Remote Transmission Method and System for Grouting Monitoring Data Based on Cloud Platform

Abstract

The present invention relates to the technical field of noise data detection, and specifically relates to a remote transmission method and system for grouting monitoring data based on a cloud platform. First, the change characteristic values corresponding to each data point in the total set of pressure time series data are obtained; in the process of constructing an isolation tree according to the binary partitioning of the data set to be processed, according to the set continuous value, set correlation value and element quantity of the pressure data segmentation set, the noise degree value of each pressure data segmentation set is obtained; the actual segmentation value of the data set to be processed is determined; and then all the noise data in the total set of pressure time series data are corrected to obtain the grouting monitoring transmission data of the total set of pressure time series data. In the embodiment of the present invention, by optimizing the selection of the actual segmentation value, the accuracy of identifying noise data and real data is improved, and the remote transmission data efficiency of the grouting pressure is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of noise data detection, and particularly to a remote transmission method and system for grouting monitoring data based on a cloud platform. Background Art

[0002] In the scenario of grouting construction on the thin toe slab of the Shitai Pumped Storage Upper Reservoir Dam, since adverse uplift deformation is extremely likely to occur during the grouting process and there is no grouting gallery under the dam toe slab, it is difficult to carry out maintenance and supplementary grouting for the project. Therefore, it is necessary to realize real-time data acquisition and remote transmission during the grouting process, and display the grouting data through a cloud platform to achieve real-time grouting monitoring. The slurry enters the cracks and holes of the rock under the action of pressure, and the grouting pressure can reflect the degree of slurry diffusion through the pressure of the slurry in the grouting hole. Therefore, it is necessary to remotely transmit the grouting pressure data to achieve real-time grouting monitoring. In order to improve the transmission efficiency of the grouting pressure data, run-length encoding is used to encode and compress the grouting pressure data. Run-length encoding has a good compression ratio for data with strong continuity. Since the grouting pressure data has noise data that affects the continuity of the grouting pressure data, it is necessary to detect the noise data in the grouting pressure data, and then eliminate the noise data to improve the data continuity, and then improve the compression ratio of the grouting pressure data.

[0003] During the process of using the isolation forest algorithm to detect noise data in the grouting pressure data, since there is noise data with a relatively low degree of noise manifestation mixed among the real data with normal changes in the grouting pressure data, the abnormal measurement values of some noise data are relatively low. Since there are also abnormal situations in the grouting pressure data, resulting in the pressure values of some real data being too high or too low, the abnormal measurement values of some real data are relatively high; this in turn leads to inaccurate identification of noise data and real data by the isolation forest algorithm, resulting in inaccurate detection of the noise data in the grouting pressure data, inaccurate correction of the noise data, inability to effectively improve the data continuity, poor improvement effect of the compression ratio of the grouting pressure data, and low efficiency of the remote transmission data of the grouting pressure. Summary of the Invention

[0004] In order to solve the technical problems of inaccurate identification of noise data and real data by the isolation forest algorithm and low efficiency of the remote transmission data of the grouting pressure, the purpose of the present invention is to provide a remote transmission method and system for grouting monitoring data based on a cloud platform, and the specific technical solutions adopted are as follows:

[0005] The present invention proposes a remote transmission method for grouting monitoring data based on a cloud platform, and the method includes:

[0006] Obtain the total set of pressure time series data, the total set of flow time series data, and the total set of density time series data during the grouting process;

[0007] In the total set of pressure time series data, according to the degree of change of the pressure values around each data point, obtain the change characteristic value corresponding to each data point in the total set of pressure time series data;

[0008] Sample the total set of pressure time series data to obtain a set of data to be processed, and construct each isolation tree according to each set of data to be processed, so as to obtain the noise data in the total set of pressure time series data. In the process of constructing the isolation tree according to the set of data to be processed, obtain each to-be-processed segmentation value of the set of data to be processed according to the change characteristic value in the set of data to be processed; perform binary partitioning on the set of data to be processed according to the to-be-processed segmentation value to obtain two pressure data segmentation sets corresponding to each to-be-processed segmentation value;

[0009] According to the time difference of the data points in the pressure data segmentation set and the change characteristic value difference, obtain the set continuity value of each pressure data segmentation set; according to the change difference of the pressure data segmentation set, the total set of flow time series data, and the total set of density time series data, obtain the set correlation value of each pressure data segmentation set;

[0010] According to the set continuity value, the set correlation value, and the number of elements of the pressure data segmentation set, obtain the noise degree value of each pressure data segmentation set; according to the difference in the noise degree values between all the pressure data segmentation sets corresponding to the to-be-processed segmentation value, obtain the preference degree value of each to-be-processed segmentation value of the set of data to be processed; according to the preference degree value of each to-be-processed segmentation value of the set of data to be processed, determine the actual segmentation value of the set of data to be processed, and construct the isolation tree of the set of data to be processed according to the actual segmentation value;

[0011] Correct all the noise data in the total set of pressure time series data to obtain the grouting monitoring transmission data of the total set of pressure time series data.

[0012] Furthermore, the acquisition formula of the change characteristic value includes:

[0013] ; where is the change characteristic value of the rd data point in the total set of pressure time series data; is the total number of neighborhood data points in the preset neighborhood range of the th data point in the total set of pressure time series data; is the first-order difference of the th neighborhood data point in the preset neighborhood range of the rd data point; is at the Among the preset neighborhood ranges of the data points, the second-order difference of the th neighborhood data point; is the number of positive and negative transformations of the first-order difference among the preset neighborhood ranges of the th data point; is a normalization function.

