78 results about "Kernel principal component analysis" patented technology

The invention provides a photovoltaic output prediction method and a distributed optical storage transformer area active / reactive support capability optimization method, and belongs to the technical field of power system automation. Comprising the following steps: constructing a prediction model based on meteorological information, extracting multi-scale features of meteorological data, reducing dimensions through kernel principal component analysis, predicting photovoltaic power through a long-short-term memory network, quantifying the uncertainty of the photovoltaic power, and forming a prediction interval. Based on a photovoltaic power prediction result, a robust nonlinear programming model is established, the sum of absolute values of active / reactive power transmitted to a main power grid by a light storage area is maximized as a target, and linear and nonlinear constraints such as energy storage operation, power balance, equipment capacity and power factors are considered at the same time. The robustness under the photovoltaic output uncertainty is ensured through a budget uncertainty set; the model is solved by adopting a sequential quadratic programming algorithm. According to the method, the photovoltaic power prediction precision is effectively improved, and organic combination of high-precision prediction and robust nonlinear optimization is realized.

According to the plateau area load prediction method and system based on the signal decomposition and the neural network, an original signal is decomposed into a plurality of intrinsic mode components with physical significance by adopting completely adaptive noise set empirical mode decomposition, mode aliasing is suppressed, and multi-scale features are extracted to effectively suppress load demand oscillation; the problem of non-stationarity of the load of the service area is solved; then, K-means clustering grouping is carried out on the components, similar modes are combined to reduce calculation redundancy, and the problems of heterogeneity and local feature redundancy in complex time series data are effectively solved; performing nonlinear dimension reduction processing on the grouped reconstructed signals by adopting kernel principal component analysis, eliminating noise interference and extracting key features, thereby effectively solving the multiple nonlinear problem; and finally, inputting the dimensionality-reduced features into a long-short-term memory network modeling time sequence dependency relationship, and realizing high-precision prediction through adaptive weighted integration of component prediction results.

The invention relates to the technical field of power system fault monitoring, and discloses a measurement switch fault identification and early warning system, which comprises a signal acquisition unit, a feature analysis unit, a fault judgment unit and an early warning trigger unit. The signal acquisition unit acquires a potential fluctuation sequence and an exciting current pulse sequence through a contact voltage sensor and a coil current transformer; the feature analysis unit extracts time-frequency domain and statistical domain feature groups by means of wavelet transform and moment function analysis; the fault judgment unit generates a fault judgment vector by adopting kernel principal component analysis fusion features; and the early warning triggering unit triggers four-level early warning by using dynamically updated grading threshold values (including historical reference values and online dynamic values). According to the system, through multi-modal feature fusion, dynamic threshold adjustment and a grading early warning strategy, accurate identification and grading early warning of measurement switch faults are achieved, the fault identification accuracy and early warning timeliness are improved, and the system is suitable for the field of electric power and industrial automation.

The invention is suitable for the technical field of radar signal processing, and provides a tomographic SAR super-resolution imaging method based on a structured sparse network, and the method comprises the steps: firstly obtaining multi-channel SAR observation data, constructing multi-channel observation vector MMV data in a data domain, and constructing a multi-pixel signal model corresponding to the MMV data based on a neighborhood pixel elevation consistency hypothesis. The method comprises the following steps: designing a kernel principal component analysis KPCA expansion network model, carrying out dimension reduction and enhancement on MMV data, introducing a norm compressed sensing model on the basis of the MMV data after dimension reduction and enhancement, solving the compressed sensing model by using an ADMM iterative algorithm, expanding the ADMM algorithm into a deep network, reconstructing an elevation spectrum, and obtaining an SAR image super-resolution three-dimensional reconstruction result. According to the method, the super-resolution performance and the solving efficiency of three-dimensional reconstruction can be effectively improved, and high-resolution three-dimensional reconstruction of a large-scale scene can be efficiently realized.

The invention discloses a fault diagnosis method and system based on improved residual error and GRF feature fusion. The method comprises the following steps: integrating signals of a multi-source sensor; reducing dimensions by using kernel principal component analysis; extracting features in parallel by adopting an improved residual network, an LSTM (Long Short Term Memory) and a Transform; after normalization and projection transformation are carried out on output of each branch, feature weighted fusion is carried out by adopting a dynamic gating weight mechanism, and timing sequence modeling capability is enhanced by utilizing a gating recursion unit; and finally, stage mark embedding and dynamic zooming processing are introduced, and diagnosis results are output through a full connection layer and Softmax classification. According to the method, the global and local information of the time-frequency domain of multiple sensors is fused, and the diagnosis precision of the complex fault of the rolling bearing is remarkably improved.

