42 results about "Rate analysis" patented technology

The invention relates to a human body health state assessment system and assessment method based on multi-point acquisition, and the system comprises an event sensing module which is used for collecting an image data signal, a physiological signal and an environment data signal of a human body, and outputting an asynchronous pulse event flow when the signal change is detected; the pulse coding module is used for generating a space-time pulse sequence through a preset adaptive threshold coding technology; the neural network processing module is used for performing event driving processing by utilizing leakage integral distribution neurons in a preset pulse neural network according to the space-time pulse sequence to obtain a current health characteristic pulse mode; and the health state decoding module is used for obtaining a health state score and a risk classification result of the current human body through a preset pulse distribution rate analysis algorithm and a preset time sequence decoding technology according to the current health characteristic pulse mode. Therefore, the problems of high power consumption, high delay, low data transmission efficiency, low early pathological feature recognition accuracy and the like of a human health state evaluation system are solved.

The invention discloses a method and a system for detecting the health state of a transformer in real time, relates to the technical field of transformer detection, is used for solving the problem of fault early warning lag, and aims at solving the problem of fault early warning lag by monitoring the output voltage of the primary side and the secondary side of the transformer, calculating the fluctuation ratio, analyzing the stable operation state and counting the abnormal duration. After entering a temperature rise analysis mode based on duration judgment, detecting oil temperature and winding temperature, calculating a difference value, calling a cooling system start-stop instruction to obtain a cooling response value, calculating a response lag rate, collecting data of gas dissolved in oil, calculating a gas anomaly ratio, and analyzing thermoelectric synergy anomaly characteristics in combination with the response lag rate; the method comprises the following steps: acquiring an insulation aging grade and an oil product replacement period from a historical maintenance database, evaluating an insulation degradation coefficient according to the insulation aging grade and the oil product replacement period, analyzing a health state in combination with thermoelectric collaborative abnormal characteristics, selecting an operation optimization or shutdown maintenance mechanism, improving fault early warning capability, optimizing a maintenance strategy, and prolonging the service life of equipment.

The invention discloses a radiation monitoring method and gate-type radiation monitoring equipment, and relates to the technical field of radiation monitoring. The method comprises the following steps: performing pulse analysis processing on a real-time gamma ray pulse signal to obtain a real-time net counting rate and a pulse PSD distribution histogram; determining a compensation coefficient according to the environmental data, and performing real-time compensation on the real-time net counting rate to obtain a compensated real-time net counting rate; performing compensation counting rate calculation according to the reference parameter, the compensated real-time net counting rate and the pulse PSD distribution histogram to obtain a compensation counting rate; and determining a target alarm type according to the compensation counting rate, the real-time net counting rate and a preset alarm condition corresponding to each alarm type, and performing alarm control according to an alarm mode corresponding to the target alarm type. Compared with the prior art, the method does not depend on single dose rate measurement or total count rate analysis, at the moment, it can be guaranteed that detection is more accurate, the method is not affected by environmental fluctuation any more, and then the problems of frequent false alarms and the high false alarm rate can be solved.

The invention discloses a software testing method and device, a medium and a product, and relates to the technical field of software test.The software testing method includes the steps that coverage rate data of unit testing, integration testing and system testing of target software are collected through a code coverage rate analysis tool, and the coverage rate data are converted into target text files; the coverage rate data comprises execution times and coverage rates of the codes; storing each target text file in a corresponding target directory, and reading and analyzing the target text file in each target directory to obtain corresponding analyzed data; the analyzed data comprises a total row number, a total branch number, a total function number, a total coverage row number, a total coverage branch number and a total coverage function number; and merging and storing the analyzed data into a preset nested dictionary, traversing the preset nested dictionary to determine the code coverage rate of the target software based on the traversed analyzed data, and determining the software defect of the target software by using the code coverage rate. The quality and stability of software can be improved through testing.

The application discloses a kind of RSF performance test methods based on Jmeter.The method includes developing Jmeter plug-in;Set the basic data of request message;Determine whether the format of request message meets the requirements, and if not, carry out request message format conversion;The content of request message is converted into the message type set;The converted request message is sent to the server registered by remote service framework RSF;Response message returned by the server is received, and the received response message is checked, if the response message meets the expectation, this request is set as success, otherwise, this request is set as failure;Using the performance statistical tool of Jmeter, the number of transactions per second and transaction response time analysis are carried out according to the received response message, and success rate analysis is carried out according to the checking result, and then the analyzed content is displayed.The application greatly reduces the test time, and reduces the learning cost of test personnel.