[0014] Furthermore, the method for obtaining the set of continuous values includes:

[0015] ; where is the set of continuous values of the pressure data segmentation set; is the total number of data points in the pressure data segmentation set; is the time of the th data point in the pressure data segmentation set; is the time of the th data point in the pressure data segmentation set; is the change characteristic value of the th data point in the pressure data segmentation set; is the change characteristic value of the th data point in the pressure data segmentation set; is an exponential function with the natural number e as the base.

[0016] Furthermore, the method for obtaining the set of correlation values includes:

[0017] Obtain the data correlation value according to the data correlation value formula, and the data correlation value formula includes:

[0018] ; where is the data correlation value of the th data point in the pressure data segmentation set; is the total number of data points in the preset reference range of the th data point in the pressure data segmentation set; is the first-order difference of the th reference data point in the preset reference range of the th data point; is the first-order difference of the density data point at the same time corresponding to the th reference data point in the preset reference range of the th data point in the total set of density time-series data; is the first-order difference of the flow data point at the same time corresponding to the th reference data point in the preset reference range of the th data point in the total set of flow time-series data; is the second-order difference of the th reference data point in the preset reference range of the th data point; is the second-order difference of the density data point at the same moment corresponding to the th reference data in the preset reference range of the th data point in the total set of density time-series data; is the second-order difference of the flow data point at the same moment corresponding to the th reference data in the preset reference range of the th data point in the total set of flow time-series data; is a normalization function;

[0019] Calculate the mean value of the data correlation values of all data points in the pressure data segmentation set to obtain the set correlation value of the pressure data segmentation set.

[0020] Further, the method for obtaining the noise degree value includes:

[0021] Obtain the noise degree value of the pressure data segmentation set according to the set continuous value, the set correlation value and the number of elements of the pressure data segmentation set;

[0022] The set continuous value and the noise degree value are negatively correlated; the set correlation value and the noise degree value are negatively correlated; the number of elements and the noise degree value are negatively correlated.

[0023] Further, the method for obtaining the preference degree value includes:

[0024] Calculate the absolute value of the difference between the noise degree values corresponding to all the pressure data segmentation sets of the to-be-processed segmentation value to obtain the first preference value of the to-be-processed segmentation value;

[0025] Normalize the first preference value to obtain the preference degree value of the to-be-processed segmentation value.

[0026] Further, the method for obtaining the to-be-processed segmentation value includes:

[0027] In the to-be-processed data set, obtain the maximum value of the change feature value and the minimum value of the change feature value; use all the change feature values between the minimum value of the change feature value and the maximum value of the change feature value, except for the maximum value and the minimum value of the change feature value, as all the to-be-processed segmentation values.

[0028] Further, the method for obtaining two pressure data segmentation sets corresponding to the to-be-processed segmentation value includes:

[0029] Perform binary partitioning on the to-be-processed data set according to the to-be-processed segmentation value, count all the data in the to-be-processed data set that is not greater than the to-be-processed segmentation value, and construct a first pressure data segmentation set corresponding to the to-be-processed segmentation value;

[0030] Count all the data in the to-be-processed data set that is greater than the to-be-processed segmentation value, and construct a second pressure data segmentation set corresponding to the to-be-processed segmentation value;

[0031] Take the first pressure data segmentation set and the second pressure data segmentation set as two pressure data segmentation sets corresponding to the to-be-processed segmentation value.

[0032] Furthermore, the method for obtaining the grouting monitoring transmission data includes:

[0033] Calculate the mean value of a preset number of data points before and after each noise data as the replacement data point for each noise data;

[0034] In the total set of pressure time series data, replace each noise data with each replacement data point to obtain the noise-reduced data of the total set of pressure time series data;

[0035] Compress the noise-reduced data based on run-length encoding to obtain the grouting monitoring transmission data of the total set of pressure time series data.

[0036] A remote transmission system for grouting monitoring data based on a cloud platform includes a memory and a processor. The processor executes the calculation program stored in the memory to implement any one of the remote transmission methods for grouting monitoring data based on a cloud platform.

[0037] The present invention has the following beneficial effects:

[0038] Since the changing trend of noise data in time series is often more rapid and unstable, while real data does not have this characteristic, according to the difference in the changing trend of data, the change characteristic values corresponding to each data point in the total set of pressure time series data are obtained, and the change characteristic values represent the possibility of data noise level. Since there is noise data with a relatively low noise manifestation degree mixed among the real data with normal changes, the abnormal measurement values of some noise data detected by using the conventional isolation forest algorithm are low; since there are also abnormal situations in the grouting pressure data, resulting in the pressure values of the real data being too high or too low, the difference between the real data and the noise data in the abnormal situation is reduced, and the difference in the abnormal measurement values of the real data and the noise data with normal changes detected by using the conventional isolation forest algorithm is also reduced; in order to accurately screen out the noise data, through the appropriate selection of binary partitioning by the isolation forest method, the difference in the abnormal measurement values between the noise data and the normal data is increased. In order to study the segmentation effect of the segmentation value to be processed, first analyze the influencing factors of the noise level of the pressure data segmentation set. Since the real data has continuity and the noise data does not have continuity, the set of continuous values can reflect the overall continuity of the pressure data segmentation set. The stronger the continuity, the smaller the overall noise level of the pressure data segmentation set. The set correlation value can reflect the correlation between the pressure data segmentation set and other grouting monitoring data, that is, when the pressure changes, the density and flow rate of the slurry also change, and the changing trends are quite different. Therefore, the smaller the set correlation value, the greater the possibility that the data in the pressure data segmentation set is noise data. According to the set of continuous values, the set correlation value and the number of elements of the pressure data segmentation set, the noise level value of each pressure data segmentation set is obtained; the noise level value can comprehensively reflect the noise possibility of the pressure data segmentation set. Furthermore, the actual segmentation value is determined, which reflects the segmentation value to be processed corresponding to the best segmentation effect, and then the difference in the abnormal measurement values between the noise data and the real data is increased, so that the result of screening the noise data is more accurate. The noise data can more accurately reflect the actual noise data. In order to improve the compression ratio during data compression and improve the remote transmission efficiency, all the noise data in the total set of pressure time series data are corrected to obtain the grouting monitoring transmission data of the total set of pressure time series data. By improving the accuracy of the isolation forest algorithm in identifying noise data and real data, the accuracy of correcting the noise data in the grouting pressure data is improved, and the data compression ratio of the grouting pressure is increased, so that the grouting monitoring transmission data of the total set of pressure time series data can be transmitted more quickly, and the remote transmission data efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0040] Figure 1 This is a flowchart of a method for remotely transmitting grouting monitoring data based on a cloud platform provided by an embodiment of the present invention. Detailed implementation manners

