30 results about "Time-series segmentation" patented technology

An adaptive calibration method and system for an electric energy meter, relating to the technical field of electric energy meter calibration. The method comprises: acquiring basic data and additional data of an electric energy meter (S1); performing time series data alignment, and performing period division on an environment data set to generate a time series segmentation result (S2); acquiring device information of the electric energy meter, performing key feature analysis within a period, and establishing an interference weight factor for each feature (S3); building a performance degradation model (S4); acquiring usage environment data of the electric energy meter by means of a monitoring sensor (S5); then generating an adaptive calibration result on the basis of the performance degradation model (S6); and finally performing electric energy meter calibration management (S7). The present invention solves the technical problem in the prior art where the accuracy of electric energy metering is affected due to untimely calibration caused by an inability to monitor performance degradation of an electric energy meter in real time, thereby implementing effective calibration management for the electric energy meter, ensuring the long-term accuracy and reliability of the electric energy meter, and improving the efficiency of calibration work.

The application provides an intelligent operation and maintenance learning method and system related to monitoring a baseline, and relates to the technical field of computer operation and maintenance. The method comprises: collecting a plurality of core index monitoring data of a target server cluster in a set historical period to obtain a historical monitoring data set; performing time series segmentation processing on the historical monitoring data set to obtain a time series data segment set; performing trend fitting on the time series data segment set by a linear regression algorithm to obtain a corresponding trend fitting straight line; determining a center data point based on the trend fitting straight line, and determining a bidirectional data fluctuation interval according to a residual error analysis result at the center data point. The application realizes dynamic matching of the baseline with the business, full-process automation, precision improvement, and reduction of false and missed reports.

This invention belongs to the field of artificial intelligence technology and discloses a traffic flow prediction method, system, and device based on intra-block convolution and axis remapping. The invention divides the input traffic flow time series into multiple data blocks; uses a temporal convolutional network as the backbone to independently extract local temporal features within each data block; transforms the dimension representing the data block index into a channel dimension through axis remapping, thereby reorganizing the discrete block features into a structured multi-channel feature tensor; maps the tensor to the target dimension through a projection head network to obtain an encoded representation; and inputs the encoded representation into a prediction decoder to generate a prediction result for future traffic flow. This invention integrates the local temporal capture capability of TCN with the strong representation capability of modern sequence models, achieving both high computational efficiency and high accuracy.

The invention relates to a method for extracting the power consumption of an individual device within a set of devices (E1, E2, E3,..., En), the method comprising transforming a stream of measurements of an overall power consumption into a time series, and dividing (S1) the time series into a set of sequences, and then, for each sequence, the steps of: - generating (S4) a positional encoding matrix (MEP); - extracting (S5) features from the sequence to form a feature matrix (MF); - projecting (S6) statistical metrics into a statistics vector (MS); - concatenating (S7) these various elements (MEP, MF, MS) to form an input matrix (ME) provided to a transformation neural network designed to infer (S8) a time sequence corresponding to the individual device.

The invention discloses a zero sample fault detection method based on weighted semantic consistency embedding, and belongs to the field of industrial fault diagnosis. The method comprises the following steps that firstly, industrial equipment is used for collecting multi-mode sensor data, and standardized data are obtained through normalization, wavelet denoising and time sequence segmentation preprocessing; determining a fault attribute dimension based on domain expert knowledge, assigning each fault class to form a semantic attribute vector, and constructing a semantic attribute weight matrix through mutual information gain between attributes and fault tags; embedding data and semantic attributes into a shared space by using a data encoder and a semantic attribute encoder, decoding and calculating reconstruction loss by combining a data decoder and a semantic attribute decoder, and introducing weight to construct weighted semantic consistency loss; combining the two types of loss to train an encoder and a decoder, and extracting training data features by using the trained encoder and training a multi-attribute classifier; and finally, in a zero sample scene, extracting embedded features without fault data, inputting the embedded features into an attribute classifier, and matching fault categories through a maximum posterior probability. According to the method, the equal-weight hypothesis defect of an existing semantic consistency embedding method is broken through, key semantics are strengthened through mutual information gain weights, consistency loss is weighted to resist noise interference, and the precision, robustness and generalization ability of zero sample fault diagnosis are remarkably improved.

