42results about How to "Consider comprehensively" patented technology

The invention provides a project cost estimation method and system, electronic equipment and a storage medium, and relates to the field of project cost. The method comprises the following steps: acquiring historical project cost sample data, and constructing a project cost knowledge graph; then, acquiring information of a project to be estimated, inputting the information into the knowledge graph to obtain a structured representation, and mapping the structured representation into the graph to obtain an association mode; thirdly, integrating the engineering information based on an association mode, extracting engineering influence factor characteristics, and generating a project cost influence factor representation vector through an attention mechanism; meanwhile, inputting the engineering information into a preset standard model, and outputting an engineering cost standard; and finally, constructing a project cost model based on the project cost influence factor representation vector, the project information and the cost standard, and inputting the project information into the model to obtain a final project cost estimation result. According to the method, through semantic understanding and correlation analysis of engineering information, the accuracy of engineering cost estimation is improved from the perspective of data driving.

The invention belongs to the technical field of geological exploration, and discloses a continental facies shale oil mobility geological control factor combined evaluation method and system.The continental facies shale oil mobility geological control factor combined evaluation method comprises the steps that a sedimentary structure type and an organic matter type are divided based on a rock sample; comprehensively dividing continental facies shale lithofacies by combining three-end-member lithofacies types; analyzing the saturation of the movable fluid according to the multiphase relaxation time spectrum of the rock sample, and evaluating the mobility of shale oil by using pyrolysis parameters; and analyzing the influence factors of the mobility of shale oil based on the geological control factors. According to the method, aiming at research requirements of continental facies shale oil mobility control factors, in order to solve the problem of how to comprehensively and completely evaluate the shale oil mobility control factors, a multi-scale multi-type qualitative observation technology and a quantitative test technology are combined, a plurality of influence factors of shale oil mobility are comprehensively considered, and a relatively comprehensive evaluation system is established.

The invention relates to the field of power grid main equipment operation and maintenance, in particular to a self-consistency enhanced large language model power equipment operation and maintenance method, system and device, and the method comprises the steps: generating a reasoning path based on a question in a question and answer pair, and carrying out the segmentation of the reasoning path, and obtaining a plurality of clauses; for each clause, extracting an operation, maintenance and overhaul entity, and obtaining a node sequence related to the operation, maintenance and overhaul entity in the power operation, maintenance and overhaul map according to the operation, maintenance and overhaul entity; by calculating similarity scores of the clauses and the node sequences, screening out associated node sequences corresponding to the clauses, and calculating implication scores and contradiction scores between the associated node sequences and the corresponding clauses; and according to the similarity score, the implication score and the contradiction score, scoring the reasoning path to obtain a reasoning path score, and reordering the reasoning path score to obtain a final operation and maintenance scheme. The method has remarkable advantages in the aspects of improving the operation and maintenance efficiency, the accuracy, the intelligent level, the user experience and the like.

The invention discloses a computer network intelligent analysis platform based on big data, relates to the technical field of computer networks, and solves the problem that network fluctuation generated by related access items of a computer is not considered, so that errors exist in abnormal evaluation conditions of the computer. Data access waveforms of a computer are considered and analyzed in advance to lock regular waveforms existing in the access waveforms of the computer, feature confirmation is directly carried out for single-wave waveforms, the regular waveforms are locked based on features of the corresponding waveforms and period interval conditions, feature confirmation is synchronously carried out for multi-wave waveforms, and the method is simple and convenient to implement. The method comprises the following steps: determining whether a waveform is a regular waveform or not, if the waveform does not belong to the regular waveform, performing waveform division, decomposing the waveform into a plurality of single-wave waveforms, and performing correlation confirmation on whether the single-wave waveforms are the regular waveforms or not, so that a more accurate correlation confirmation effect of the regular waveform can be achieved through a correlation mode of analysis, confirmation and processing.

The invention discloses an imaging spectrometer resolution optimization method based on a PGP model, and the method comprises the steps: respectively calculating a first emergent angle of a maximum incident angle passing through the PGP model and a second emergent angle of a minimum incident angle passing through the PGP model corresponding to a central wavelength spectral line, and determining a bending value of the central wavelength spectral line according to a difference value between the first emergent angle and the second emergent angle, and circularly calculating the angle values of the angle ranges of the first prism and the second prism according to the above steps according to a preset step length to obtain bending values of a plurality of central wavelength spectral lines, and then screening the bending values of the plurality of central wavelength spectral lines to obtain a minimum bending value. And finally, designing and manufacturing a PGP model according to the vertex angle value of the first prism and the vertex angle value of the second prism corresponding to the minimum bending value. According to the method, the condition that the top angles of the two prisms are the same can be calculated, and the condition that the top angles of the two prisms are different can also be calculated, so that the imaging spectrometer can still realize higher spectral resolution in a wider wavelength range.