[0041] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following describes in detail a method and system for remotely transmitting grouting monitoring data based on a cloud platform proposed by the present invention, including its specific implementation manners, structures, features and effects, with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0043] The following specifically describes the specific solutions of a method and system for remotely transmitting grouting monitoring data based on a cloud platform provided by the present invention with reference to the accompanying drawings.

[0044] Please refer to Figure 1 , which shows a flowchart of a method for remotely transmitting grouting monitoring data based on a cloud platform provided by an embodiment of the present invention. The method includes the following steps:

[0045] Step S1, obtaining the total set of pressure time-series data, the total set of flow rate time-series data, and the total set of density time-series data during the grouting process.

[0046] The slurry enters the cracks and holes of the rock under the action of pressure. The grouting pressure can reflect the degree of slurry diffusion through the pressure of the slurry in the grouting hole. Therefore, it is necessary to remotely transmit the grouting pressure data to achieve real-time grouting monitoring. There are noise data in the total set of pressure time-series data during the grouting process. Since the flow rate and density will also change correspondingly when the pressure changes, the changes in the total set of pressure time-series data, the total set of flow rate time-series data, and the total set of density time-series data during the grouting process are similar. The total set of flow rate time-series data and the total set of density time-series data are obtained for subsequent screening of the noise data in the total set of pressure time-series data.

[0047] Specifically, during the detection process, a grouting recorder, a sensor module, and high-precision operations are utilized. During acquisition, the pressure value, flow rate value, density value, and time of the grouting process are synchronously acquired according to a fixed set frequency. The pressure values at all acquired moments are statistically analyzed to obtain the total set of pressure time-series data; the flow rate values at all acquired moments are statistically analyzed to obtain the total set of flow rate time-series data; the density values at all acquired moments are statistically analyzed to obtain the total set of density time-series data. In an embodiment of the present invention, the fixed set frequency is once per minute, and the implementer can set it by himself according to the implementation scenario.

[0048] It should be noted that, for the convenience of calculation, all index data participating in the calculation in the embodiments of the present invention have undergone data preprocessing, thereby canceling the influence of the dimension. The specific means of canceling the dimension influence are well-known technical means to those skilled in the art and will not be limited herein.

[0049] Step S2, in the total set of pressure time-series data, according to the degree of change of the pressure values around each data point, obtain the change characteristic value corresponding to each data point in the total set of pressure time-series data.

[0050] During the grouting construction of the upper reservoir dam of Shitai Pumped Storage on the thin toe slab, the grouting pressure value usually stabilizes within a certain range, but the pressure value may also change due to abnormal situations, such as leakage, blockage, material supply interruption, or valve problems. Although these abnormalities will cause pressure changes, they also follow a certain regularity, and the change trend is usually not too large, and the change direction is also relatively stable. And the noise data generated due to environmental interference or equipment failure is highly random, with a large change trend and low stability in time series. Therefore, in the total set of pressure time-series data, there is noise data generated due to environmental interference or equipment failure, and there is also real data that truly reflects the real pressure value. Since the change trend of noise data in time series is often more rapid and unstable, and real data does not have this characteristic, according to the degree of change of the pressure values around each data point, obtain the change characteristic value corresponding to each data point in the total set of pressure time-series data. The change characteristic value can reflect the difference in the change trend of the data, and thus characterize the possibility of the data noise level.

[0051] Preferably, in an embodiment of the present invention, the method for obtaining the change characteristic value includes:

[0052] To distinguish noise data from real data, through the change trend around the data point at the th moment in the total set of pressure time-series data, obtain the change characteristic value of the data point at the th moment. In an embodiment of the present invention, the formula for obtaining the change characteristic value includes:

[0053] ; where is in the total set of pressure time-series data, at the The change characteristic value of a data point; In the total set of pressure time-series data, the Total number of neighborhood data points within the preset neighborhood range of the th data point; In the preset neighborhood range of the th data point, the first-order difference of the th neighborhood data point; In the preset neighborhood range of the th data point, the second-order difference of the th neighborhood data point; In the preset neighborhood range of the th data point, the number of positive and negative transformations of the first-order difference; is a normalization function. It should be noted that In the total set of pressure time-series data, the difference between the pressure value of the data point after the th neighborhood data point and the pressure value of the th neighborhood data point; In the total set of pressure time-series data, the difference between the first-order difference of the data point after the th neighborhood data point and the first-order difference of the th neighborhood data point. It should be noted that in the preset neighborhood range of the data point at the th moment, the first-order difference corresponding to each data point is obtained, and the number of times the positive and negative signs of the first-order difference corresponding to the data points in the preset neighborhood range are changed is used as

[0054] Specifically, a preset neighborhood range is constructed with the data point as the center point, and the center of the preset neighborhood range is the center point; the size of the preset neighborhood range is 5*1; each data point in the preset neighborhood range is used as each neighborhood data point, that is, the total number of neighborhood data points in the preset neighborhood range is usually 5.