A method for extracting the power consumption of an individual piece of equipment within a set of equipment (E1, E2, E3, …, En), comprising transforming a stream of measurements of overall power consumption into a time series, and dividing this series (S1) into a set of sequences. Then, for each sequence, the following steps are performed: generation (S4) of a positional encoding matrix (PEM); extraction (S5) of features from the sequence to form a feature matrix (FM); projection (S6) of statistical metrics into a statistics vector (SM); and concatenation (S7) of these different elements (PEM, FM, SM) to form an input matrix (IM) provided to a transformation neural network designed to infer (S8) a time sequence corresponding to the individual piece of equipment. Figure for the abstract: Fig. 1

The application provides a kind of tunnel boring machine's measured data outlier detection and correction method, belongs to the field of outlier data detection, and the measured data of tunnel boring machine under different working conditions is detected by the method of sliding window, and the outlier point is corrected and filled up.The method first divides the original time series into multiple sub-time series by sliding window, and extracts the confidence interval radius of sub-time series slope by fast calculation and identifies abnormal sub-time series, then further determines the outliers using local outlier factor algorithm, and finally uses regression technique to reasonably fill the outliers removed.The application can effectively identify outliers in tunnel boring machine measured data, and reasonably correct and fill the outliers, ensuring the engineering usability of tunnel boring machine measured data and providing good conditions for further data analysis.

The application provides a time series segmented ship model resistance test data evaluation method, which is a method for evaluating single ship model resistance test data by using a segmented analysis method of time series segmentation considering the periodicity of the ship model in the running process. The method can be applied to single test of any ship type, eliminates the influence of artificial factors in the ship model resistance test, and makes the determination result of the single ship model resistance test more reliable.

The application provides a shield tail sealing system pressure abnormal state detection method, and belongs to the technical field of shield machines. Specifically, the following steps are included: S1, collecting pressure data recorded in a shield tail sealing system during shield machine construction; S2, preprocessing data recorded by a sensor; S3, using a time sequence segmentation method to segment and extract the pressure data and establishing a data set; S4, performing offline anomaly detection by using a density clustering DBSCAN method in unsupervised learning; S5, expanding abnormal sample points by using a Smote algorithm; and S6, training an online anomaly detection model by using the expanded data set by using a RUSBoost algorithm. The application provides a shield tail sealing pressure data processing method, so that shield tail sealing rear cavity pressure data recorded during construction can be fully utilized, so that whether the shield tail sealing rear cavity pressure state is abnormal can be quickly found during shield machine tunneling, and a certain guarantee is provided for the safe construction of the shield machine.

The invention relates to the related technical field of application performance monitoring, in particular to a low-Token consumption time sequence analysis method based on time sequence data visualization and a multi-modal large model, which comprises the following steps: step S1, cleaning, normalizing and feature extraction are performed on original time sequence data, and features comprise trend features, periodic features and abnormal features; s2, a customized time sequence trend chart is generated according to the features extracted in the step S1, the trend chart comprises a customized coordinate axis, a multivariable visual distinguishing identifier, a feature highlight identifier and a long time sequence segmentation identifier, and the method relates to a long time sequence data efficient analysis technology under the scenes of industrial equipment operation and maintenance, financial risk control, urban traffic scheduling and the like. By customizing the time sequence visual coding compression data Token amount and combining with the time sequence enhancement fine tuning optimization multi-modal model, low-cost and high-precision analysis of long time sequence data is realized, and technical support is provided for time sequence abnormity early warning, trend prediction and correlation diagnosis.

