38 results about "Trend extraction" patented technology

The invention relates to the technical field of power system operation and control, and particularly discloses a time-sharing electric quantity prediction method based on a logarithmic load density growth curve, which comprises the following steps of: firstly, performing causal detection and dynamic time-delay optimization on historical load and multivariate external data through convergence cross mapping and mutual information technologies, and constructing a causal time-delay feature set; and the problems of multi-element coupling and time-delay effect quantization are solved. Secondly, fitting a load trend by using time-frequency decomposition in cooperation with a segmented logistic model, extracting dynamic parameters representing a growth rate and a saturation capacity, and endowing the model with a sensing ability for a load evolution stage; then, causal features, growth parameters and load components are deeply fused through cross-domain modulation and a gating mechanism, the nonlinear modulation effect of an external environment on a load mode is explicitly modeled, and finally, a probability interval is generated in combination with quantile regression and residual error correction. According to the scheme, accurate and probabilistic prediction of the time-sharing electric quantity in a complex scene is realized, and the scientificity of an agent electricity purchase decision is improved.

The invention discloses an oil chromatography data trend classification method and system based on depth feature enhancement and an attention mechanism, and the method comprises the steps: carrying out numeralization conversion, deletion detection and grouping trend calculation on oil chromatography original gas component data, and generating a basic feature vector; executing multi-scale sliding statistics, change rate and subsequence feature enhancement, and calculating comprehensive similarity and attention weight based on a template library to generate a weighted similarity vector; splicing the enhanced feature and the weighted similarity vector into a time sequence input sequence, and outputting an oil chromatogram trend classification result after attention expansion and long and short term memory network processing. According to the method, structured processing and basic trend extraction of data are realized, adaptive matching and weighted aggregation of historical operation modes are realized, and a multi-dimensional dependency relationship and time sequence dynamic change are captured, so that accurate classification of oil chromatogram trends is realized.

The invention relates to the technical field of storage intelligence, and discloses an early warning and allocation system and method for intelligent inventory of inspection and quarantine reagent consumables, and the system comprises a consumption trend extraction module, a demand prediction module, a risk early warning module, an early warning trigger module, a supply feasibility evaluation module and an intelligent allocation module. Performing seasonal decomposition on the historical consumption record to obtain a consumption trend component; periodically superposing the consumption trend component to obtain a demand prediction result; based on the demand prediction result, performing inventory risk research and judgment on the current inventory to obtain an inventory early warning index; when the inventory early warning index exceeds a safety threshold value, inventory early warning information is generated; according to the consumable demand information, carrying out feasibility evaluation on inventory data and logistics constraint conditions of the supply point to obtain a feasible allocation point; generating a deployment scheme according to the safety inventory standard of the feasible deployment point; according to the invention, the perspectiveness and accuracy of early warning and allocation of the intelligent inventory of the inspection and quarantine reagent consumables can be improved.

The invention discloses a tunnel construction safety risk management and control method and system. The method comprises the following steps: collecting multi-source sensor data of a construction site, constructing a continuous time sequence monitoring data set, introducing a Hampel model to detect abrupt change and abnormity and identify sudden risk features, and introducing a CUSUM model to detect a slow change trend and extract evolutionary risk features. Time synchronization and vector combination are carried out on the two types of risk features, a joint risk feature matrix is constructed, dominant risk types are further judged, dynamic weighted fusion is implemented, a fusion risk index is generated, risk levels are divided according to the fusion risk index value, and the risk level is determined by combining the dominant risk types and the risk trend. And an informatization risk response result including a response level, a response strategy and trend prediction is generated, and accurate identification and intelligent management and control of tunnel construction risks are realized. The invention is suitable for dynamic risk early warning and management of a tunnel engineering construction site, and belongs to the technical field of tunnel engineering safety management.

The invention relates to the technical field of post-operation follow-up visit management, in particular to a post-operation follow-up visit management system based on artificial intelligence, which performs standardization and time alignment processing on information such as electronic medical records, vital signs, inspection and examination data, medication records and post-operation symptom texts through a data acquisition module to construct a unified initial data matrix; and the time sequence continuity of modeling is ensured. And introducing a medical natural language processing flow, extracting keywords of symptoms, parts and discomfort properties, and realizing structured expression. The symptom keywords and the related physiological indexes are dynamically matched through the feedforward neural network, the limitation of a fixed template is avoided, and the pertinence of feature extraction is improved. And finally, inputting the symptom associated data into a time sequence model comprising a gating circulation unit and an attention mechanism, modeling a characteristic evolution trend, extracting key time period changes, and effectively revealing potential anomalies. Intelligent analysis, index focusing and trend prediction of postoperative discomfort feedback are realized, and follow-up management efficiency and predecessibility are improved.