[0055] In the change characteristic value formula, is the average value of the first-order difference, reflecting the magnitude of the change speed of the data point on the total set of pressure time-series data. The faster the change speed, the larger the change characteristic value; is the average value of the second-order difference, reflecting the degree of change instability of the data point on the total set of pressure time-series data. The greater the degree of change instability, the larger the change characteristic value; Let \(n\) be the number of sign changes of the first-order difference within the preset neighborhood range of the data point, which reflects the directional instability of the change speed. The greater the directional instability of the change speed, the larger the change eigenvalue. The change eigenvalue reflects the rapid and unstable change trend of the data point by comprehensively considering the change speed, the degree of change instability, and the directional instability of the change speed. Since the change trend of noise data in time series is often more rapid and unstable, while real data does not have this characteristic, the larger the change eigenvalue, the stronger the noise manifestation of the data point, and the more likely it is a noise point.

[0056] Step S3: Sample the total set of pressure time series data to obtain a set of data to be processed, and construct individual isolation trees based on each set of data to be processed, thereby obtaining the noise data in the total set of pressure time series data. During the process of constructing an isolation tree based on the set of data to be processed, obtain each segmentation value to be processed of the set of data to be processed according to the change eigenvalue in the set of data to be processed; perform binary partitioning on the set of data to be processed according to the segmentation value to be processed, and obtain two segmented sets of pressure data corresponding to each segmentation value to be processed.

[0057] Due to the existence of noise data with a relatively low degree of noise manifestation mixed among the real data with normal changes, the abnormal metric values of some noise data detected by the conventional isolation forest algorithm are relatively low; due to the existence of abnormal situations in the grouting pressure data, the pressure values of the real data are too high or too low, resulting in a reduction in the difference between the real data with abnormal situations and the noise data, and also leading to a reduction in the difference in abnormal metric values between the real data with normal changes and the noise data detected by the conventional isolation forest algorithm; in order to accurately screen out the noise data, through the appropriate selection of binary partitioning by the isolation forest method, the difference in abnormal metric values between the noise data and the normal data is increased. During the process of appropriately selecting binary partitioning by the isolation forest method, first sample the total set of pressure time series data to obtain a set of data to be processed, and construct individual isolation trees based on each set of data to be processed, thereby obtaining the noise data in the total set of pressure time series data. During the process of constructing an isolation tree based on the set of data to be processed, use the set of data to be processed as the data set for binary partitioning, and the segmentation value to be processed is the value that can be used for binary partitioning, and then analyze the segmentation effect of the segmentation value to be processed. In order to study the segmentation effect of the segmentation value to be processed, first study that during the construction of an isolation tree in the process of binary partitioning based on the set of data to be processed, obtain each segmentation value to be processed of the set of data to be processed according to the change eigenvalue in the set of data to be processed; perform binary partitioning on the set of data to be processed according to the segmentation value to be processed, and obtain two segmented sets of pressure data corresponding to each segmentation value to be processed. Then analyze the noise difference between the two segmented sets of pressure data corresponding to the segmentation value to be processed, for subsequent determination of the segmentation value to be processed that can best segment out the noise data.

[0058] Specifically, the isolation forest algorithm needs to train multiple isolation trees. By constructing multiple isolation trees, it is possible to better capture noise data and improve the accuracy and robustness of the noise data. Sample the total set of pressure time series data to obtain a set of data to be processed with a preset sample size. The set of data to be processed is 80% of the number of data in the total set of pressure time series data. In one embodiment of the present invention, the preset sample size is 100, and the implementer can set it according to the implementation scenario. And construct each isolation tree according to each set of data to be processed, so as to obtain the noise data in the total set of pressure time series data. The embodiment of the present invention focuses on improving the proper selection of binary partitioning, thereby improving the segmentation effect. The process of training isolation trees by the isolation forest algorithm is a well-known technical means in the art and will not be elaborated here.

[0059] Preferably, in one embodiment of the present invention, the method for obtaining the segmentation value to be processed includes:

[0060] In the set of data to be processed, obtain the maximum value of the change feature value and the minimum value of the change feature value; take all the change feature values that are between the minimum value of the change feature value and the maximum value of the change feature value, except for the maximum value of the change feature value and the minimum value of the change feature value, as all the segmentation values to be processed. The segmentation values to be processed are all the change feature values that can be used for binary partitioning, for subsequent determination of the segmentation value to be processed that can best segment out the noise data.

[0061] Preferably, in one embodiment of the present invention, the method for obtaining the two pressure data segmentation sets corresponding to the segmentation value to be processed includes:

[0062] To analyze the noise difference between the two pressure data segmentation sets corresponding to the segmentation value to be processed, perform binary partitioning on the set of data to be processed according to the segmentation value to be processed, and count all the data in the set of data to be processed that is not greater than the segmentation value to be processed, and construct the first pressure data segmentation set corresponding to the segmentation value to be processed;

[0063] Count all the data in the set of data to be processed that is greater than the segmentation value to be processed, and construct the second pressure data segmentation set corresponding to the segmentation value to be processed;

[0064] Take the first pressure data segmentation set and the second pressure data segmentation set as the two pressure data segmentation sets corresponding to the segmentation value to be processed.

[0065] Step S4, according to the time difference and change feature value difference of the data points in the pressure data segmentation set, obtain the set continuity value of each pressure data segmentation set; according to the change difference of the pressure data segmentation set, the total set of flow time series data, and the total set of density time series data, obtain the set correlation value of each pressure data segmentation set.