The invention relates to the technical field of warehouse management, in particular to an AGV scheduling and warehouse stock decision-making system and method based on data analysis. Comprising an inventory management module used for obtaining storage data, sales data and to-be-warehoused data, the storage data comprises real-time inventory data, and the storage quantity and the storage amount of each type of goods, the sales data comprises the sales amount of each type of goods, and the to-be-warehoused data comprises to-be-warehoused goods and the to-be-warehoused quantity; the area analysis module is used for analyzing the occupancy rate of each current inventory area according to the real-time inventory data, and the inventory area with the occupancy rate lower than a preset occupancy threshold value is marked as a low-occupancy area mark; and the turnover rate analysis module is used for calculating the stock turnover rate warehousing grade of the goods in the preset period according to the storage amount and the sales amount of the goods in the preset period, and is used for determining the warehousing grade of the goods to be warehoused according to the to-be-warehoused quantity and the storage amount turnover rate of the goods to be warehoused.

The application discloses an InSAR surface deformation multi-source vertical contribution analytical method based on deformation signal spectrum decomposition and belongs to the technical field of surface deformation monitoring. The steps are as follows: first, long-time SAR images are acquired, vertical deformation data and related rates and deformation variables are obtained through SBAS-InSAR technology processing; then, a plurality of auxiliary conditions are fused, regional spatial partitioning is realized through K-Means clustering; subsequently, CEEMDAN decomposition and mean threshold multi-scale reconstruction are performed on deformation signals of each partition, high-frequency items, low-frequency items and trend items are obtained; finally, ICA blind source separation and variance contribution rate analysis are performed, and shallow, medium and deep surface deformation information is inversed through directional reorganization; the K-Means clustering algorithm is adopted to perform spatial partitioning on the regional deformation field, the homogeneity and precision of subsequent signal processing are improved; and the CEEMDAN decomposition and multi-scale reconstruction method is introduced, adaptive and fine scale separation of deformation time series signals is realized; the blind source separation technology is applied to multi-scale component reorganization, and the physical attribution of the vertical position of the subsidence source is realized.

The invention discloses an exchange rate time sequence determination method and device based on a lightweight generative adversarial network, and relates to the financial science and technology field, and the method comprises the steps: obtaining source domain data and target domain data according to a preset window time length, employing a maximum mean value difference algorithm to quantify a distribution difference parameter, and obtaining an exchange rate time sequence; under the condition that the distribution difference parameter is greater than a preset detection threshold value, judging that a drift condition occurs, freezing a main body network weight of the basic exchange rate time sequence model, and adjusting a tail linear layer of the basic exchange rate time sequence model by utilizing a fine adjustment data set, the fine tuning data set is obtained by merging a pseudo drift sample generated by the lightweight generative adversarial network with target domain data; and outputting exchange rate time sequence parameters of a preset time step length by utilizing the adjusted basic exchange rate time sequence model. The technical problem that the accuracy of a model output result is reduced due to the fact that market data in a drifting condition cannot be adaptively adjusted in an exchange rate analysis mode in related technologies is solved.

The invention discloses a MoE model intensified training method based on classification data, and relates to the technical field of MoE model intensified training, and the method comprises the following steps: collecting the correctly distributed number of various samples in each batch in the MoE model intensified training process, and carrying out the preprocessing to obtain sample correct rate data and expert matching rate data; performing sample accuracy analysis to obtain first weight feature data; performing expert matching rate analysis to obtain second weight feature data; based on the first weight feature data and the second weight feature data, dynamically adjusting the proportion of various samples in the batch of enhanced training; the method is used for solving the problem that when an existing MoE model intensive training technology carries out intensive training on a MoE model used for classifying data, the proportion of various samples in each training batch cannot be dynamically adjusted according to the number of correctly distributed samples in each training batch, the efficiency of intensive training of the MoE model is improved, and meanwhile the training efficiency of the MoE model is improved. And the classification precision of the MoE model is improved.