The invention provides a cellular data low-frequency trend self-adaptive extraction method, which is used for automatically extracting an evolution trend from nonlinear and strong-fluctuation cellular flow data. According to the method, firstly, missing values in an original sequence are restored through Akima spline interpolation, and continuity and smoothness of a data structure are restored; and then constructing a joint objective function to minimize the permutation entropy of the candidate trend and maximize the Pearson correlation coefficient of the candidate trend and the original sequence, and then guiding the adaptive confirmation of the order and modal quantity parameters of a B spline in TVFEMD decomposition, thereby extracting a low-frequency trend component with the highest representation force, and realizing unsupervised and automatic trend extraction. Finally, the provided method can perform adaptive modeling under various time scales such as day, week, month, year and the like, outputs trend components of corresponding time scales, and can be applied to application scenes such as capacity planning, holiday and festival load analysis, energy-saving scheduling and service prediction in a cellular network. The method has the advantages of generalization, full-automatic parameter adjustment and the like, the trend extraction efficiency and stability of the cellular data are remarkably improved, and reliable trend information support is provided for intelligent operation and maintenance and decision optimization of the cellular data.

A business management system with integrated business intelligence and IT-supported optimization, consisting of: a computer device comprising at least one processor and a memory arrangement for storing executable instructions; a data acquisition and integration unit that is operationally connected to the at least one processor and is configured to receive heterogeneous enterprise data from a variety of source systems, including financial transaction systems, operational control systems, human resource information systems, supply chain systems, customer interaction systems, and IT infrastructure monitoring systems, wherein the data acquisition and integration unit is further configured to normalize syntactic formats, resolve temporal inconsistencies, and map source-specific attributes into a unified enterprise data representation stored in the storage system; a business intelligence processing unit that is operationally coupled with the data acquisition and integration unit and is configured to calculate business performance indicators through multidimensional aggregation, correlation analysis, and trend extraction over the unified business data representation, with the calculated business performance indicators being time-stamped and permanently stored; an optimization unit that is operationally coupled with the business intelligence processing unit and is configured to compare the calculated business performance indicators with stored business objectives and constraints and to determine optimized business control parameters that correspond to resource allocation, operational planning, cost efficiency or performance improvement; a control and orchestration unit that is operationally linked to the data acquisition and integration unit, the business intelligence processing unit, and the optimization unit, wherein the control and orchestration unit is configured to coordinate the execution sequence, forward the results of the business intelligence analysis to the optimization unit, and manage feedback by incorporating the effects of applied optimized business control parameters into subsequent data acquisition cycles; and a communication interface that is connected to at least one processor and configured to exchange data and control signals with external enterprise systems and user terminals.

The application relates to a charging pile intelligent operation and maintenance management method and device, computer equipment and a medium. The method comprises collecting battery state parameters and charging pile operation state parameters; trend extraction is performed on the time sequence change of the battery state parameters and the charging pile operation state parameters, and correlation analysis is performed to generate a multi-dimensional state index representing the current operation state of the charging object and the charging equipment; based on the multi-dimensional state index, a corresponding control strategy is selected from a charging control strategy set, and stage control parameters set for different residual capacity intervals in the control strategy are adaptively evaluated with the current state index; and according to the adaptivity evaluation result, a charging operation of the corresponding strategy is performed. The application has the effect of improving the charging control efficiency of the charging pile.

The invention discloses a business demand analysis method and device, computer equipment and a storage medium. According to the invention, the purchase data, the transaction information and the rejection record in the hotel resource management system can be uniformly mapped to the labeled performance index set, and the system can accurately identify the dependency relationship between the purchase behavior and the transaction result based on the indexes. The abnormal transaction relationship is identified through the order rejection record, and the dependency relationship model is divided into a plurality of relationship model subgroups. Key operation trend characteristics are extracted through trend clustering and analysis of the multi-dimensional integrated intersection points, and the trend characteristics are mapped to all links of the resource operation process. Based on the operation demand results, the system can identify a key optimization path and generate an executable set of demand adjustment rules. On the whole, closed-loop management from data acquisition, dependency relationship identification, anomaly analysis, trend extraction to demand optimization rule generation is realized, and the intelligent level of resource operation is improved.