The invention relates to the technical field of natural disaster risk assessment, in particular to a multi-disaster disaster-inducing factor internal correlation analysis method, which comprises four steps of data preprocessing, Bayesian network correlation model construction, space-time dynamic evolution analysis and risk early warning threshold model construction. And deep coupling analysis of disaster-inducing factors of multiple disasters such as mountain torrent-debris flow and the like is realized. The method comprises the following steps: firstly, collecting 17 disaster-inducing factor data such as terrain and rainfall, performing dimensionality reduction through kernel principal component analysis, and calculating a dynamic weight by adopting an entropy evaluation method; then constructing a Bayesian network model to quantify conditional probability association; introducing a space-time attention mechanism to optimize a multi-scale coupling weight; and finally, establishing a coupling risk early warning threshold value and outputting a relevance intensity matrix. According to the method, the problem that a traditional single-disaster analysis method cannot capture the space-time linkage effect of disaster-inducing factors is solved, the risk early warning accuracy of multiple disasters is improved, and scientific decision support is provided for regional disaster prevention and reduction.

The invention discloses a distributed energy storage system cooperative control method based on data fusion, and belongs to the field of power system energy storage control. The edge layer performs dimensionality reduction on multi-source data through kernel principal component analysis, the cloud constructs a space-time correlation model by combining a graph neural network with a long-short-term memory network, and data privacy is guaranteed by means of federal learning distributed updating. An upper layer model prediction control framework is combined with an improved whale algorithm to solve multi-target optimization of system economy, energy storage life and power grid stability; the lower-layer power type energy storage adopts self-adaptive droop control, and the energy type energy storage model predicts, controls and quantifies the degradation cost of the battery. When communication is interrupted, the analytic hierarchy process and the consistency algorithm cooperate to disperse power distribution; and when the equipment fails, convolutional neural network diagnosis is combined with alliance chain redundancy switching. Multi-source deep correlation, characteristic differentiation regulation and control and high robustness are realized, the energy storage efficiency is improved, the service life of equipment is prolonged, and the method is suitable for micro-grids, smart grids and other scenes.

The invention relates to an aero-engine fault diagnosis method and device based on tensor decomposition and a medium, and the method comprises the steps: collecting the gas path parameter time sequence data of a plurality of parts of an aero-engine, and carrying out the time alignment and working condition label labeling; for the multi-parameter data of each component, a kernel principal component analysis and automatic encoder fused feature extraction method is adopted, and parameters of different components are unified into fixed dimension features; the feature matrixes obtained after feature extraction of all the components are combined, and a time-component-feature third-order tensor is constructed; training a high-order singular value decomposition model based on the engine health monitoring data; fault diagnosis is carried out by calculating the reconstruction error of the test data and the monitoring data; and fault positioning is realized by combining core tensor difference analysis. According to the method, the problem of feature fusion of engine multi-source heterogeneous data under variable working conditions is solved, and the engine fault detection sensitivity and positioning precision are improved.

The application relates to a method and device for monitoring over-error of a capacitive voltage transformer, electronic equipment and a computer readable storage medium. The method comprises: performing kernel principal component analysis based on historical sample data to determine an SPE statistical quantity threshold; calculating the SPE statistical quantity of current sample data collected in real time; if the SPE statistical quantity is less than or equal to the SPE statistical quantity threshold, determining that the operation state of the transformer group is normal; and if the SPE statistical quantity is greater than the SPE statistical quantity threshold, calculating the contribution rate of each element in the current sample data to the square prediction error statistical quantity, determining a fault element, and locating an over-error transformer in the transformer group. The application effectively processes the nonlinear problem of data by using KPCA, more accurately extracts data features, realizes fault element separation, accurately locates the over-error CVT, performs online real-time monitoring on the CVT error, and improves the stability and reliability of the power system.

