Compression transmission method for recording sampling data

Abstract

The invention belongs to the technical field of compression transmission, and particularly relates to a compression transmission method for recording sampling data, which comprises the following steps of: compressing and transmitting a real-time waveform data sequence by taking a time length as a parameter of an SAX algorithm, the process of calculating the preference comprises the steps of predicting a compression ratio according to the size of a pre-selected length, performing multi-dimensional Gaussian fitting on a plurality of historical waveform data sequences, calculating a first advantage of the pre-selected length according to a fitting effect and the consistency of Gaussian parameters of all dimensions, decomposing each historical waveform data sequence to obtain a trend term sequence, and calculating the preference of each historical waveform data sequence. And obtaining all segment points and all expected points of the trend term sequence according to the pre-selected length and the extreme points in the trend term sequence, so as to calculate a second advantage of the pre-selected length according to the number and the distribution condition of the segment points and the expected points, and calculating the preference of the pre-selected length according to the first advantage, the second advantage and the compression ratio. According to the invention, the data transmission efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of compression transmission, and more specifically, to a method for compressing and transmitting wave recording sampling data. Background Art

[0002] In the power system, real-time monitoring is a key means to ensure stable system operation and rapid fault diagnosis. As an important device for real-time monitoring, the acquisition module can collect and record the waveform data of the power system in real time. In practical applications, in order to reduce the burden of the system and improve the efficiency of data transmission, an efficient data compression method is needed.

[0003] In the related technology, for example, the Chinese patent application document with application publication number CN111930725A discloses a method and device for compressing and fusing distribution and utilization power data, including: using a parallelized EEMD algorithm to decompose the original distribution and utilization power information data, symbolizing the decomposed main component distribution and utilization power information data time series based on the feature-based symbolic clustering approximation (SAX), and then reconstructing the distribution and utilization power information data of the symbolized distribution and utilization power information time series to complete the cleaning of the collected original distribution and utilization power information data; compressing the distribution and utilization power information data based on the SPIHT compression algorithm, and completing the classification and fusion of the compressed distribution and utilization power collection information data using a multi-input-single-output network structure; the invention can complete the accurate collection of distribution and utilization power information on the basis of realizing the cleaning, compression and fusion of the original distribution and utilization power information data, fully explore the value of information data, and provide an important scientific basis for the operation of the power system.

[0004] In related technologies, the SAX (Symbolic Aggregate approximation) algorithm is an efficient data compression method; the SAX algorithm segments the time series by time length and symbolizes each segment to achieve data dimensionality reduction, among which the time length is an important factor affecting the compression effect of the SAX algorithm; therefore, in order to reduce the burden of the system and improve the efficiency of data transmission, it is necessary to reasonably select the time length according to the data characteristics of the waveform data. Summary of the invention

[0005] In order to solve the above technical problem that the time length needs to be reasonably selected according to the data characteristics of the waveform data in order to reduce the burden of the system and improve the efficiency of data transmission, the present invention provides a compression transmission method for recording sampling data, comprising: each data point in the collected real-time waveform data sequence has a corresponding time and value; the time length is used as The algorithm parameters compress and transmit the real-time waveform data sequence; the time length is the pre-selected length with the highest preference, and the process of calculating the preference is: taking each integer within the preset range as a pre-selected length; predicting the length of the pre-selected length according to the size of the pre-selected length. The algorithm compresses the real-time waveform data sequence to obtain the compression rate; performs multi-dimensional Gaussian fitting on multiple historical waveform data sequences, and calculates the first advantage of the pre-selected length based on the fitting effect of the multi-dimensional Gaussian distribution and the consistency of the Gaussian parameters of all dimensions; decomposes each historical waveform data sequence to obtain a trend item sequence; obtains all segmentation points and all expected points of the trend item sequence based on the pre-selected length and the extreme points in the trend item sequence; and The first The fitting slope of all data points between segment points is calculated. The weight of the segment point ; Calculate the second advantage of the preselected length , , is the natural exponential function, are the number of all segmentation points and all expected points respectively, For the The time corresponding to the segment point is For the The time corresponding to the expected point closest to each segment point at the corresponding time; according to the first advantage, the second advantage and the compression rate, the preference of the preselected length is calculated.

