Method, device and medium for visual playback of time series data based on key point recognition
By analyzing and aggregating the device timing data, determining key points and performing dynamic video playback, the problem of visual playback of timing data in the prior art based on abnormal changes in the device is solved, and the efficiency and user experience of visual playback are improved.
Patent Information
- Application Number
- CN202411296987.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-18
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-18
AI Technical Summary
The existing timing data playback methods cannot perform timing data visual playback based on abnormal changes in the device, and there are limitations in data display and visualization, and the user experience and interactivity are insufficient.
By obtaining device timing data, data analysis and pre-configuration, trend detection, PELT variable point analysis, abnormal node monitoring and key segment aggregation, timing key points are determined, and dynamic video is visualized and played back according to these key points to match appropriate visualization Elements to achieve effective display of time series data.
It realizes visual playback of timing data based on timing key points, improves the efficiency and user experience of visual playback, and can dynamically display abnormal changes in the device, satisfy users' needs for intuitive data understanding.
Smart Images

Figure CN118820739B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of electronic digital processing technology, and in particular to a method, device and medium for visualizing and replaying time series data based on key point recognition. Background Art
[0002] The rapid development of IoT technology has led to the interconnection of a large number of devices, which has led to the generation of massive time series data. These data contain information that is crucial for equipment status monitoring, fault analysis, and historical trend analysis, and are particularly valuable in the fields of industrial automation and intelligent manufacturing. However, existing technologies and products have obvious defects in processing and replaying these data.
[0003] Traditional time series data playback methods mostly use linear playback of data points one by one. This method can still cope with small data volumes, but when faced with large data volumes, its efficiency and user experience are difficult to meet the needs of modern industry. In addition, these products have limitations in data presentation and visualization. Existing tools usually only display limited data when displaying time series data, lack animation effects, and lack flexibility in customizing the visualization interface, failing to meet users' needs for intuitive understanding of data, resulting in insufficient user experience and interactivity. In existing time series data playback methods, it is impossible to perform visualization playback of time series data based on abnormal changes in equipment. Summary of the invention
[0004] The embodiments of the present application provide a method, device and medium for visual playback of time series data based on key point recognition, which solves the technical problem that the existing time series data playback method cannot perform visual playback of time series data based on abnormal changes in the device.
[0005] In a first aspect, an embodiment of the present application provides a method for visualizing and replaying time series data based on key point recognition, the method comprising: acquiring device time series data, and pre-configuring data analysis on the device time series data to obtain device time series parameter data to be analyzed; performing trend detection on the device time series parameter data to be analyzed to determine the device time series trend data; determining the change point analysis data based on the device time series parameter data to be analyzed, and performing PELT change point analysis on the change point analysis data to obtain the time series change point position; performing abnormal node monitoring on the device time series parameter data to be analyzed to obtain the device time series abnormal time node; determining the device time series analysis output data through key segment aggregation based on the device time series trend data, the time series change point position and the device time series abnormal time node; matching the device time series analysis output data with visualization elements to determine the device time series visualization type, and obtaining the time series data visualization playback data through visualization playback according to the device time series visualization type.
[0006] In one implementation of the present application, data analysis pre-configuration is performed on device timing data to obtain device timing parameter data to be analyzed, specifically including: data cleaning of the device timing data to obtain device timing data to be configured; configuring the analysis time range of the device timing data to be configured to determine the device timing analysis time interval; based on the device timing analysis time interval, determining the number of device timing analysis data points through a first analysis deviation evaluation; based on the device timing analysis time interval and the number of device timing analysis data points, obtaining the device timing parameter data to be analyzed through a second analysis deviation evaluation.
[0007] In one implementation of the present application, a trend detection is performed on the timing parameter data of the device to be analyzed to determine the device timing trend data, specifically including: determining the maximum number of line segments for trend detection based on the timing parameter data of the device to be analyzed; determining the time point of the device timing trend change according to the maximum number of line segments by iterating the area enclosed between the line segments; and performing a trend slope analysis on the time point of the device timing trend change to determine the device timing trend data.
[0008] In one implementation of the present application, a PELT change point analysis is performed on the change point analysis data to obtain the timing change point position, specifically including: determining the PELT penalty parameter based on the change point analysis data; wherein the change point analysis data includes: a background data group, a cost function; performing an initial minimum cost configuration on the PELT penalty parameter, and determining the PELT timing change point set through minimum cost iteration based on the minimum cost configuration; and screening the PELT timing change point set for the optimal solution through preset pruning conditions to obtain the timing change point position.
[0009] In one implementation of the present application, abnormal node monitoring is performed on the timing parameter data of the device to be analyzed to obtain the device timing abnormal time node, specifically including: performing abnormal threshold detection on the timing parameter data of the device to be analyzed to obtain the device timing abnormal data; according to the device timing abnormal data, the device timing abnormal time node is determined through a preset alarm extraction mechanism; wherein the device timing abnormal time node includes: timing abnormal value point, alarm event occurrence time point, and alarm event end time point.
[0010] In one implementation of the present application, before performing abnormal node monitoring on the device timing parameter data to be analyzed to obtain the device timing abnormal time node, the method also includes: inputting the device timing parameter data to be analyzed into a preset Pearson correlation coefficient calculation formula to obtain device attribute correlation values; according to a preset device attribute correlation threshold, performing correlation division on the device attribute correlation values to obtain device attribute correlation data, and determining the device attribute correlation based on the device attribute correlation data.