A kind of oil pipeline abnormal early warning monitoring method and system based on multi-source data fusion attention mechanism, oil pipeline abnormal early warning technology, to realize the early small leakage anomaly detection of oil pipeline by multi-source data fusion technology and deep learning technology.Technical points: collect pipeline operation state historical data from acoustic wave, temperature, negative pressure wave, vibration sensor, construct data set;Get time series segmentation code for algorithm training;Build fusion attention module for algorithm training;Based on the obtained multi-source data fusion attention Transformer model, adopt the training method of adversarial learning to predict the future operation state of oil pipeline;Use the Gaussian distribution with learnable scale parameter to train the oil pipeline early warning model;Use the trained multi-source data fusion attention Transformer model as the oil pipeline abnormal early warning model to identify the abnormal behavior of future operation state and obtain the failure probability.The present application effectively improves the accuracy and reliability of oil pipeline abnormal early warning by combining multi-source data analysis technology and deep learning technology, and using innovative Gaussian kernel scaling parameter and failure probability calculation method, which has important application value in small leakage detection.

Provided are an information processing method, a computer program, and an information processing device which can be expected to accurately detect similar objects from a plurality of non-time series image data items relating to substrate processing.　An information processing method according to the present embodiment involves an information processing device that comprises a time-series segmentation model that sequentially receives time-series image data items as input and sequentially outputs, for each image item, detection results for a region in which an object is captured. The information processing device acquires a plurality of non-time-series image items obtained by imaging an object relating to substrate processing. An acquired image data item is input to the time-series segmentation model to acquire detection results, and the acquired detection results are output.

The application discloses a cloud server performance degradation prediction method based on time sequence segmentation, first extracts performance resource time sequence data on a cloud server, decomposes the obtained time sequence data by adopting a DTW-BU time sequence segmentation algorithm, respectively constructs LSTM models for segmented subsequences, and predicts cloud server resource time sequence data, verifies model precision by using root mean square error and average absolute percentage error, predicts system performance degradation trend according to time sequence prediction data of the LSTM model, tests data fitting degree, and determines a software regeneration time node according to prediction data threshold value; the application can improve the precision of cloud server performance degradation prediction results, avoid overfitting phenomenon in the prediction process, and solves the problem of how to perform software regeneration at an optimal time point for cloud server performance degradation.

The invention discloses a new energy power generation power prediction method based on multiple meteorological conditions, and the method comprises the following steps: 1, selecting historical multi-source meteorological data and power data in the same period, and carrying out the format standardization and temporal-spatial resolution unification of the data; the method comprises the steps of 1, extracting basic features and cross features of multi-source meteorological data and carrying out feature screening and dimension reduction processing, 4, carrying out weighted fusion processing on the extracted meteorological features, 5, constructing a power generation power prediction model based on fused meteorological features and preprocessed power data, 6, training a data set based on time sequence segmentation and verifying the prediction model, and 7, carrying out prediction. 7, applying the prediction model, namely inputting fused meteorological characteristics in a prediction time range, and outputting generated power data in the prediction time range, and 8, performing closed-loop optimization management on the prediction model; the method has the advantages of being high in data reliability, high in prediction precision and high in project landing performance, and effective data support can be provided for planning of a power grid dispatching strategy.

The invention provides an oil leakage fault detection method based on data change of an oil sensor, and the method comprises the steps: collecting oil parameters and working condition factor data, building a normal fluctuation rule model through a time sequence segmentation and clustering algorithm, and generating a dynamic alarm boundary in combination with regression prediction; in the real-time monitoring process, real leakage events are accurately recognized by means of sliding window analysis and historical data fusion, meanwhile, model parameters are dynamically updated to optimize the detection precision, a reasonable alarm boundary can be automatically adjusted along with the real-time change of the equipment operation state, and the detection precision is improved.

The invention discloses a quasi-periodic time sequence segmentation method and system, a storage medium and equipment, and belongs to the technical field of time sequence processing, and the method comprises the following steps: S1, carrying out the self-feedback directional compression of an input original quasi-periodic time sequence, and outputting a key point index; s2, executing multi-view collaborative clustering based on the key point index and the original quasi-periodic time sequence, identifying periodic segmentation points in the original quasi-periodic time sequence, and outputting segmented quasi-periodic fragments according to the periodic segmentation points; and S3, evaluating the segmentation result in the step S2 by using the segmentation integrity factor, considering the period accuracy rate and the coverage rate, and iteratively optimizing the step S1 and / or the step S2 according to the evaluation result. The method is remarkably superior to the prior art in the aspects of cross-domain adaptability, large-scale data processing efficiency and segmentation precision, and is suitable for automatic segmentation of various QTS data such as electrocardiosignals, motion data and industrial vibration signals.