[0066] To study the segmentation effect of the segmentation values to be processed, first analyze the influencing factors of the noise level of the pressure data segmentation set. Since the occurrence of noise data is highly random and the changes in normal data caused by abnormal situations are uncontrollable, noise data often lacks continuity. The true grouting pressure of the data usually stabilizes within a certain range and changes continuously. However, abnormal situations can cause the pressure to be too high or too low, such as leakage, blockage, material supply interruption, or valve problems. Although the true data of these abnormal situations will cause pressure changes, the change trend usually is not too large and is continuous. Since the true data is continuous and the noise data is not continuous, calculate the continuity of the data corresponding time and change characteristic values in the pressure data segmentation set. According to the time difference and change characteristic value difference of the data points in the pressure data segmentation set, obtain the set continuity value of each pressure data segmentation set. The set continuity value can reflect the overall continuity of the pressure data segmentation set. Stronger continuity indicates a smaller overall noise level of the pressure data segmentation set, while weaker continuity indicates a larger overall noise level of the pressure data segmentation set. Since the change of pressure data is related to other grouting monitoring data to a certain extent. For example, an increase in pressure will cause an increase in the density of the slurry and also an increase in the flow rate of the slurry. To analyze the authenticity of the pressure data segmentation set, according to the change differences among the pressure data segmentation set, the total set of flow time series data, and the total set of density time series data, obtain the set correlation value of each pressure data segmentation set. The set correlation value can reflect the correlation between the pressure data segmentation set and other grouting monitoring data, that is, when the pressure changes, the density and flow rate of the slurry also change accordingly, and the change trends are quite different. Therefore, the smaller the set correlation value, the greater the possibility that the data in the pressure data segmentation set is noise data.

[0067] Preferably, in an embodiment of the present invention, the method for obtaining the set continuity value includes:

[0068] By calculating the differences of the data corresponding time and change characteristic values in the pressure data segmentation set, analyze the continuity of the overall data, and obtain the set continuity value of the pressure data segmentation set. In an embodiment of the present invention, the formula for obtaining the set continuity value includes:

[0069] ; where is the set continuity value of the pressure data segmentation set; is the total number of data points in the pressure data segmentation set; in the pressure data segmentation set, the th moment of the th data point; in the pressure data segmentation set, the th moment of the th data point; is the change characteristic value of the th data point in the pressure data segmentation set; is the exponential function with the natural number e as the base.

[0070] In the set continuous value formula, is the time interval between adjacent data points in the pressure data segmentation set in time series. The smaller the time interval, the stronger the time continuity of the data, and the larger the set continuous value; is the difference in the change characteristic values of adjacent data points in the pressure data segmentation set. The smaller the difference in the change characteristic values, the stronger the change continuity of the data, and the larger the set continuous value; the set continuous value reflects the continuity of the overall data by taking the average of the time continuity of all data and the change continuity of the data. Since the real data is continuous and the noise data is not continuous, the larger the set continuous value, the smaller the overall noise level of the pressure data segmentation set.

[0071] Preferably, in an embodiment of the present invention, the method for obtaining the set correlation value includes:

[0072] By analyzing the differences in the changes corresponding to the pressure data segmentation set, the total flow time series data set, and the total density time series data set at the same moment, the overall data correlation is analyzed, and the data correlation value of the th data point in the pressure data segmentation set is obtained. In an embodiment of the present invention, the method for obtaining the data correlation value includes:

[0073] ; where is the data correlation value of the th data point in the pressure data segmentation set; is the total number of data points in the preset reference range of the th data point in the pressure data segmentation set; is the first-order difference of the th reference data point in the preset reference range of the th data point; is the first-order difference of the density data point corresponding to the th reference data point in the preset reference range of the th data point at the same moment in the total density time series data set; is the first-order difference of the flow data point corresponding to the th reference data point in the preset reference range of the th data point at the same moment in the total flow time series data set; is the second-order difference of the th reference data point in the preset reference range of the th data point; is the In the preset reference range of a data point, the second-order difference of the density data point at the same moment in the total set of density time-series data corresponding to the th reference data; In the preset reference range of a data point, the second-order difference of the flow data point at the same moment in the total set of flow time-series data corresponding to the Normalization function. It should be noted that, in the total set of pressure time-series data, by calculating the difference between the pressure value of the data point after the th reference data point and the pressure value of the th reference data point; in the total set of pressure time-series data, by calculating the difference between the first-order difference of the data point after the th reference data point and the first-order difference of the th reference data point; in the total set of density time-series data, by calculating the difference between the density value of the density data point after the density data point corresponding to the th reference data point at the same moment and the density value of the current density data point; in the total set of density time-series data, by calculating the difference between the first-order difference of the density data point after the density data point corresponding to the th reference data at the same moment and the first-order difference of the current density data point; in the total set of flow time-series data, by calculating the difference between the flow value of the data point after the flow data point corresponding to the th reference data at the same moment and the flow value of the current data point; in the total set of flow time-series data, by calculating the difference between the first-order difference of the data point after the flow data point corresponding to the th reference data at the same moment and the first-order difference of the current data point.

[0074] Specifically, a preset reference range is constructed with the data point as the central data point, and the center of the preset reference range is the central data point; the total number of data in the preset reference range is 7, the size of the preset reference range is adaptively determined, and each data point in the preset reference range is used as each reference data point.

[0075] In the data correlation value formula, reflects the difference in the change trends of pressure and density. The smaller the difference, the more consistent the change trends of pressure and density, and the greater the data correlation; the difference in the change trends of pressure and flow. The smaller the difference, the more consistent the change trends of pressure and flow, and the greater the data correlation; Reflect the stability difference between the pressure and density change trends. The smaller the difference, the more consistent the change trends of pressure and density, and the greater the data correlation; Reflect the stability difference between the pressure and flow rate change trends. The smaller the difference, the more consistent the change trends of pressure and flow rate, and the greater the data correlation; The data correlation value comprehensively reflects the change consistency of pressure, density, and flow rate data at the same moment by considering the data change differences and stability differences around the data points. The greater the change consistency, the greater the data correlation value; Since an increase in pressure will cause an increase in the density of the slurry and also an increase in the flow rate of the slurry, the greater the data correlation value, the greater the possibility that the data points are real data.