The invention provides a third industry user-oriented load adjustable characteristic evaluation and interactive regulation and control method, and aims to improve the demand side resource adjustable capability and support supply and demand coordinated regulation and control under a novel power system. The method comprises the following steps: firstly, collecting typical three-product user loads and multi-source heterogeneous data, after preprocessing and feature extraction, realizing feature dimension reduction by adopting kernel principal component analysis, and completing typical user load clustering through a K-Means clustering algorithm; and secondly, constructing a multi-dimensional user portrait evaluation system based on a clustering result, and quantitatively evaluating the adjustable potential of the typical user load in combination with a multi-dimensional time sequence load model. And finally, establishing a multi-constraint optimization scheduling model, introducing a particle swarm optimization algorithm for solving, and outputting targeted regulation and control strategies and regulation instructions of different time periods and different devices. According to the method, refined evaluation and rapid scheduling optimization of the three-production user load adjustable potential are effectively realized.

The invention discloses a customer group classification method and device based on machine learning and electronic equipment, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the dimension reduction processing of an original data set through employing a kernel principal component analysis algorithm, obtaining a data point set after the dimension reduction, calculating the outlier factor of each data point through employing a local density factor algorithm, and obtaining an outlier classification result; removing abnormal data values by using outlier factors, calculating connection strength between nodes according to a relationship between data points, establishing a linear equation set, solving the linear equation set, obtaining an initial clustering center point, performing clustering analysis based on the initial clustering center point, classifying each data point in a data point set after dimension reduction into a cluster to which the closest centroid belongs, and performing clustering analysis based on the clustering center point. And obtaining a financial management customer group classification result until a clustering convergence condition is met. The technical problems that the real-time updating demand of large-scale customer data cannot be adapted during financial management customer classification in the related technology, the recommended product does not meet the demand, and the customer satisfaction is reduced are solved.

The invention relates to the technical field of medical information, and discloses a psychological disease treatment effect evaluation method and system adopting data analysis. The method includes: classifying standardized physiological parameters based on the model; performing kernel principal component analysis on the sleep quality parameters to obtain dimensionality reduction vectors, and performing nonlinear mapping on the dimensionality reduction vectors and the voice emotion parameters to obtain cross-modal correlation features; if the correlation coefficient exceeds the threshold value, kernel function parameters are adjusted in combination with emotion stress parameters, and features are extracted again; a dynamic weight coefficient is obtained according to the cross-modal correlation features and a weight mechanism, and a treatment effect index is calculated in a weighted mode; and identifying an index abnormal mode through abnormal detection, and generating a personalized evaluation report. Through multi-dimensional physiological data analysis and dynamic adjustment, the problem that evaluation lacks objective multi-dimensional continuous monitoring and dynamic personalized analysis is solved, and the evaluation precision and the personalized level are improved.

This invention provides a method for predicting aging based on metabolite profiles using nonlocal variance enhancement, relating to the fields of medical data analysis and metabolomics. The method acquires LC-MS and / or GC-MS metabolite profile data and the actual age of the subject to be predicted. It performs missing data checks and intra-class normalization on the data, constructs a class-constrained nonlocal variance enhancement feature extraction model, and combines multi-kernel principal component analysis and linear multi-view fusion to obtain fused metabolite profile features. The fused features are then input into a support vector regression model to output predicted age values. The aging degree is determined based on the difference between the predicted and actual age values, which improves the feature expression ability and interpretability of metabolite profile age prediction.

The invention relates to the field of reliability analysis of aircraft hydraulic systems, in particular to a two-stage residual life prediction method of an aircraft hydraulic system based on data dimension reduction, which comprises the following steps: acquiring state data of the aircraft hydraulic system and preprocessing the state data; sequentially carrying out principal component analysis and kernel principal component analysis on the preprocessed data to obtain dimension reduction data; establishing a dual-target prediction model; inputting the trend term and the periodic term into a double-target prediction model to obtain a preliminary life prediction value; inputting the dimension reduction data into a residual prediction network to obtain a life residual prediction value; the residual prediction network is composed of a one-dimensional convolutional neural network and a bidirectional long-short term memory network which are connected in sequence; adding the initial life prediction value and the life residual prediction value to obtain a final residual life prediction value; according to the method, the residual life prediction efficiency and robustness of the aircraft hydraulic system can be improved.