[0006] The present invention calculates the first advantage, the second advantage and the compression rate when each integer within a preset range is used as the pre-selected length according to the data characteristics of the waveform data, calculates the preference of the pre-selected length according to the first advantage, the second advantage and the compression rate, and reasonably selects the time length according to the preference, ensuring that the selected time length can not only maximize the compression efficiency, but also ensure that the compressed data maintains sufficient information volume and accuracy, thereby achieving optimal performance in data storage and transmission; by selecting a suitable time length, the compression efficiency of the pre-selected length is improved. The algorithm compresses the real-time waveform data sequence to reduce the burden on the system and improve the efficiency of data transmission.

[0007] Preferably, the compression ratio is equal to the inverse of the preselected length.

[0008] The present invention can help evaluate the impact of different pre-selected lengths on data compression efficiency by predicting the compression rate, thereby selecting a time length that can maximize the compression efficiency while maintaining data integrity.

[0009] Preferably, the multidimensional Gaussian fitting of multiple historical waveform data sequences includes: dividing each historical waveform data sequence into multiple subsequences with a length equal to a preselected length; taking all subsequences of the multiple historical waveform data sequences as samples; performing multidimensional Gaussian fitting on the samples, requiring that the number of all dimensions in the multidimensional Gaussian fitting is equal to the preselected length; the multidimensional Gaussian distribution obtained by the multidimensional Gaussian fitting includes an isotropic multidimensional Gaussian distribution and an anisotropic multidimensional Gaussian distribution; in the obtained multidimensional Gaussian distribution, the Gaussian parameters of each dimension include a mean and a variance, which are respectively recorded as a first parameter and a second parameter.

[0010] The present invention reveals the statistical characteristics of waveform data by performing multi-dimensional Gaussian fitting on historical waveform data sequences, and provides a mathematical model for subsequent compression and analysis.

[0011] Preferably, the first advantage of the preselected length satisfies the expression: ; In the formula, The first advantage of preselecting the length, is the fitting effect of multidimensional Gaussian distribution, is the consistency of the Gaussian parameters of all dimensions of a multidimensional Gaussian distribution.

[0012] The first advantage of the present invention is that the preselected length is calculated based on the consistency of the fitting effect and the Gaussian parameters, and the applicability of different preselected lengths can be quantified, and the time length that can maintain the consistency and stability of data features can be preferentially selected.

[0013] Preferably, the method for calculating the consistency of Gaussian parameters of all dimensions comprises: when the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is: Satisfies the expression: When the multidimensional Gaussian distribution is anisotropic, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is Satisfies the expression: ; In the formula, is the first parameter of the Gaussian parameters of all dimensions The variance of The second parameter of the Gaussian parameters for all dimensions The variance of is a natural exponential function.

[0014] Preferably, the fitting effect of the multidimensional Gaussian distribution is obtained by a chi-square test, in which: when the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, the degrees of freedom in calculating the chi-square test are The number of estimated Gaussian parameters is Equal to the preselected length; when the multidimensional Gaussian distribution is anisotropic multidimensional Gaussian distribution, the degrees of freedom in calculating the chi-square test The number of estimated Gaussian parameters is Equal to twice the preselected length.

[0015] Preferably, the step of obtaining all segmentation points and all expected points of the trend item sequence according to the preselected length and the extreme value points in the trend item sequence respectively includes: the extreme value points include maximum value points and minimum value points; for any extreme value point, when When it is greater than the preset threshold, the extreme point is used as the segmentation point, where To obtain the maximum value function, is the value corresponding to the extreme point, is the value corresponding to the extreme point located to the left of the extreme point and closest to the extreme point, is the value corresponding to the extreme point located to the right of the extreme point and closest to the extreme point, To obtain the absolute value, the data in the trend item sequence whose sequence number is equal to an integer multiple of the pre-selected length is taken as the expected point.

[0016] The present invention decomposes the historical waveform data sequence to obtain the trend item sequence, which helps to identify and remove non-periodic changes in the waveform data and provide a clearer trend pattern for subsequent compression; all segmentation points and expected points are obtained according to the pre-selected length and extreme points in the trend item sequence, which helps to retain key information points during the compression process and ensure that the compressed data can accurately reflect the important characteristics and change trends of the original waveform data.

[0017] Preferably, the The weight of the segment point Satisfies the expression: ; In the formula, is the first The trend strength of the segment points, and the The trend strength of the segment point is equal to the Segment point to The slope of the linear fitting results of all data points between segment points, To take the absolute value, is the first The trend strength of each segment point, The number of all segmentation points in the trend item sequence.