The invention discloses a WeChat applet operation trend prediction method based on digital twinning, and the method comprises the following steps: collecting WeChat applet multi-source operation data, and generating structured input data; constructing a digital twin structure map; injecting a strategy disturbance event into the digital twin structure map through a strategy disturbance simulation method, and generating a structure variant sequence and a behavior variant sequence; performing fusion coding on the structured input data and the variant sequence; using an improved Giraffe model to output an operation trend prediction result and a confidence interval; carrying out trend extraction and strategy influence analysis by adopting a sensitivity scoring method and an inflection point detection method; and selecting an optimization strategy parameter according to the strategy sensitivity score, updating the digital twin structure map through a strategy backfilling and prediction verification method, and performing re-prediction. According to the method, precise modeling and strategy optimization of the WeChat applet operation trend are realized, and the strategy decision efficiency and the operation prediction reliability are improved.

The invention relates to the technical field of self-adaptive temperature control of an SSD, and discloses a self-adaptive temperature control method and system based on the SSD.The method comprises the steps that internal data and external parameters of the SSD are collected, trend extraction is carried out, and a load change trend is obtained; according to the load change trend, identifying a heat accumulation risk point, carrying out heat dissipation adjustment, collecting temperature data, and generating a temperature distribution diagram; performing regional division on the temperature distribution diagram, and generating a load distribution scheme in combination with preset load adjustment parameters; screening tasks in the load distribution scheme, and determining an initial task list; if the continuous high-intensity read-write operation exists in the initial task list, redistributing the continuous high-intensity read-write operation to other partitions of the SSD to generate an optimized task list; and recording temperature and performance fluctuation conditions in the operation process of the optimization task list, and generating a temperature stable state record. According to the method, the temperature can be controlled, and the SSD performance is improved.

This application relates to the field of computer-aided process optimization technology, and particularly to a method and apparatus for optimizing and analyzing automotive bolt forming processes. The method includes: acquiring a dataset of process parameters for an automotive bolt forming task and optimization trend information for bolt forming; extracting trends from the fluctuation data of each parameter in the process parameter set to obtain parameter characteristic information of the process parameter set; determining parameter fluctuation information for each process parameter set based on the parameter characteristic information; performing optimization calculations based on the parameter fluctuation information and optimization requirements in the process parameter set to obtain a target process set; determining process influence information from the parameter information corresponding to the target process set; and calculating the optimal parameter combination for the bolt forming process based on the parameter adjustment range of the process influence information and optimization trend information. The method provided in this application can solve the problem of deviations in process optimization direction, which in turn affect product quality.

The invention provides an urban heat island effect space-time evolution mode mining method and system fusing time sequence clustering and trend extraction. The method comprises the following steps: acquiring remote sensing image data of a research region in a research year section; dividing urban areas in the research area based on the land coverage data; dividing a rural reference area based on the urban area in the research area, the land coverage data, the normalized vegetation index and the elevation data; calculating the surface urban heat island intensity of each pixel in the urban area in the research year section in combination with the surface high temperature data and the country reference area; performing clustering processing on each pixel in the urban area based on a calculation result of the surface urban heat island intensity by adopting a time sequence clustering algorithm to obtain a plurality of space partitions; and constructing an average earth surface urban heat island intensity time sequence of each space partition and carrying out trend analysis to obtain urban heat island effect space-time evolution mode characteristics, including earth surface urban heat island intensity evolution trends and significance levels of the earth surface urban heat island intensity evolution trends of each space partition.

The application discloses an injection mold human-computer interaction optimization system and particularly relates to the field of fault prediction and health management of injection mold monitoring, and comprises a data normalization module, a trend extraction module, a tensor construction module, a score identification module, a risk generation module and an early warning triggering module; the data normalization module collects cavity pressure curve, guide pillar displacement, mold locking force fluctuation curve and mold surface temperature state data in an injection cycle, performs cycle segmentation and synchronous normalization processing, and constructs a state data window; the application constructs a behavior mutual exclusion relationship tensor based on state data combination under the condition of edge calculation, dynamically identifies and predicts deviation areas in combination with confidence change trends, realizes the perception response of the conflict relationship of multiple state parameters, and thus solves the problems that the existing fault prediction and health management system cannot identify the conflict between state variables, causes the distortion of health determination and the delay of intervention.