The invention belongs to the technical field of poultry farming, and particularly relates to a compound feed for effectively improving sternum calcification of cherry valley hang white meat ducks. The feed contains a nutrient supplement, and every 1 part by weight of the nutrient supplement comprises 50 g of zinc methionine, 50 g of manganese methionine, 3000 IU of 25-hydroxyvitamin D3, 100 g of magnesium oxide, 300 g of blue kadsura and 500 g of eggshell powder. The functional nutrition enhancer can improve sternum calcification of cherry valley hang white meat ducks. In the invention, the conventional nutritional indexes and the required amount of trace elements and vitamins of the cherry valley-hang white meat ducks are researched, and the formed compound feed preparation technology is more scientific and more comprehensive in consideration; the functional nutrition enhancer is adopted, so that the sternum calcification degree of the meat ducks is effectively improved, the tibia strength is improved, the occurrence rate of leg diseases is improved, and the breeding benefits are improved.

The invention discloses an on-orbit failure probability estimation method for a single-particle hard error of a wide bandgap semiconductor device, and aims to support on-orbit prediction of the single-particle hard error failure probability of the wide bandgap semiconductor device. Wide bandgap semiconductor devices such as silicon carbide (SiC), gallium nitride (GaN) and the like are easy to generate hard errors such as single-particle burnout and single-particle gate-through under the radiation of high-energy particles, and the risk of the hard errors in aerospace application needs to be comprehensively predicted and evaluated. Therefore, the invention provides a method for estimating the in-orbit failure probability of the single-particle hard error of the wide bandgap semiconductor device, which considers the influence of a failure section angle and particle distribution characteristics in different orbit environments, and provides data support for model selection.

The invention provides a remote interference management method and system based on user equipment, satellite equipment and a base station, and the method comprises the steps: demodulating a reference signal from interference through a configured reference signal when same-frequency interference exists between a ground network and a non-ground network, and obtaining a same-frequency interference source based on the currently demodulated reference signal; and the current interference condition and the same-frequency interference source are reported, so that the network can make corresponding adjustments including scheduling, configuration or resource management, and interference is avoided.

The invention is applicable to the field of games, and discloses a game difficulty adjusting method, system and device and a medium, the method comprises the following steps: acquiring first meteorological data of the current position of a player, and adjusting the movement track of the player according to the first meteorological data; acquiring second meteorological data of the current position of the player, and adjusting the moving speed and physical exhaustion speed of the player in combination with character attribute data of the player; and acquiring third meteorological data and a scene model of the current position of the player, and adjusting the communication quality of the player according to the third meteorological data and the scene model. According to the method, different meteorological data of the current position of a player are obtained, and the moving track, the moving speed, the physical exhaustion speed and the communication quality of the player are adjusted respectively. The adjustment basis comprises the role attribute data of the player and the game scene model, so that the game difficulty is personalized and dynamic, and the game experience and interactivity are improved.

The invention relates to the technical field of water conservancy and hydropower engineering construction, discloses an optimal design method for the length of a tunnel plug in the main flood season of extra-high dam construction, and solves the problems that in a design scheme of the length of the tunnel plug in the traditional technology, the project investment is increased, the construction period is prolonged, and the reliability analysis of the tunnel plug is not comprehensive enough. The method comprises the following steps: firstly, determining a tunnel control sliding interface through engineering basic parameter analysis, and initially simulating the length of a tunnel plug; then building an anti-sliding stable failure risk model of the tunnel plug in the main flood season of the ultra-high dam construction; then determining distribution parameters of main random factors influencing the water load and the anti-sliding force; then, based on the initially planned length of the tunnel plug, in combination with distribution parameters of main random factors influencing the water load and the anti-sliding force, random simulation of the water load generated by the maximum flood level of dam flood control in the main flood season and the total anti-sliding force generated by the tunnel plug is carried out; according to the stochastic simulation result, the anti-sliding stability failure risk rate of the plug is estimated; and finally, according to the estimated anti-sliding stability failure risk rate, optimizing the length of the tunnel plug until the convergence condition is met.

