and pattern recognition on the released impurity signals in combination with wavelet packet transformation and a neural network algorithm; according to the method, a low-temperature thermal cycle process optimized by a self-adaptive genetic algorithm is constructed in the pre-assembly treatment stage, ferromagnetic impurities hidden in microcracks are accurately induced to be released in advance, and accurate removal of the impurities before entering the assembly link is realized by combining wavelet packet transformation and targeted pulse demagnetization driven by a neural network; the limitation that impurities in the microdefects cannot be touched by a traditional cleaning and demagnetizing method is broken through, and the hidden danger of magnetic pollution which cannot be found by initial detection is eliminated from the source.

This invention relates to the field of nuclear radiation detection technology, and more particularly to an energy spectrum processing method for a quasi-hemispherical CdZnTe detector. The method includes: acquiring the raw signal collected by the quasi-hemispherical CdZnTe detector; normalizing the amplitude of the raw signal from the baseline to the highest point during the rise phase; dynamically selecting multiple energy thresholds; performing rise time discrimination on the normalized signal within each energy threshold; filtering out signals with rise times greater than RTflag and determining the rise time range corresponding to that energy threshold; constructing an intermediate energy spectrum corresponding to each energy threshold based on the signals with rise times greater than RTflag and the rise time range; and post-processing the intermediate energy spectrum to obtain the target energy spectrum corresponding to each energy threshold. This invention combines energy threshold discrimination, rise time discrimination, and energy spectrum post-processing algorithms to stably improve the energy resolution of the quasi-hemispherical CdZnTe detector for 662 keV gamma rays from the conventional 2%~3% to less than 1%.

The application discloses a kind of pulse Doppler radar weather cloud clutter automatic filtering methods, belong to radar data processing technical field.The method utilizes the statistical characteristics of weather cloud clutter on radar echo spectrum diagram, and the spatial position and motion characteristics of radar track in monitoring area, realizes the automatic identification and effective elimination of weather cloud clutter, reduces false track, and improves the detection performance of radar to "low, slow and small" target.

The present application relates to the technical field of road maintenance, and particularly relates to an asphalt pavement base crack identification method and system, which comprises: acquiring an asphalt pavement base image and preprocessing the image to obtain a gray image, then performing edge detection on the gray image by using a Canny operator to obtain all edge connected domains, calculating the corresponding crack similarity based on the contour shape features of each edge connected domain, screening out suspected crack edges and non-crack edges according to the crack similarity, then extracting the HOG feature values of each edge point in the two types of edges and calculating the crack confidence value of the suspected crack edges, determining the suspected crack edges with the crack confidence value greater than a preset threshold as target crack edges, and finally completing crack positioning in the asphalt pavement base image according to the target crack edges to realize accurate identification of the asphalt pavement base cracks. The present application effectively eliminates the noise interference of material particle edges in the asphalt pavement image through the double technical mechanisms of coarse screening of crack similarity and fine screening of crack cluster degree.

The application relates to the technical field of battery rubberizing, and discloses a lithium battery pole piece rubberizing device which comprises a battery pole piece conveying assembly, a buffer assembly, a power assembly, an adjusting assembly and a burr cutting assembly. The battery pole piece conveying assembly conveys the battery pole piece. The buffer assembly and the power assembly are coordinated to realize seamless adaptation of continuous conveying and intermittent rubberizing, greatly improve production efficiency, the burr cutting assembly is integrated in a pre-rubberizing process, burrs on the pole piece are accurately removed, the risk of film breakage and battery short circuit is avoided, product quality is guaranteed, the adjusting assembly can flexibly adjust the cutting width, the pole piece is prevented from deviating, different specifications of pole pieces are adapted for processing, a waste material collecting structure realizes closed-loop recovery of burrs, and the working environment is kept clean. The overall device has high integration, is convenient to operate, considers precision and efficiency, significantly reduces production cost, and adapts to the automatic production demand of lithium batteries.