The invention discloses a time series data prediction method based on adaptive patch selection and dynamic recombination, which comprises the following steps of: segmenting a time series of input data into adjacent patches by adopting an adjacent patch division strategy; performing filling alignment on the time sequence according to a fixed step length and a patch length determined by an adjacent patch division strategy; extracting potential patches in the aligned time sequence under the conditions of patch length and step length 1; and scoring the potential patches by using a first scoring device based on the MLP to obtain n potential patches with the highest score, and taking the n potential patches as the selected potential patches. Generating a sequence score for the selected potential patches by using a second scoring device based on the MLP, and sorting to obtain an optimal arrangement sequence of the potential patches; respectively coding the optimal arrangement sequences of the adjacent patches and the potential patches, and carrying out weighted combination to obtain fusion embedded coding representation; and finally, predicting according to the fused embedded coding representation. According to the invention, the accuracy of time series data prediction is improved.

The application discloses a kind of data center refrigeration system prediction and shutdown temperature setting optimization method, comprising: obtaining the time series data when refrigeration system runs;Time series segmentation clustering model is used to identify and classify the time series data obtained in stage using manual labeling method, and sample data set is constructed;The start-stop process of refrigeration system under given heat load and the refrigeration power consumption of refrigeration system are predicted, and the refrigeration system prediction model is constructed, and the refrigeration system prediction model constructed is trained using the sample data set;The operation of refrigeration system is simulated using the trained refrigeration system prediction model, to determine the shutdown temperature setting value that makes the cumulative energy consumption lowest, to realize the optimization of shutdown temperature setting.The application can predict the start-stop process and power consumption of refrigeration system in a period of time in the future, and reduce the overall energy consumption of refrigeration system by optimizing the shutdown temperature configuration of refrigeration system.

This invention discloses a natural gas demand forecasting method based on GPT-driven pattern-aware segmentation and matching. The method comprises the following steps: S10, acquiring a natural gas demand time series and inputting it into a GPT-TempFeat time series feature learning model fine-tuned using low-rank adaptive LoRA to capture potential patterns in the natural gas demand time series while preserving the original knowledge of the natural gas demand time series; S20, using the time features extracted by GPT-TempFeat, dynamically segmenting the time series into different segments using a rolling window segmentation algorithm; S30, for each identified segment, using LoRA to fine-tune and train an individual GPT-Forecast model to obtain a segment model; S40, in the prediction stage, using a cosine similarity-based pattern matching method to identify the most suitable segment model for each input window; and then using the selected segment model for prediction to obtain the natural gas demand forecast result.

The invention discloses a noise tag-containing time sequence classification method and system based on multi-instance learning, electronic equipment and a storable medium, which are used for classifying noise tag-containing time sequences. The method comprises the following steps: firstly, preprocessing a time sequence data set to obtain a time sequence data set containing a noise label, then segmenting each time sequence in the data set into a plurality of instances, encoding the instances, then obtaining a prediction vector of each instance through a classifier, and obtaining a prediction vector of each instance; and a clean sample and a noise sample are selected by using a Gaussian mixture model. And for noise samples, calculating the label correlation score of each instance by using the prediction vector of the classifier, implementing two types of enhancement on the instance with the highest label correlation score in the samples, and inputting the enhanced noise samples into the classifier again to obtain a final classification result.

The invention relates to the technical field of safety control methods and systems, in particular to an electric energy meter automatic calibrating device fault rule base establishing method and system, and the system comprises a data fusion module, a time sequence segmentation module, a clustering analysis module, a probability correction module, a rule screening module and a man-machine interaction module. According to the invention, by introducing a multi-source data fusion and probability correction module, the coverage range of the rule base is obviously expanded, and the method can better adapt to a complex fault scene in an actual operation environment; through combined use of a distributed computing architecture and a task scheduling module, the time cost of large-scale data processing is greatly reduced, the fault diagnosis process is more efficient, an intelligent rule screening mechanism ensures dynamic updating and continuous optimization of a rule base, and rule redundancy and outdated problems are avoided.