The invention discloses an advertisement creation failure reason diagnosis method, system and device and a medium. The method specifically comprises the steps of collecting multi-source data in an advertisement creation process; constructing a knowledge graph of the advertisement creation process based on the multi-source data; node embedding and relation reasoning are carried out on the knowledge graph through a graph neural network, and associated features in the advertisement creation process are extracted; based on the association features, performing cross-modal alignment and fusion on the text description, the system log and the user operation video by using a Transform model, and generating a fine-grained reason diagnosis result of the failed task; and constructing a success rate prediction model according to historical data and real-time features based on the associated features and the fine-grained reason diagnosis result, and dynamically adjusting advertisement budget allocation based on a prediction result of the success rate prediction model. According to the method, automatic and intelligent diagnosis of the advertisement creation failure reason is realized, the efficiency and accuracy of advertisement creation success rate analysis and optimization are improved, and advertisement putting decision scientization is facilitated.

The invention relates to the technical field of corrosion layer inversion, in particular to a steel corrosion layer depth inversion method based on LIBS spectral correlation analysis, which comprises the following steps: S1, acquiring a spectral sequence from a covering corrosion layer to a metal matrix, performing layer-by-layer variation rate analysis on each wave band spectral line, and constructing a spectral depth variation path set changing along with the denudation process; s2, identifying positions with sudden change or stability turning of spectral line features, marking the corresponding denudation step number as a transition position between corrosion layers, and constructing a transition index set comprising the transition position, the mutation rate sudden enhancement degree and the spectral line group linkage degree dimension according to the transition position, the mutation rate sudden enhancement degree and the spectral line group linkage degree dimension; and S3, inputting the transition index set into a trained structure mapping model, and outputting the total corrosion layer depth, the thickness estimation value of each sub-layer and the confidence level of the target corrosion area in combination with the multi-type corrosion sample library. The method improves the detection resolution and process adaptability, and can be widely applied to industrial scenes sensitive to the corrosion layered structure.

The invention provides a reference axis precision control method from a contraction section to a diffusion section, and belongs to the technical field of wind tunnel construction. A stainless steel plate control system is established as a bottom plate profile control reference, and a discontinuous point matrix and a contribution rate calculation model are constructed to realize key control point identification; a step point matrix and a virtual-real data comparison system are established to carry out precision control on a template profile, a multi-point joint observation network system is arranged to realize omnibearing monitoring, an intelligent contribution rate analysis system is utilized to carry out template installation correction, and a point stability matrix and a dynamic monitoring optimization equation set are established to realize real-time monitoring of a pouring process. A virtual-real data fusion analysis system is adopted for hole body axis fitting and optimization, and the technical problem that the terminal precision is poor due to the fact that the distance from a contraction section to a first diffusion section is long and measurement errors can be accumulated and amplified in the axis transmission process is solved.

The application discloses a low-voltage transformer area line loss rate analysis method, device, equipment and storage medium, the method comprises the following steps: obtaining the ideal line loss rate of the transformer in the ideal operation state of three-phase balance; using the preset phase equivalent resistance method to calculate the theoretical line loss rate according to the preset transformer area line loss related data; based on the ideal line loss rate, the theoretical line loss rate and the preset statistical line loss rate, the statistical deviation index and the technical deviation index are calculated respectively; according to the statistical deviation index, the technical deviation index and the deviation threshold, the management line loss measure and the technical line loss measure analysis are carried out, and the analysis result is obtained. The application can solve the technical problems that the existing technology cannot quantitatively analyze the line loss anomaly of the low-voltage transformer area, and the line loss analysis effect and the treatment scheme are poor.

The invention provides a background management system and method for an intelligent coffee station, and the method comprises the steps: collecting data in each intelligent coffee station in real time, so as to calculate and analyze raw material depletion time, equipment maintenance urgency, a sales trend report and user preference data; raw material depletion time, equipment maintenance urgency, a sales trend report and user preference data are input into a preset background management system to generate an operation decision scheme, and the operation decision scheme is analyzed through the steps of real-time data acquisition, inventory consumption rate analysis, equipment health assessment, sales trend identification, user preference mining and the like. The background management system can carry out dynamic, accurate and automatic operation decision making on replenishment, maintenance and sales strategies of the intelligent coffee station, and the problems that replenishment is not timely, equipment faults are not timely processed and the sales strategies are difficult to adjust when the number of equipment is large, the distribution area is wide or user requirements change frequently in the prior art are solved.