This utility model discloses a nozzle device for a centrifuge containing blast furnace molten slag, including a centrifuge roller, a rubber ring, multiple nozzles, and an air ring. The rubber ring is configured to be extended to cover the cotton-like area under the centrifuge roller. An adjustable nozzle angle mechanism is provided between the rubber ring and the air ring. This mechanism includes multiple adjustable supports, one end of which is fixed to the rubber ring, and the other end is connected to the nozzle. This utility model, by extending the rubber ring, increasing the number of nozzles to at least 12, and combining the adjustable nozzle angle mechanism with an anti-clogging device, clears the blind spots of glue coverage in the cotton-like area under the centrifuge roller, reducing the waste cotton rate. Simultaneously, combined with the air ring groove interlocking structure and auxiliary air duct system, it improves the uniformity of cotton dispersion, reduces nozzle maintenance frequency, increases daily output per line, saves glue costs and waste disposal fees, improves the stability of continuous production line operation and resource utilization efficiency, and meets the requirements of efficient utilization and high-cleanliness production of blast furnace molten slag in the metallurgical industry.

The application discloses a kind of inventory optimization methods of charging facility spare parts replacement cycle and failure rate analysis, and the core process of this method includes building data layer, multidimensional data is cleaned and optimized, to realize standardized storage;Based on the LSTM algorithm, a multidimensional linkage prediction model is built, the loss effect of design defects and environmental factors is quantified to optimize the model, and the accurate spare parts replacement cycle threshold and equipment failure rate are output;According to the flow properties of spare parts, a differentiated inventory management mode is designed, and the optimal management mode of slow-moving parts is selected;A full-cost quantification model is built to calculate the total cost and service level under each mode;Finally, with the goal of minimizing cost, an intelligent inventory control model is built, which dynamically outputs inventory warning thresholds, replenishment quantities and times;Through the algorithmization and parameter quantification of the whole process, the application realizes fine and intelligent management of inventory, significantly reduces operating costs and stockout risk, and improves the stability of charging facilities.

The utility model relates to a recycling device for waste electrolyte of a lithium ion battery, and belongs to the technical field of lithium battery processing. A cover plate is detachably mounted at the top end of the filter, a long rod is inserted into the inner wall of the filter, an insertion block is inserted into the inner wall of the long rod, a guide rod is fixedly connected to the inner wall of the insertion block, a limiting mechanism is arranged on the inner wall of the insertion block, a clamping groove is formed in the inner wall of the long rod, and a filter screen is fixedly connected to the outer wall of the insertion block. An adjusting assembly is arranged on the outer wall of the filter, and the limiting mechanism comprises an inclined plate. According to the utility model, a plurality of filter screens are adopted to carry out graded filtration on the lithium battery electrolyte, so that the filtration precision can be effectively improved, impurities can be accurately removed, and when the filter screens need to be replaced or cleaned, the long rod, the separation inclined plate and the clamping groove can be easily moved out, so that the rapid replacement of the single filter screen is realized, and the maintenance time and cost are greatly reduced; the continuous and efficient operation of the filtering process is ensured.

The utility model relates to a drainage pipe technical field especially relates to a dredge water seal bottom drainage pipe anti -block structure. Its technical scheme includes water seal, and the outer wall of water seal is connected with first communicating pipe, and one end of first communicating pipe is connected with one -way valve, and one end of one -way valve is connected with second communicating pipe, and dredge pipe is set through second communicating pipe, and one end of dredge pipe is connected with valve, and one end of valve is connected with third communicating pipe, and one end of third communicating pipe is connected with connector, and the spray head of being set in one end of dredge pipe is arranged, and the inside of dredge pipe is provided with the adjusting assembly for adjusting the water pressure of spray head. The utility model embodiment can more effectively clear the binding force of hard scale layer and pipe wall, accurately remove the deposited sewage, tar and other sundries in water seal bottom drainage pipe, avoid the water seal from being unable to normally drain due to blockage, ensure smooth gas delivery, and reduce the influence on production.

The application discloses a kind of optical fiber side light extraction structure and method, belong to optical fiber side light extraction technical field, structure has including femtosecond grating light extraction component, filter component including filter layer and with slit light shielding layer, and detection component.The method includes: with grating diffraction separation optical signal in fiber core, with filter component filters optical signal, with detector receives signal.The application is by femtosecond laser in fiber core and writes small refractive index modulation area, with diffraction effect will transport light be introduced with specific angle to side, again cooperate external filter and metal slit and constitute spatial filter, physically block the stray light deviating from preset angle.This internal diffraction+external filter synergic mechanism avoids physical damage optical fiber surface, and realizes high purity, high directional signal extraction under micro size.