The application discloses a disturbance detection method and device based on trend extraction, equipment and a control system, and belongs to the field of process industry control. After obtaining time series data, a segmented polynomial fitting method is used to extract the trend of the time series data to obtain a target segmentation result. Then, a trend index and a normalized mean absolute error of the time series data are calculated based on the target segmentation result. Finally, when the trend index is greater than or equal to a first threshold value and the normalized mean absolute error is less than a second threshold value, it is determined that a disturbance exists. When the trend index is greater than or equal to the first threshold value and the normalized mean absolute error is less than the second threshold value, it indicates that the time series data has a significant trend, that is, a certain disturbance exists. With the double criteria of the trend index and the normalized error, the small disturbance hidden in the trend can be effectively identified, the noise resistance is high, and the accuracy of disturbance detection is greatly improved.

The invention relates to the technical field of learning methods, in particular to an AI-based broadcast television network fault prediction and self-healing method, which comprises the following steps of: extracting a channel quality difference value through edge detection, recombining to form a change trend set, performing three-point windowing to execute differential analysis, identifying a confidence interval compression trend, and extracting a prediction and detection symbol relationship. And identifying an error reversal direction, and finally generating a repair trigger signal label set. According to the method, a dynamic trend structure of a prediction difference value is constructed, a label index is extracted and triggered in combination with symbol consistency and direction reversal features, aggregation and label binding of time sequence features are completed, a training sample set capable of being used for learning is generated, the association identification capability between the features is enhanced, the sensing precision of weak change signals is improved, and the early warning range is expanded. Abnormal misjudgment is reduced, a judgment basis based on trend evolution is formed, and reliable data support is provided for fault prediction and self-healing response.

The invention relates to the technical field of cloud collaboration, and particularly discloses a heart rate data intelligent processing method and system based on feature domain dynamic decoupling, and the method comprises the steps: an edge calculation port constructs a credibility curve reflecting data reliability, and facilitates the distinguishing of stable sampling and abnormal sampling at a data level; then, the edge calculation port dynamically decouples the credibility curve to form an evidence chain quantization driving table set for representing the heart rate steady state evolution process, and a heart rate steady state data trend line is generated, so that the robustness of the trend extraction process to short-time disturbance is enhanced; on the basis, an edge computing port is combined with an evidence chain quantization driving table set and a heart rate steady state data trend line, heart rate data interpretation results with interpretability are extracted and uploaded to a user side and a cloud processing computing side, and stable processing and result collaboration of the heart rate data are achieved.

The application provides a photovoltaic power day-ahead prediction method and system based on time period-multiple task learning, which predicts a general power trend according to irradiance and temperature prediction values, predicts a power trend with adjacency based on features of adjacent days, extracts historical similar features under the same weather type to predict a power trend with similarity, and solves the optimal ratio of each power trend by using a genetic algorithm according to the general power trend, the adjacent power trend and the similar power trend, and obtains the final prediction result by integrating each power trend through a time period integration method. The application is based on a deep learning and multiple task learning-time period integration framework, so that the prediction is more targeted and adaptive, and the power fluctuation details can be described under extreme weather types.

The invention provides a photovoltaic power generation-oriented combined transformer intelligent adaptation control method and system. The method comprises the following steps of: obtaining an operation state vector of a transformer system; the running state vectors at each moment form a running state vector sequence according to a time sequence. And performing trend extraction according to the running state vector sequence to obtain trend feature representation. And performing penalty calculation on the trend characteristic representation according to power slope constraint and temperature inertia to obtain a trend index set. And calculating according to a preset mapping function and the trend index set to obtain the contribution degree weight of each control channel. And performing strategy generation according to a preset control strategy optimization function and the contribution degree weight to obtain a control activation combination, and controlling the transformer system based on the control activation combination. Wherein the control activation combination is used for indicating whether the corresponding control channel is activated or not. Dynamic feed-forward adjustment of a transformer system is achieved, and stability is improved.

To provide trend information to the appropriate users. [Solution] The information processing device according to the present invention is characterized by having: an identification unit that identifies trends based on user behavior information in a predetermined service; an extraction unit that extracts accounts in a predetermined messaging service that correspond to the trends identified by the identification unit; and a provision unit that provides trend information on a predetermined messaging service to target persons linked to the accounts extracted by the extraction unit, via a provision account different from the account.

