418 results about "Data slice" patented technology

The invention discloses an intelligent report generation method and system based on multi-source heterogeneous data fusion, and particularly relates to the technical field of data fusion, and the method comprises the steps: constructing a fact anchor point containing a stable identifier in a fusion layer, and solidifying an aperture version signature and a time interval; packaging a screening condition, a sorting rule, an access plan, time window limitation and the like into a controlled context packet and freezing the controlled context packet into a unique fact source; in the range, generating sentences by using a limited template and writing sentence-level reference marks, so that each sentence can be traced back to a data slice and an evidence fragment; performing consistency verification according to a source, an aperture and time, performing conservative degradation on the attribution expression according to a consistency threshold value, and outputting a degradation strategy description; and finally, page two-way pointer and cross-sentence consistency recording and audit playback are realized through an evidence chain pointer table and a published version signature, and the problems of caliber drift, cross-window access and difficulty in evidence tracing are solved.

The invention discloses a data slice recombination order-preserving method based on a Beidou short message, which relates to the technical field of data transmission and recombination optimization, and ensures the stability and low error rate of data in the transmission process by dynamically optimizing the size and quantity of slices according to channel bandwidth and delay through a self-adaptive slice adjustment mechanism. A unique serial number identifier and a first check code attached to each fragment are beneficial to identifying data under a complex channel condition, and errors and chaos in a recombination process are reduced. At a receiving end, a conflict analysis model is constructed by analyzing sequence offset and delay conditions of fragment arrival in real time, and a multi-layer correction instruction is provided according to an overall conflict rate and a time sequence consistency coefficient. The order-preserving mechanism ensures that the original order of the data can be maintained even in a complex network environment by preferentially processing the unlocked fragment and generating a retransmission instruction of the lost fragment. The combination of dynamic priority adjustment and a retransmission mechanism promotes the structure of the original data after recombination.

The invention relates to a large model data protection method and device based on a trusted environment, equipment and a medium. In the scheme, after a server side obtains original large model data, the original large model data needs to be encrypted and signed, and then the original large model data and an encrypted information table are issued to a target application side; when the target application end calls the large model data, performing exception check according to a credible strategy; if the check is passed, performing signature verification on each encrypted data slice; if the signature verification succeeds, decrypting the encrypted slice data by using the encrypted information table to obtain original large model data; wherein the server side and the target application side are both in a trusted environment. It can be seen that the trusted environment and the key system are combined, and a dual protection mechanism for large model data is achieved; moreover, before the large model data is called, the operation environment of the target application end needs to be subjected to exception check, so that an attacker is prevented from acquiring or tampering the sensitive data, and all-around data protection is provided for the large model data.

The invention discloses an intelligent monitoring and adjusting method and system for a steel structure, and the method comprises the steps: deploying a plurality of types of sensors at key parts of the steel structure, continuously collecting various types of physical quantity data, recognizing potential abnormal events in real time through a dynamic threshold detection mechanism, and responding to the abnormality of a certain sensor. The system automatically associates and extracts the time sequence data of other sensors associated with the spatial position of the sensor before and after the exception to form a multi-dimensional and time-space coupled data slice, and then performs time-space feature extraction on the data to extract the mode representation of the complex exception state; and based on this, accurate matching and diagnosis of the causes of the abnormal events are realized. Through the mode, the intelligent sensing and analysis capability of the health state of the complex, multi-source and multi-dimensional steel structure is improved, solid data support and decision basis are provided for subsequent scientific adjustment measures, and the operation safety and management intelligence level of large-scale steel structure engineering are better guaranteed.

The invention discloses a railway tunnel foreign matter detection method based on radar point cloud and RGB image fusion. The method comprises the following steps: acquiring a laser radar point cloud and an RGB image in a railway tunnel; obtaining an RGB image by using an adversarial network model GAN to judge the abnormal occurrence state of the foreign matter in the railway tunnel, and if the abnormal state of the foreign matter occurs, projecting the obtained laser radar point cloud to an image plane to obtain a dense depth map; processing through a sliding window to obtain data slices which are approximately square; and carrying out data feature extraction, fusion and foreign object target detection on the data slices of the RGB image and the dense depth map by using a fusion target detection network based on the OfficientNet-FPN, and outputting a foreign object target information detection result containing a target category, a position, a confidence coefficient and a distance. According to the railway tunnel foreign matter detection method based on radar point cloud and RGB image fusion, high precision, high efficiency and high reliability of railway tunnel foreign matter detection are realized.