The invention relates to the technical field of electric power communication, in particular to a substation wireless access point anomaly detection method, system and device and a medium, and the method comprises the steps: collecting multivariable time sequence data of the operation of a wireless access point (AP), and constructing a directed graph structure by taking a performance index in the data as a node and variable dependence as an edge; a time convolutional network (TCN) is used to extract time features, a graph convolutional network (GCN) is used to carry out convolution combination on correlation information between variables to obtain time-space fusion features, prediction errors and change rate analysis are introduced to enhance fault recognition, and finally an anomaly detection result is obtained. According to the method, a multivariable time sequence directed graph structure is constructed, the time dependence and structural relationship of equipment performance indexes is fused, long and short time features are extracted in combination with a time convolutional network (TCN), variable space correlation modeling is enhanced by a graph convolutional network (GCN), the problem of spatial-temporal feature splitting in a traditional method is solved, and precise recognition of a complex fault mode is achieved.

The invention discloses a Tibet region atmosphere water substance conveying amount and change fine evaluation method, and relates to the technical field of atmosphere diagnosis. The invention relates to a Tibet region atmospheric water substance conveying amount and change refined evaluation method. The method comprises the following steps of boundary lattice point screening, a boundary lattice point conveying method, regional atmospheric water substance conveying amount, regional conveying change trend and result verification. The method aims at improving the precision and applicability of estimation of the regional atmospheric water substance conveying amount, and particularly has remarkable advantages for complex terrain regions such as Tibet. Performing cross validation in combination with high-resolution reanalysis data and rainfall grid point data, and proving the reliability of the method in the aspects of time sequence change trend and spatial distribution characteristics; standardized processing and linear tendency rate analysis can clearly reveal the long-term change rule of water vapor, cloud liquid water, cloud ice water and the like, and a scientific basis is provided for regional aerial water resource management and climate evolution research.

The invention provides a single-cell RNA (Ribonucleic Acid) rate automatic analysis method for a modular assembly line, which comprises the following steps of: reading initial sample data under a specified path, and generating a corresponding loom file from each initial sample data by utilizing a velocyto tool; reading initial sample data with a specified name, and converting each initial sample data into target sample data in a target format; extracting unique identification information in the initial sample data; on the basis of the unique identification information, matching is carried out on a loom file and a csv file corresponding to the initial sample data; combining the matched loom files and storing the combined loom files as a total loom file in a target format; the target sample data and the total loom file are merged, and scVelo rate analysis and visualization processing are performed on the merged summary file, so that manual intervention in the RNA rate automatic analysis process can be reduced, and the analysis efficiency and accuracy are effectively improved while the automatic analysis speed is improved.

The application provides a method, system and related equipment for assisting IC verification in analyzing defect rate, the method comprising connecting a preset defect management server to obtain all original defect data under a specified project; receiving at least one set of filter parameters including filter parameters for positive matching and NFilter parameters for negative filtering at a webpage client; screening all original defect data based on the at least one set of filter parameters to obtain a matched defect data set in which data items meet the filter parameters and do not meet the NFilter parameters; performing data analysis and statistics on the matched defect data set to generate at least one preset-dimension defect rate analysis result; and visually rendering and displaying the defect rate analysis result at the webpage client. The application realizes full-process automation from data acquisition to visual display, and users can conveniently perform real-time and dynamic multi-dimension data screening and analysis through the webpage, thereby significantly improving defect analysis efficiency and convenience in the IC verification process.

The application discloses a kind of based on execution rate prediction's flight time whole life cycle dynamic management method, including constructing core table system and date interval week enumeration formula, according to rule split and calculate and process airline application information, process flight time cancellation, process flight time return and approved quota, process flight time return and approved quota, daily update flight actual operation execution state, season operation in real time and estimate and predict time execution rate, after season end, the actual execution rate of whole time is counted;The application solves the problems that the existing system lacks flight time whole life cycle closed-loop management and control capability, time execution rate monitoring lags behind, lacks pre-warning capability, breaks the limitation that traditional mode can only rely on manual checking to carry out post-execution rate analysis after season ends, builds whole-process execution rate prediction and early warning capability relying on real-time operation data, helps airlines to adjust operation strategy in time, reduces the idle and waste of flight time resources.