This invention provides a wireless communication link selection method and a half-duplex communication system for selecting multiple physical wireless communication links between a transmitter and a receiver in a half-duplex communication system. Both the transmitter and receiver are equipped with switching switches and antennas. The method includes: the transmitter and receiver synchronously and sequentially switching their respective antennas into working states according to their pre-stored switching switch timing sequences; when any transmitter antenna and any receiver antenna are simultaneously in working states, a wireless communication link is formed, and each wireless communication link is traversed at least once; during the transmission period of any wireless communication link, the receiver transmits unidirectional heartbeat packets or time-division bidirectional heartbeat packets with the transmitter through the current wireless communication link; the transmitter and / or receiver that receives the heartbeat packet selects the optimal wireless communication link from the multiple wireless communication links based on the signal strength and packet error rate during the heartbeat packet transmission process.

This invention relates to the field of image analysis technology, specifically to a machine vision-based quality inspection system for new energy vehicle wiring harnesses and cable ties assembly. The system includes: an image acquisition module that uses polarimetric stereo technology to acquire multi-angle images; a geometric reconstruction and reference extraction module that calculates the surface normal field of the insulation layer and separates the background field of the cable's cylindrical curvature and the micro-indentation residual field, thereby extracting the global axial and radial curvature moduli of the cable; a feature decoupling analysis module that calculates the contour-shaping migration flow field of the indentation micro-residual field relative to the standard contour, and performs anisotropic decomposition of the flow field based on the global axial direction of the cable to obtain the axial slip index; and a quality inspection module that determines an adaptive slip judgment threshold based on the cable's radial curvature modulus, thereby generating inspection results. This invention solves the problems of difficulty in sensing micro-deformation on reflective cylindrical surfaces and the difficulty in decoupling slip risk from assembly tolerances, thus improving the accuracy of wiring harness and cable tie assembly quality inspection.

The invention provides a radar point cloud-based foreign object intrusion detection method and system in a working area, and the method comprises the steps: setting a detection area range according to the working space of a mechanical connecting rod, and dynamically generating a direction bounding box (OBB) of the connecting rod through obtaining the posture state information and size data of the working connecting rod in real time; spatial screening is carried out on original radar point cloud data, only point cloud data which are located in a detection area and are not located in an OBB are reserved, and background point clouds outside the detection area and point clouds generated by a connecting rod are filtered out; and performing multi-dimensional anomaly detection on the filtered point cloud data, and comprehensively judging whether a continuously existing foreign object intrusion event exists or not. Through a linkage mechanism of point cloud filtering, background modeling and clustering verification, high-precision, low-false-alarm and real-time industrial intrusion detection is realized, and the problems of poor adaptability, high false alarm rate and the like of a traditional method in a complex environment are solved.

The application discloses a kind of gap foreign matter detection methods based on visual image feature matching, comprising the following steps: obtaining the video of gap detection area, feature extraction and initial matching are carried out to reference image and image to be detected;Using the nearest neighbor distance ratio NNDR algorithm carries out preliminary elimination of mis-matching point;Using local deletion method is further eliminated to the mis-matching point of one-to-many for the matching of a feature point in reference image and multiple feature points of image to be detected, to obtain effective matching point;Foreign matter detection is carried out by the number and distribution area ratio of effective matching point.The method can effectively eliminate mis-matching point when feature matching, improve matching accuracy;Foreign matter detection is carried out to gap using the relative number of effective matching feature point and the distribution area of effective matching point and other information, with the advantages of fast and accurate.

The invention relates to the field of paint spraying, in particular to a curved surface paint spraying boundary technology which comprises the following steps that S1, a first color layer is sprayed to a curved surface structure; s2, removing the first color layer on the second side of the color block boundary; s3, spraying a release interlayer on the curved surface structure, wherein the release interlayer is sprayed on the second side of the color block boundary and the surface of the first color layer; s4, removing the release interlayer on the second side of the color block boundary; s5, spraying a second color layer on the curved surface structure, wherein the second color layer is sprayed on the second side of the color block boundary and the surface of the release interlayer; and S6, removing the release interlayer so as to remove the second color layer sprayed on the surface of the release interlayer, and respectively forming a first color block and a second color block on the first side and the second side of the color block boundary. The multi-color paint spraying device has the effect of conveniently carrying out boundary multi-color paint spraying on the curved surface structure.