The present invention relates to the field of health information processing technology, and specifically discloses an artificial intelligence-based student psychological data analysis and early warning system. The system sets a data set complexity determination module, a data slice determination module, a GPU compatible matching module, and a psychological data analysis and early warning module. The psychological health monitoring platform collects multi-source student psychological data sets, evaluates their comprehensive complexity, and determines the number of psychological data slices and their respective complexities based on this. The platform monitors the execution data of each GPU in real time, evaluates its initial execution quality indicators and compatible complexity. Subsequently, through data processing, the comprehensive complexity of the psychological data slices is matched with the compatible complexity of each GPU, and the allocation of psychological data slices to GPUs is completed. Finally, the platform uses this allocation relationship to analyze the students' psychological data, and ultimately achieves analysis and early warning of the students' psychological data.

The invention discloses an archive information processing method and system based on digital security, and relates to the technical field of computer application, and the method comprises the following steps: S1, constructing a time sequence consistency profile, and collecting the actual arrival time and the corresponding position index of each archive data fragment, a network jitter fingerprint model reflecting instantaneous delay fluctuation characteristics is generated based on multi-round transmission data; s2, in a data inbound process, injecting a unique position label for each archive data fragment, and enabling the position labels and fragment contents to jointly participate in Hash calculation to generate a content Hash and position Hash parallel dual verification baseline; according to the method, sequence verification and correction are realized by constructing a time sequence consistency profile and a dual Hash mechanism, the Hash collision risk is prevented in combination with disturbance verification, result write-back and model updating are introduced, a closed-loop control link is formed, and the safety, accuracy and traceability of file synchronization are remarkably improved.

The invention discloses a cloud storage service implementation method and system for unified management of multiple types of storage, and belongs to the technical field of cloud storage, and the method specifically comprises the steps: building a unified abstract model oriented to heterogeneous storage media, mapping distributed files, object storage and block storage devices into standardized storage units, deploying a cloud block chain, and carrying out uplink, a storage index table under a chain is constructed, an index state is dynamically updated through a loose coupling consensus mechanism, the loose coupling consensus mechanism comprises a weighted voting algorithm based on a storage node reputation value and an asynchronous verification mechanism based on space-time certification, and data is divided into data pieces and verification pieces; fragmentation information is written into a storage metadata registration chain and a service quality proof chain at the same time, cross-chain storage is carried out on data, and cloud storage resources are dynamically distributed; according to the method, the storage management function is decoupled through the block chain, so that the retrieval accuracy is improved, the retrieval response time is effectively shortened, and the data transmission cost is reduced.

The invention relates to the technical field of communication quality, and particularly discloses an acquisition terminal for communication quality monitoring, which comprises an electric energy meter interface module for reading an instrument value from an electric energy meter; the time synchronization module is used for aligning the terminal clock to the coordinated world time and generating a time frame with 1s as granularity; the communication quality monitoring module is used for measuring a communication quality index and calculating a value at the 95% position within a preset statistical time period as a statistical result preset statistical time period; the data slicing and counting module is used for forming 10T0 fragments according to 10 fundamental wave periods in each time frame; the caching and out-of-order correction module is used for performing grouping, out-of-order rearrangement and missing compensation on instrument values and communication quality indexes according to the same time frame before uploading and completing framing in a waiting window; and the alarm and uploading module is used for outputting a local alarm by taking the value of the 95% position as a statistical result. According to the invention, the overall end-to-end time delay of the acquisition terminal is reduced.

The invention discloses a high-speed rendering method and system for a three-dimensional large-scale dynamic net rack of a one-graph of a power grid. The method comprises the following steps: generating three-dimensional model parameters of a power transmission line according to tower account data; dynamically scheduling rendering of the three-dimensional model of the power transmission line according to the visible area range; slicing the power transformation multi-source heterogeneous three-dimensional data; optimizing a three-dimensional rendering pipeline of the power grid graph; and three-dimensional net rack high-speed rendering is carried out through visual cone removal, LOD and instantiation rendering technologies. According to the invention, based on the machine account data of the power transmission line, accurate rendering of the three-dimensional line on one graph of the power grid is realized through parametric modeling; for a multi-source heterogeneous transformer substation model, three-dimensional slices in a unified format are formed through a scene division technology, and support of a power grid graph on a three-dimensional slice rendering pipeline is expanded; in order to improve the rendering performance, dynamic rendering scheduling is carried out on the three-dimensional net rack in the viewport, repeated drawing of a large number of same devices is avoided, the performance overhead is reduced, and the performance bottleneck of loading the large-scale three-dimensional net rack is effectively solved.