[0076] Calculate the mean value of the data correlation values of all data points in the pressure data segmentation set to obtain the set correlation value of the pressure data segmentation set. The set correlation value comprehensively reflects the change consistency of all data in the pressure data segmentation set. The greater the set correlation value, the greater the possibility that the data in the pressure data segmentation set are real data.

[0077] In step S5, according to the set continuity value, set correlation value, and number of elements of the pressure data segmentation set, obtain the noise degree value of each pressure data segmentation set; According to the difference in the noise degree values between all pressure data segmentation sets corresponding to the to-be-processed segmentation value, obtain the preference degree value of each to-be-processed segmentation value of the to-be-processed data set; According to the preference degree value of each to-be-processed segmentation value of the to-be-processed data set, determine the actual segmentation value of the to-be-processed data set, and construct an isolation tree for the to-be-processed data set according to the actual segmentation value.

[0078] In order to study the segmentation effect of the segmentation value to be processed, so that the noise level difference between the data on both sides after binary partitioning is large, thereby increasing the difference in the anomaly measurement values between the noise data and the normal data. First, analyze the noise level of the pressure data segmentation set. According to the set continuous value, set correlation value, and number of elements of the pressure data segmentation set, obtain the noise level value of each pressure data segmentation set; since the set continuous value can reflect the overall continuity of the pressure data segmentation set, stronger continuity indicates that the overall noise level of the pressure data segmentation set is smaller; the set correlation value can reflect the correlation between the overall data of the pressure data segmentation set and other grouting monitoring data, that is, when the pressure changes, the density and flow rate of the slurry also change, and the change trends are less different, so the larger the set correlation value, the smaller the possibility that the data in the pressure data segmentation set is noise data. Since real data often shows a clustered distribution, the more the number of elements in the pressure data segmentation set, the greater the possibility that the data in the pressure data segmentation set is real data. The noise level value can comprehensively reflect the noise possibility of the pressure data segmentation set. The greater the difference in the noise level values between the pressure data segmentation sets on both sides of the segmentation value to be processed, the greater the possibility that the segmentation value to be processed can segment out noise. According to the difference in the noise level values between all the pressure data segmentation sets corresponding to the segmentation value to be processed, obtain the preference degree value of each segmentation value to be processed in the data set to be processed; the preference degree value can reflect the difference in the noise level values between all the pressure data segmentation sets corresponding to the segmentation value to be processed, and the greater the difference, the better the segmentation effect. According to the preference degree value of each segmentation value to be processed in the data set to be processed, determine the actual segmentation value of the data set to be processed. The actual segmentation value reflects the segmentation value to be processed corresponding to the best segmentation effect of the binary partitioning, thereby increasing the difference in the anomaly measurement values between the noise data and the real data. By appropriately selecting the binary partitioning, an isolation tree of the data set to be processed is constructed. Determine the data anomaly score value through the isolation tree, and screen out a more accurate noise data result.

[0079] Preferably, in one embodiment of the present invention, the method for obtaining the noise level value includes:

[0080] According to the set continuous value, set correlation value, and number of elements of the pressure data segmentation set, obtain the noise level value of the pressure data segmentation set;

[0081] The set continuous value and the noise level value are negatively correlated; the set correlation value and the noise level value are negatively correlated; the number of elements and the noise level value are negatively correlated.

[0082] In one embodiment of the present invention, the formula for obtaining the noise level value includes:

[0083] ; where, is the noise level value of the pressure data segmentation set; is the set correlation value of the pressure data segmentation set; The set continuous value of the pressure data segmentation set; The total number of data points of the pressure data segmentation set; Is the normalization function.

[0084] In the noise level value formula, due to the continuity of the real data, the set continuous value can reflect the overall continuity of the pressure data segmentation set. The stronger the continuity, the smaller the noise level value; because the real data is often correlated with other monitoring data, the set correlation value can reflect the correlation between the overall data of the pressure data segmentation set and other grouting monitoring data. The greater the correlation, the smaller the noise level value; because the real data often shows an aggregated distribution, the more elements in the pressure data segmentation set, the smaller the noise level value; the greater the possibility that the data in the pressure data segmentation set is real data. The noise level value comprehensively considers the correlation between the pressure data segmentation set and other grouting monitoring data, the overall continuity of the pressure data segmentation set, and the amount of element data in the pressure data segmentation set, and comprehensively reflects the overall noise level of the pressure data segmentation set.

[0085] Preferably, in an embodiment of the present invention, the method for obtaining the preference degree value includes:

[0086] Calculate the difference in the noise level values between all the pressure data segmentation sets corresponding to the to-be-processed segmentation value, and obtain the first preference value of the to-be-processed segmentation value;

[0087] Normalize the first preference value to obtain the preference degree value of the to-be-processed segmentation value.

[0088] The formula for obtaining the preference degree value in an embodiment of the invention includes:

[0089] ; where Is the Preference degree value of the th to-be-processed segmentation value; Is the Noise level value of the th to-be-processed segmentation value corresponding to the first pressure data segmentation set; Is the Noise level value of the th to-be-processed segmentation value corresponding to the second pressure data segmentation set.

[0090] In the preference degree value formula, Reflects the difference in the noise level values between the pressure data segmentation sets. The greater the difference in the noise level values, the better the segmentation effect. The greater the possibility that the to-be-processed segmentation value divides out noise, and the corresponding higher its preference degree. The preference degree value reflects that the better the segmentation effect of the to-be-processed segmentation value, and the to-be-processed segmentation value is selected according to the preference degree value when constructing the isolation tree.