[0018] Preferably, the preference of the preselected length satisfies the expression: ; is the preference of the preselected length, are the first and second advantages of the preselected length, is the compression ratio, are the preset first coefficient, second coefficient and third coefficient respectively, and .

[0019] The present invention calculates the optimization degree of the preselected length by combining the first advantage, the second advantage and the compression ratio, and provides a comprehensive evaluation index to help select the best time and compression balance point, thereby achieving effective data compression and efficient transmission.

[0020] Preferably, the time length is used as The parameters of the algorithm are used to compress the real-time waveform data sequence, including: converting data points in the real-time waveform data sequence into PAA features, including: segmenting the real-time waveform data sequence by time length, and taking the mean of the numerical values ​​corresponding to all data points contained in each segment as the PAA features of each segment; converting the PAA features into characters: performing Gaussian fitting on the PAA features of all segments to obtain a Gaussian distribution; obtaining a breakpoint list that divides the Gaussian distribution into any number of equally probable intervals, and then converting the PAA features of each segment into characters through the breakpoint list and the PAA features to complete symbolization, wherein the Gaussian distribution probability values ​​corresponding to any two adjacent breakpoints in the breakpoint list are equal; and taking a sequence composed of characters corresponding to all segments as the compression result of the real-time waveform data sequence.

[0021] The beneficial effects of the present invention are: The present invention calculates the first advantage, the second advantage and the compression rate when each integer within a preset range is used as the pre-selected length according to the data characteristics of the waveform data, calculates the preference of the pre-selected length according to the first advantage, the second advantage and the compression rate, and reasonably selects the time length according to the preference, ensuring that the selected time length can not only maximize the compression efficiency, but also ensure that the compressed data maintains sufficient information volume and accuracy, thereby achieving optimal performance in data storage and transmission; by selecting a suitable time length, the compression efficiency of the pre-selected length is improved. The algorithm compresses the real-time waveform data sequence to reduce the burden on the system and improve the efficiency of data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below through the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flow chart schematically illustrating a method for compressing and transmitting wave recording sampling data in the present invention; Figure 2is a flowchart schematically showing step S2; Figure 3 is a schematic diagram schematically illustrating the acquisition of PAA features of each segment; Figure 4 FIG. 4 is a schematic diagram schematically showing how to convert the PAA features of each segment into characters. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] The embodiment of the present invention discloses a method for compressing and transmitting wave recording sampling data, referring to Figure 1 , including steps S1 to S3: S1. Collect real-time waveform data sequence.

[0026] It should be noted that in the power system, real-time monitoring is a key means to ensure stable system operation and rapid fault diagnosis. The acquisition module, as an important device for real-time monitoring, can collect and record the waveform data of the power system in real time.

[0027] Specifically, a GPS receiver or a network time synchronization device is installed on each acquisition module to ensure that all acquisition modules are connected to a unified clock source; the sampling rate of the acquisition module is configured to be 1000 samples per second, and the channel is configured as dual channels, corresponding to positive and negative waveform acquisition respectively; the acquisition module is started to start real-time acquisition of waveform data to ensure the real-time and accuracy of data acquisition; wherein, time synchronization calibration is automatically performed with each acquisition module every minute through a network protocol (such as NTP) to ensure that the time deviation is within milliseconds.

[0028] Furthermore, the collected real-time waveform data sequence is a time series sequence, and each data point in the real-time waveform data sequence has a corresponding time and value.

[0029] S2. Determine the time length according to the preference of each pre-selected length within the preset range.

[0030] It should be noted that in practical applications, in order to reduce the burden of the system and improve the efficiency of data transmission, an efficient data compression method is needed; the SAX (Symbolic Aggregate approximation) algorithm is an efficient data compression method. The SAX algorithm maps a time series in a high-dimensional space to a low-dimensional space represented by symbols according to the time length, thereby achieving data compression.

[0031] Step S2 Flowchart Reference Figure 2 , including steps S201 to S204, specifically: S201, according to the size of the pre-selected length, predict the The compression ratio of the algorithm to compress the real-time waveform data sequence.