The invention relates to a cross-network domain security data interaction method and device based on dynamic multi-channel and intelligent verification, computer equipment, a computer readable storage medium and a computer program product, and is applied to an external network domain. The method comprises the following steps: acquiring network state data of each available transmission channel between an internal network and an external network, and determining a transmission priority score of each available transmission channel so as to screen out a main transmission channel and a standby transmission channel; dynamically distributing to-be-transmitted data fragments to the main transmission channel and the standby transmission channel through a reinforcement learning algorithm; transmitting the data fragments distributed to the main transmission channel and the standby transmission channel to an isolation domain; the isolation domain is used for filtering the received data fragments through a hardware filtering engine and sending the filtered data fragments to the internal network domain; the internal network domain is used for verifying the integrity of the data fragments according to the fragment hash values corresponding to the data fragments. By adopting the method, the intelligence of data transmission in the internal and external network interaction process can be improved.

The invention belongs to the technical field of digital monitoring, and provides a three-party environmental protection management system and method suitable for an industrial park, environmental protection monitoring equipment is utilized to collect monitoring data in the park in real time, a blockchain network based on an alliance chain is established, Hash processing is performed before all monitoring data are transmitted, Hash values are recorded to the blockchain, and the monitoring data are stored in the blockchain. A secure transmission protocol and an encryption technology are adopted, the security and integrity of monitoring data in the transmission process are optimized, a large monitoring data block is divided into a plurality of small data fragments at a sending end to be transmitted, and a unique identifier and verification information are added to each monitoring data fragment. Verifying the integrity of each monitoring data fragment through verification information at a receiving end, reassembling the monitoring data, processing and analyzing the collected monitoring data by using the edge computing node, generating a monitoring data abstract, processing the monitoring data processed and analyzed by the edge computing node by using the cloud computing node, and sending the monitoring data abstract to the cloud computing node. And verifying the integrity of the monitoring data.

The invention relates to an electric power multi-source data slicing processing method and system based on reinforcement learning, and the method comprises the following steps: S1, obtaining electric power multi-source data, and carrying out the preprocessing of the electric power multi-source data, and obtaining a structured original data stream; s2, obtaining a multi-dimensional feature vector through feature extraction; s3, adopting a clustering algorithm to identify a behavior pattern to obtain a clustering label, calculating the distance from a sample to each clustering center, and when the distance is obviously higher than the average value plus threshold value beta times standard deviation of the cluster, additionally marking as abnormal; s4, intelligent slicing decision making is carried out based on reinforcement learning, and a slicing strategy, data and labels are obtained; s5, according to the obtained slicing strategy, data and label, through knowledge graph service rule verification, it is ensured that domain constraints are met; and S6, sending the verified strategy to distributed slice storage and management, and carrying out data slice storage and index construction according to the strategy. According to the method, the structured governance level and the high-value utilization capability of the power data assets are effectively improved.

The invention provides a key frame extraction method for a multi-modal large model. The multi-modal large model video data processing efficiency and generation quality are improved through cooperative computing of a user side and a cloud side. According to the technical scheme, the method comprises the following steps: 1) after a client application submits data, an interface service performs primary processing on the data, and video frames and audio frames are divided and merged into data pieces according to a time sequence; 2) dynamically extracting n key video frames in each data piece of the user side, and compressing, packaging and transmitting the n key video frames and corresponding audio frames to the cloud side; 3) the cloud performs coding fusion on the multi-modal data and then inputs a large model generation result; and 4) the cloud evaluates the generation quality, and the user side adaptively adjusts data preprocessing and key frame extraction parameter configuration according to the generation quality and the response time delay. By reducing redundant data transmission and dynamically optimizing calculation and transmission resources, the response speed and the output quality of the multi-mode large-model cloud service are remarkably improved.