This invention relates to the field of power grid monitoring technology, specifically a port energy monitoring and analysis system. The system includes an equipment electrical parameter acquisition module, a path association and partitioning module, an energy consumption trend extraction module, an energy consumption structure modeling module, and a monitoring tag output module. In this invention, by dividing the three-phase current sequence of equipment into slope segments to extract action segments, refined identification of electrical parameter behavior is achieved, enhancing data continuity and behavioral recognition. Combined with GNSS positioning, the operation trajectory is divided into specific areas and matched with corresponding electrical parameter behaviors, establishing a spatial linkage relationship between the path and the load, improving the accuracy of operation scenario identification, constructing typical energy consumption response profiles of equipment, and achieving structured modeling of energy consumption patterns under different operation scenarios. Based on the current response, load anomalies are judged and the operation segment status is marked, forming a closed-loop link from acquisition, identification, analysis to early warning, significantly improving remote monitoring accuracy and energy consumption identification capabilities.

The application relates to the technical field of cloud collaboration, and particularly discloses a heart rate data intelligent processing method and system based on feature domain dynamic decoupling, which comprises the following steps: an edge computing port constructs a reliability curve reflecting data reliability, which is beneficial to distinguishing stable sampling from abnormal sampling at the data level; subsequently, the edge computing port dynamically decouples the reliability curve to form an evidence chain quantitative driving representation set for representing a heart rate steady-state evolution process, and generates a heart rate steady-state data trend line, so that the robustness of a trend extraction process to short-time disturbance is enhanced; on this basis, the edge computing port combines the evidence chain quantitative driving representation set and the heart rate steady-state data trend line to extract a heart rate data interpretation result with interpretability and uploads the result to a user end and a cloud processing computing end respectively, so that stable processing of heart rate data and result collaboration are realized.

The application relates to the technical field of learning path planning, in particular to a personalized learning path planning method and system based on LSTM fusion Attention, which comprises the following steps: constructing a question-answering feature time sequence based on information of question-answering records, wrong question tables and knowledge point sets, and generating question-answering performance features as hidden states through a time sequence correlation trend extraction model, wherein the time sequence correlation trend extraction model is constructed based on an LSTM architecture; calculating dynamic attention weights through an Attention model by setting a time decay term, a continuous question-answering difficulty feedback term and a knowledge point correlation term, and adjusting iterations of the hidden states based on the dynamic attention weights; generating a predicted knowledge point mastery degree through a fusion mapping model, wherein the fusion mapping model is constructed based on a full connection layer and an activation function; and generating a personalized recommended learning path based on the predicted knowledge point mastery degree. The application realizes personalized learning path planning that fits the individual needs of students and improves the intelligentization of a college teaching system.

The invention discloses a method and a system for prejudging a regulation window of a Sagomean area based on spot price fluctuation, and the method comprises the steps: constructing a spot price sequence, collecting day-ahead / intra-day clearing prices, and carrying out the abnormality correction and trend extraction; calculating a price change rate, and screening a historical high-order interval as a candidate excitation section; evaluating the transmission capability of the regional section, and constructing a flexible adjustment model in combination with the sensitivity matrix and the adjustment cost coefficient; calculating a comprehensive excitation index by combining price excitation and adjustment capability, and determining an effective regulation window; generating a new energy output instruction, screening stations which respond quickly, and calculating a reference output value; implementing closed-loop correction, dynamically adjusting the execution deviation and optimizing a subsequent strategy; according to the invention, through dynamic matching of price excitation and adjustment capability, a high-income regulation and control window can be accurately identified, the response economy of a new energy station is improved, and invalid regulation and control are avoided.

The invention relates to the technical field of data spectral measurement and data correction, in particular to a reflectivity reconstruction method based on simplified multi-core support vector regression, which comprises the following steps of: acquiring the difference between a light spot and a reference reflectivity and dividing sections, judging the change of a wave band and identifying abrupt change points, and counting the difference to generate a label group. According to the method, the response capability to the discontinuous band difference can be improved by identifying the variation trend of the spectral reflectivity deviation in a segmented manner, extracting the sudden change position and distinguishing different change intervals, the mapping structure type is dynamically set in combination with statistical analysis, and the reconstruction efficiency is improved. According to the method, the accuracy of complex deviation modeling is enhanced, a support vector regression structure is established in each section according to a kernel function rule, diversity mapping is realized, the nonlinear fitting capability is improved, the robustness and stability of a reconstruction result are effectively improved under a small sample condition, and the consistency processing and contrastive analysis efficiency of multi-device spectral data is enhanced.

