A large building structure safety monitoring parameter configuration method and system, and a monitoring method and system

Abstract

The application discloses a large building structure safety monitoring parameter configuration method and system and a monitoring method and system, which comprises the following steps: collecting and preprocessing historical data of monitoring items; dividing the preprocessed historical data into multiple time windows, and acquiring the maximum change rate and the maximum change trend of the historical data; collecting real-time monitoring data of the monitoring items, comparing and analyzing the change rate of the real-time monitoring data and the maximum change rate of the historical data, and judging whether the parameter configuration value of the monitoring items is abnormal based on the comparison and analysis result; acquiring parameter configuration information that needs to be adjusted based on the abnormal judgment result and the maximum change trend of the historical data; and updating the parameter configuration of the monitoring items. The application realizes data trend change judgment, dynamic parameter configuration identification and automatic intelligent parameter configuration based on big data, and can solve the problems of abnormal parameter configuration relying on manual operation and poor adaptability in the prior automatic monitoring technology.

Description

Technical Field

[0001] This invention relates to the field of automatic monitoring technology, and in particular to a method and system for configuring safety monitoring parameters for large building structures, as well as a monitoring method and system. Background Technology

[0002] In the field of existing large-scale building structure safety automatic monitoring technology, the methods used in the industry for monitoring key monitoring items such as strain, surface displacement, internal displacement, tilt, and deflection, as well as configuring abnormal parameters, have certain limitations.

[0003] In identifying abnormal parameter configurations, the industry often relies on manual experience to set fixed judgment rules. Taking a bridge monitoring project as an example, technicians, based on experience from similar past projects, manually judge the normal range of monitored data. When the monitored strain data exceeds this range, it is judged as abnormal. However, the actual monitoring environment is complex and variable. Factors such as seasonal temperature changes and traffic flow differences can alter the normal strain range of bridge structures. Fixed judgment rules cannot adapt to these changes in a timely manner, easily leading to misjudgments or missed judgments.

[0004] Furthermore, parameter configuration updates currently rely heavily on manual operation. For example, in a large-scale building structural safety monitoring project, when data analysts discover abnormal fluctuations in displacement data, they need to manually log into the monitoring platform, find the corresponding monitoring parameters among numerous parameter setting options, and then modify the parameter values ​​based on their own experience to adjust the abnormal parameter configurations for that monitoring item (spurt threshold, moving average window, moving median window, derivative threshold, number of transition points, transition threshold, etc.). This method is extremely inefficient, and when faced with a large number of parameters and frequent data changes, manual operation is prone to errors, seriously affecting the real-time performance and accuracy of the monitoring system.

[0005] In summary, existing methods for configuring safety monitoring parameters or for monitoring large building structures have the following problems or shortcomings:

[0006] 1. Poor adaptability: Fixed judgment rules based on human experience cannot adapt to complex and ever-changing actual monitoring environments. For example, in bridge monitoring, factors such as seasonal temperature changes and traffic flow differences will dynamically change the normal strain range of the bridge structure. However, fixed rules cannot respond to these changes in real time. As a result, in situations such as sudden temperature changes causing thermal expansion and contraction of the bridge or sudden increases in traffic load, normal fluctuations are easily misjudged as abnormalities, or truly abnormal data is missed, leading to misjudgments or omissions and reducing the reliability of monitoring results.

[0007] 2. Inefficiency: Manual parameter configuration and updates are cumbersome and time-consuming when dealing with a large number of monitoring parameters and frequent data changes. Taking a large-scale building structure safety monitoring project as an example, when abnormal fluctuations in displacement data occur, data analysts need to manually log in to the monitoring platform and find and modify the corresponding parameters one by one from numerous options. In large-scale monitoring scenarios, this operation is difficult to meet the needs of real-time monitoring, resulting in delayed anomaly handling and an inability to respond to potential risks in a timely manner.

[0008] 3. Prone to Errors: Manual operation is greatly affected by subjective factors and fatigue, making it easy to make mistakes during complex parameter configuration and updates. For example, incorrect parameter selection or incorrect value input can lead to errors. These errors not only fail to effectively resolve abnormal monitoring data but may also trigger new false alarms or missed alarms, interfering with normal monitoring work. Incorrect parameter configuration may even cause the monitoring system to malfunction, seriously misleading the safety assessment of the monitored objects.

[0009] 4. Lack of self-optimization capability: Existing methods rely entirely on manual intervention, and the system itself lacks the ability to automatically optimize abnormal parameter configurations based on monitoring data characteristics and environmental changes. It cannot proactively learn historical data patterns, nor can it autonomously adjust judgment rules and parameter configurations when new monitoring scenarios or data patterns emerge, making it difficult to meet the growing demands for intelligence and precision in the field of automated monitoring. Summary of the Invention

[0010] The purpose of this invention is to address the shortcomings of the prior art by providing a method and system for configuring safety monitoring parameters for large building structures, as well as a monitoring method and system, to solve at least one of the problems of the prior art.

[0011] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0012] A method for configuring safety monitoring parameters for large building structures, characterized by including the following steps:

[0013] Step 1: Collect and preprocess historical structural safety monitoring data;

[0014] Step 2: Divide the preprocessed historical data into multiple time windows; based on the rate of change and trend of change of historical data within each time window, obtain the maximum rate of change and the maximum trend of change of historical data.

[0015] Step 3: Collect real-time monitoring data of the monitored items, divide the real-time monitoring data into time windows, and obtain the rate of change and trend information of the real-time monitoring data within the time windows; compare and analyze the rate of change of the real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determine whether there are any abnormalities in the parameter configuration values ​​of the monitored items; based on the abnormality judgment results and the maximum trend of historical data, obtain the parameter configuration information that needs to be adjusted.