[0091] Specifically, the to-be-processed segmentation value corresponding to the maximum preference degree value is used as the actual segmentation value of the to-be-processed data set. The actual segmentation value reflects the to-be-processed segmentation value corresponding to the best segmentation effect, thereby increasing the difference in the anomaly metric values between the noise data and the real data, so that the result of screening the noise data is more accurate.

[0092] According to the actual segmentation value, an isolation tree of the to-be-processed data set is constructed. It should be noted that the isolation forest algorithm is a well-known prior art means for those skilled in the art. Here, only the process of obtaining the isolation tree of the to-be-processed data set according to the actual segmentation value is briefly described:

[0093] In the process of first performing binary partitioning on the to-be-processed data set, according to the change characteristic value in the to-be-processed data set, each to-be-processed segmentation value of the to-be-processed data set is obtained; then the actual segmentation value is determined; according to the actual segmentation value, the to-be-processed data set is binary partitioned, and the two pressure data segmentation sets are respectively used as new to-be-processed data sets; it should be noted that the above steps have already detailed the specific steps of determining the actual segmentation value according to the change characteristic value in the to-be-processed data set, which will not be elaborated here.

[0094] In the process of each binary partitioning of the to-be-processed data set, according to the change characteristic value in the new to-be-processed data set, each new to-be-processed segmentation value of the new to-be-processed data set is obtained; then the new actual segmentation value is determined; according to the new actual segmentation value, the new to-be-processed data set is binary partitioned, and the two new pressure data segmentation sets are respectively used as the updated to-be-processed data sets;

[0095] Until the preset dynamic adjustment condition is met, the isolation tree of the to-be-processed data set is obtained. The isolation tree can reflect the partitioning method of the to-be-processed data set.

[0096] In an embodiment of the present invention, the preset dynamic adjustment condition is that the maximum depth of the isolation tree is 8, and the implementer can set it according to the implementation scenario.

[0097] The isolation forest algorithm needs to train to obtain multiple isolation trees. By constructing multiple isolation trees, noise data can be better captured. All the isolation trees of the total set of pressure time series data are obtained, the anomaly metric value of each data is comprehensively calculated, and according to the anomaly score threshold, the abnormal data in the total set of pressure time series data is screened out, and the abnormal data is used as noise data. The anomaly score threshold is 0.75. In an embodiment of the present invention, the preset sample quantity is 100, and the implementer can set it according to the implementation scenario. It should be noted that determining abnormal data through the isolation tree is a well-known prior art means for those skilled in the art, which will not be elaborated here.

[0098] Step S6: Correct all the noise data in the total set of pressure time series data to obtain the grouting monitoring transmission data of the total set of pressure time series data.

[0099] The actual segmentation value reflects the segmentation value to be processed corresponding to the best segmentation effect, increasing the difference between the noise data and the real data, so that the result of screening the noise data is more accurate. The noise data can more accurately reflect the actual noise data. In order to improve the compression ratio during data compression and improve the remote transmission efficiency, correct all the noise data in the total set of pressure time series data, thereby improving the continuity of the grouting pressure data, and further improving the data compression ratio of the grouting pressure, so that the grouting monitoring transmission data of the total set of pressure time series data can be transmitted more quickly and the remote transmission data efficiency can be improved.

[0100] After screening out all the noise data, it is necessary to correct the noise data to improve the compression ratio during data compression and improve the remote transmission efficiency. Based on the interpolation method, calculate the mean values of a preset number of data points before and after each noise data as the replacement data points for each noise data; in one embodiment of the present invention, the preset number is 3, and the implementer can set it according to the implementation scenario.

[0101] In the total set of pressure time series data, replace each noise data with each replacement data point to obtain the noise-reduced data of the total set of pressure time series data; compress the noise-reduced data based on run-length encoding to obtain the grouting monitoring transmission data of the total set of pressure time series data. Since interpolation operations are performed on the noise data, the continuity of the encoding is enhanced, and the grouting monitoring transmission data obtained at this time is smaller than the data directly compressed before denoising, effectively improving the efficiency and accuracy of remote transmission.

[0102] In summary, the embodiment of the present invention provides a method for remotely transmitting grouting monitoring data based on a cloud platform. First, obtain the change characteristic values corresponding to each data point in the total set of pressure time series data; in the process of constructing an isolation tree according to the binary partition of the data set to be processed, obtain the noise degree values of each pressure data partition set according to the set continuous value, set association value, and number of elements of the pressure data partition set; determine the actual segmentation value of the data set to be processed; correct all the noise data in the total set of pressure time series data to obtain the grouting monitoring transmission data of the total set of pressure time series data. In the embodiment of the present invention, by optimizing the selection of the actual segmentation value, the accuracy of identifying noise data and real data is improved, and the remote transmission data efficiency of the grouting pressure is improved.

[0103] The present invention provides a system for remotely transmitting grouting monitoring data based on a cloud platform, including a memory and a processor. The processor executes the calculation program stored in the memory to implement the method for remotely transmitting grouting monitoring data based on a cloud platform as described above.