The invention discloses an attention learning mechanism-based inverter model control prediction method, system and device, and a storage medium, and the method comprises the steps: pre-constructing a three-phase inverter simulation model, carrying out the real-time sampling of a state signal through employing an MPC controller, and obtaining a signal characteristic set; performing feature selection on the signal feature set, and extracting feature data; setting a screening threshold to screen the feature data, and taking the screened feature data as optimal features; inputting the optimal feature into a pre-trained machine learning model module to obtain a prediction result of the control signal of the three-phase inverter; the control performance can be improved while more calculation amount and communication traffic are reduced.

The invention relates to an intelligent internet of things non-intrusive sensor controller. The system comprises a data acquisition unit, a data processing unit, a communication transmission unit and a remote operation and maintenance control unit, collecting first pipeline data, second signal data and third terminal data, and performing processor edge calculation, result screening and abnormal value elimination on the three types of data to obtain regular data; judging the battery voltage and the switching value according to a preset threshold value, and generating an alarm identifier; classifying and caching the regular data and the alarm identifier to obtain a standardized data frame; acquiring the data frame by adopting a modulation mode; synchronously transmitting the second signal data and generating a gateway core jump packet, and simultaneously completing terminal power-on and obtaining cloud platform instruction information; corresponding data parameters are adjusted according to the instruction, a cloud platform control instruction is received by means of an alarm identifier, a relay is driven to act after the instruction is transmitted through a LoRa channel, and the equipment state is read in real time based on regular data; and low-power-consumption data acquisition and remote control in a scattered scene are realized.

This application discloses a foreign fiber removal device for cotton web, relating to the field of textile machinery and equipment. It includes a frame and a feeding conveyor belt, a foreign fiber detection component, and a spray valve assembly mounted on the frame. An impurity collection channel is fixedly provided at the feeding end of the feeding conveyor belt. The impurity collection channel has an upper opening as its inlet and a lower opening as its outlet, with the inlet plane located below the conveying surface of the feeding conveyor belt. The foreign fiber detection component and the spray valve assembly are located above the feeding conveyor belt and the impurity collection channel, with the detection end of the foreign fiber detection component and the ejection end of the spray valve assembly both facing the inlet of the impurity collection channel. This device can perform online foreign fiber detection and positioning on the cotton web output from the carding machine, and spray airflow to accurately eject foreign fibers. Combined with the fan and impurity collection channel, it quickly guides foreign fibers away from the production line, thereby effectively removing foreign fiber impurities from the cotton web and achieving online removal of foreign fibers from the cotton web.

The invention discloses an educational academic literature tracing method and system based on retrieval enhancement generation. The method comprises the following steps: performing fine-grained analysis on the multi-source literature, constructing a document object containing fields such as a title, an abstract, a methodology, an experimental result and a conclusion, and establishing a field-level index and a citation network feature score; identifying an academic intention queried by a user by utilizing a large model, dynamically loading a weight mapping table according to the academic intention, and performing weighted rearrangement on multiple paths of retrieval results to obtain candidate core literatures; generating a traceability text with an explicit label based on the candidate literature; and executing reference consistency verification by using the natural language inference model, and executing illusion correction according to a verification result. According to the method, through structured field analysis and dynamic rearrangement of intention driving, the problems of semantic fragmentation and reference fictition when a general RAG is used for processing complex academic literatures are solved, and the academic traceability preciseness and accuracy are remarkably improved.