[0016] Step 4: Based on the parameter configuration information that needs adjustment obtained in Step 3, update the parameter configuration of the monitoring items according to preset rules. The preset rules are as follows:

[0017] If no anomalies are found, the parameter configuration information to be adjusted is obtained based on the maximum change trend of historical data. The parameters to be configured include spurt threshold, moving average window and moving median window, derivative threshold, number of gradient points and gradient threshold.

[0018] If an anomaly is found, the correlation between the parameters of the monitored item and the anomaly monitoring results is analyzed, and the parameter configuration information that needs adjustment is optimized based on the correlation analysis results. As a preferred method, the adjusted parameter configuration information includes:

[0019] The spurious threshold is set to n times the standard deviation of historical data, where n is any positive number. The moving average window and the moving median window are set to a% of the number of granular windows corresponding to the range of the largest historical trend, where a is any positive number.

[0020] The derivative threshold is adjusted according to the slope k. If the rate of change of the slope k exceeds the preset value, the derivative threshold is reduced.

[0021] The number of gradient points is configured as the number of granular windows corresponding to the range of change of the largest historical trend of normal data.

[0022] The gradient threshold is configured as the difference between the first and last values ​​corresponding to the range of the largest historical change trend in normal data.

[0023] As a preferred approach, the analysis obtains the correlation between each parameter of the monitoring item and the abnormal monitoring results of the monitoring item, and the parameter configuration information that needs to be adjusted is obtained based on the correlation analysis results, including:

[0024] Analyze the correlation between the parameters of the monitoring items and the abnormal monitoring results of the monitoring items;

[0025] The fitness function F is set based on the results of correlation analysis, the misjudgment of normal data, and the omission of abnormal data;

[0026] The fitness function F is optimized, and the parameter configuration information corresponding to the minimum value of F is taken as the parameter configuration information that needs to be adjusted.

[0027] As a preferred method, the rate of change of historical data or the rate of change of real-time monitored data within the time window is obtained in the following way:

[0028] Fit the curve to obtain the trend of data changes within the time window;

[0029] Calculate the rate of change m based on the trend curve, where , This represents the maximum change in data within the time window. The length of the time window;

[0030] The data within the time window corresponds to historical data or real-time monitoring data. As a preferred embodiment, in step 2, obtaining the maximum trend of change in the historical data includes:

[0031] Step 201: Starting from the first historical data, continuously slide the time window backward, perform linear fitting on the historical data in each time window, and obtain the slope k of the fitted line of the historical data in each time window;

[0032] Step 202: Based on the sign of the slope k corresponding to the first time window and the difference between the slope k values ​​corresponding to the first time window and subsequent time windows, determine the maximum trend of change in historical data.

[0033] As a preferred embodiment, step 202 includes:

[0034] Determine whether the compliance rate meets the requirements. If so, continue sliding the time window forward and determine whether the compliance rate meets the requirements. Otherwise, take the time period from the first historical data to the historical data where the slope begins to change in the opposite direction as a trend window, and update the historical data where the slope begins to change in the opposite direction to the first historical data, and jump to step 201. Here, the slope change in the opposite direction means that the slope change trend changes from continuous increase to decrease, or the slope change trend changes from continuous decrease to increase.

[0035] After step 201 processes all historical data, the maximum value of all trend windows is compared and taken as the maximum trend Z of the historical data. max ;

[0036] The methods for determining whether the compliance rate meets the requirements include:

[0037] Take N sets of monitoring item values ​​corresponding to the continuous change of slope in the opposite direction, and calculate N sets of differences between the N sets of monitoring item values ​​and the last set of monitoring item values ​​before the slope changes in the opposite direction.

[0038] When k > 0, if the proportion of the number of differences greater than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements.

[0039] When k < 0, if the proportion of the number of differences less than 0 in the N groups of differences reaches the target proportion, then the compliance rate meets the requirements; otherwise, it does not meet the requirements.

[0040] As a preferred approach, in step 3, the change rate of real-time monitoring data and the maximum change rate of historical data are compared and analyzed. Based on the comparison and analysis results, it is determined whether the parameter configuration value of the monitoring item is abnormal. This includes: if the change rates of the real-time monitoring data for L1 consecutive periods all exceed A times the maximum change rate of the historical data, then it is determined that the parameter configuration value of the monitoring item is abnormal; otherwise, the parameter configuration value of the monitoring item is not abnormal; where L1 and A are both preset values.

[0041] Based on the same inventive concept, the present invention also provides a parameter configuration system, characterized by comprising:

[0042] Data acquisition module: used to collect historical data and real-time monitoring data of the monitored items;

[0043] Data preprocessing module: used to preprocess historical data of monitoring items;

[0044] Data Analysis Module:

[0045] This is used to divide preprocessed historical data into multiple time windows; based on the rate of change and trend of change of historical data within each time window, the maximum rate of change and maximum trend of change of historical data are obtained.

[0046] This tool is used to divide real-time monitoring data into time windows and obtain the rate of change and trend information of real-time monitoring data within the time window; it compares and analyzes the rate of change of real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determines whether the parameter configuration values ​​of the monitoring items are abnormal; based on the abnormality judgment results and the maximum change trend of historical data, it obtains the parameter configuration information that needs to be adjusted.

[0047] Parameter configuration adjustment module: Used to update the parameter configuration of the monitoring items based on the acquired parameter configuration information that needs to be adjusted.

[0048] Based on the same inventive concept, the present invention also provides a monitoring method, characterized in that the monitoring method includes updating the parameter configuration of the monitoring items using the large building structure safety monitoring parameter configuration method described above.