[0032] Specifically, each integer within the preset range is used as a preselected length; according to the size of the preselected length, the prediction is passed The compression ratio of the algorithm to compress the real-time waveform data sequence, including: , the real-time waveform data sequence is divided into multiple segments, each segment is represented by the mean value, which is equivalent to dividing the real-time waveform data sequence into data points are compressed into 1 data point, so the compression ratio is equal to the inverse of the preselected length; where the preset range is [10,50].

[0033] Among them, the smaller the compression rate, the better the compression effect.

[0034] It should be noted that the present invention can help evaluate the impact of different pre-selected lengths on data compression efficiency by predicting the compression rate, so as to select a time length that can maximize the compression efficiency while maintaining data integrity.

[0035] S202, performing multi-dimensional Gaussian fitting on multiple historical waveform data sequences, and calculating the first advantage of the pre-selected length according to the fitting effect of the obtained multi-dimensional Gaussian distribution and the consistency of Gaussian parameters of all dimensions.

[0036] Specifically, each historical waveform data sequence is divided into multiple subsequences with a length equal to a preselected length; all subsequences of the multiple historical waveform data sequences are taken as samples, and multidimensional Gaussian fitting is performed on the samples, requiring that the number of all dimensions in the multidimensional Gaussian fitting is equal to the preselected length, and the first advantage of the preselected length is calculated based on the fitting effect of the obtained multidimensional Gaussian distribution and the consistency of the Gaussian parameters of all dimensions.

[0037] It should be noted that the present invention reveals the statistical characteristics of waveform data by performing multi-dimensional Gaussian fitting on the historical waveform data sequence, and provides a mathematical model for subsequent compression and analysis.

[0038] Among them, the Gaussian parameters of each dimension include mean and variance, which are respectively recorded as the first parameter and the second parameter; the multidimensional Gaussian distribution obtained by the multidimensional Gaussian fitting includes an isotropic multidimensional Gaussian distribution and an anisotropic multidimensional Gaussian distribution.

[0039] Among them, the isotropic multidimensional Gaussian distribution refers to a multidimensional Gaussian distribution in which the variance of each dimension is the same. When fitting, it is only necessary to make the length of each dimension the same to obtain isotropy, that is, the distribution density value is only related to the distance from the data point to the mean of the Gaussian parameter, but has nothing to do with the direction. At the same time, each dimension of the isotropic multidimensional Gaussian distribution is also independent of each other. The number of parameters of this type of Gaussian distribution increases linearly with the dimension, only the mean increases, and the variance is a scalar. Therefore, the isotropic multidimensional Gaussian fitting does not require much computation and storage.

[0040] Among them, the anisotropic multidimensional Gaussian distribution refers to a multidimensional Gaussian distribution with different variances in each dimension. When fitting, the length of each dimension is different, that is, the distribution density value is not only related to the distance from the data point to the mean in the Gaussian parameter, but also related to the direction. At the same time, each dimension of the anisotropic multidimensional Gaussian distribution is independent of each other. Therefore, the anisotropic multidimensional Gaussian fitting has a better fitting effect.

[0041] Further, the fitting effect of the multidimensional Gaussian distribution is calculated by a chi-square test; the consistency of the Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is calculated; and the first advantage of the preselected length is calculated based on the fitting effect of the obtained multidimensional Gaussian distribution and the consistency of the Gaussian parameters of all dimensions.

[0042] The Chi-Square Test is used to calculate the fitting effect of the multidimensional Gaussian distribution, including: (1) dividing the distribution range of the sample into multiple intervals, and calculating the expected frequency in each interval. The expected frequency is calculated based on the multidimensional Gaussian distribution and represents the number of data that should appear in the interval; (2) calculating the actual frequency in each interval, that is, the number of data that actually appears in the interval; (3) calculating the Chi-Square statistic based on the expected frequency and the actual frequency in all intervals. , which is used to measure the difference between the actual frequency and the expected frequency, then , the number of all intervals is equal to , Respectively represent The expected frequency and actual frequency within an interval; (4) Determine the degrees of freedom , , is the number of estimated Gaussian parameters; (5) According to the degrees of freedom and significance level, find the critical value from the chi-square distribution table, denoted as The specific value of the significance level can be set according to the actual application scenario and requirements, and the value of the significance level is [0.01, 0.1]. The significance level is set to 0.05 in this invention; (6) According to the chi-square statistic and the critical value, the fitting effect of the multidimensional Gaussian distribution is calculated. , .