[0011] In one implementation of the present application, based on the equipment timing trend data, the timing change point position and the equipment timing abnormal time node, the equipment timing analysis output data is determined by key segment aggregation, specifically including: performing fast Fourier transform on the equipment timing parameter data to be analyzed, and performing amplitude critical sampling on the equipment timing parameter data to be analyzed after the fast Fourier transform to obtain the equipment timing periodic parameters; performing statistical feature analysis on the equipment timing periodic parameters to determine the equipment timing maximum point; determining the timing key points based on the equipment timing trend data, the timing change point position, the equipment timing abnormal time node and the equipment timing maximum point; wherein the timing key points include: trend turning points, change points, maximum value points, minimum value points, timing abnormal value points, alarm start and end points; according to the timing key points, determining the timing key segments by segmenting the timing data; according to the equipment timing periodic parameters, performing key segment aggregation on the timing key segments to determine the equipment timing analysis output data.
[0012] In one implementation of the present application, the device timing analysis output data is matched with visualization elements to determine the device timing visualization type, specifically including: based on the device timing analysis output data, converting it through a visualization engine to obtain a device timing status data group; matching the backup timing status data group with visualization elements to determine the device timing visualization type; wherein the visualization elements include: line charts, heat maps, pie charts, bar charts, tables, and box plots.
[0013] In a second aspect, an embodiment of the present application also provides a time series data visualization playback device based on key point recognition, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can: obtain device time series data, and pre-configure data analysis on the device time series data to obtain device time series parameter data to be analyzed; perform trend detection on the device time series parameter data to be analyzed to determine the device time series trend data; determine the change point analysis data to be analyzed based on the device time series parameter data to be analyzed, and perform PELT change point analysis on the change point analysis data to obtain the time series change point position; perform abnormal node monitoring on the device time series parameter data to be analyzed to obtain the device time series abnormal time node; determine the device time series analysis output data through key segment aggregation based on the device time series trend data, the time series change point position and the device time series abnormal time node; match the device time series analysis output data with visualization elements to determine the device time series visualization type, and obtain the time series data visualization playback data through visualization playback according to the device time series visualization type.
[0014] In a third aspect, an embodiment of the present application further provides a non-volatile computer storage medium for visual playback of time series data based on key point identification, storing computer executable instructions, wherein the computer executable instructions are configured to: obtain device time series data, and pre-configure data analysis on the device time series data to obtain device time series parameter data to be analyzed; perform trend detection on the device time series parameter data to be analyzed to determine the device time series trend data; determine the change point analysis data based on the device time series parameter data to be analyzed, and perform PELT change point analysis on the change point analysis data to obtain the time series change point position; perform abnormal node monitoring on the device time series parameter data to be analyzed to obtain the device time series abnormal time node; determine the device time series analysis output data through key segment aggregation based on the device time series trend data, the time series change point position and the device time series abnormal time node; match the device time series output analysis data with visualization elements to determine the device time series visualization type, and obtain the time series data visualization playback data through visualization playback according to the device time series visualization type.
[0015] The embodiments of the present application provide a method, device and medium for visual playback of time series data based on key point recognition. By performing comprehensive trend analysis on device time series trends, time series change point locations, device attribute correlations, outlier detection, alarm event recognition, etc., the technical problem of the existing time series data playback method that it is impossible to perform visual playback of time series data based on abnormal changes in the device is solved. The visual playback of time series data based on time series key points is realized, and playback is performed in a limited time in the form of dynamic video, and the efficiency of visual playback is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0017] Figure 1 A flow chart of a method for visualizing and replaying time series data based on key point recognition provided in an embodiment of the present application;
[0018] Figure 2 A schematic diagram of the internal structure of a time series data visualization playback device based on key point recognition provided in an embodiment of the present application. DETAILED DESCRIPTION
[0019] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.
[0020] The embodiments of the present application provide a method, device and medium for visual playback of time series data based on key point recognition. By performing comprehensive trend analysis on device time series trends, time series change point locations, device attribute correlations, outlier detection, alarm event recognition, etc., the technical problem of the existing time series data playback method that it is impossible to perform visual playback of time series data based on abnormal changes in the device is solved. The visual playback of time series data based on time series key points is realized, and playback is performed in a limited time in the form of dynamic video, and the efficiency of visual playback is improved.
[0021] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.
[0022] Figure 1 A flowchart of a method for visualizing and replaying time series data based on key point recognition provided by an embodiment of the present application. Figure 1 As shown, the embodiment of the present application provides a method for visualizing and replaying time series data based on key point recognition, which specifically includes the following steps:
[0023] Step 101: Acquire device timing data, and pre-configure data analysis on the device timing data to obtain device timing parameter data to be analyzed.
[0024] Specifically, the method includes: cleaning the device timing data to obtain the timing data of the device to be configured; configuring the analysis time range of the timing data of the device to be configured to determine the device timing analysis time interval; determining the number of device timing analysis data points through a first analysis deviation evaluation based on the device timing analysis time interval; and obtaining the timing parameter data of the device to be analyzed through a second analysis deviation evaluation based on the device timing analysis time interval and the number of device timing analysis data points.