The invention belongs to the technical field of artificial intelligence, and discloses a traffic flow prediction method, system and device based on intra-block convolution and axis remapping. The method comprises the following steps: segmenting an input traffic flow time sequence into a plurality of data blocks; a time convolutional network is adopted as a trunk, and local time sequence features in each data block are independently extracted; the dimension representing the data block index is converted into a channel dimension through an axis remapping operation, so that discrete block features are recombined into a structured multi-channel feature tensor; mapping the tensor to a target dimension through a projection head network to obtain a coded representation; and inputting the coded representation into a prediction decoder to generate a prediction result of the future traffic flow. According to the method, the local time sequence capture capability of the TCN and the strong representation capability of the modern sequence model can be fused, and the calculation is efficient and high in precision.

The application discloses a kind of time series data alignment method and system based on cross-modal semantic consistency related to cross-modal data processing, sequence modeling and artificial intelligence technical field.The method comprises obtaining target video data and corresponding target text label data;Based on pre-trained alignment model, the target video data is executed multi-scale time series segmentation to extract visual features, and the target text label data is executed standardization representation and global semantic coding to extract semantic features;By dynamic programming algorithm, the visual features and semantic features are adaptively aligned, and the optimal alignment path is obtained;According to the optimal alignment path, the semantic matching degree between video and text is output.The application solves the problems of long sequence modeling, cross-scale alignment and dynamic fusion through three-level technical architecture of visual feature pyramid modeling, text semantic unit extraction and cross-modal adaptive alignment and global semantic consistency constraint.

The invention discloses a flood risk identification method and system based on machine learning, and the method comprises the steps: obtaining a time series segmentation training frame from historical flood data through progressive learning, and inputting events of different years in batches in order to obtain a model parameter fine adjustment optimization result; a structured knowledge base is obtained from key elements stored in a risk precipitation mechanism, similar risk modes are classified through a clustering algorithm to obtain a risk feature vector library, and a graph neural network is adopted to perform drainage basin connectivity analysis. Using the river tributary flood storage and detention areas as nodes and water flow directions as edges to train and recognize a flood peak transfer time attenuation rule to obtain flood routing path prediction parameters; a clustering analysis method is adopted to identify risk value natural groups according to classification response anomalies, and the classification boundary discretization continuous risk probability is adjusted to obtain risk levels. Through combination of data driving and an intelligent algorithm, the flood prediction precision and response timeliness are remarkably improved, and a scientific basis is provided for flood control decision making.

This invention provides a method for identifying the year of open-pit coal mine disturbance by integrating multiple land cover features, belonging to the field of mine ecological monitoring technology. It involves acquiring long-term remote sensing images of open-pit coal mines, performing radiometric calibration and cloud removal preprocessing to establish a high-quality annual dataset; calculating the spectral indices of vegetation, bare soil, and bare coal; and constructing a novel open-pit coal mine disturbance index, MDI_NBC, by comprehensively considering the changes in the spectral indices of these three land cover features during open-pit coal mining; comparing the index changes in undisturbed areas to determine the MDI_NBC threshold for distinguishing mining disturbances; and applying a time-series segmentation algorithm with optimized parameters to conduct long-term monitoring of open-pit coal mine disturbances in different vegetation gradient zones, identifying the year of open-pit coal mine disturbances. This invention achieves universal monitoring of open-pit coal mine disturbances across vegetation gradient zones, effectively distinguishing vegetation disturbances caused by climate change, and improving the efficiency and accuracy of identifying the year of open-pit coal mine disturbances.