[0043] In the above process of calculating the fitting effect of the multidimensional Gaussian distribution by the chi-square test: when the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, it is only necessary to determine the mean of the Gaussian parameters of each dimension, and the number of all dimensions is equal to the preselected length. Therefore, in calculating the degrees of freedom The number of estimated Gaussian parameters is When the multidimensional Gaussian distribution is anisotropic, it is necessary to determine the mean and variance of the Gaussian parameters of each dimension, and the number of all dimensions is equal to the preselected length. Therefore, when calculating the degrees of freedom The number of estimated Gaussian parameters is Equal to twice the preselected length.

[0044] The method for calculating the consistency of Gaussian parameters of all dimensions includes: when the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is: Satisfies the expression: ; In the formula, is the first parameter of the Gaussian parameters of all dimensions The square, is a natural exponential function.

[0045] When the multidimensional Gaussian distribution is an anisotropic multidimensional Gaussian distribution, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is Satisfies the expression: ; In the formula, is the first parameter of the Gaussian parameters of all dimensions The variance of The second parameter of the Gaussian parameters for all dimensions The variance of is a natural exponential function.

[0046] Then the first advantage of the preselected length satisfies the expression: ; In the formula, The first advantage of preselecting the length, is the fitting effect of multidimensional Gaussian distribution, is the consistency of the Gaussian parameters of all dimensions of a multidimensional Gaussian distribution.

[0047] It should be noted that the first advantage of the present invention of calculating the preselected length according to the consistency of the fitting effect and the Gaussian parameters is that the applicability of different preselected lengths can be quantified, and the time length that can maintain the consistency and stability of data features is preferentially selected.

[0048] S203, decomposing each historical waveform data sequence to obtain a trend item sequence, respectively obtaining all segmentation points and all expected points of the trend item sequence, and calculating the second advantage of the preselected length according to the number and distribution of all segmentation points and all expected points.

[0049] Specifically, each historical waveform data sequence is subjected to STL decomposition to obtain a trend item sequence, a period item sequence and a residual item sequence of each historical waveform data sequence; the trend item sequence includes the trend item of each historical waveform data in the historical waveform data sequence; the period item sequence includes the period item of each historical waveform data in the historical waveform data sequence; the residual item sequence includes the residual item of each historical waveform data in the historical waveform data sequence; STL (Seasonal-Trend Decomposition using LOESS) decomposition is a decomposition technology for time series, which can decompose time series into three parts: trend (Trend), seasonality (Seasonal) and residual (Residual); STL decomposition is a well-known technology and will not be described in detail here.

[0050] It should be noted that the present invention decomposes the historical waveform data sequence to obtain a trend item sequence, which helps to identify and remove non-periodic changes in the waveform data and provide a clearer trend pattern for subsequent compression.

[0051] Further, all segmentation points of the trend item sequence are obtained, including: obtaining all extreme value points in the trend item sequence, wherein the extreme value points include maximum value points and minimum value points; for any extreme value point, when When it is greater than the preset threshold, the extreme point is used as the segmentation point, where To obtain the maximum value function, is the value corresponding to the extreme point, is the value corresponding to the extreme point located to the left of the extreme point and closest to the extreme point, is the value corresponding to the extreme point located to the right of the extreme point and closest to the extreme point, To take the absolute value.

[0052] The specific value of the preset threshold can be set according to the actual application scenario and requirements, and the method for obtaining the value of the preset threshold is as follows: , the present invention sets the preset threshold value to , They are the maximum and minimum values ​​in the trend item sequence respectively.

[0053] Furthermore, data in the trend item sequence whose sequence number is equal to an integer multiple of a preselected length are taken as expected points, thereby obtaining all expected points of the trend item sequence; wherein, an integer multiple of a number refers to the number multiplied by an integer, where an integer refers to an integer greater than or equal to 1; illustratively, integer multiples of 100 include 100, 200, 300, 1000, etc., while 98, 1050, and 3985 are not integer multiples of 100.

[0054] It should be noted that the present invention obtains all segmentation points and expected points based on the extreme points in the pre-selected length and trend item sequence, which helps to retain key information points during the compression process and ensure that the compressed data can accurately reflect the important characteristics and changing trends of the original waveform data.