[0025] The present application obtains device timing data and pre-configures data analysis on the device timing data to obtain device timing parameter data to be analyzed, thereby achieving accuracy and practicality of device timing data analysis and improving data positioning efficiency.
[0026] In the examples of the present application, a detailed explanation is given by the following Example 1.
[0027] Example 1: Complete device access and data collection by configuring the IoT data collection module. It is necessary to configure products, devices, communication topics, and data flow solutions in the IoT platform console. The device establishes a connection with the platform through multiple protocols supported by the IoT platform (such as MQTT, CoAP, HTTPS, etc.), and sends data to the IoT platform through the corresponding protocol. The device is the device object under study, representing industrial machines and equipment connected to the IoT platform, such as CNC machine tools, robots, conveyors, pumps, compressors, electric meters, water meters, gas meters, etc. The device attributes are the types of time series data obtained by sensors, etc., reflecting the operating status of the device, including temperature, humidity, pressure, flow, vibration, speed, acceleration and other attributes. A device can contain multiple device attributes, and the collected data is the device attribute data of the IoT device.
[0028] The received raw data is preprocessed through the IoT data processing module. Data preprocessing includes data cleaning, data formatting, and data standardization, etc., establishing the critical value for outlier detection, and configuring the alarm mechanism according to the user-customized standard.
[0029] Data storage is to store the processed data in the IoT data storage module in chronological order, and store data from different data sources separately according to the "device-device attribute" index format to facilitate subsequent query and analysis. The stored data includes the simulated values of each device attribute of the IoT device, the threshold value for abnormal value detection, and the alarm mechanism for abnormal conditions.
[0030] First, according to the preset index in the data storage module, a joint list of devices and device attributes is formed, so that users can select the devices and device attribute objects to be analyzed by checking. After the check is completed, a new data playback analysis object list will be formed.
[0031] Then, set the time range for analysis, that is, the time interval for equipment timing analysis. The system will extract the start and end time points of the corresponding data in the data storage module based on the list of data playback analysis objects, thereby determining the time range of available timing data. Users can further select the time range for analysis within the optional time range. For example, the time range of available timing data for "Wind turbine - power" is from 0 to 24 o'clock, and users can select 12 to 18 o'clock as the time range to be analyzed according to their needs.
[0032] In order to ensure the reliability and effectiveness of the analysis results, the number of data points in the selected time range will be verified. In this application, the number of data points for equipment timing analysis is set to 1000 to avoid serious deviations in the analysis results. If the number of data points is insufficient, the system will prompt the user to reselect the time range. For example, continuing with the previous example, it is identified that the time range for analysis of "wind turbine-power" is from 12 o'clock to 18 o'clock, and there are only 800 data points, which is less than 1000, so the user needs to reselect the time range to continue with the subsequent steps.
[0033] According to the index of "device + device attribute + analysis time range", go to the data storage module to check whether the playback video, intelligent report and analysis results already exist. If yes, it means that the user has completed the analysis of the same content once and does not need to analyze it again. The system will tell the user that "the data playback analysis results already exist", directly call the existing analysis results, playback video and intelligent report, and enter the data visualization playback stage in step 106 for playback; otherwise, continue with the following steps.
[0034] Finally, based on the data playback analysis object list and analysis time range selected by the user, the "device-device attribute" index intelligently identifies and accurately extracts data from the data storage module.
[0035] Step 102: Perform trend detection on the timing parameter data of the device to be analyzed to determine the timing trend data of the device.
[0036] Specifically, it includes: determining the maximum number of line segments for trend detection based on the timing parameter data of the device to be analyzed; determining the time point when the device timing trend changes based on the maximum number of line segments by iterating the area enclosed between the line segments; and performing trend slope analysis on the time point when the device timing trend changes to determine the device timing trend data.
[0037] The present application determines the device timing trend data by performing trend detection on the device timing parameter data to be analyzed, thereby realizing the change trend analysis of the device timing parameter data and improving the accuracy of the change trend analysis of the device timing parameter data.
[0038] In the examples of the present application, detailed explanation is given by the following Example 2.
[0039] Example 2: Since this application needs to select valuable information from a large amount of time series data, play it back in a limited time in the form of dynamic video, and provide users with a medium to read data stories in the form of intelligent reports, it is necessary to accurately identify key data points and aggregate these key information to ensure efficient playback of all key information in a limited time.
[0040] The first part is to analyze the changing trend of the equipment timing parameter data, with the aim of determining the trend of the equipment timing parameter data in each time period and finding the trend turning point in the timing data.
[0041] The device time series is divided into multiple segments through the PLA (Piecewise Linear Approximation) method, and each segment is approximated by a straight line, so that division and approximation are performed under the criteria of optimizing storage and compressing data quality.
[0042] First, we need to set the maximum number of line segments, then start initialization, connect each pair of adjacent points into a straight line, calculate the cost (the area enclosed by the fitting line segment and the original two lines) of merging adjacent line segments (connecting the first and last points into a line), and select the method with the lowest cost to merge. Then, we iterate until the number of line segments is reduced to less than the set maximum number of line segments and then stop iterating. By detecting the slope of each segment after the linear approximation of the original data and the time points of the first and last points, we can determine whether the trend in each time period is rising, falling, or stable.