Aiming at simultaneous optimization of compression quality and approximate quality in time sequence segmentation, the invention provides a time sequence segmentation method based on hybrid multi-objective particle swarm optimization, which comprises the following steps of: constructing a time sequence segmentation model for simultaneously optimizing the compression quality and the approximate quality; the method comprises the following steps of: improving a Bare-bones particle swarm optimization algorithm; constructing a group of genetic operators for exploring the compression scale of the segmentation scheme, and optimizing the improved particle swarm optimization algorithm; and solving the time sequence segmentation model by using the optimized particle swarm optimization algorithm to obtain an optimized time sequence segmentation scheme set. According to the method, the time sequence with linear approximation characteristics can be segmented, the compression quality and the approximation quality are considered in a balanced manner, and the segmentation schemes under different compression degrees are output.

This invention discloses a video anomaly detection method based on a spatiotemporal integrated perception diffusion model, comprising the following steps: S1, training data acquisition and skeleton sequence construction; S2, time series segmentation and multi-granularity sampling; S3, bidirectional temporal diffusion modeling training; S4, after model training, inputting the video skeleton sequence to be detected into the trained diffusion model, and during the diffusion sampling process, generating a predicted skeleton sequence step by step based on the conditional sequence to obtain the prediction result for the corresponding time step; S5, calculating the anomaly score for each frame based on the deviation between the prediction result in S4 and the actual skeleton sequence. This invention improves the accuracy and robustness of anomaly detection by introducing a dual-granularity full-time-series perception mechanism and a global-local collaborative spatial perception mechanism into the diffusion modeling framework. This maintains the diffusion model's strong modeling ability for normal motion distribution while explicitly enhancing its perception ability for abnormal behavior in the temporal and spatial dimensions, thereby improving the accuracy and robustness of anomaly detection.

The present application provides a kind of real-time online early warning method of supersonic inlet non-start state based on deep learning, combined algorithm can receive sensor historical pressure signal, automatically segmented pressure signal by WCA-PCCD, constructs training data set, trains in single channel and multichannel WPT-CNN network, obtains the classifier for real-time online early warning, receives real-time dynamic test data by classifier, realizes inlet non-start early warning.The present application proposes two innovations for real-time online supersonic inlet non-start state early warning, respectively WCA-PCCD and MC-WPT-CNN, for different stages of pattern recognition.Compared with previous time series segmentation and pattern recognition, with the help of optimization algorithm and multidimensional deep learning model, early warning task is more simple and fast, with stronger accuracy, relatively simple implementation process, suitable for different types of inlet and various working conditions.Compared with traditional early warning method, the difficulty and complexity of early warning work are greatly reduced.

The present application provides a kind of fusion multi-feature open coal mine disturbance year identification method, belong to the technical field of mine ecology monitoring.The long time series remote sensing image of open coal mine is obtained, after radiation calibration, cloud removal preprocessing, annual high-quality data set is established;The spectral index of vegetation, bare soil and bare coal is calculated, the spectral index change of three ground objects in the process of open coal mining is integrated, and a new open coal mining disturbance index MDI_NBC is constructed;By comparing the index change of the undisturbed area, the MDI_NBC threshold value distinguishing coal mining disturbance is determined;The time series segmentation algorithm after parameter optimization is applied, and the long time series mining disturbance monitoring of open coal mine in different vegetation gradient area is carried out, and the open coal mining disturbance year information is identified.The present application realizes the universal monitoring of open coal mining disturbance across vegetation gradient area, can effectively distinguish the vegetation disturbance caused by climate change, and improves the identification efficiency and accuracy of open coal mining disturbance year.

The application provides a kind of wetland vegetation change detection method and device, the method comprises: obtaining the Landsat image and high time resolution MODIS image in the set time period of the area to be monitored, and pretreatment is carried out;Reconstruct long time series, high density Landsat surface reflectivity data set using space-time fusion algorithm;By analyzing the measured spectral characteristics of wetland vegetation, an optimized wetland vegetation change detection index based on spectral angle mapping principle is constructed;Through the LandTrendr time series segmentation algorithm of fusing spatial context features, the accurate and automatic detection of large-scale, long time series wetland vegetation change is realized.The method uses cloud computing capability to process massive remote sensing data, effectively solves the problem of data missing and insufficient algorithm stability in large-scale, long time series wetland vegetation monitoring, and provides efficient technical support for wetland resource management, carbon sink function evaluation and "carbon neutralization" strategy.