The present invention discloses a computer network intelligent analysis platform based on big data, which relates to the field of computer network technology and solves the problem that the network fluctuations generated by the relevant access items of the computer itself are not taken into account, resulting in errors in the abnormal situation assessed. The present invention sets two processing stages to consider and analyze the data access waveform of the computer itself in advance to lock the regular waveform existing in the access waveform. For a single-wave waveform, feature confirmation is directly performed, and the regular waveform is locked based on the features of the corresponding waveform and the periodic interval. For a multi-wave waveform, feature confirmation is performed synchronously to identify whether it is a regular waveform. If it is not a regular waveform, waveform division is performed to decompose it into a number of single-wave waveforms, and relevant confirmation is performed on whether the single-wave waveform is a regular waveform. This related method of analysis and confirmation processing can achieve a more accurate regular waveform related confirmation effect.

The invention discloses a software testing device for an escalator printing plate. The software testing device comprises a relay output module, an IO input board card, an IO output board card, a pulse output board card, a CAN communication board card, an RS232 communication board card, a relay and a testing system. The test system is connected with the relay output module, the IO input board card, the IO output board card, the pulse output board card, the CAN communication board card and the RS232 communication board card. The IO input board card is connected with the relay, and a level output signal of an escalator printing plate is transferred by the relay and then is input into the IO input board card; the IO output board card and the relay output module are connected with the relay, and level output signals of the IO output board card and the relay output module are transferred through the relay and then input into an escalator printing plate level input interface. And the pulse output board card is connected with a pulse input interface of the escalator printing plate. All external interfaces of an escalator hardware system can be replaced, more functions and scenes are covered, and the automation degree of software testing is improved.

This invention discloses a three-impulse control method and system applicable to chemical waste heat boilers, belonging to the technical field of boiler automatic regulation. It includes the following steps: acquiring the operating parameters of the chemical waste heat boiler; collecting and storing the boiler drum water level, steam flow rate, and feedwater flow rate; calculating the steam flow rate change rate; determining the current operating state; setting the boiler drum set water level; calculating the target feedwater flow rate using the set water level, current boiler drum water level, steam flow rate, steam enthalpy, and energy conversion coefficient; selecting a comprehensive state coefficient based on the current operation; calculating the feedwater regulation amount using the selected comprehensive state coefficient combined with the target feedwater flow rate; generating a feedwater regulation command based on the feedwater regulation amount and transmitting it to the chemical waste heat boiler. This invention has the effect of automatically and timely controlling the water level of the chemical waste heat boiler with good control stability.

The invention relates to the technical field of computers, and provides a termination judgment method, system and device of a program and a medium, the method comprises the following steps: after obtaining a program code, carrying out grammar analysis, and constructing an abstract syntax tree based on grammar nodes; in the abstract syntax tree, on the basis of a dependency relationship of an expression or a statement, constructing a one-way edge between syntax nodes; for a grammar node A of which the grammar is a loop meaning, establishing a return edge between the grammar node A and a sub-grammar node in a loop body of the grammar node A; further obtaining a directed abstract syntax tree graph; encoding the directed abstract syntax tree into a feature graph, inputting the feature graph into a graph attention network, and outputting a first vector; inputting the adjacent matrix of the feature map into a convolutional neural network, and outputting a second vector; and carrying out weighted summation and normalization on the two vectors to obtain a non-termination probability estimation result. According to the method, the loop structure in the program can be effectively identified, and the accuracy of judging the termination is improved.

This design is a multi-task automatic EEG detection and prediction system for epilepsy. The system first receives the EEG signals from epilepsy patients to be identified or the collected patient EEG signals used for training the module through a preprocessing module. The preprocessing module performs noise reduction, filtering, segmentation, and normalization on the signals, and outputs the processed patient EEG signals to the training module. In the training module, the collected patient EEG signals are used for training and testing. First, feature extraction is performed using a multi-scale convolutional network. Then, the detection and prediction branches incorporating a dual attention mechanism are trained, and the trained model parameters are saved for use in the detection and prediction module. Finally, the detection and prediction module provides detection and prediction results for the EEG signals from epilepsy patients to be identified. This invention employs a modular design, achieving real-time detection and prediction of noisy, unbalanced EEG signals, providing a reliable system for practical applications.