[0043] Taking the equipment water flow data of Plant A as an example, the fluctuation of water flow closely corresponds to the fluctuation of water demand corresponding to the operation of the equipment. Especially during the peak water usage period of the equipment, the increase in flow is particularly significant, showing obvious periodic characteristics.
[0044] Therefore, when conducting trend analysis, we selected traffic data within 24 hours as samples and set the maximum number of segments to 24 to match the time period of one day. The first step of the analysis is to perform a time series analysis on the traffic data, connecting consecutive time points to form multiple initial segments. The above segments are merged and gradually optimized through an iterative process until the total number of segments is reduced to no more than the preset maximum value, that is, 24 segments, for the first time.
[0045] When this condition is reached, the algorithm stops iterating, and the retained line segments and their corresponding time spans constitute the fitting curve and trend time period of the water flow trend. In each trend time period, by analyzing the slope of the line segment, the specific trend of the flow change can be accurately determined, whether it is rising, falling or remaining stable.
[0046] It should be noted that trend detection of other parameters of the device can also be performed using the above method.
[0047] Step 103: Based on the timing parameter data of the device to be analyzed, determine the change point analysis data to be analyzed, and perform PELT change point analysis on the change point analysis data to obtain the timing change point position.
[0048] Specifically, it includes: determining PELT penalty parameters based on the change point analysis data; wherein, the change point analysis data includes: background data group, cost function; performing initial minimum cost configuration on the PELT penalty parameters, and determining the PELT time series change point set through minimum cost iteration based on the minimum cost configuration; screening the PELT time series change point set for the optimal solution through preset pruning conditions to obtain the PELT time series change point position.
[0049] The present application determines the change point analysis data to be analyzed based on the timing parameter data of the device to be analyzed, and performs PELT change point analysis on the change point analysis data to obtain the PELT timing change point position, thereby realizing the change point position determination of the device timing parameter data and improving the accuracy and efficiency of the change point position determination.
[0050] In the examples of the present application, this is explained in detail through the following Example 3.
[0051] Example 3: A change point is a point in a time series where the statistical characteristics of the data change significantly.
[0052] The present application adopts a change point detection method based on PELT (Pruned Exact Linear Time) using a linear kernel; wherein the linear kernel function is usually defined as the sum of the absolute values of the difference between two data points.
[0053] K represents the value of the kernel function. In the case of a linear kernel, the kernel function is defined as:
[0054] First, define the change point analysis data. Read the background data set to be analyzed ,
[0055] Then a penalty function is constructed, which is used to measure the cost of introducing a new change point. It usually consists of data fitting cost and change point introduction cost.
[0056] set up Indicates at a point in time The total cost of Indicates the time interval The data fitting cost within represents the cost of introducing the change point, represents the number of change points introduced before the time point. .
[0057] For a linear kernel, the data fitting cost can be calculated using the least squares method. Assume that in the time interval The data can be modeled using a linear Fitting, where is the observed value. By using the least squares method, we can get the parameter and The estimated value of is the residual sum of squares: .
[0058] Then, we use the idea of dynamic programming to find the optimal combination of change points. Define an array , represents the minimum total cost at a point in time. Initially, For each time point ,calculate Values: , that is, traverse all possible change point positions and select the position that minimizes the total cost.
[0059] According to the current optimal cost and penalty function, some parts that cannot be the optimal solution are excluded. For example, if at a certain time point t, the current minimum total cost is greater than a preset threshold, then the search after that time point can be stopped.
[0060] Finally, determine the change point, after traversing the entire time series, This is the final minimum total cost. Starting from time point T, we step back through the array , determine the locations of all change points.
[0061] For each time point t, find The s that holds is a change point. Repeat this process until you go back to time point 1. Finally, apply pruning conditions to remove those change points that cannot be the optimal solution, reduce unnecessary calculations, and finally determine the change point position that minimizes the cost function.
[0062] Taking the change point analysis of the vibration properties of the mining ventilator equipment in Mine B as an example, the cost function C based on the linear kernel is set initially to a large value, usually positive infinity.
[0063] The first data point is used as the initial split point. For each time point, the cost of using it as a change point is calculated. The cost can be calculated by using a linear kernel function to calculate the degree of data change between the current split point and the previous split point.
[0064] By updating the minimum cost and the corresponding segmentation point, after traversing all data points, the final change point segmentation scheme is determined according to the minimum cost. The change point divides the time series data of the vibration properties of the mining ventilator equipment into different intervals, and the data in each interval has relatively stable characteristics.
[0065] Step 104: Monitor abnormal nodes on the timing parameter data of the device to be analyzed to obtain abnormal time nodes of the device timing.
[0066] Specifically, it includes: performing abnormal threshold detection on the timing parameter data of the device to be analyzed to obtain the device timing abnormality data; according to the device timing abnormality data, determining the device timing abnormality time node through a preset alarm extraction mechanism; wherein the device timing abnormality time node includes: timing abnormality value point, alarm event occurrence time point, and alarm event end time point.