[0055] Further, the second advantage of the preselected length is calculated according to all segmentation points and all expected points of the trend item sequence; ; In the formula, The second advantage of preselecting the length is is the natural exponential function, To obtain the maximum value function, is the number of all segmentation points in the trend item sequence, is the number of all expected points in the trend term sequence, is the first The weight of the segment point, is the first The time corresponding to the segment point is is the first The time corresponding to the nearest expected point of each segment point at the corresponding time, To take the absolute value.

[0056] It should be noted that the second advantage of the present invention of calculating the pre-selected length according to the number and distribution of segmentation points and expected points can further optimize the selection of time length, ensuring that the compressed data has a high degree of information retention and can effectively reduce the amount of data.

[0057] Among them, the trend term sequence The weight of the segment point Satisfies the expression: ; In the formula, is the first The trend strength of the segment points, and the The trend strength of the segment point is equal to the Segment point to The slope of the linear fitting results of all data points between segment points, To take the absolute value, is the first The trend strength of each segment point, The number of all segmentation points in the trend item sequence.

[0058] It should be noted that when At that time, The trend strength of the segment point is equal to the trend item sequence Data points to The slope of the linear fit results of all data points between the segment points.

[0059] It should be noted that through When the algorithm performs compression, for each segment, the compression result is closely related to the mean of the value of the segment; in the trend item sequence, there is a trend between two adjacent extreme points, and a trend can be divided into multiple segments, so that the compression result of each segment can still maintain the numerical characteristics of different parts of a trend, but multiple trends cannot be summarized by one segment, otherwise, after the mean processing, the compression result of this segment destroys the numerical characteristics of the trend, resulting in the loss of the numerical characteristics of multiple trends in the received compression result; therefore, Can be less than or equal to ; However, when Greater than hour, The larger it is, the smaller the secondary advantage of the preselected length.

[0060] S204. Calculate the preference of the preselected length according to the first advantage and the second advantage of the preselected length and the compression rate; and obtain the time length according to the preference.

[0061] Specifically, the preference of the preselected length satisfies the expression: ; In the formula, is the preference of the preselected length, are the first and second advantages of the preselected length, is the compression ratio, are the preset first coefficient, second coefficient and third coefficient respectively, and .

[0062] Among them, the preset first coefficient , the second coefficient and the third coefficient The specific value of can be set according to the actual application scenario and requirements, and , the present invention will preset the first coefficient , the second coefficient and the third coefficient Set to , , .

[0063] It should be noted that the present invention combines the first advantage, the second advantage and the compression ratio to calculate the preference of the preselected length, and provides a comprehensive evaluation index to help select the best time and compression balance point to achieve effective data compression and efficient transmission.

[0064] Furthermore, the preselected length with the highest preference within the preset range is used as the time length.

[0065] It should be noted that the present invention reasonably selects the time length through optimization, ensuring that the selected time length can not only maximize the compression efficiency, but also ensure that the compressed data retains sufficient information volume and accuracy, thereby achieving optimal performance in data storage and transmission.

[0066] S3, taking the time length as The algorithm parameters are used to compress and transmit the real-time waveform data sequence.

[0067] Specifically, at the sending end, according to the length of time, The algorithm compresses the real-time waveform data sequence to obtain the compression result of the real-time waveform data sequence; and packages the timestamp, time length and the compression result of the real-time waveform data sequence together to obtain a data packet, and transmits the data packet.

[0068] Among them, the The algorithm includes two steps: (1) converting data points in the real-time waveform data sequence into PAA features, including: segmenting the real-time waveform data sequence by time length, and taking the mean of the values ​​corresponding to all data points contained in each segment as the PAA feature of each segment; illustratively, the schematic diagram of obtaining the PAA features of each segment is as follows: Figure 3As shown; (2) Convert PAA features into characters: Gaussian fitting is performed on the PAA features of all segments to obtain a Gaussian distribution; a breakpoint list is obtained that divides the Gaussian distribution into any number of equally probable intervals, and then the PAA features of each segment are converted into characters through the breakpoint list and the PAA features to complete the symbolization, and a sequence of characters corresponding to all segments is used as the compression result of the real-time waveform data sequence; wherein, the Gaussian distribution probability values ​​corresponding to any two adjacent breakpoints in the breakpoint list are equal; illustratively, when there are two breakpoints and the breakpoints are -0.43 and 0.43 respectively, for all segments with PAA features less than -0.43, the corresponding character is "a", for all segments with PAA features greater than or equal to -0.43 and less than 0.43, the corresponding character is "b", and for all segments with PAA features greater than or equal to 0.43, the corresponding character is "c"; a schematic diagram of converting the PAA features of each segment into characters is shown in Figure 4 shown.