The invention discloses an asymmetric multi-fiber wavelength division multiplexing optical network blocking rate analysis method, which belongs to the technical field of optical communication networks, and comprises the following steps of: dividing a call request into a leaving call, a continuous call and an entering call; according to the occupancy states of various call requests on a first hop link and a second hop link, wavelengths are divided into wavelengths which are only occupied by continuous calls on the first hop link and are not idle, and wavelengths which are jointly occupied by continuous calls and leaving calls on the first hop link and are not idle; the wavelength is idle on the first-hop link and is occupied by the continuous call and the incoming call on the second-hop link without idle wavelength, and the wavelength is idle on both the first-hop link and the second-hop link; and constructing a three-dimensional Markov chain model according to the arrival rates of various call requests, and calculating the probability that the number of idle wavelengths on the first hop link and the second hop link is zero as the blocking rate of the optical path. According to the method, the accuracy of blocking rate estimation can be improved while the calculation efficiency is maintained.

The invention relates to an information processing system based on a block chain and big data, in particular to the technical field of information processing, and the system comprises an information acquisition module used for acquiring employee basic information, employee job application information, employer basic information, employer position information and distance information, an information storage module used for authenticating the information and storing the information to a server. An employee analysis module, an employee recommendation degree analysis module, an adjustment module, an employee recommendation degree analysis process adjustment module, an employer analysis module, an employer recommendation degree analysis module, a matching module, employer recommendation management for the employee, employee recommendation management for the employer, and employee recommendation management for the employer. The management module is used for managing contract signing of employees and an employer according to a matching result, the optimization module is used for optimizing the process of performing employee recommendation management on the employer, and the feedback module is used for correcting the analysis process of the employee recommendation degree in the next management period.

The invention provides a conversion rate analysis method and device for a business payment process and electronic equipment, and the method comprises the steps: distributing a first identifier for each payment link in a business process of a target type, and distributing a second identifier for the business process of the target type; based on the first identifier and the second identifier, collecting behavior data of a user in each payment link in the business process of the target type; wherein a plurality of pieces of behavior data are provided; and calculating the conversion rate of the business process of the target type based on the behavior data. By adopting the method, the accuracy of a conversion rate analysis result can be improved, and a more accurate data basis is provided for subsequent data analysis.

The present application relates to the technical field of coal storage monitoring, in particular to a method for monitoring coal bunker sticky wall, the existing coal bunker sticky wall monitoring method cannot be quantified, has low safety, poor real-time performance and is difficult to adapt to digital operation demand; the method obtains the coal accumulation and the coal discharge accumulation under the empty state of the coal bunker, calculates the difference to obtain the sticky wall amount, compares the sticky wall amount with the preset threshold and combines the sticky wall rate analysis to realize the sticky wall diagnosis and early warning; at the same time, the data can be updated by regular calibration to ensure the calculation accuracy; the present application is a non-invasive online real-time monitoring, fully utilizes the existing measurement data, realizes the accurate quantification of the sticky wall amount, can complete the historical tracking and trend analysis, can be integrated with the existing management system, provides data support for the preventive maintenance of the coal bunker, and effectively improves the safety and intelligent level of the coal bunker operation.

The application discloses a rate analysis method of an intelligent metasurface system based on random geometry distribution, and comprises the following steps: constructing an initial optimization problem, in which a robustness problem of a distributed intelligent metasurface system under the influence of an initial mobile user is constructed; a receiver receives superimposed signals from multiple reflection links, and a determination formula of an initial reachable rate of the system is given; and the determination formula is further converted into an upper limit solution according to the Jesse inequality and a probability density function; the time intensity of vertical switching and beam selection caused by user movement is obtained; the total time overhead is further obtained; and an expression of a final area traversal reachable rate is obtained. The application is applied to an actual random distribution intelligent metasurface assisted communication system, an expression of an area spectrum rate limit is given, and the system performance can be effectively evaluated.

The invention discloses a low-voltage drawer cabinet remote operation and maintenance management platform and method, and relates to the technical field of low-voltage drawer cabinet remote operation and maintenance, and the method specifically comprises the following steps: respectively executing unit behavior normalization processing and dynamic change rate analysis based on a collected voltage change sequence and a collected current response sequence, constructing a three-dimensional characteristic data body containing a voltage change rate, a current response slope and a fluctuation integral index; and inputting the constructed three-dimensional feature data body into an identification structure, extracting pollution indication factors used for representing the pollution state of the conductive sliding contact, and carrying out fusion calculation on the pollution indication factors to generate judgment parameters used for judging the contact performance of the sliding contact. According to the invention, the problem that conductive sliding contact pollution is difficult to be monitored and identified by traditional current after remote switching on is solved, and accurate identification and intelligent regulation and control of a contact performance reduction state are realized.