[0067] Before monitoring the abnormal nodes of the device timing parameter data to be analyzed to obtain the abnormal time nodes of the device timing, the method also includes: inputting the device timing parameter data to be analyzed into a preset Pearson correlation coefficient calculation formula to obtain device attribute correlation values; dividing the device attribute correlation values according to the preset device attribute correlation threshold to obtain device attribute correlation data, and determining the device attribute correlation based on the device attribute correlation data.
[0068] This application implements abnormal node analysis of device timing parameters by monitoring abnormal nodes of device timing parameter data to be analyzed to obtain abnormal time nodes of device timing, and improves the timeliness of device abnormal status alarm according to a preset alarm extraction mechanism, thus providing a data basis for the generation of device timing analysis output data.
[0069] In the examples of the present application, a detailed explanation is given by the following Example 4.
[0070] Example 4: For predicting equipment behavior and optimizing equipment operation and maintenance, it is necessary to analyze the correlation between different equipment attributes to reveal the potential connections between them.
[0071] Pearson correlation coefficient To perform paired correlation analysis on the device attribute data of the same IoT device, the formula is expressed by the following formula:
[0072] (1)
[0073] in, and The observed values of the two device attribute variables of the device attribute data set;
[0074] and It is the average value of two device attribute variables in the device attribute data set.
[0075] Perform abnormal threshold detection on the timing parameter data of the device to be analyzed. Values between 0.8 and 1.0 indicate a very strong correlation, 0.6 to 0.8 indicate a strong correlation, 0.4 to 0.6 indicate a moderate correlation, 0.2 to 0.4 indicate a weak correlation, and 0.0 to 0.2 indicate a very weak correlation or no correlation.
[0076] Taking the mining ventilation fan equipment of Mine B as an example, a variety of sensors are installed in key parts of the equipment such as the rotating shaft and the fan outlet, including temperature sensors, vibration sensors and speed sensors.
[0077] A total of 24 characteristic sensors and calculated values are collected, including 5 temperature values, 3 voltage values, 2 current values, 2 pressure values, etc.
[0078] When conducting a correlation analysis on the vibration attributes of the equipment, the Person correlation coefficients between each attribute are first calculated. Then, based on the results of the correlation analysis, the attributes with an absolute value greater than or equal to 0.6 are selected as relevant features, thereby representing the correlation index of the mining ventilator equipment.
[0079] The outlier detection threshold is used to detect outliers in historical data and predicted data within the selected time range, and the occurrence and end time points of the alarm events are extracted based on the user-configured alarm mechanism.
[0080] It should be noted that the time series anomaly point is not the same as the alarm event occurrence time point and the alarm event end time point. In the process of determining the start and end points of the alarm time, the alarm point needs to be set according to actual needs. Generally, the start and end points of the alarm need to be earlier or later than the time series anomaly point.
[0081] The above method can quantify and accurately describe the relationship between equipment attribute data, providing a basis for in-depth analysis and decision-making.
[0082] Step 105: Based on the equipment timing trend data, the timing change point position and the equipment timing abnormal time node, the equipment timing analysis output data is determined through key segment aggregation.
[0083] Specifically, it includes: performing fast Fourier transform on the timing parameter data of the equipment to be analyzed, and performing amplitude critical sampling on the timing parameter data of the equipment to be analyzed after the fast Fourier transform to obtain the equipment timing periodic parameters; performing statistical characteristic analysis on the equipment timing periodic parameters to determine the equipment timing maximum point; determining the timing key points based on the equipment timing trend data, the timing change point position, the equipment timing abnormal time node and the equipment timing maximum point; wherein, the timing key points include: trend turning point, change point, maximum value point, minimum value point, timing abnormal value point, alarm start and end points; according to the timing key points, determine the timing key segments by segmenting the timing data; according to the equipment timing periodic parameters, perform key segment aggregation on the timing key segments to determine the equipment timing analysis output data.
[0084] This application determines the equipment timing analysis output data through key segment aggregation based on equipment timing trend data, timing change point locations and equipment timing abnormal time nodes, thereby realizing key point integration of timing data based on timing key points and improving the efficiency of extracting and integrating important parameters of timing data.
[0085] In the examples of the present application, this is explained in detail through the following Example 5.
[0086] Example 5: First, in order to determine whether there is a periodic change pattern in the device timing data, this application uses fast Fourier transform (FFT) and selects the frequencies with the highest amplitude values to determine the period of the timing data.
[0087] For a given time series data ;in , FFT calculates the corresponding frequency domain representation ,in .
[0088] in, Indicates the frequency The complex amplitude at is the sampling frequency.
[0089] Typically, a signal with significant periodicity will show larger amplitudes at the corresponding frequency points. By identifying the frequency with the largest amplitude, we can infer the main periodic components in the data. This is because periodic signals have higher energy than noise and random fluctuations, which is reflected as larger amplitudes in the frequency domain.
[0090] After FFT analysis, we can select the first three frequencies with the highest amplitude values, and record them as, and from large to small. The frequency and period are inversely proportional to each other, that is, the period. We can determine that there are three significant periods in the time series data, corresponding to the above frequencies, from small to large. , and .
[0091] For example, for a temperature time series data collected every minute for a month, FFT analysis may find significant amplitude peaks at the frequency points corresponding to one hour, one day, and one week. By selecting these frequencies with the top amplitudes, we can determine that the main periodic components in the data are one hour, one day, and one week, respectively.