[0069] The data packet format is: a packet header, a compression result of a real-time waveform data sequence, and a checksum. According to the defined packet format, the compression result of the real-time waveform data sequence is encapsulated into a data packet to ensure the integrity and identifiability of the data. A sequence number is added to the packet header part of the data packet so that the receiving end can reassemble the data in sequence. The packet header contains the type of real-time waveform data, the source of the real-time waveform data, the time length, the breakpoint list, and the timestamp of the first data point in the real-time waveform data sequence. The timestamp is accurate to milliseconds and uses the UTC time format.

[0070] The transmission includes: establishing a TCP connection between the sending end and the receiving end to ensure the reliability and stability of data transmission; continuously sending encapsulated data packets to the receiving end through the established TCP connection; during the transmission process, a confirmation mechanism is used to ensure that each data packet is correctly received by the receiving end. If packet loss or error occurs, the sending end retransmits according to the feedback from the receiving end. During the data transmission process, the network status is regularly monitored, such as bandwidth utilization, packet loss rate, etc., so as to timely adjust the data transmission strategy or remind users to optimize the network.

[0071] It should be noted that the present invention improves the time by selecting a suitable time length. The algorithm compresses the real-time waveform data sequence to reduce the burden on the system and improve the efficiency of data transmission.

[0072] On the receiving device, a corresponding number of data receiving channels are configured according to the number of the sending acquisition modules; each channel has independent data processing capabilities and can receive and process data packets from a specific sending acquisition module; when the receiving device is started, it automatically detects and connects to the data streams of each sending acquisition module; once the connection is successful, the receiving channel begins to continuously monitor and receive data packets; after receiving the data packet, the receiving channel first parses it according to the header information of the data packet. The header information contains the type of real-time waveform data, the source of the real-time waveform data, the time length, and the timestamp of the first data point in the real-time waveform data sequence, which is used to guide the subsequent unpacking operation; during the parsing process, the compression result and time length of the real-time waveform data sequence are extracted for subsequent decompression.

[0073] On the receiving device, the decompression algorithm is pre-installed and configured; when the compression result of the real-time waveform data sequence is received, the compression result of the extracted real-time waveform data sequence is decompressed using the same decompression algorithm as the sending end; after the decompression is completed, check whether the decompressed real-time waveform data sequence is complete and lossless; if there is any abnormality, record the error information and try to re-receive and decompress the compression result of the real-time waveform data sequence.

[0074] It should be noted that the present invention calculates the first advantage, the second advantage and the compression rate when each integer within a preset range is used as the pre-selected length according to the data characteristics of the waveform data, calculates the preference of the pre-selected length according to the first advantage, the second advantage and the compression rate, and reasonably selects the time length according to the preference, ensuring that the selected time length can not only maximize the compression efficiency, but also ensure that the compressed data maintains sufficient information volume and accuracy, thereby achieving optimal performance in data storage and transmission; by selecting a suitable time length, the compression efficiency is improved. The algorithm compresses the real-time waveform data sequence to reduce the burden on the system and improve the efficiency of data transmission.

[0075] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0076] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A compression transmission method for recording sampling data, characterized in that: include: Each data point in the real-time waveform data sequence has a corresponding time and value; the time length is used as The parameters of the algorithm compress and transmit the real-time waveform data sequence; The time length is the pre-selected length with the highest preference, and the process of calculating the preference is: Each integer in the preset range is used as a preselected length; according to the size of the preselected length, the prediction is passed The compression ratio of the algorithm to compress the real-time waveform data sequence; Perform multi-dimensional Gaussian fitting on multiple historical waveform data sequences, and calculate the first advantage of the pre-selected length according to the fitting effect of the obtained multi-dimensional Gaussian distribution and the consistency of Gaussian parameters of all dimensions; Decompose each historical waveform data sequence to obtain a trend item sequence; obtain all segmentation points and all expected points of the trend item sequence according to the preselected length and the extreme value points in the trend item sequence; The first The fitting slope of all data points between segment points is calculated. The weight of the segment point ; Calculate the second advantage of the preselected length , , is the natural exponential function, are the number of all segmentation points and all expected points respectively, For the The time corresponding to the segment point is For the The time corresponding to the nearest expected point of each segment point at the corresponding time; The preference of the preselected length is calculated based on the first advantage, the second advantage and the compression ratio.

2. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The compression ratio is equal to the inverse of the preselected length.

3. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The multi-dimensional Gaussian fitting of multiple historical waveform data sequences includes: Dividing each historical waveform data sequence into a plurality of subsequences having a length equal to a preselected length; taking all subsequences of the plurality of historical waveform data sequences as samples; performing multidimensional Gaussian fitting on the samples, requiring that the number of all dimensions in the multidimensional Gaussian fitting is equal to the preselected length; The multidimensional Gaussian distribution obtained by the multidimensional Gaussian fitting includes an isotropic multidimensional Gaussian distribution and an anisotropic multidimensional Gaussian distribution; in the obtained multidimensional Gaussian distribution, the Gaussian parameters of each dimension include a mean and a variance, which are respectively recorded as a first parameter and a second parameter.

4. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The first advantage of the preselected length satisfies the expression: ； In the formula, The first advantage of preselecting the length, is the fitting effect of multidimensional Gaussian distribution, is the consistency of the Gaussian parameters of all dimensions of a multidimensional Gaussian distribution.

5. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The calculation method of the consistency of Gaussian parameters of all dimensions includes: When the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is Satisfies the expression: ； When the multidimensional Gaussian distribution is an anisotropic multidimensional Gaussian distribution, the consistency of Gaussian parameters of all dimensions of the obtained multidimensional Gaussian distribution is Satisfies the expression: ； In the formula, is the first parameter of the Gaussian parameters of all dimensions The variance of The second parameter of the Gaussian parameters for all dimensions The variance of is a natural exponential function.

6. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The fitting effect of the multidimensional Gaussian distribution is obtained by the chi-square test, in which: When the multidimensional Gaussian distribution is an isotropic multidimensional Gaussian distribution, the degrees of freedom in calculating the chi-square test The number of estimated Gaussian parameters is Equal to the preselected length; When the multidimensional Gaussian distribution is anisotropic, the degrees of freedom in calculating the chi-square test The number of estimated Gaussian parameters is Equal to twice the preselected length.

7. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The method of obtaining all segmentation points and all expected points of the trend item sequence according to the preselected length and the extreme value points in the trend item sequence respectively includes: The extreme value points include maximum value points and minimum value points; for any extreme value point, when When it is greater than the preset threshold, the extreme point is used as the segmentation point, where To obtain the maximum value function, is the value corresponding to the extreme point, is the value corresponding to the extreme point located to the left of the extreme point and closest to the extreme point, is the value corresponding to the extreme point located to the right of the extreme point and closest to the extreme point, To take the absolute value; The data in the trend item sequence whose sequence number is an integer multiple of the pre-selected length is taken as the expected point.

8. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The said The weight of the segment point Satisfies the expression: ； In the formula, is the first The trend strength of the segment points, and the The trend strength of the segment point is equal to the Segment point to The slope of the linear fitting results of all data points between segment points, To take the absolute value, is the first The trend strength of each segment point, The number of all segmentation points in the trend item sequence.

9. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The preferred degree of the preselected length satisfies the expression: ； In the formula, is the preference of the preselected length, are the first and second advantages of the preselected length, is the compression ratio, are the preset first coefficient, second coefficient and third coefficient respectively, and .

10. The method for compressing and transmitting wave recording sampling data according to claim 1, characterized in that: The time length is taken as The parameters of the algorithm to compress the real-time waveform data sequence include: Converting data points in the real-time waveform data sequence into PAA features, including: segmenting the real-time waveform data sequence by time length, and taking the average of the values ​​corresponding to all data points contained in each segment as the PAA feature of each segment; Convert PAA features to characters: Gaussian fitting is performed on all segmented PAA features to obtain Gaussian distribution; a breakpoint list is obtained that divides the Gaussian distribution into any number of equally probable intervals, and then the PAA features of each segment are converted into characters through the breakpoint list and PAA features to complete symbolization, wherein the Gaussian distribution probability values ​​corresponding to any two adjacent breakpoints in the breakpoint list are equal; The sequence composed of characters corresponding to all segments is taken as the compression result of the real-time waveform data sequence.

Images