The invention discloses a high-voltage circuit breaker health state assessment method and system based on multi-source data fusion, and the method comprises the steps: constructing a health index model based on multi-source data fusion, and carrying out the effective dimension reduction of complex features in the multi-source data of a high-voltage circuit breaker through the kernel principal component analysis; and the dimensionality-reduced features are fused through a weighted mahalanobis distance to obtain a one-dimensional health index capable of comprehensively reflecting the operation state of the equipment. On the basis, the degradation starting point is identified based on the degradation rate, and the normal state and the potential fault state of the equipment are divided, so that the problems of dependence on a single data source, incomplete evaluation, incapability of early identification of performance degradation and the like in the prior art are solved, and the change of equipment performance can be identified earlier; and accurate evaluation and predictive maintenance of the health state of the high-voltage circuit breaker are realized.

The invention discloses a high-reliability planetary gearbox fault diagnosis method, system, medium and equipment, and the method comprises the steps: carrying out the feature extraction and feature selection based on a random forest algorithm, and carrying out the nonlinear feature extraction based on multi-scale permutation entropy, performing multi-scale coarse graining processing on each vibration signal, calculating permutation entropy values of coarse graining sequences under different scales, forming a multi-dimensional phase space, and obtaining a feature matrix of permutation entropy; performing feature dimension reduction and feature embedding based on kernel principal component analysis; according to the fault diagnosis based on the bidirectional long-short term memory neural network, a classifier of the bidirectional long-short term memory neural network is used, a feature set is divided into a training set and a test set to serve as input layers, and then a classification result is output through two LSTM units of a hidden layer, a full connection layer and Softmax function mapping. And high-reliability planetary gearbox fault diagnosis is realized.

This invention discloses a method and system for detecting fertilizers, relating to the field of fertilizer detection technology using image extraction. The method includes: segmenting particle images using a multi-scale sliding window, extracting hue, saturation, and brightness components, and combining this with an illumination compensation model to eliminate illumination effects and noise interference. Then, high-dimensional feature vectors are compressed using kernel principal component analysis to construct gradient direction distribution and texture consistency analysis, accurately locating defect areas. Subsequently, by comparing the local features of defective and normal areas, local features of the defective area are extracted, and the texture consistency and color component differences of adjacent sub-blocks are analyzed to output the defect location coordinates and a comprehensive quality rating. This invention establishes a fertilizer visual feature quantification system adapted to complex working conditions, achieving a stable mapping between particle physical properties and imaging features, and improving the accuracy of automated detection.

The present application relates to a kind of coal mine driving face gas emission quantity prediction method based on KPCA-POA-LSTM model, belong to coal mine driving face gas prediction technical field.The present application includes: using kernel principal component analysis to the nonlinear coal mine driving face coal mine driving face gas outburst data of coal mine driving face is reduced dimension and initialization;Again using peacock optimization algorithm optimization long short-term memory network node, make its parameter regenerate continuous distribution;Multi-dimensional state matrix is constructed, using long short-term memory network is mapped to multi-dimensional state matrix and selects sigmoid as activation function, Adam is solver;Again using optimized long short-term memory network to the coal mine driving face gas data is predicted.The present application can accurately predict the coal mine driving face gas emission quantity, and the prediction method is faster, and error rate is lower.

The application discloses an IGBT residual life prediction method based on multi-feature fusion and KPCA optimization. The method first collects IGBT collector current and voltage and other multi-source signals, extracts time domain, frequency domain, time-frequency domain and derived statistical domain features after pretreatment; then adopts a two-stage strategy of comprehensive evaluation index preliminary screening and mutual information regression fine screening to eliminate redundant features and retain high correlation features. On this basis, nonlinear dimension reduction is carried out by using kernel principal component analysis to construct a high-robustness health index, and the exponential weighted moving average and adaptive gradient detection are combined to accurately divide the degradation into three stages. Finally, the CNN-BiLSTM model is used to deeply mine the time sequence degradation features, and the MC Dropout algorithm is introduced to realize the accurate prediction and uncertainty quantification of the residual life. The application effectively solves the one-sidedness of single-source signal representation and the feature redundancy interference problem, and significantly improves the prediction accuracy and reliability.