[0049] Based on the same inventive concept, this invention also provides a monitoring system, characterized in that the monitoring system includes the aforementioned parameter configuration system, which is used to update the parameter configuration of the monitoring items of the monitoring system. Compared with the prior art, this invention exhibits significant advantages in several aspects:

[0050] 1. Improve monitoring accuracy: Traditional technologies rely on simple filtering and manual judgment, which are insufficient for complex data and changing environments, leading to frequent misjudgments or missed detections. This invention incorporates a data preprocessing module to preprocess the data (e.g., using adaptive filtering combined with wavelet transform for noise reduction), which can more accurately remove data noise and ensure data quality. Furthermore, the anomaly detection method based on the maximum rate of change and the maximum trend of change in historical data can dynamically adjust the judgment criteria according to the actual operating status of the monitored items, significantly improving the accuracy of anomaly identification.

[0051] 2. Enhanced System Adaptability: Existing technologies rely on manual experience to set fixed rules, which cannot adapt to changes in the monitoring environment. This invention, through in-depth analysis of historical big data, automatically learns the changing patterns and trend characteristics of monitoring data. When the data characteristics change due to factors such as seasonal changes, traffic flow variations, or construction progress, this invention can automatically adjust abnormal parameter configurations (such as glitch thresholds, sliding window sizes, derivative thresholds, etc.), ensuring that the monitoring system always remains in optimal working condition, greatly enhancing its adaptability to different working conditions and environments.

[0052] 3. Improved Work Efficiency: In traditional monitoring methods, parameter configuration updates rely on manual operation, which is cumbersome and time-consuming. This invention achieves fully automated operation from data acquisition, processing, anomaly identification to parameter configuration updates. Once a data anomaly is detected, the system can quickly and automatically analyze and optimize the parameter configuration, and immediately send the new parameter configuration information to the monitoring platform to complete the update. This meets the high efficiency requirements of real-time monitoring and is especially suitable for large-scale monitoring projects with large amounts of data and frequent changes.

[0053] 4. Reduce the risk of human error: Manual operation is susceptible to subjective factors and fatigue, making errors prone to occur during complex parameter configuration. This invention completely eliminates manual intervention, automatically optimizing parameter configuration through algorithms and programs. This avoids operational errors such as incorrect parameter selection and incorrect value input, ensuring the accuracy and consistency of parameter configuration, effectively reducing monitoring misjudgments caused by human error, and improving the stability and reliability of the monitoring system.

[0054] 5. Achieving Intelligent Self-Optimization: Existing technologies lack self-optimization capabilities and cannot proactively adapt to new monitoring scenarios and data patterns. This invention, based on big data analysis and optimization algorithms, possesses self-learning and optimization capabilities. As monitoring data accumulates, it continuously improves anomaly detection and parameter configuration information, thereby continuously enhancing the performance of the monitoring system and providing crucial technical support for the intelligent development of automated monitoring. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of one implementation method of the method for configuring safety monitoring parameters for large building structures according to the present invention.

[0056] Figure 2 This is a block diagram of an embodiment of the parameter configuration system described in this invention. In the diagram, 1 is a data acquisition module, 2 is a data preprocessing module, 3 is a data analysis module, and 4 is a parameter configuration adjustment module.

[0057] Figure 3 This is a schematic diagram of historical data for bridge strain monitoring, showing the maximum rate of change in historical data identified using this invention.

[0058] Figure 4 This is a schematic diagram of historical data for bridge strain monitoring, showing the maximum trend of change in historical data identified using this invention.

[0059] Figure 5 To illustrate the identification of bridge strain monitoring data before and after parameter configuration updates using this invention, the figure shows the data before the update process.

[0060] Figure 6 To illustrate the identification of bridge strain monitoring data before and after parameter configuration updates using this invention, the figure shows the data after the update process.

[0061] Figure 7 To illustrate the identification of bridge strain monitoring data before and after parameter configuration updates using this invention, the figure shows the abnormal data identified using this invention. Detailed Implementation

[0062] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments are clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0063] This invention aims to solve the technical problems in the field of existing automatic monitoring, such as excessive reliance on human experience for abnormal parameter configuration, poor adaptability, low efficiency, and susceptibility to errors. In traditional methods, fixed judgment rules are difficult to adapt to complex and ever-changing monitoring environments. Manually updating parameter configurations is not only inefficient but also prone to errors. Furthermore, the system lacks self-optimization capabilities and cannot meet the growing intelligent demands of automated monitoring.

[0064] This invention is based on big data to determine data trend changes, dynamically identify parameter configurations, and automatically and intelligently configure parameters, aiming to solve the problems of existing automatic monitoring technologies, such as reliance on manual configuration of abnormal parameters and poor adaptability.

[0065] like Figure 1 and Figure 2 As shown, the first aspect of this invention provides a method for configuring safety monitoring parameters for large building structures and a corresponding parameter configuration system.

[0066] One embodiment of the method for configuring safety monitoring parameters for large building structures according to the present invention includes the following steps:

[0067] Step 1: Collect and preprocess historical data for structural safety monitoring items.

[0068] Data collection:

[0069] The monitoring system includes various sensors, such as strain sensors, displacement sensors, and tilt sensors, which collect real-time monitoring data on strain, surface displacement, internal displacement, tilt, and deflection. The sensors transmit data to the data acquisition module 1 via wired or wireless communication.