The invention discloses a coal mine electrical load multi-step interval prediction method based on trend extraction, and the method comprises the steps: obtaining original time sequence data of a coal mine electrical load, carrying out the preprocessing, obtaining the preprocessed electrical load time sequence data, and dividing the data into a first training set, a verification set and a first test set according to the time sequence; applying a trend extraction method to the preprocessed electrical load time sequence data to extract a smooth trend component sequence; training and verifying the sPatchTST model for single-variable optimization according to the trend component sequence of the first training set and the verification set to obtain a first sPatchTST point prediction model, and obtaining a trend prediction value of a time step corresponding to the verification set; calculating a residual error sequence of the preprocessed electrical load time sequence data corresponding to the verification set and the trend prediction value; performing interval generation on the residual error sequence by using a conformal prediction method to obtain a residual error fluctuation interval; and inputting the first test set into the first sPatchTST point prediction model to obtain a multi-step trend prediction value under the first test set in a future preset duration, and merging the multi-step trend prediction value with the residual fluctuation interval to generate a final coal mine electrical load multi-step prediction interval. According to the method, a prediction interval with strict statistical coverage rate guarantee is constructed for a prediction result, so that reliable quantification of uncertainty is realized.

This invention discloses a method and system for assessing catheter tip displacement risk. It collects patient positional change frequency data and historical catheter positioning deviation records. Risk factor sets are extracted through rhythm disorder feature analysis. Combining the risk factor sets with historical deviation records, displacement risk characteristic factors are extracted and risk contribution assessments are performed to generate a risk intervention plan. Scoring indicators are extracted from the risk intervention plan to construct a displacement scoring rule library. A set of sensitive displacement parameters is identified and archived to generate a displacement risk profile. Scoring item association analysis is performed on the displacement risk profile, and a graded displacement assessment sequence is formed through drift trend extraction and weighted labeling. Displacement probability calculation is performed on the graded displacement assessment sequence. Catheter tip displacement risk assessment results are generated through stratified monitoring of fluctuation amplitude and matching of re-examination frequency, achieving dynamic assessment of catheter tip displacement risk throughout the entire process.

The invention relates to the technical field of rail transit, and provides an anomaly recognition method and device based on electrical analog quantity curve characteristics, and the method comprises the steps: carrying out the index moving average processing of an original analog quantity data sequence, and obtaining a target analog quantity data sequence; carrying out data change trend extraction on the target analog quantity data sequence by adopting a sliding window to obtain a window slope; and comparing the window slope with a preset slope threshold range, and when the window slope exceeds the preset slope threshold range, determining that the analog quantity is abnormal and performing early warning. According to the method, random noise in the original analog quantity data sequence is removed, the data change trend is extracted from the target analog quantity data sequence in the sliding window, the window slope reflecting the rate and direction capable of quantifying analog quantity trend change is obtained, and the window slope reflects the change trend of the analog quantity after denoising, so that the accuracy of the analog quantity trend change is improved. Simulation anomalies can be accurately and timely determined through the window slope, and the fault risk can be pre-judged in advance.

The invention relates to the technical field of data processing, in particular to an ecological environment detection data processing method based on big data acquisition, which comprises the following steps: acquiring a monitoring field and identifying medium attributes, continuously constructing a merging sequence, extracting a numerical value change direction, and extracting a pollution range in combination with a continuous field. And comparing direction differences to generate a field path collaboration and differentiation structure identification result, and combining with spatio-temporal information to promote trend identification to obtain a monitoring data structure evolution trend identification result. According to the method, a structure sequence is continuously constructed through field medium attribute marking and a source sequence, a trend classification basis is extracted in combination with numerical direction changes, the change recognition capability of continuous data is enhanced, structural relations between fields are constructed by means of trend differentiation comparison, and construction of trend paths between the fields is promoted in cooperation with spatial positions and attribute mapping. The trend extraction, the structure partitioning and the direction distribution identification of the pollution monitoring data are completed, and the data analysis depth under the attribute is expanded.

This invention discloses a method and system for fault diagnosis of electric actuators, belonging to the field of fault diagnosis technology. The method specifically includes: collecting current, voltage, displacement, temperature rise, and electromagnetic noise signals of the electric actuator during operation to form a multi-dimensional operating dataset; obtaining feature vectors by time-frequency domain fusion processing of the multi-dimensional operating data; constructing a health state trajectory of the electric actuator; introducing an electromagnetic noise coupling factor into the health state trajectory; performing evolutionary analysis on the health state trajectory; extracting trend information related to faults; and when the trend information exceeds a set threshold, determining the fault type and severity by comparing it with pre-stored actuator fault evolution samples. This application can achieve early identification and classification of potential faults through trend extraction and adaptive comparison, thereby improving the comprehensiveness and accuracy of diagnosis.