The invention belongs to the technical field of electric power station centralized control systems, and particularly relates to a signal segmentation-based centralized control system alarm signal matching method and system. The method comprises the following steps: extracting feature words similar to various feature words of any typical alarm signal from a real-time alarm signal, wherein the various feature words of the typical alarm signal are formed by segmenting the typical alarm signal; matching the typical alarm signal with the highest total similarity with the real-time alarm signal with the real-time alarm signal; wherein the mode of calculating the total similarity between a certain typical alarm signal and the real-time alarm signal is that the comprehensive similarity between any feature word of the typical alarm signal and feature words of the same kind of the real-time alarm signal is calculated, and the comprehensive similarity is obtained according to the pinyin similarity and the character string similarity; and obtaining the total similarity of the typical alarm signal and the real-time alarm signal according to the comprehensive similarity of all feature words of the typical alarm signal and the real-time alarm signal. And the alarm signal matching precision is high.

The invention relates to the field of power safety, in particular to an industrial power utilization safety situation analysis method and system based on power data, and the method comprises the steps: screening the abnormal types of power utilization abnormal enterprises to obtain a target abnormal type, and scoring the alarm rate of the target abnormal type to obtain an alarm rate score; comparing the power data of the enterprises in the current time period with the power consumption data of the same time period in the previous year to obtain the proportion of the number of enterprises with obviously increased power consumption, the proportion of the number of enterprises with obviously increased power consumption being greater than a preset value compared with the power consumption of the same time period in the previous year, and scoring the proportion of the number of enterprises with obviously increased power consumption to obtain the score of the proportion of the number of enterprises with obviously increased power consumption; and carrying out weighted summation on the alarm rate score and / or the proportion score of the number of enterprises with obviously increased power consumption and the power consumption change scores under different dimensions to obtain focus targets under different dimensions, thereby realizing enterprise security situation analysis. According to the invention, the verification and analysis of the safety situation of industrial electricity under different dimensions are realized, the consideration is comprehensive, and the calculation is simple.

The application discloses a power battery monomer abnormality early warning method and device, electronic equipment and storage medium. Including: acquiring full working condition state parameters of a power battery pack; respectively storing and accumulating voltage state data of monomer batteries in the power battery pack during charging and discharging; establishing a monomer battery abnormality detection model; inputting the voltage state data into the monomer battery abnormality detection model, combining the full working condition state parameters of the power battery pack, and determining monomer batteries in an abnormal state. The early warning information can be sent before the power battery cannot be repaired, the efficiency of inspection and repair is improved, the monomer battery abnormality alarm lag is avoided, and the advance and accuracy of the new energy automobile power battery safety early warning are improved.

The intelligent rehabilitation training auxiliary system comprises a remote cloud processing center and a plurality of sub rehabilitation systems, the remote cloud processing center is used for a rehabilitative or a doctor to manage and monitor the physical function level and the training condition of each patient and distribute a training scheme for each patient, and the sub-rehabilitation system is used for rehabilitation training of the patients and inputting of new actions by the rehabilitative; the system makes full use of artificial intelligence, edge computing and other technologies, gives full play to the advantages of deep learning, and achieves the purpose that a rehabilitation teacher checks patient data, formulates a training scheme and records training actions; evaluating and correcting rehabilitation actions of the patient, evaluating a physical state and monitoring an abnormal state; interaction with a patient is realized through a voice interaction technology and a human-computer interaction interface, the patient is stimulated and danced, and the rehabilitation enthusiasm of the patient is enhanced; the system integrates guidance and protection, provides real-time guidance and safety guarantee for rehabilitation training of patients in hospitals or at home, and has high practical value and wide application prospect.

The invention relates to the field of meteorological prediction, and provides a multi-mode north pole climate change prediction and regulation and control mechanism analysis method. The purpose of the invention is to predict the climate change of the arctic region through the relationship among NPO, AMOC and arctic near-surface air temperature. According to the method, by constructing and verifying a large sample data set containing AMOC substitute data and north pole SAT historical observation data, the performance and deviation of the CMIP6 mode in simulation of AMOC and north pole SAT changes are evaluated. Then, a physical mechanism of the influence of the AMOC on the north pole SAT is disclosed, and the regulation and control effect of the NPO internal change rate on the relationship between the AMOC and the north pole SAT is explored; based on this, the future climate change scene is simulated, the mode parameters are adjusted to accurately reflect the influence of AMOC and NPO on the north pole SAT, and the change trend of the future north pole SAT and AMOC is predicted to evaluate the reliability of mode prediction. The method is used for predicting the climate change of the arctic region, and the prediction accuracy of the future change of the arctic climate is improved by deeply understanding the interaction between different climate modes.