[0070] Data preprocessing:

[0071] According to the system's data verification rules, the data undergoes rigorous checks, including both completeness and validity. Checking data completeness confirms the absence of missing values; checking data validity determines whether data values ​​are within a reasonable range—for example, strain values ​​must not exceed the range of the corresponding sensor. Invalid data is immediately discarded. Subsequently, the data undergoes de-glitching, a crucial step in ensuring data quality. De-glitching preferably employs, but is not limited to, adaptive filtering methods, dynamically adjusting the data filtering range based on the data's inherent characteristics to remove interference from other factors. In a preferred embodiment, de-glitching includes processing all data in the acquired data set... Perform data difference calculation, where, For the first data group to be collected One data point, The value range is from 1 to n. The data difference is calculated by comparing the difference between the next granularity of data and the previous granularity of data. For the nth... Data points Calculate its relationship with the previous data point The difference between them, that is, The data groups with data differences can be obtained. , If the difference between two consecutive data points All in Besides ( The average of the data difference groups. If the standard deviation of the data group is used, the corresponding data is identified as abnormal spike data. After filtering and processing using an optimized algorithm, the cleaned data is obtained, and the spike threshold is recorded. Step 2: Divide the preprocessed historical data into multiple time windows; based on the rate of change and trend of change of historical data within each time window, obtain the maximum rate of change and the maximum trend of change of historical data.

[0072] Step 2 preferably includes, but is not limited to:

[0073] Time window division:

[0074] Collect preprocessed historical data (such as preprocessed data from the previous year), and divide this data into multiple time windows according to chronological order and set time periods (which can be freely divided, such as minutes, hours, days, etc.). For example, using days, the data is divided into 365 time windows in chronological order. Each time window contains a certain number of data points. During actual monitoring, the size of the time window can be flexibly adjusted according to actual monitoring needs.

[0075] Polynomial fitting: For the data within each time window, a polynomial fitting method is used to construct a trend curve of the data changes. The polynomial fitting process preferably includes, but is not limited to, performing multiple polynomial fittings using the least squares method for the data within each time window. Taking a 3rd-order polynomial fitting as an example, assuming the number of data points within the time window is 200... The data point value corresponding to time is The fitted polynomial is ,in, For a moment The corresponding fitted value. This is achieved by minimizing the error function. ( for (Corresponding fitted values), calculate the polynomial coefficients a0, a1, a2, a3, and obtain the data change trend curve. Rate of change calculation:

[0076] The rate of change *m* of the data within each time window is calculated based on the trend curve. This rate of change reflects how quickly the data changes within that time period. This represents the maximum change in data within the time window. This represents the length of the time window. Compare the rates of change across all time windows and find the maximum value; this represents the maximum rate of change of the data. Identifying the largest trends in historical data:

[0077] The slope k of the changing curve is used to determine the increase or decrease of data, thereby determining the maximum change trend of historical data. This trend will serve as an important reference for subsequent abnormal parameter configuration.

[0078] Identifying the largest trends in historical data preferably includes, but is not limited to:

[0079] Step 201: Take historical preprocessed data (such as historical preprocessed data from the previous year), and continuously slide the time window backward from the first historical data point, performing linear fitting on the historical data within each time window. ,in, Historical data within each time window The corresponding linear fit value, The slope of the straight line fitted to historical data within each time window. Step 202 involves fitting the intercept of a straight line to the historical data within each time window. Based on the sign of the slope k corresponding to the first time window (k < 0, function decreasing; k > 0, function increasing), and the difference between the slope k values ​​corresponding to the first and subsequent time windows, the maximum trend of change in the historical data is determined. When the slope continuously shifts in one direction, but the slope changes in the opposite direction for N consecutive time windows, and the first slope reverses, a comparison is made based on the sign of the slope k.

[0080] Step 202 preferably includes, but is not limited to, determining whether the compliance rate meets the requirements. If so, the time window is continued to slide backward and the compliance rate is determined again. Otherwise, the time period between the first historical data and the historical data where the slope begins to change in the opposite direction is taken as a trend window, and the historical data where the slope begins to change in the opposite direction is updated to the first historical data, and the process jumps to step 201. Here, the slope change in the opposite direction means that the slope change trend changes from continuously increasing to decreasing, or the slope change trend changes from continuously decreasing to increasing.

[0081] After step 201 processes all historical data, the maximum value of all trend windows is compared and taken as the maximum trend Z of the historical data. max The methods for determining whether the compliance rate meets the requirements include:

[0082] Take N sets of monitoring item values ​​corresponding to the continuous change of slope in the opposite direction, and calculate N sets of differences between the N sets of monitoring item values ​​and the last set of monitoring item values ​​before the slope changes in the opposite direction.

[0083] When k > 0, if the proportion of the number of differences greater than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements.

[0084] When k < 0, if the proportion of differences less than 0 in the N groups of differences reaches the target proportion, the compliance rate meets the requirements; otherwise, it does not. The following example of strain monitoring illustrates how to obtain the maximum trend of historical data in a specific application case: the slope k continuously changes in one direction within the previous time window. When k changes in the opposite direction for five consecutive groups, such as the currently fitted slope k = 0.678 (k increasing direction), the slope decreases for the next five consecutive granularities (adjustable granularity windows), such as 0.665, 0.659, 0.643, 0.631, 0.622 (k continuously decreasing). If the slope k > 0 / k < 0, then the strain value corresponding to the particle size when the slope changes in the opposite direction minus the strain value corresponding to the last group of particle size changes in the same direction as the slope is > 0 / < 0 (5 groups of data reach 60%, where the number of groups, i.e., the target percentage, can be adjusted according to actual needs), then the trend identification continues; if the target rate is not met, then the first data point until the slope changes in the opposite direction is taken as a trend window. The next trend window begins from the first time window where the slope changes in the opposite direction, and so on. ... Find the biggest historical trend. It will automatically select the corresponding number of time windows (data granularity) and data range. Specifically, the processing procedure and effect of this invention will be illustrated using bridge strain monitoring as an example, such as... Figure 3-4 As shown, Figure 3 This is a schematic diagram of historical data for bridge strain monitoring, showing the maximum rate of change in historical data identified using this invention. , Figure 4 The figure shows the maximum change trend of historical data identified using the present invention. .