[0104] It should be noted that the above order of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0105] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A remote transmission method for grouting monitoring data based on a cloud platform, characterized in that, The method includes: Obtaining the total set of pressure time series data, the total set of flow rate time series data, and the total set of density time series data during the grouting process; In the total set of pressure time series data, obtaining the change characteristic value corresponding to each data point in the total set of pressure time series data according to the degree of change of the pressure values around each data point; Sampling the total set of pressure time series data to obtain a set of data to be processed, and constructing each isolation tree according to each set of data to be processed, so as to obtain the noise data in the total set of pressure time series data. During the process of constructing the isolation tree according to the set of data to be processed, obtaining each to-be-processed segmentation value of the set of data to be processed according to the change characteristic value in the set of data to be processed; performing binary partitioning on the set of data to be processed according to the to-be-processed segmentation value to obtain two pressure data segmentation sets corresponding to each to-be-processed segmentation value; Obtaining the set continuity value of each of the pressure data segmentation sets according to the time difference of the data points in the pressure data segmentation set and the difference of the change characteristic values; obtaining the set correlation value of each of the pressure data segmentation sets according to the change differences among the pressure data segmentation set, the total set of flow rate time series data, and the total set of density time series data; Obtaining the noise degree value of each of the pressure data segmentation sets according to the set continuity value, the set correlation value, and the number of elements of the pressure data segmentation set; obtaining the preference degree value of each of the to-be-processed segmentation values of the set of data to be processed according to the difference in the noise degree values among all the pressure data segmentation sets corresponding to the to-be-processed segmentation value; determining the actual segmentation value of the set of data to be processed according to the preference degree value of each of the to-be-processed segmentation values of the set of data to be processed, and constructing the isolation tree of the set of data to be processed according to the actual segmentation value; Correcting all the noise data in the total set of pressure time series data to obtain the grouting monitoring transmission data of the total set of pressure time series data.

2. The remote transmission method for grouting monitoring data based on a cloud platform according to claim 1, characterized in that, The acquisition formula of the change characteristic value includes: Among them, is the change feature value of the s-th data point in the total set of pressure time series data; n is the total number of neighborhood data points in the preset neighborhood range of the s-th data point in the total set of pressure time series data; is the first-order difference of the i-th neighborhood data point in the preset neighborhood range of the s-th data point; is the second-order difference of the i-th neighborhood data point in the preset neighborhood range of the s-th data point; D is the number of positive and negative transformations of the first-order difference in the preset neighborhood range of the s-th data point; norm() is a normalization function.

3. The remote transmission method for grouting monitoring data based on a cloud platform according to claim 1, characterized in that, The acquisition method of the set continuity value includes: where ω is the set continuous value of the pressure data segmentation set; M is the total number of data points in the pressure data segmentation set; T j+1 is the time of the (j + 1)-th data point in the pressure data segmentation set; T j is the time of the j-th data point in the pressure data segmentation set; is the change characteristic value of the (j + 1)-th data point in the pressure data segmentation set; is the change characteristic value of the j-th data point in the pressure data segmentation set; exp() is the exponential function with the natural number e as the base.

4. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, characterized in that, The acquisition method of the set correlation value includes: Obtaining the data correlation value according to the data correlation value formula, and the data correlation value formula includes: wherein, δ r is the data association value of the r-th data point in the pressure data segmentation set; m is the total number of data points in the preset reference range of the r-th data point in the pressure data segmentation set; is the first-order difference of the q-th reference data point in the preset reference range of the r-th data point; is the first-order difference of the density data point at the same moment corresponding to the q-th reference data in the total set of density time series data in the preset reference range of the r-th data point; is the first-order difference of the flow data point at the same moment corresponding to the q-th reference data in the total set of flow time series data in the preset reference range of the r-th data point; is the second-order difference of the q-th reference data point in the preset reference range of the r-th data point; is the second-order difference of the density data point at the same moment corresponding to the q-th reference data in the total set of density time series data in the preset reference range of the r-th data point; is the second-order difference of the flow data point at the same moment corresponding to the q-th reference data in the total set of flow time series data in the preset reference range of the r-th data point; Calculating the mean value of the data correlation values of all the data points in the pressure data segmentation set to obtain the set correlation value of the pressure data segmentation set.

5. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, wherein, The acquisition method of the noise degree value includes: Obtaining the noise degree value of the pressure data segmentation set according to the set continuity value, the set correlation value, and the number of elements of the pressure data segmentation set; The set continuity value and the noise degree value are negatively correlated; the set correlation value and the noise degree value are negatively correlated; the number of elements and the noise degree value are negatively correlated.

6. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, characterized in that The acquisition method of the preference degree value includes: Calculating the absolute value of the difference in the noise degree values among all the pressure data segmentation sets corresponding to the to-be-processed segmentation value to obtain the first preference value of the to-be-processed segmentation value; Normalizing the first preference value to obtain the preference degree value of the to-be-processed segmentation value.

7. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, characterized in that, The acquisition method of the to-be-processed segmentation value includes: In the to-be-processed data set, obtain the maximum value of the change feature value and the minimum value of the change feature value; use all the change feature values that are between the minimum value of the change feature value and the maximum value of the change feature value, excluding the maximum value and the minimum value of the change feature value, as all the to-be-processed segmentation values.

8. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, characterized in that, The method for obtaining the two pressure data segmentation sets corresponding to the to-be-processed segmentation values includes: Perform binary partitioning on the to-be-processed data set according to the to-be-processed segmentation value, count all the data in the to-be-processed data set that is not greater than the to-be-processed segmentation value, and construct the first pressure data segmentation set corresponding to the to-be-processed segmentation value; Count all the data in the to-be-processed data set that is greater than the to-be-processed segmentation value, and construct the second pressure data segmentation set corresponding to the to-be-processed segmentation value; Use the first pressure data segmentation set and the second pressure data segmentation set as the two pressure data segmentation sets corresponding to the to-be-processed segmentation value.

9. The remote transmission method of grouting monitoring data based on a cloud platform according to claim 1, characterized in that The method for obtaining the grouting monitoring transmission data includes: Calculate the mean values of a preset number of data points before and after each noise data as the replacement data points for each noise data; In the total set of pressure time series data, use each replacement data point to replace each noise data to obtain the noise-reduced data of the total set of pressure time series data; Compress the noise-reduced data based on run-length encoding to obtain the grouting monitoring transmission data of the total set of pressure time series data.

10. A remote transmission system for grouting monitoring data based on a cloud platform, comprising a memory and a processor, characterized in that, The processor executes the calculation program stored in the memory to implement a remote transmission method for grouting monitoring data based on a cloud platform as described in any one of claims 1-9.

Images