The invention belongs to the technical field of feed breeding, and particularly relates to a yak low-altitude barn feeding fattening method. The method comprises the following steps: step 1) transitional period: 0-5 days after entering a fence, wherein the daily ration structure comprises 1-2 jin of corn straw grass and 10 g of salt; 6-10 days after entering the fence, the daily ration structure comprises 2-4 jin of corn straw grass, 10 g of salt and 2 liang of concentrate supplement; 11-15 days after the yak enters the fence, the daily ration structure comprises 4 jin of corn straw grass, 1 jin of straw, 10 g of salt and 2-4 liang of yak concentrate supplement; (2) in the early stage of fattening, putting into a fence for 16-30 days, wherein the daily ration structure comprises 0.5-2 jin of concentrated feed, 0.5 jin of corn straw grass, 0.5 jin of rice straw and the like; according to the method, the growth performance of the yaks can be effectively improved, the nutritional requirements and growth curves of the yaks are effectively combined, the adaptability of the yaks to the low-altitude environment is fully exerted, and the concept of short-term rapid fattening of the yaks is achieved.

The invention discloses a reel type sprinkling irrigation machine dispatching method for soil mechanical compaction and reduction, and relates to the field of agricultural machinery dispatching. The problem that the influence of the machine on the soil compaction effect is not considered in an existing sprinkling irrigation scheduling method is solved. According to the method, the total scheduling target function is constructed with the lowest total scheduling cost of the reel type sprinkling irrigation machine and the highest total compaction satisfaction degree of the field, and the minimum scheduling target function is constrained according to the set constraint condition, so that optimization solution is carried out, and the optimization solution efficiency is improved. And distributing corresponding reel type sprinkling machines and service paths of the reel type sprinkling machines to the field parcels corresponding to the water source points. The method and the device are mainly used for scheduling the reel sprinkling machine.

The invention belongs to the technical field of natural gas exploration and development, and discloses a fault control type karst residual hill weathering crust gas reservoir favorable area optimization method and related device.The method comprises the steps that well seismic data in a target area are obtained, marker beds are selected according to the well seismic data for layer position calibration, and the stratum thickness of each layer is counted; according to the horizon calibration and the stratum thickness, an impression method and a trend surface removing method are adopted to recover ancient landform and micro-ancient landform forms, and a target area weathering crust interface top surface structure and a fracture distribution diagram are compiled; and calculating an ancient landform evaluation index, a structure evaluation index and a fault evaluation index at the deployment well point according to the ancient landform, the micro-ancient landform, the weathering crust interface top surface structure and the fracture distribution diagram, solving a comprehensive evaluation index of a favorable region, and screening a favorable target region. According to the method, the relevance and interaction of different control factors are explored, a multi-factor fusion method is provided to solve the favorable area, and the method has better evaluation means and accuracy.

This invention belongs to the field of pedestrian trajectory prediction technology, and particularly relates to a method for predicting pedestrian trajectories at unsignalized right-turn intersections based on game theory. The method includes the following steps: S1, acquiring historical data on pedestrians and vehicles at unsignalized right-turn intersections; S2, analyzing the human-vehicle game factors at unsignalized right-turn intersections and constructing a corresponding human-vehicle game model; S3, inserting the human-vehicle game model into a pre-set S-GAN model to obtain an SDG-GAN model for predicting pedestrian trajectories; S4, training the SDG-GAN model using the historical data acquired in S1; S5, using the trained SDG-GAN model to predict pedestrian trajectories at unsignalized right-turn intersections in real time. This invention ensures the accuracy of pedestrian trajectory prediction at unsignalized right-turn intersections and the effectiveness of assisted driving decisions at such intersections, thus balancing the efficiency and safety of vehicles passing through unsignalized right-turn intersections.

The invention provides a verifiable robustness optimization method in an adversarial environment based on a mask auto-encoder, and the method comprises the following steps: firstly, obtaining an image data set, and carrying out the random attack and preprocessing of the data set; then, performing distribution boundary curvature evaluation on the confrontation samples in the data set after random attack, and rejecting confrontation samples with low purification potential according to an evaluation result; secondly, reconstructing the purified and rejected adversarial sample by using a mask auto-encoder to obtain a purified sample; and finally, carrying out global smooth optimization on the purified sample to improve the global representativeness. According to the method, the samples after attack can be effectively purified, the problem that a purification model is a black box and cannot be theoretically proved is avoided through process design, and multiple methods and processes of the random smoothing technology are optimized.