[0085] Step 3: Collect real-time monitoring data of the monitored items, divide the real-time monitoring data into time windows, and obtain the rate of change and trend information of the real-time monitoring data within the time windows; compare and analyze the rate of change of the real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determine whether there are any abnormalities in the parameter configuration values ​​of the monitored items; based on the abnormality judgment results and the maximum change trend of historical data, obtain the parameter configuration information that needs to be adjusted.

[0086] In step 3, based on the anomaly detection results and the maximum change trend of historical data, the parameter configuration information that needs to be adjusted is obtained, preferably including, but not limited to:

[0087] If no anomalies are found, the parameter configuration information that needs to be adjusted is obtained based on the maximum change trend of historical data;

[0088] If an anomaly is found, the correlation between the parameters of the monitored item and the abnormal monitoring results of the monitored item is analyzed, and the parameter configuration information that needs to be adjusted is obtained based on the correlation analysis results.

[0089] The analysis obtains the correlation between each parameter of the monitoring item and the abnormal monitoring results of the monitoring item. Based on the correlation analysis results, the parameter configuration information that needs to be adjusted is preferably, but not limited to, the following:

[0090] Analyze the correlation between the parameters of the monitoring items and the abnormal monitoring results of the monitoring items;

[0091] The fitness function F is set based on the results of correlation analysis, the misjudgment of normal data, and the omission of abnormal data;

[0092] The fitness function F is optimized, and the parameter configuration information corresponding to the minimum value of F is taken as the parameter configuration information that needs to be adjusted.

[0093] In step 3, the rate of change of real-time monitoring data and the maximum rate of change of historical data are compared and analyzed. Based on the comparison and analysis results, it is determined whether the parameter configuration value of the monitoring item is abnormal. Preferably, but not limited to, the following is determined: if the rate of change of L1 consecutive real-time monitoring data exceeds A times the maximum rate of change of historical data, then the parameter configuration value of the monitoring item is determined to be abnormal; otherwise, the parameter configuration value of the monitoring item is not abnormal; wherein, L1 and A are preset values.

[0094] Step 3 involves identifying abnormal parameter configurations. Based on the parameter change characteristics presented by the maximum change trend in historical data, such as the range and rate of change of strain under the maximum change trend, and combined with preset adjustment rules, the abnormal parameter configuration items that need to be adjusted are determined, including spur threshold, moving average window, moving median window, derivative threshold, number of gradient points, and gradient threshold. Then, these parameters are adjusted and optimized using a specific optimization algorithm to obtain the parameter configuration scheme most suitable for the current monitoring situation. In a specific application case, step 3 includes:

[0095] Real-time data processing: For real-time monitoring data, the same processing method as for historical data is applied, dividing it into time windows and performing polynomial fitting and rate of change calculation. Anomaly detection: When the rate of change of real-time data exceeds the maximum rate of change of historical data... If the change trend is A times (the multiple A is adjustable, e.g., A is 1.2), and this trend continues for L1 time windows (the number of time windows L1 is adjustable, e.g., L1 is 5), then the current data is considered abnormal. Parameter configuration adjustment (when the current data is not abnormal): Based on the parameter change characteristics involved in the maximum change trend of historical data (e.g., the range of strain change under the maximum change trend to determine the corresponding number of gradient points and gradient threshold), combined with preset rules (e.g., the faster the change rate, the smaller the derivative threshold; the larger the change range, the larger the gradient threshold), determine the abnormal parameter configuration items to be adjusted (spur threshold 3σ, moving average window, moving median window, derivative threshold, number of gradient points, gradient threshold, etc.) and the adjustment direction. Specific adjustment rules can be: 1) Spur threshold

[0096] Rule: Use three times the standard deviation of historical data as the spurious threshold.

[0097] Setting method: Burr threshold = 3σ

[0098] 3σ means that 99.7% of normal data will fall within this range. Any sudden, transient abrupt change exceeding this threshold will be identified as a "spurt" anomaly, meaning that point anomalies are captured.

[0099] 2 / 3) Moving mean window and moving median window

[0100] Rule: 5% (or other percentage value) of the granularity window corresponding to the range of the historical maximum change trend.

[0101] Setting method: Sliding window size = N_zmax * α (where α is a scaling factor, for example, 5%, i.e., 0.05)

[0102] N_zmax represents the time span required for the system to complete one "maximum normal change". The window size must be much smaller than this time span (e.g., 5%) to ensure: Sensitivity: The window is small enough to keenly capture the trend of the change starting to occur.

[0103] Noise resistance: The window cannot be too small (otherwise it will be the original data), and it can effectively smooth out small, meaningless fluctuations.

[0104] This window is used to filter random noise, providing a smoother data basis for calculating derivatives and gradients.

[0105] 4) Derivative threshold

[0106] Rule: Adjust according to the slope k (the faster the change, the smaller the derivative threshold).

[0107] Setting method: Derivative threshold = ω / |k_zmax| (ω is a coefficient adjusted based on experience. To reduce false alarms, ω is generally between 1.2 and 1.8, and can be adjusted according to the actual situation)

[0108] k_zmax represents the largest historical trend. The fastest rate of change. The faster the rate of change, the lower the derivative threshold should be set (for greater sensitivity). This is because if normal changes are very fast, then a change slightly slower than k_zmax might also be normal; however, if a change's rate is close to or even exceeds the historical fastest rate k_zmax, it is highly suspicious and needs to be detected. Lowering the threshold can improve the sensitivity of detecting abnormally high-speed changes.

[0109] 5) Number of gradient points

[0110] Rule: The number of granular windows corresponding to the range of the largest historical change trend of normal data.