The utility model discloses a double-shaft coating stirring barrel, which relates to the technical related field of coating stirring barrels and comprises a device body, the device body comprises a workbench, a stirring barrel, a stirring component and a filtering component, the stirring barrel is fixed on the workbench, an upper cover is arranged at the upper end of the stirring barrel, a discharging pipe is arranged on one side of the lower end of the stirring barrel, and the stirring component is arranged at the upper end of the upper cover. The filtering assembly is connected to the middle of the discharging pipe, the stirring assembly comprises a first motor and a second motor, the lower end of the first motor is connected with a gearbox, the lower end of the gearbox is connected with a main stirring shaft, the lower end of the second motor is connected with a side stirring shaft, and the filtering assembly comprises a filtering box communicated to the middle of the discharging pipe. The coating is prevented from being adhered to the inner wall of the barrel, the stirring efficiency and quality of the coating are effectively improved, the stirring blades are convenient to disassemble, assemble, maintain or replace, and different requirements are conveniently met.

This invention provides a method for detecting rotating debris in coal flow based on physical fuzzy modeling and scene-customized attention. The method includes: acquiring a clear coal flow image and then labeling non-coal debris targets in the image; establishing a dataset; constructing a debris detection network; constructing a physics-driven motion fuzzy mapping model; using the motion fuzzy mapping model to dynamically blur the clear coal flow image, generating training samples simulating motion fuzziness phenomena in a conveyor belt; training and optimizing the debris detection network; inputting a real-time coal flow image into the optimized network to obtain candidate detection results for non-coal debris targets; post-processing the candidate detection results and outputting the final filtered detection results. This invention can improve the stability and accuracy of debris detection under complex conditions of high-speed coal flow, reduce the influence of motion fuzziness and background interference, and provide technical support for the safety monitoring and automatic sorting of coal mine conveying systems.

The application provides a power spot balance clearing planning model construction method, medium and system, belongs to the technical field of power model, and through the construction of the timestamp alignment of the multi-source heterogeneous measurement data based on the Apache Kafka distributed stream processing pipeline, the weather field, voltage, power flow and load data after the alignment are encoded into multi-channel heterogeneous graph signals, the flexible adjustment capacity boundary is output by the demand prediction model after the flexible adjustment of multiple time scales, the artificial intelligence assisted power spot clearing optimization main model is established, the layered progressive plane solver is solved by the Kirchhoff current law propagation mechanism graph, the constraint iteration solution is activated layer by layer from shallow to deep, and the relaxation tightness is re-verified in the closed loop by the node voltage deviation comprehensive index, so that the problems that the power spot balance clearing model is mutually fragmented in the three links of multi-source heterogeneous measurement fusion, source and load uncertainty prediction and large-scale non-convex optimization solution, and cannot form a physically self-consistent closed loop are solved.

This application provides a method, apparatus, device, and storage medium for detecting abnormal battery charging and discharging. The method includes: performing pressure calibration on the battery's clamping state to obtain calibrated clamping state parameters; correcting the original battery data collected by the detection circuit based on the calibrated clamping state parameters to generate reference characteristic data; performing differential calculation between the reference characteristic data and real-time detection data during the charging and discharging test to determine dynamic offset parameters; and aligning the dynamic offset parameters with a standard charging and discharging curve to determine abnormal battery state parameters. By performing pressure calibration on the battery clamping state to obtain calibration parameters and correcting the original data to generate reference characteristic data, and then performing differential calculation and database comparison with real-time detection data, data deviations caused by poor clamp contact can be effectively eliminated, thereby accurately identifying abnormal battery states and improving the reliability of charging and discharging test data and the real-time nature of anomaly determination.

The application provides a medical question and answer method and system based on cooperative dual-source retrieval enhancement generation, which comprises the following steps: according to an input original medical question and candidate items, performing task self-adaptive query rewriting through a large language model to generate differential diagnosis keywords, entity enhanced queries and hypothetical medical abstracts; for the differential diagnosis keywords, performing an iterative network retrieval with a reflection mechanism to obtain network candidate evidence; for the entity enhanced queries and the hypothetical medical abstracts, performing a hybrid retrieval in a local medical knowledge base to obtain local candidate evidence; aggregating and deduplicating the network candidate evidence and the local candidate evidence, obtaining the aggregated heterogeneous evidence, performing deep semantic correlation scoring through a cross-encoder model, screening out a target evidence set, and splicing the original medical question as context input into a generative large language model to generate a final medical answer. The application improves the accuracy, timeliness and explainability of the medical question and answer system.