The embodiment of the invention provides a vehicle controller upgrading method, electronic equipment, a storage medium and a program product. The method comprises the steps of obtaining battery information of a vehicle in response to an upgrading request of a vehicle controller; wherein the battery information comprises high-voltage battery information and low-voltage battery information; the upgrading request of the vehicle controller comprises a software upgrading package; determining and displaying a controller upgrading mode of the vehicle controller according to the high-voltage battery information and the low-voltage battery information; wherein the controller upgrading mode indicates a power supply type when software upgrading is carried out on the vehicle controller; and upgrading the vehicle controller according to the controller upgrading mode of the vehicle controller and the software upgrading package. The method is used for achieving the effect of improving the success rate of upgrading the vehicle controller and the user experience.

The invention relates to a prognosis prediction model construction method and device, electronic equipment and a storage medium. The prognosis prediction model construction method comprises the following steps: determining key distinguishing features capable of distinguishing a normal region from a focus region based on a brain medical image; determining a hidden damaged tissue boundary around the focus area according to the key distinguishing features; wherein the brain tissue between the focus area and the damaged tissue boundary is a hidden abnormal blood flow tissue area which is difficult to directly detect based on a brain medical image; and constructing the prognosis prediction model based on the brain tissue characteristics in the damaged boundary. The embodiment of the invention can improve the prediction level.

The invention discloses a buried PE pipeline reliability calculation method, belongs to the field of pipeline safety calculation, and aims at solving the problem that an existing pipeline reliability calculation method is single in evaluation condition and improving the reliability calculation accuracy. Calculating the circumferential stress reliability P1 according to the vertical additional stress and the circumferential additional stress; the axial stress reliability P2 is calculated by considering the factors of the burial depth and the pipe diameter; and obtaining the reliability P0 of the pipeline according to the reliability P1 of the circumferential stress and the reliability P2 of the axial stress. According to the method, the stress response of the buried pipeline is obtained under the condition that the large load F is applied above the soil and the burial depth and the pipe diameter of the pipeline are comprehensively considered, the consideration is more comprehensive, the operation safety and reliability calculation of the buried pipeline is more accurate, and safety evaluation guidance suggestions can be more accurately and effectively proposed. Moreover, the adopted calculation index large load F, the measurement and acquisition of the burial depth and the pipe diameter of the pipeline are relatively easy, the acquisition time is short, and the reliable calculation can be quickly carried out.

The invention discloses a drug resistance gene detection system and program for beta-lactam drugs for RA, and relates to the technical field of gene detection. The drug resistance gene detection system comprises: a gene library containing drug resistance markers of beta-lactam drugs for RA; the gene sequencing equipment is used for performing gene detection on whether the to-be-detected sample contains the drug resistance marker or not and the serum type of the to-be-detected sample; and the data processing module is used for giving a drug resistance analysis report based on a detection result of the gene sequencing equipment. According to the invention, drug resistance markers and serum detection are combined, the prediction accuracy of drug resistance is high, and the method has good reference value for predicting pathological changes and prognosis after infection; the method can be applied to the fields of virus detection, risk assessment, drug research and development and the like.

The invention discloses an active vibration suppression method for random frequency mixing of a three-phase surface-mounted permanent magnet motor, and the method comprises the steps: firstly carrying out the LMS and peak estimation of a real domain signal collected by a vibration acceleration sensor in a segmented manner with the step length of 1Hz, and finally obtaining a series of peak radial vibration frequencies x1, x2, x3... xk; establishing a current injection model under a rotating coordinate system, and for each peak radial vibration frequency xk, calculating a positive-sequence symmetric current of xk + f frequency and injecting the positive-sequence symmetric current into the model; combining and injecting all obtained frequency current injection models into the rear end of a current loop PI regulator, and obtaining the vibration of each peak frequency; multi-target vibration deep learning is carried out on the vibration of each peak frequency to obtain an optimal frequency mixing current, and the motor frequency is adjusted to realize active vibration suppression; according to the comprehensive vibration suppression algorithm based on vibration feedback, the coupling problem of vibration of each frequency generated by multi-frequency current is solved, the vibration suppression effect is improved, the stability and the anti-interference capability of the system are enhanced, and the method is easy to implement.