[0111] Setting method: Number of gradient points = N_zmax

[0112] This configuration rule aims to detect "duration anomalies." For an anomalous gradual change to "masquerade" as a normal change, it needs to last at least N_zmax to produce an impact comparable to the historical maximum normal change (Δ_zmax). Therefore, setting the detection window directly to N_zmax means the system will continuously monitor for a period of N_zmax. If, within this entire period, the cumulative change in data exceeds the expected value based on historical learning (Δ_zmax), an anomaly is reported.

[0113] 6) Gradient threshold

[0114] Rule: The difference between the first and last values ​​corresponding to the range of the largest historical change trend of normal data.

[0115] Setting method: Gradient threshold = Δ_zmax

[0116] This is used in conjunction with the previous configuration rule, defining the maximum cumulative change allowed within N_zmax time windows. Any trend with a change exceeding Δ_zmax within the same or shorter time frame will be considered an abnormal change. Specifically, the processing procedure and effects of this invention will be illustrated using bridge strain monitoring as an example. Figure 5 To illustrate the identification of bridge strain monitoring data before and after parameter configuration updates using this invention, the data before the update process is shown. Figure 6 The image shows the data after the update process. Figure 7 The image shows anomaly data identified using this invention.

[0117] Parameter configuration optimization and adjustment (when current data contains anomalies): After discovering data anomalies, conduct correlation analysis. By directly linking parameters with anomaly results, focus on the impact of individual parameters on anomaly identification. Quantitative analysis clarifies the correlation between "parameter adjustment" and "number of false positives / false negatives." For example, use the Pearson correlation coefficient to calculate the correlation strength between the adjustment amount of the derivative threshold and the number of false negatives. If the absolute value of the correlation coefficient is ≥0.7, it indicates a strong correlation, meaning the derivative threshold is a core parameter affecting false negatives, indicating a high false negative weight index. If not, continue to correlate with other parameters.

[0118] Parameters are optimized using genetic algorithms, and a fitness function is defined. Among them, the number of times normal data was misjudged and the number of times abnormal data was missed were both determined by manual review. , This is the weighting coefficient, which is adjusted according to the correlation. For example, if the absolute value of the correlation coefficient between the parameter and the number of missed detections is ≥0.7, then it can be set to the value corresponding to the parameter. Specifically, this means paying more attention to the number of missed detections. Through genetic operations such as selection, crossover, and mutation, the system continuously iterates and optimizes, comparing the size of F (selecting the best combination with the smaller F) to obtain the optimal parameter configuration.

[0119] Specifically,

[0120] 1) First stage: Correlation analysis (diagnostic stage)

[0121] Objective: To identify the key parameters that cause the "false positive / false negative" problem and prioritize them for subsequent optimization.

[0122] Specific steps:

[0123] Data collection: Run the system with the current parameter configuration, record the anomaly identification results over a period of time, and compare them with the actual results of manual review. Record the "number of false positives" and "number of false negatives" at each time point.

[0124] Perturbation parameters: Systematically and slightly adjust each parameter to be optimized (e.g., increase or decrease the derivative threshold by 5%; increase or decrease the window size by 10 granularities, etc.). Only one parameter is changed at a time, keeping the other parameters unchanged, and record the number of false positives / false negatives after each adjustment.

[0125] Quantitative correlation: For each parameter (e.g., derivative threshold), calculate the Pearson correlation coefficient between its adjustment amount (the proportion of change relative to the original value) and the change in the number of misjudgments / missed judgments.

[0126] Pearson correlation coefficient: measures the degree of linear correlation between two variables, with a range of [-1, 1].

[0127] This is a preferred embodiment of the present invention, but it does not mean that the present invention can only be set in this way. For example, it can be set that when the absolute value is ≥ 0.7: strong correlation.

[0128] Absolute value around 0.5: moderate correlation.

[0129] Absolute value ≤ 0.3: weak correlation.

[0130] Diagnostic conclusion:

[0131] If the absolute value of the correlation coefficient between the adjustment of the derivative threshold and the change in the number of missed detections is ≥ 0.7 (strong negative correlation), this means that lowering the derivative threshold significantly reduces the number of missed detections. Conversely, raising the threshold also reduces the number of missed detections. This confirms that the derivative threshold is the core parameter causing the missed detection problem and needs to be optimized.

[0132] 2) Second stage: Genetic algorithm optimization (solution stage)

[0133] Objective: Based on the diagnostic results of the first phase, automatically find the optimal set of parameter configurations.

[0134] Individual: A specific set of parameter configurations. For example: Individual 1 = {Spurt threshold: 3.1σ, Sliding window: 15, Derivative threshold: 0.08, ...}

[0135] Population: A collection of multiple sets of different parameter configurations.

[0136] Fitness function: This is the core of the genetic algorithm, used to evaluate the quality of a set of parameters. The smaller the function value F, the better the set of parameters.

[0137] F =

[0138] Key point: Weighting coefficient and It is not fixed, but dynamically set based on the correlation analysis results of the first stage.

[0139] If a strong correlation is found between the derivative threshold and the number of missed detections (|r|≥0.7), then during optimization, we should assign a higher weight to the number of missed detections (i.e., increase the weight of r). This allows the algorithm to prioritize finding solutions that reduce missed detections.

[0140] Similarly, if the glitch threshold is strongly correlated with the number of false positives, then increase it. .

[0141] In this way, the fitness function F becomes a focused and targeted optimization objective.

[0142] Genetic manipulation:

[0143] Selection: Select superior individuals with low F-values ​​(high fitness) from the current population, giving them a greater probability of "reproducing" the next generation.

[0144] Crossover: mimicking gene recombination. For example, combining parameters (such as the window size of the parent and the derivative threshold of the mother) of two superior parent individuals to create a new offspring individual.