The invention provides a UCIe standard protocol layer data transmission circuit, a related method and a chip, and the circuit comprises a slicing circuit which is configured to segment a data packet from a service port; sending the slice frame and slice description information associated with the slice frame to a framing device; the framing device is configured to fill a flow control unit Flit with the slice frames from the slice devices and slice description information associated with the slice frames, and send the Flit to the deframing device; the frame deframer is configured to determine slice frames belonging to the same service port from the Flit, and send the slice frames and slice description information associated with the slice frames to a packing circuit corresponding to the service port; and the packing circuit is configured to splice the data slices belonging to the same data packet in the slice frame based on the slice description information and output the data packet. The technical problem that extra delay is introduced due to beat insertion due to the fact that large multiplexers need to be introduced in existing protocol layer data transmission of the UCIe standard is solved.

The invention provides an environment monitoring method and system based on big data analysis, and belongs to the technical field of data processing, and the method comprises the steps: obtaining an original data flow corresponding to at least one monitoring node in an environment monitoring region, determining a target storage node of each data fragment according to the data type of the monitoring data in each data stream in the original data stream, and storing each data fragment based on each target storage node; processing the monitoring data in all the target storage nodes to obtain a data cube corresponding to the original data stream; determining a monitoring result of each environment state in the environment monitoring area according to the space-time grid data in the data cube; the monitoring result comprises distribution information corresponding to each environment state and pollution source positioning information, and the environment states comprise a normal state and an abnormal state. According to the environment monitoring method and system based on big data analysis provided by the invention, real-time and accurate monitoring of the environment can be realized.

The invention discloses an RAG knowledge base construction method and device and a medium, and the method comprises the steps: carrying out the analysis of the format, structure and semantic features of input data through a multi-level priority judgment logic, and automatically recognizing the specific data type. And calling a corresponding adaptive segmentation strategy for processing according to the identified data type, and generating data slices with semantic coherence. And asynchronously calling the large language model to generate a keyword set for each data slice, and uniformly storing the data slices, the keyword sets and associated multi-modal resource links by adopting a mixed storage mode to construct an enhanced RAG knowledge base. In the retrieval stage, query keywords are generated for user query according to the same logic, and related results are accurately retrieved from the knowledge base through a keyword matching mechanism. According to the method, the semantic segmentation problem is effectively avoided through intelligent segmentation with adaptive types, multi-modal data access is supported, and unified management and efficient retrieval of heterogeneous resources are realized.

The invention discloses a method and device for automatic roasting and process intelligent control of a belt type roasting machine, and relates to the technical field of pellet roasting of the belt type roasting machine in the metallurgical iron and steel industry. The method comprises the four steps of historical data slice acquisition and modeling, target process parameter determination, feedback verification and online correction and rolling prediction control, a model historical library is established through service slices and time slices, modeling is performed based on a three-layer feed-forward neural network, flexible weighted deviation calculation combined with dynamic weight and rolling optimization of a softening coefficient are combined, and the rolling prediction control is performed. And accurate regulation and control of equipment parameters are realized. The automatic control system is mainly used for automatically controlling the pellet production process of the belt type roasting machine, the product quality stability is improved, the energy consumption is reduced, the labor intensity of workers is relieved, and support is provided for unmanned and intelligent production.

The invention relates to the technical field of health information processing, and particularly discloses a student psychological data analysis and early warning system based on artificial intelligence, which is provided with a data set complexity judgment module, a data slice judgment module, a GPU compatible matching module and a psychological data analysis and early warning module. The psychological health monitoring platform evaluates the comprehensive complexity by collecting a multi-source psychological data set of a student, and determines the number of psychological data slices and the respective complexity according to the comprehensive complexity. The platform monitors execution data of each GPU in real time, and evaluates an execution quality initial index and compatible complexity of the GPU. Then, through data processing, the comprehensive complexity of the psychological data slices is matched with the compatible complexity of each GPU, and distribution of the psychological data slices and the GPUs is completed. And finally, the platform analyzes the psychological data of the students by using the distribution relation, and finally realizes analysis and early warning of the psychological data of the students.

The present invention relates to a satellite-based method and system for processing ocean observation data. The method includes: obtaining ocean observation data of an ocean observation device within a set time; formatting the ocean observation data to obtain a plurality of data slices; arranging the plurality of data slices based on timestamps, and for each data slice, calculating the difference between this data slice and an adjacent data slice with a smaller timestamp, and defining the difference as the differential slice data corresponding to this data slice; for each differential slice data, performing: preprocessing the differential slice data to obtain preprocessed data, and determining whether the preprocessed data meets the similarity requirement. If so, defining this differential slice data as data to be processed, otherwise defining the preprocessed data as data to be processed; performing compression, encryption, and channel coding on the data to be processed to obtain data to be sent; and sending the data to be sent to a land observation station. The present invention can simplify ocean observation data and improve the security of transmitting ocean observation data.