[0092] After determining the data period transformation law in the backup time series data, the statistical feature analysis is performed by numerical analysis. Among them, the key statistical indicators include: difference group, interquartile range, maximum value, minimum value, and average value. The difference group includes: variance, standard deviation, and range.
[0093] Stability refers to the degree to which the distribution pattern of a time series remains consistent over time, and the coefficient of variation is used as a key indicator to judge the stability of time series data.
[0094] The coefficient of variation is the percentage of the standard deviation to the mean. By directly observing the value of the coefficient of variation, we can make an intuitive judgment on the stability of the data and characterize it as "very stable (coefficient of variation greater than 0.9)", "stable (coefficient of variation greater than 0.6)" or "unstable (coefficient of variation less than 0.6)".
[0095] Finally, the trend turning points, change points, maximum points, minimum points, timing abnormal value points, alarm event occurrence time points, and alarm event end time points obtained from the equipment timing trend data, PELT timing change point positions, and timing abnormal time nodes are set as key points for analysis.
[0096] Starting from the starting point of the time series data, according to the smallest time period found in the periodicity analysis , divide the entire time series data into In each segment, the time point in the middle is taken as the aggregation point of all data points in the time series segment, then the set of aggregation points is , .
[0097] The above method of segmenting the time series according to the period and these aggregation points will be used in the subsequent process of generating dynamic videos. When the minimum period is 1 hour and the total time of the time series is 30 days, the time series data will be divided into 720 segments according to the time span of 1 hour, and 720 aggregation points will be generated.
[0098] Output the results of data analysis to the data playback engine and generative AI, including the duration of the rising, falling and stable trends of the time series data within the selected time range and their proportion in the total duration, trend turning points, change points, time series abnormal value points, alarm event occurrence time points, alarm event end time points, correlation coefficient values and correlation strength between various device attributes under the same device, time series data periodicity, time series data aggregation point location, statistical characteristic values and stability test results. At the same time, these results are saved in the data storage module according to the index of "device + device attribute + analysis time range" for subsequent retrieval.
[0099] Step 106: Match the device timing analysis output data with visualization elements to determine the device timing visualization type, and obtain the timing data visualization playback data through visualization playback according to the device timing visualization type.
[0100] Specifically, it includes: based on the output data of the equipment timing analysis, it is converted through the visualization engine to obtain the equipment timing status data group; the equipment timing status data group is matched with the visualization elements to determine the equipment timing visualization type; wherein the visualization elements include: line chart, heat map, pie chart, bar chart, table, box plot.
[0101] The present application determines the device timing visualization type by matching the device timing stability data with visualization elements, and obtains the timing data visualization playback data through visualization playback according to the device timing visualization type, thereby achieving the matching of visualization elements for different types of device timing data, improving the revisit effect of the device timing visualization playback data and the user experience.
[0102] In the examples of the present application, this is explained in detail through the following Example 6.
[0103] Example 6: Existing text-to-video generation technologies are usually limited to being able to generate only up to 1 minute of coherent video content, which may not fully meet users' demands for video length. This problem limits the original patent's method of combining generative AI and high-precision prompt words to generate playback videos.
[0104] Therefore, according to the requirement of displaying all analyzed key points and the overall characteristics of time series data within the specified time, manual setting of dynamic video playback scripts and visualization elements is adopted.
[0105] The analysis results are converted into intuitive visual content through the data playback engine, using static charts to comprehensively display the macro characteristics of the analyzed data and dynamic videos to display information at key points in the data evolution process.
[0106] According to the preset playback script, after the user clicks the "Start" button, the initial static chart is displayed for 15 seconds, and then smoothly transitions to the dynamic video. In the dynamic video, we adopted different playback time strategies based on the periodic characteristics of the analyzed time series data. When the total playback time t of the dynamic video is fixed, the time spent at each aggregation point is For example, if the user requires t=600S and after aggregation processing D=30, the dynamic video will proceed smoothly at a speed of 20s at each aggregation point.
[0107] Static charts include six visualization elements: line chart, heat map, pie chart, bar chart, table and box chart. Line chart is used to show the overall appearance of time series data within the selected time range. Users can focus on the time range of their interest by dragging the time axis below the line chart, so as to observe and analyze the data of a specific period in depth; heat map is used to show the size of the correlation between the attributes of each device under the current device; pie chart is used to show the proportion of the three trends of rising, falling and stable in the total time; bar chart is used to show the number of trend turning points, the number of change points, the number of alarm events, the number of abnormal values, and the number of maximum points respectively; table is used to show the calculation results of statistical data and the results of periodic analysis; box chart is used to describe the statistical characteristics within the selected time range.
[0108] In addition to static graphs, dynamic videos include two visualization elements: dynamic line graphs and scrolling tables. The dynamic line graph shows data fluctuations over time. The dynamic line graph shows the situation in each period segment in turn. The aggregation point in each period segment is always located in the center of the dynamic line graph window. The left and right sides are respectively the leftmost and rightmost sides of the period in which the aggregation point is located. The key points in the period segment are distinguished by different color labels in the line graph. There is a timeline below the line graph to explain to users the position of the currently displayed content in the entire time series data. Users can simply drag the timeline below the line graph to quickly locate the specific time point they are concerned about and realize instant viewing of the data status at that moment. The table details the type and time of occurrence of the key points in the current period segment, and scrolls and refreshes in real time as the dynamic video plays.