[0145] Mutation: By randomly changing a parameter value of an individual with a small probability (for example, fine-tuning the number of gradient points of an individual from 50 to 52), diversity is introduced to avoid the algorithm getting stuck in local optima.

[0146] Iteration: The process of "calculating fitness -> selection -> crossover -> mutation" is repeated to produce generation after generation of the population. Each generation tends to find a better solution.

[0147] Termination and Output: The algorithm stops when the maximum number of iterations is reached, or when the F-value of the optimal individual no longer improves significantly over several generations. The algorithm outputs the individual with the smallest F-value in the current population, representing the optimal parameter configuration.

[0148] Step 4: Based on the parameter configuration information requiring adjustment obtained in Step 3, update the parameter configuration of the monitoring items. In some preferred embodiments, this specifically involves sending the optimized abnormal parameter configuration information to the monitoring platform and accurately storing the new parameter configuration information in the database. Simultaneously, the monitoring platform will update the front-end display interface and data processing logic, enabling the entire monitoring system to operate based on the new parameter configuration scheme, achieving automatic parameter configuration updates, and thus more accurately monitoring the monitored objects in real time.

[0149] like Figure 2 As shown, the parameter configuration system provided by this invention corresponds to the inventive concept of the above-mentioned parameter configuration method for safety monitoring of large building structures. The parameter configuration system preferably includes, but is not limited to, the following:

[0150] Data acquisition module 1: Used to collect historical data and real-time monitoring data of the monitored items;

[0151] Data preprocessing module 2: Used to preprocess historical data of monitoring items;

[0152] Data Analysis Module 3: This module divides the preprocessed historical data into multiple time windows; based on the rate of change and trend of change of the historical data within each time window, it obtains the maximum rate of change and the maximum trend of change of the historical data.

[0153] This tool is used to divide real-time monitoring data into time windows and obtain the rate of change and trend information of real-time monitoring data within the time window; it compares and analyzes the rate of change of real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determines whether the parameter configuration values ​​of the monitoring items are abnormal; based on the abnormality judgment results and the maximum change trend of historical data, it obtains the parameter configuration information that needs to be adjusted.

[0154] Parameter configuration adjustment module 4: Used to update the parameter configuration of the monitoring items based on the acquired parameter configuration information that needs to be adjusted.

[0155] The aforementioned data acquisition module 1, data preprocessing module 2, data analysis module 3, and parameter configuration adjustment module 4 are used to implement the above-mentioned method for configuring parameters for safety monitoring of large building structures.

[0156] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0157] This invention also provides a monitoring method, which includes updating the parameter configuration of monitoring items using the above-described method for configuring safety monitoring parameters of large building structures.

[0158] This invention also provides a monitoring system, which includes the parameter configuration system described above. The parameter configuration system is used to update the parameter configuration of the monitoring items of the monitoring system.

[0159] The parts of the parameter configuration system, monitoring method, and monitoring system that correspond to the inventive points of the parameter configuration method for safety monitoring of large building structures will not be elaborated here, but this will not affect the understanding and implementation of the present invention by those skilled in the art.

[0160] The concept of this invention is based on big data technology to construct an intelligent parameter configuration system. The system deeply mines and analyzes massive amounts of historical monitoring data, using data processing algorithms to remove spikes and ensure data quality; it employs data recognition algorithms to accurately identify abnormal patterns in the data; and based on the maximum change trend of historical data, it automatically determines and adjusts abnormal parameter configurations (including spike thresholds, moving average windows, moving median windows, derivative thresholds, number of transition points, transition thresholds, etc.) without manual intervention, achieving dynamic optimization and intelligent updating of parameter configurations.

[0161] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0162] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0163] In the embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the method and system embodiments described above are merely illustrative. For instance, the division of modules or units is merely a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0164] The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0165] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0166] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0167] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for configuring safety monitoring parameters for large building structures, characterized in that, Includes the following steps: Step 1: Collect and preprocess historical data for structural safety monitoring items; Step 2: Divide the preprocessed historical data into multiple time windows; Based on the rate of change and trend of historical data within each time window, obtain the maximum rate of change and maximum trend of historical data, including: Step 201: Starting from the first historical data, continuously slide the time window backward, perform linear fitting on the historical data in each time window, and obtain the slope k of the fitted line of the historical data in each time window; Step 202: Based on the sign of the slope k corresponding to the first time window and the difference between the slope k values ​​corresponding to the first and subsequent time windows, determine the maximum trend of change in historical data; including: Determining whether the compliance rate meets the requirements includes: taking N sets of monitoring item values ​​corresponding to the continuous change of the slope in the opposite direction, and calculating N sets of differences between the N sets of monitoring item values ​​and the last set of monitoring item values ​​before the slope changes in the opposite direction; When k > 0, if the proportion of the number of differences greater than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements. When k < 0, if the proportion of the number of differences less than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements. If yes, continue sliding the time window backward and determine whether the target rate is met; otherwise, take the time period between the first historical data and the historical data where the slope begins to change in the opposite direction as a trend window, update the historical data where the slope begins to change in the opposite direction to the first historical data, and jump to step 201. After step 201 processes all historical data, the maximum value of all trend windows is compared and taken as the maximum trend Z of the historical data. max ; Step 3: Collect real-time monitoring data of the monitored items, divide the real-time monitoring data into time windows, and obtain the rate of change and trend information of the real-time monitoring data within the time windows; compare and analyze the rate of change of the real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determine whether there are any abnormalities in the parameter configuration values ​​of the monitored items; based on the abnormality judgment results and the maximum trend of historical data, obtain the parameter configuration information that needs to be adjusted. Step 4: Based on the parameter configuration information that needs adjustment obtained in Step 3, update the parameter configuration of the monitoring items according to preset rules. The preset rules are as follows: If no abnormality is found, the parameter configuration information to be adjusted is updated based on the maximum change trend of historical data. The parameter configuration information to be adjusted includes spurt threshold, moving average window and moving median window, derivative threshold, number of gradient points and gradient threshold. If an anomalies are found, the correlation between the parameters of the monitored items and the anomaly monitoring results is analyzed. Based on the correlation analysis results, the misjudgment of normal data, and the omission of anomaly data, a fitness function F = is set. , in These are the weighting coefficients; The fitness function F is optimized, and the parameter configuration information corresponding to the minimum value of F is taken as the parameter configuration information that needs to be adjusted.