The invention relates to the technical field of computers, in particular to a coverage rate report generation method, a test platform and equipment. The method comprises the steps of obtaining a full code coverage rate report of a current version branch of a target application; performing code difference analysis on the current version branch and the main branch of the target application by adopting a preset code hosting platform to obtain a difference code and target change information; cleaning the full code coverage report according to the difference code and the target change information, and generating and displaying incremental code coverage data of the difference code; and performing difference marking on the full code coverage rate report according to the difference code and the target change information, and generating and displaying the full code coverage rate data of the difference marking, so that the problem that information in the coverage rate report output by a native code coverage rate analysis tool is not rich enough can be effectively solved.

The invention discloses a lock sink strategy generation method and device based on a generative agent and electronic equipment, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the preprocessing and fusion processing of multi-modal data, obtaining a multi-dimensional feature vector set, inputting the multi-dimensional feature vector set into the generative agent, and carrying out the fusion processing of the multi-dimensional feature vector set; the generative agent comprises a diffusion model and an adversarial network model, the reference path set and the extreme path set are input into an exchange rate simulation environment, a target lock sink strategy and lock sink strategy income data are generated, and the exchange rate simulation environment comprises a defined multi-dimensional state space, an action space and a reward function; the action space outputs a plurality of lock sink actions of the target lock sink strategy, and the reward function is used for determining expected revenue data corresponding to each lock sink action. The technical problem of low accuracy due to adoption of a single exchange rate analysis mode in a financial exchange rate risk management and control process in the prior art is solved.

The invention relates to the technical field of artificial intelligence prediction, and discloses a caliper brake service life prediction method and system based on machine learning, and the method comprises the steps: carrying out the pixel-level segmentation based on a corrected image set, calculating a wear rate, and obtaining a wear trend curve graph. The U-2-Net model automatically extracts multi-scale semantic features and optimizes boundary segmentation precision by nesting a U-shaped framework and a side output supervision mechanism, pixel-level accurate recognition of a wear area is realized, an improved Paris formula is combined with a temperature-stress coupling coefficient, geometric factors and crack growth rate are dynamically calibrated through machine learning, and the accuracy of the wear area is improved. A nonlinear degradation behavior under the combined action of temperature and stress is quantified, a machine learning driven rule engine can fuse dynamic data such as wear rate, crack length and hot spot area in real time, and through LOESS smoothing and fractal dimension change rate analysis, an early warning threshold value is adaptively adjusted and a three-dimensional feature map is generated. And accurate division from an initial damage stage to a later damage stage is realized.

The invention provides an equipment early warning method and device based on business data association, electronic equipment and a storage medium. The method comprises the following steps: acquiring real-time operation data and current business scene data of target equipment; screening out historical effective operation data matched with the current service scene from the historical data, and constructing an operation baseline model and a rate change model according to the historical effective operation data; comparing the real-time operation data with the operation baseline model to obtain a first deviation result; calculating a real-time change rate of the target equipment according to the real-time operation data, and comparing the real-time change rate with a rate change model to obtain a second deviation result; and generating target equipment early warning information according to the first deviation result and the second deviation result. According to the method and the device, the association relationship between the business data and the equipment operation is established, and the historical data analysis based on business filtering and the data change rate analysis based on business association are comprehensively judged, so that the accuracy and the reliability of equipment early warning are improved.

The application discloses a kind of methods and systems for detecting the health state of transformer in real time, it is related to detecting transformer technical field, for solving the problem of fault early warning lag, by monitoring transformer primary side and secondary side output voltage and calculating fluctuation rate, analysis operation stable state, statistics abnormal duration, based on duration judgment enters temperature rise analysis mode, detect oil temperature and winding temperature and calculate difference, call cooling system start-stop instruction and obtain cooling response value, calculate response lag rate, collect oil dissolved gas data and calculate gas anomaly ratio, combined with response lag rate analysis thermal power collaborative abnormal characteristics, obtain insulation aging grade and oil replacement cycle from historical maintenance database, according to insulation aging grade and oil replacement cycle evaluation insulation degradation coefficient, combined with thermal power collaborative abnormal characteristics analysis health state, select operation optimization or shutdown maintenance mechanism, improve fault early warning capability, optimize maintenance strategy, prolong equipment life.