The invention relates to the technical field of biological feature recognition, and discloses a gait recognition method and device, a storage medium and computer equipment, and the method comprises the steps: firstly obtaining a gait video sequence, extracting a key frame of the gait video sequence, converting the key frame into a space-time energy diagram, carrying out the kernel principal component analysis of the space-time energy diagram, and obtaining a gait recognition result; the method comprises the steps of obtaining low-dimensional gait features, obtaining corresponding personnel identity tags, inputting the low-dimensional gait features and the personnel identity tags into an initial gait recognition model, training the initial gait recognition model based on a preset machine learning algorithm to obtain a target gait recognition model, and finally compressing the target gait recognition model to obtain a target gait recognition result. And obtaining a lightweight gait recognition model, and performing gait recognition in a scene to be recognized by using the lightweight gait recognition model to obtain a gait recognition result. According to the method, high precision and stability in the recognition process are ensured, the lightweight and deployment efficiency of the model are considered, and the method has a wide application scene.

The invention provides a seismic data filtering method and device based on kernel principal component analysis and a medium, and belongs to the field of seismic data processing. The method comprises the following steps: extracting a trend time difference attribute of an input three-dimensional seismic data volume; setting a surface element size and extracting a corresponding line data volume according to the surface element size; selecting a target point and calculating surface element coordinates according to the trend time difference data; setting the size of a time window and extracting a small three-dimensional data volume in combination with surface element coordinates; and performing kernel principal component analysis on the small three-dimensional data volume to obtain a first principal component, and taking central point data of the first principal component as filtered data of the target point. According to the method, original data are mapped into a high-dimensional space through nonlinear mapping by adopting kernel principal component analysis, so that nonlinear structures and characteristics in the original data can be better reserved; in addition, the trend time difference attribute of seismic data is also considered, geologic structure data can be more accurately obtained by opening up a time window along the event trend, the problem of data discontinuity caused by the influence of stratum inclination is improved, and the filtering effect is improved.

The invention discloses a construction method for a CO2 emission prediction model in the iron and steel industry. The construction method comprises the steps of data collection, cleaning and normalization processing; normalized CO2 emission data is input into an empirical mode decomposition (EMD) module, the non-stationarity of the data is reduced, and kernel principal component analysis (KPCA) processing is carried out on each IMF component to extract nonlinear features of the IMF component and reduce the dimensionality of the IMF component. And taking the IMF component after dimension reduction and the residual term as input, constructing a long short-term memory (LSTM) network for training and prediction, modeling and predicting the CO2 emission data after dimension reduction through an LSTM model obtained through training, and finally obtaining a prediction result of the CO2 emission. According to the method, the EMD method, the KPCA method and the LSTM method are combined, and the problems that according to an existing prediction method, data non-stationarity and non-linear feature mining is not deep, the feature extraction efficiency is low, and the long sequence data processing capacity is limited are effectively solved.

The application relates to the field of medical information technology and discloses a psychological disease treatment effect evaluation method and system adopting data analysis. The method comprises the following steps: classifying and standardizing physiological parameters based on a model; obtaining a dimension reduction vector of a sleep quality parameter through kernel principal component analysis, and then obtaining a cross-modal correlation feature through nonlinear mapping of the sleep quality parameter and a voice emotion parameter; calculating a correlation coefficient, adjusting a kernel function parameter in combination with an emotional stress parameter if the correlation coefficient exceeds a threshold value, and then reextracting the feature; obtaining a dynamic weight coefficient through the cross-modal correlation feature and a weight mechanism, and weighting a treatment effect index; and identifying an index abnormal mode through abnormal detection to generate a personalized evaluation report. Through multi-dimensional physiological data analysis and dynamic adjustment, the application solves the problems of lacking objective multi-dimensional continuous monitoring and dynamic personalized analysis in evaluation, and improves the evaluation accuracy and the personalized level.

The application relates to a kind of industrial robot position error and its interval prediction method and electronic equipment, the method comprises the following steps: constructing data set containing joint angle vector and position error;Using kernel principal component analysis, joint angle vector is projected to high-dimensional space, feature dimension is upgraded;After feature dimension is upgraded, feature and corresponding position error are respectively used as the input and label of bayesian neural network, and bayesian neural network is trained;Test set data is input to trained bayesian neural network after being upgraded according to the same method, and the prediction of position error and its confidence interval is output.Compared with prior art, the application has the advantages of realizing the prediction of robot position error and interval, more accurate prediction value and good applicability.

The invention provides a coal-fired unit group comprehensive energy consumption curve generation method based on kernel principal component analysis. According to the coal-fired unit group comprehensive energy consumption curve generation method based on kernel principal component analysis, the real energy consumption characteristics of a pure condensing type coal-fired unit group can be reflected more accurately, the stability and reliability of the comprehensive energy consumption curve are improved, and scientific measurement and calculation of variable cost and market mechanism design in the electricity market are effectively supported.