[0109] The generated playback video is saved in the data storage module according to the index of "device + device attribute + analysis time range" to facilitate subsequent retrieval.
[0110] Although the method of combining generative AI with prompt words mentioned in the original patent is difficult to implement in dynamic video generation, this method shows advantages in generating text reports, especially for the part of generating intelligent reports; first, the analysis results are output from the data storage module in Excel format and uploaded to the generative AI.
[0111] Then, AI is used to generate the basic content of the report, including four parts: summary, method, results, and conclusion. The summary is used to outline the analysis objectives. The method is used to describe the method and process of data analysis. The results are used to list in detail the results obtained by the data analysis method of this patent, such as trend duration, proportion, turning point, change point, outliers, alarm events, correlation analysis, periodicity, aggregation point, statistical characteristics, and stability assessment. In the conclusion, generative AI is used to summarize the key findings, propose possible causes and impacts, and make recommendations or next steps.
[0112] Finally, the final intelligent report is output to the required format, such as PDF, Word document, etc. The generated intelligent report is saved in the data storage module according to the index of "device + device attribute + analysis time range" for subsequent retrieval.
[0113] The above is an embodiment of the method proposed in this application. Based on the same inventive concept, the embodiment of this application also provides a time series data visualization playback device based on key point recognition, and its structure is as follows: Figure 2 shown.
[0114] Figure 2 A schematic diagram of the internal structure of a time series data visualization playback device based on key point recognition provided by an embodiment of the present application. Figure 2 As shown, the device includes:
[0115] at least one processor 201;
[0116] and, a memory 202 communicatively connected to the at least one processor;
[0117] The memory 202 stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor 201 to enable at least one processor 201 to:
[0118] Acquire device timing data, and perform data analysis pre-configuration on the device timing data to obtain the device timing parameter data to be analyzed; perform trend detection on the device timing parameter data to be analyzed to determine the device timing trend data; determine the change point analysis data based on the device timing parameter data to be analyzed, and perform PELT change point analysis on the change point analysis data to obtain the timing change point position; perform abnormal node monitoring on the device timing parameter data to be analyzed to obtain the device timing abnormal time node; determine the device timing analysis output data through key segment aggregation based on the device timing trend data, the timing change point position and the device timing abnormal time node; match the device timing output analysis data with visualization elements to determine the device timing visualization type, and obtain the timing data visualization playback data through visualization playback according to the device timing visualization type.
[0119] Some embodiments of the present application provide corresponding Figure 1 A non-volatile computer storage medium for visual playback of time series data based on key point recognition, storing computer executable instructions, wherein the computer executable instructions are set as follows:
[0120] Acquire device timing data, and perform data analysis pre-configuration on the device timing data to obtain the device timing parameter data to be analyzed; perform trend detection on the device timing parameter data to be analyzed to determine the device timing trend data; determine the change point analysis data based on the device timing parameter data to be analyzed, and perform PELT change point analysis on the change point analysis data to obtain the timing change point position; perform abnormal node monitoring on the device timing parameter data to be analyzed to obtain the device timing abnormal time node; determine the device timing analysis output data through key segment aggregation based on the device timing trend data, the timing change point position and the device timing abnormal time node; match the device timing output analysis data with visualization elements to determine the device timing visualization type, and obtain the timing data visualization playback data through visualization playback according to the device timing visualization type.
[0121] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the IoT device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.
[0122] The system and medium provided in the embodiments of the present application correspond one-to-one to the method. Therefore, the system and medium also have similar beneficial technical effects to the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the system and medium will not be repeated here.
[0123] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0124] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0125] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0127] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0128] Memory may include non-permanent storage in a computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.
[0129] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0130] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0131] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A method for visualizing and replaying time series data based on key point recognition, characterized in that: The method comprises: Acquire device timing data, and perform data analysis preconfiguration on the device timing data to obtain device timing parameter data to be analyzed; Performing trend detection on the timing parameter data of the device to be analyzed to determine the timing trend data of the device; Based on the timing parameter data of the device to be analyzed, determining the change point analysis data to be analyzed, and performing PELT change point analysis on the change point analysis data to obtain the timing change point position; Perform abnormal node monitoring on the timing parameter data of the device to be analyzed to obtain the abnormal time node of the device timing; Based on the equipment timing trend data, the timing change point position and the equipment timing abnormality time node, the equipment timing analysis output data is determined through key segment aggregation; The device timing analysis output data is matched with visualization elements to determine the device timing visualization type, and according to the device timing visualization type, the timing data visualization playback data is obtained through visualization playback.
2. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: The device timing data is pre-configured for data analysis to obtain the device timing parameter data to be analyzed, specifically including: Cleaning the device timing data to obtain the timing data of the device to be configured; Configuring the time range for analyzing the timing data of the device to be configured to determine the time interval for device timing analysis; Based on the device timing analysis time interval, determining the number of device timing analysis data points through a first analysis deviation evaluation; According to the device timing analysis time interval and the device timing analysis data points, the device timing parameter data to be analyzed is obtained through a second analysis deviation evaluation.
3. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: Performing trend detection on the timing parameter data of the device to be analyzed to determine the timing trend data of the device, specifically including: Based on the timing parameter data of the device to be analyzed, determining the maximum number of line segments for the trend detection; According to the maximum number of line segments, the time point of the device timing trend change is determined by iterating the area enclosed by the line segments; A trend slope analysis is performed on the equipment timing trend change time points to determine the equipment timing trend data.
4. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: The PELT change point analysis is performed on the change point analysis data to obtain the time series change point position, specifically including: Determine a PELT penalty parameter based on the pending change point analysis data; wherein the pending change point analysis data includes: a background data group and a cost function; Performing an initial minimum cost configuration on the PELT penalty parameter, and determining a PELT time series change point set through minimum cost iteration according to the minimum cost configuration; The optimal solution of the PELT time series change point set is screened through preset pruning conditions to obtain the time series change point positions.
5. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: The abnormal node monitoring is performed on the timing parameter data of the device to be analyzed to obtain the abnormal time node of the device timing, specifically including: Performing abnormal threshold detection on the timing parameter data of the device to be analyzed to obtain abnormal timing data of the device; According to the device timing abnormality data, the device timing abnormality time node is determined through a preset alarm extraction mechanism; wherein the device timing abnormality time node includes: a timing abnormality value point, an alarm event occurrence time point, and an alarm event end time point.
6. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: Before performing abnormal node monitoring on the timing parameter data of the device to be analyzed to obtain the abnormal time node of the device timing, the method further includes: Input the timing parameter data of the device to be analyzed into a preset Pearson correlation coefficient calculation formula to obtain a device attribute correlation value; According to a preset device attribute correlation threshold, the device attribute correlation value is divided into correlation levels to obtain the device attribute correlation data, and the device attribute correlation level is determined based on the device attribute correlation data.
7. The method for visualizing and replaying time series data based on key point recognition according to claim 1, characterized in that: Based on the equipment timing trend data, the timing change point position and the equipment timing abnormal time node, the equipment timing analysis output data is determined through key segment aggregation, specifically including: Performing fast Fourier transformation on the timing parameter data of the device to be analyzed, and performing amplitude critical sampling on the timing parameter data of the device to be analyzed after the fast Fourier transformation, so as to obtain the timing periodicity parameters of the device; Performing statistical characteristic analysis on the periodic parameters of the device timing to determine the maximum value point of the device timing; Determine the key points of the timing based on the equipment timing trend data, the timing change point position, the equipment timing abnormal time node and the equipment timing maximum value point; wherein the key points of the timing include: trend turning point, change point, maximum value point, minimum value point, timing abnormal value point, alarm start and end points; According to the timing key points, determining the timing key segments by segmenting the timing data; According to the device timing periodicity parameters, the timing key segments are aggregated to determine the device timing analysis output data.
8. The method for visual playback of time series data based on key point recognition according to claim 1, characterized in that: Matching the device timing analysis output data with a visualization element to determine a device timing visualization type, specifically including: Based on the output data of the device timing analysis, the device timing state data group is obtained by converting it through a visualization engine; The backup timing state data group is matched with the visualization element to determine the device timing visualization type; wherein the visualization element includes: a line chart, a heat map, a pie chart, a bar chart, a table, and a box plot.
9. A time series data visualization playback device based on key point recognition, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: Acquire device timing data, and perform data analysis preconfiguration on the device timing data to obtain device timing parameter data to be analyzed; Performing trend detection on the timing parameter data of the device to be analyzed to determine the timing trend data of the device; Based on the timing parameter data of the device to be analyzed, determining the change point analysis data to be analyzed, and performing PELT change point analysis on the change point analysis data to obtain the timing change point position; Perform abnormal node monitoring on the timing parameter data of the device to be analyzed to obtain the abnormal time node of the device timing; Based on the device timing trend data, the timing change point position and the device timing abnormal time node, the device timing analysis output data is determined through key segment aggregation; the device timing analysis output data is matched with visualization elements to determine the device timing visualization type, and according to the device timing visualization type, the timing data visualization playback data is obtained through visualization playback.
10. A non-volatile computer storage medium for visual playback of time series data based on key point recognition, storing computer executable instructions, characterized in that: The computer executable instructions are configured to: Acquire device timing data, and perform data analysis preconfiguration on the device timing data to obtain device timing parameter data to be analyzed; Performing trend detection on the timing parameter data of the device to be analyzed to determine the timing trend data of the device; Based on the timing parameter data of the device to be analyzed, determining the change point analysis data to be analyzed, and performing PELT change point analysis on the change point analysis data to obtain the timing change point position; Perform abnormal node monitoring on the timing parameter data of the device to be analyzed to obtain the abnormal time node of the device timing; Based on the equipment timing trend data, the timing change point position and the equipment timing abnormality time node, the equipment timing analysis output data is determined through key segment aggregation; The device timing analysis output data is matched with visualization elements to determine the device timing visualization type, and according to the device timing visualization type, the timing data visualization playback data is obtained through visualization playback.
Citation Information
Patent Citations
Association rule mining method for multivariate time series monitoring data
CN111046084A
Cloud platform time sequence monitoring data management method and device, equipment and storage medium
CN113688004A
Storm surge analysis method and system for long-time-sequence sea level data
CN117540561A
Time-Series Segmentation and Anomaly Detection
US20240184283A1