2. The method for configuring safety monitoring parameters for large building structures according to claim 1, characterized in that, The steps for updating the parameter configuration information that needs adjustment based on the maximum change trend of historical data specifically include: The spurious threshold is to use n times the standard deviation of historical data as the new spurious threshold, where n is any positive number; The moving average window and the moving median window represent a% of the number of granular windows corresponding to the range of the historical maximum change trend, where a is any positive number. The derivative threshold is adjusted according to the slope k. If the rate of change of the slope k exceeds the preset value, the derivative threshold is reduced. The number of gradient points is configured as the number of granular windows corresponding to the range of change of the largest historical trend of normal data. The gradient threshold is configured as the difference between the first and last values ​​corresponding to the range of the largest historical change trend in normal data.

3. The method for configuring safety monitoring parameters for large building structures according to claim 2, characterized in that, The rate of change of historical data or the rate of change of real-time monitored data within a time window is obtained in the following ways: Fit the curve to obtain the trend of data changes within the time window; Calculate the rate of change m based on the trend curve, where , This represents the maximum change in data within the time window. The time window is the length of the time window; the data within the time window corresponds to historical data or real-time monitoring data.

4. The method for configuring safety monitoring parameters for large building structures according to any one of claims 2 to 3, characterized in that, In step 3, the rate of change of real-time monitoring data and the maximum rate of change of historical data are compared and analyzed. Based on the comparison and analysis results, it is determined whether the parameter configuration values ​​of the monitoring item are abnormal, including: If the L1 consecutive rates of change of the real-time monitoring data all exceed A times the maximum rate of change of the historical data, then it is determined that the parameter configuration value of the monitoring item is abnormal; otherwise, the parameter configuration value of the monitoring item is not abnormal; where L1 and A are both preset values.

5. A large-scale building structure safety monitoring parameter configuration system, characterized in that, include: Data acquisition module: used to collect historical data and real-time monitoring data of the monitored items; Data preprocessing module: used to preprocess historical data of monitoring items; Data Analysis Module: Used to divide preprocessed historical data into multiple time windows; Based on the rate of change and trend of historical data within each time window, obtain the maximum rate of change and maximum trend of historical data, including: Starting from the first historical data point, continuously slide the time window backward, perform linear fitting on the historical data within each time window, and obtain the slope k of the fitted line for the historical data within each time window; Based on the sign of the slope k corresponding to the first time window, and the difference between the slope k values ​​corresponding to the first and subsequent time windows, determine the maximum trend of historical data; including: Determining whether the compliance rate meets the requirements includes: taking N sets of monitoring item values ​​corresponding to the continuous change of the slope in the opposite direction, and calculating N sets of differences between the N sets of monitoring item values ​​and the last set of monitoring item values ​​before the slope changes in the opposite direction; When k > 0, if the proportion of the number of differences greater than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements. When k < 0, if the proportion of the number of differences less than 0 in the N groups of differences reaches the target proportion in the total number of differences, then the compliance rate meets the requirements; otherwise, it does not meet the requirements. If yes, continue sliding the time window backward and determine whether the target rate is met; otherwise, take the time period between the first historical data and the historical data where the slope begins to change in the opposite direction as a trend window, update the historical data where the slope begins to change in the opposite direction to the first historical data, and jump to the linear fitting step. After processing all historical data, the maximum value of all trend windows is compared and taken as the maximum trend Z of the historical data. max ; This is used to collect real-time monitoring data of the monitored items, divide the real-time monitoring data into time windows, and obtain the rate of change and trend information of the real-time monitoring data within the time windows; compare and analyze the rate of change of the real-time monitoring data with the maximum rate of change of historical data, and based on the comparison and analysis results, determine whether there are any abnormalities in the parameter configuration values ​​of the monitored items; based on the abnormality judgment results and the maximum trend of historical data, obtain the parameter configuration information that needs to be adjusted. Parameter configuration adjustment module: used to update the parameter configuration of the monitoring items according to preset rules based on the acquired parameter configuration information that needs to be adjusted. The preset rules are as follows: If no anomalies are found, the parameter configuration information to be adjusted is updated based on the maximum change trend of historical data. This adjusted parameter configuration information includes the glitch threshold, moving average window and moving median window, derivative threshold, number of gradient points, and gradient threshold. If anomalies are found, the correlation between each parameter of the monitored item and the abnormal monitoring results of the monitored item is analyzed. Based on the correlation analysis results, the misjudgment of normal data, and the omission of abnormal data, a fitness function F = ... ,in The weighting coefficients are used to optimize the fitness function F, and the parameter configuration information corresponding to the minimum value of F is taken as the parameter configuration information that needs to be adjusted.

6. A method for monitoring the safety of large building structures, characterized in that, The monitoring method includes updating the parameter configuration of the monitoring items using the parameter configuration method for safety monitoring of large building structures as described in any one of claims 1 to 4.

7. A large-scale building structure safety monitoring system, characterized in that, The monitoring system includes the parameter configuration system as described in claim 5, which is used to update the parameter configuration of the monitoring items of the monitoring system.

Images