The invention discloses a microelectronic system sensor anomaly detection method based on improved kernel principal component analysis, and relates to the technical field of microelectronic sensing, and the method comprises the following steps: S1, obtaining steady-state normal operation data and fault data of a target sensor, and constructing an input matrix according to the steady-state normal operation data; s2, standardizing the input matrix, and training the improved KPCA model by using a standardized training sample; s3, obtaining an orthogonality enhanced feature vector matrix; s4, calculating the T2 statistic of the weighted mahalanobis distance and the SPE statistic of the local density; and S5, calculating a self-adaptive empirical threshold, and carrying out anomaly detection by using the self-adaptive empirical threshold. According to the method, the complete normalized training feature matrix is stored in the training stage, the consistency of kernel space construction in the prediction stage is ensured, and the engineering applicability and deployment stability of the model are improved.

This application discloses a method for determining model parameters, a method for classifying vehicle abnormal noises, an apparatus, and a device. The method includes: acquiring a training dataset of vehicle abnormal noise features; generating an initial population based on the training dataset; the population containing multiple individuals; constructing a kernel principal component analysis (KPC) model based on preset model parameters; the KPC model is used to extract target abnormal noise feature vectors from the vehicle's abnormal noise features for vehicle abnormal noise classification; determining an objective function based on the KPC model; iteratively updating the initial population based on the objective function to obtain an updated population; determining the optimal individual from the updated population; and determining the target model parameters of the KPC model based on the optimal individual and a preset parameter range. This method can determine the target model parameters of the KPC model most suitable for the current abnormal noise environment, thereby improving the classification accuracy of vehicle abnormal noise classification.

The invention provides a fire range extraction method based on a nuclear combustion index and adaptive spatial spectrum optimization, and relates to the technical field of remote sensing image processing, and the method comprises the steps: collecting remote sensing image data; mapping a spectral vector corresponding to the fire image data set to a high-dimensional space based on a Gaussian radial basis kernel function, and performing kernel principal component analysis on the spectral vector to construct a kernel difference combustion index; according to kernel density estimation, performing adaptive threshold segmentation on the kernel difference combustion index to obtain an initial fire pixel detection result; constructing a spectrum space composite kernel similarity matrix, and optimizing the initial fire pixel detection result in a conditional random field according to the spectrum space composite kernel similarity matrix to obtain a transition fire pixel detection result; and performing morphological operation and small patch area filtering on the transition fire pixel detection result to obtain a predicted fire pixel detection result. According to the invention, the overall precision of fire range extraction can be improved.

The invention provides a data dimension reduction method based on quantum mechanical characteristics. The method comprises the following steps: S1, performing nonlinear dimension reduction processing on high-dimensional data through quantum kernel principal component analysis; s2, performing linear dimensionality reduction on the output of the S1 based on a quantum neighborhood preserving embedding algorithm of mahalanobis distance; and S3, mapping the output of the S2 to a low-dimensional space by adopting a quantum variational manifold learning algorithm, and generating a final dimension reduction result. According to the method, efficient dimension reduction of high-dimensional data can be realized, the efficiency and precision of quantum machine learning preprocessing are improved, the calculation complexity is reduced, and the accuracy and reliability of a dimension reduction result are improved.

The invention discloses a gear remaining life prediction method based on multi-feature fusion and an Informer model, and belongs to the technical field of mechanical equipment fault prediction and health management. The method comprises the following steps: firstly, acquiring data to obtain a full-life-cycle vibration signal of a gear; then, extracting an initial feature set capable of representing performance degradation from multiple dimensions of a time domain, a frequency domain and a time-frequency domain; then, a feature fusion method based on kernel principal component analysis is adopted to carry out dimension reduction and fusion on the high-dimensional initial features, and single, sensitive and monotonous health indexes are constructed; and finally, the constructed health index sequence is used as input, and the residual service life of the gear is directly predicted by using the strong long sequence time sequence prediction capability of the Informer model. According to the method, the characterization capability of health indexes is effectively improved through multi-feature fusion, the limitation of a traditional recurrent neural network in long sequence prediction is overcome by using the Informer model, and more accurate and more stable prediction of the gear remaining life is realized.