The invention belongs to the field of computer vision, relates to a rockfall monitoring method and system based on visual identification, and aims to solve the technical problems that an existing scheme is difficult to achieve good balance among reliability, accuracy and economy and is poor in application effect. The method comprises the following steps: acquiring video stream data and vibration waveform data of a target area; monitoring according to the vibration waveform data, and when the vibration energy exceeds a preset first threshold value, generating a primary vibration event trigger and extracting video data slices from the video stream data; inputting the video data slices into a target detection network to identify candidate moving targets in the video frames and extract kinematic features; performing spectral analysis on the vibration waveform data to determine a comprehensive confidence coefficient; and counting the number of candidate targets of which the comprehensive confidence exceeds a threshold value in a specific time window, and determining a current rockfall risk level. Based on the method, video detection and vibration detection are combined, and the actual application effect is better.

The invention relates to the technical field of intelligent manufacturing, and discloses a silicon carbide smelting whole process data linkage ERP management method, which comprises the following steps: collecting multi-source data of a silicon carbide smelting whole process in real time; constructing a dynamic tensor model, and updating node features based on a dynamic graph convolutional network; embedding a silicon carbide smelting process equation as a constraint condition into the tensor decomposition process, and solving a tensor factor matrix through a constraint optimization algorithm; performing incremental tensor decomposition on the newly added data slices, and updating real-time data change; dynamically adjusting weight coefficients of the process target and the business target according to real-time business requirements; and issuing the optimization instruction to production equipment. By constructing a four-dimensional dynamic tensor model, multi-dimensional data such as equipment operation states, process parameters and business requirements are uniformly mapped into structured feature representation, the problem of local optimization caused by data islands in a traditional method is solved, and high-precision data support is provided for global decision-making in the silicon carbide smelting process.

The invention discloses a quantum random encryption and distributed synchronization information transmission method and device, and is applied to the technical field of power systems. Acquiring initial data of the state of the power equipment; generating an unpredictable quantum true random number; encrypting the initial data by using the quantum true random number to obtain target data; synchronizing the clock synchronization precision of the plurality of nodes; generating data slices according to the target data and distributing the data slices to the plurality of nodes; and synchronizing the data slices of the plurality of nodes. According to the invention, the technical problem of safe and reliable information transmission of the power system is solved at least to a certain extent.

The invention discloses a six-dimensional sensor noise covariance matrix adjusting system based on impact response, and relates to the technical field of data processing, and the system comprises a base covariance construction module which is used for obtaining multi-sensor channel impact data, calculating a channel-level indication quantity, and fusing the channel-level indication quantity to obtain an impact score; determining an allowed adjustment channel set and an adjustable sub-block index, generating a diagonal priority mark, and obtaining a base covariance through weighted smoothing; the increment calculation verification module intercepts data, constructs a spatio-temporal data slice matrix, carries out shrinkage robustness estimation, generates candidate trial increment by combining channel weight and characteristic decomposition, and outputs unique trial increment after verification of an improved gate and a certificate margin gate; the adjustment fusion optimization module superposes the trial writing increment to the base covariance, after stability and consistency certificate inspection, self-adaptive adjustment is performed according to a three-section strategy, and a current noise covariance matrix is output through forgetting fusion; and the problems of numerical instability and distortion caused by redundancy adjustment in an impact environment are avoided.

Methods for a machine-learning network that provide efficient, scalable, and granular analyses during validation of a machine learning model are disclosed. Validation of models depends upon many factors, including the real-world application of the model, the type of model being trained, and the types of data samples it is being trained on. In order to provide relevant edge case information to users that pertains to their specific model, data slice finding techniques may be used to identify subsets of the dataset that are particularly problematic. By limiting a length of the slice description that the algorithm searches and by configuring the algorithm to target specific types of errors, users are provided with a more granular analysis that then allows them to determine how or if they need to retrain the model.

The invention relates to the field of capacitor monitoring, and provides a capacitor quasi-online state monitoring method and system for a wind power converter, and the method comprises the steps: extracting a data slice comprising DC bus voltage, capacitor branch current, housing temperature and output power from a data buffer area; furthermore, output power data is utilized, intelligent screening is carried out through a quasi-steady-state slice trigger, and only effective data with the power average value large enough and the fluctuation small enough are reserved. And for the screened effective data slices, extracting the DC bus voltage and the capacitance branch current, and accurately identifying the voltage, current amplitude and phase information of the main harmonic through fast Fourier transform. Based on the key harmonic data, the complex impedance of the capacitor is calculated, and then the original ESR and capacitance estimation value are calculated. And finally, in combination with the capacitor shell temperature extracted from the same data slice, performing temperature compensation on the original parameters, and outputting a final result capable of truly reflecting the aging state of the capacitor.

The invention provides a physical frame structure design method suitable for satellite-ground high-order high-bit-rate wireless communication, which comprises the following steps of: slicing service data according to a preset length, and generating a baseband frame based on the obtained data slices; performing scrambling processing on the baseband frame to obtain a scrambled baseband frame; sequentially executing preset error correction coding processing on the scrambled baseband frames to obtain coded frames; performing internal interleaving processing on the coding frame to obtain internal interleaving data, and forming a first physical frame; performing external interleaving processing on the first physical frame to obtain external interleaving data; mapping the outer interleaved data into a symbol stream; performing physical frame framing based on the symbol stream to obtain a second physical frame; and performing scrambling processing on the to-be-scrambled symbols except the physical frame header in the second physical frame to obtain a scrambled physical frame, and performing modulation based on the scrambled physical frame to perform data transmission, thereby improving phase noise resistance and channel distortion resistance, enhancing Doppler frequency shift resistance, and optimizing transmission efficiency and reliability.

The invention discloses a method and a system for detecting vertical floating of duct pieces in a shield tunnel splicing stage, and belongs to the field of tunnel engineering monitoring. The method comprises the following steps: acquiring a laser point cloud and an apparent image of the inner surface of a segment, and constructing a ring position data slice set indexed according to ring numbers; performing refined section fitting considering segment local deformation on the slice set, and extracting an actual measurement center of each ring through physical rigid body separation or statistical joint optimization; self-adaptive parameterization interpolation is executed based on the fitting sensitivity features, and an actual measurement tunnel center line is generated; constructing a self-calibration optimization objective function containing multi-source constraints, and performing incremental updating on external parameters and ring-by-ring attitudes of the sensor by using a sliding window; and establishing a multidirectional component unified coordinate system, decomposing spatial deviation and judging a segment floating mode. The problem of coupling of rigid body displacement and elastic deformation is solved, and high-precision and anti-drifting duct piece floating closed-loop detection is achieved.

The invention discloses an OTFS joint channel estimation and data detection method and model based on deep learning in a satellite scene. The method comprises the steps of constructing a data set, designing a pilot frequency design scheme, constructing a joint channel estimation and data detection deep learning model, repeatedly training the deep learning model, obtaining channel state information and data symbols through prediction data, comparing a network prediction result with the corresponding channel state information and data symbol labels, and determining the channel state information and the data symbols. Evaluating the output of the deep learning model through the minimum mean square error and the cross entropy loss; the obtained optimized deep learning model is used for obtaining channel state information and data symbols through verification data. According to the method, OTFS received signals are learned through the deep learning network so as to complete channel estimation and data detection tasks at the same time, a pilot frequency design scheme and a data slicing processing scheme are adopted at the same time, high performance is kept, the accuracy of channel estimation and data detection is remarkably improved, and the overhead of pilot frequency resources is reduced at the same time.

The invention relates to the technical field of data processing, and provides a gas-liquid two-phase rotational flow pattern identification method and system based on multi-module deep learning, and the method comprises the steps: obtaining a pressure drop signal, and carrying out the data slicing of the pressure drop signal; and performing classification prediction on the flow pattern of the spiral airflow through the trained classification prediction model, wherein the method specifically comprises the steps of extracting time domain features and frequency domain features of the pressure drop signal after data slicing processing through a CNN network; performing enhancement processing on the time domain features and the frequency domain features to obtain enhanced time domain features and frequency domain features; performing feature fusion processing on the enhanced time domain feature and the frequency domain feature to obtain a fusion feature vector; and performing classification prediction on the flow pattern of the spiral airflow according to the fusion feature vector and the full connection layer. According to the method, combined modeling of multi-dimensional and multi-scale features can be realized, so that the flow pattern recognition accuracy and the system robustness under the complex rotational flow working condition are improved, and the method has good engineering practicability and engineering value.