Foundation pit safety early-warning method based on multi-source monitoring data fusion

Abstract

Disclosed in the present invention is a foundation pit safety early-warning method based on multi-source monitoring data fusion. The method comprises: first constructing a foundation pit safety evaluation indicator system, and acquiring cumulative values and change rates of multi-source time-series monitoring indicators; then determining a foundation pit safety grade identification framework, using multi-source time-series monitoring data as different evidence, and performing normalization processing; then, on the basis of the multi-source time-series monitoring data, determining basic probability assignment values in the identification framework; then, separately calculating weights of multi-source monitoring indicators, and credibility; and finally, fusing multi-source monitoring data of a foundation pit, in order to obtain a foundation pit safety evaluation grade. The present invention is characterized in that the importance of different monitoring indicators and the credibility of evidence are taken into consideration while making full use of actually measured monitoring data of a foundation pit, thereby ensuring that foundation pit safety state identification is objective, avoiding interference from subjective human factors, and effectively overcoming the defects of a traditional D-S evidence theory.

Description

A method for early warning of foundation pit safety through multi-source monitoring data fusion Technical Field

[0001] This invention belongs to the field of foundation pit safety early warning technology, and more specifically, it relates to a foundation pit safety early warning method that integrates multi-source monitoring data. Background Technology

[0002] As a crucial component of civil engineering construction, the safety of foundation pits directly impacts the lives of construction workers and the stability of the surrounding environment. With rapid urbanization, the scale and depth of foundation pit projects are constantly increasing, making foundation pit safety issues increasingly prominent. Due to the influence of factors such as site location, surrounding environment, and soil variations, the internal forces of the foundation pit support change during excavation, leading to phenomena such as soil heave at the bottom of the pit, deformation of roads and pipelines, and uneven settlement and lateral displacement of surrounding buildings. When the deformation reaches a certain value, the support structure fails, the foundation pit becomes unstable and collapses, causing related economic losses and casualties.

[0003] Safety monitoring is one of the most commonly used and effective risk management methods in foundation pit engineering. However, the phenomenon of "emphasizing data collection but neglecting analysis" has always existed in actual projects, resulting in monitoring work failing to play its due role. Furthermore, the monitoring information currently obtained from foundation pit engineering not only exhibits significant spatiotemporal differences but also varies greatly in type and source, thus displaying multi-source heterogeneous characteristics. Traditional data analysis methods often employ threshold analysis based on single indicators (such as displacement or deformation rate of key points), reflecting only a local performance of the foundation pit from a single perspective, which is one-sided and leads to poor accuracy in foundation pit safety early warning.

[0004] To address the above problems, this invention proposes a multi-source monitoring data fusion method for early warning of foundation pit safety. This method integrates and applies different monitoring data to comprehensively and rationally assess the safety status of the foundation pit, thereby avoiding information blind spots and improving information integrity.

[0005] Summary of the Invention

[0006] To address the problems existing in current technologies, the purpose of this invention is to provide a method for early warning of foundation pit safety based on multi-source monitoring data fusion. By employing the interquartile range (ICM) method to calculate the weights of multi-source monitoring indicators and using Chebyshev distance to assess the credibility of different pieces of evidence, this method combines the advantages of both approaches, improving the overall accuracy and robustness of the assessment. Weight allocation ensures that the importance of different monitoring indicators is reasonably reflected in subsequent analysis; while credibility assessment helps to screen reliable evidence sources, avoiding the influence of erroneous or misleading information on the analysis results.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for early warning of foundation pit safety based on multi-source monitoring data fusion;

[0008] S1. Based on the characteristics of foundation pit engineering and the features of safety monitoring, construct a foundation pit safety evaluation index system and obtain the cumulative value and rate of change of multi-source time series monitoring indicators;

[0009] S2. Determine the framework for identifying the safety level of the foundation pit, and use multi-source time series monitoring data as different evidence and perform normalization processing;

[0010] S3. Based on multi-source time series monitoring data, determine the basic probability allocation values ​​of different focal elements in the identification framework;

[0011] S4. Calculate the weights of multi-source monitoring indicators using the interquartile range method, and obtain the credibility of different evidence using the Chebyshev distance.

[0012] S5. Based on the weights of the multi-source monitoring indicators and the credibility of the evidence, adjust the basic probability allocation values ​​of different focal elements.

[0013] S6. The multi-source monitoring data of the foundation pit are fused using the DS evidence theory synthesis rules to obtain the foundation pit safety evaluation level.

[0014] Optionally, the construction pit safety evaluation index system includes:

[0015] S11. Determine the evaluation index system. Based on the characteristics of the foundation pit project, such as geological conditions, support structure type, excavation depth, and surrounding environmental impact, determine the key safety evaluation indicators.

[0016] S12. Data collection: Real-time monitoring of various indicators through sensors to obtain time series data;

[0017] S13. Data preprocessing: Cleaning, denoising, and interpolating the data to obtain cumulative values ​​and rates of change.

[0018] The determination index system consists of monitoring indicators of the surrounding environment and monitoring indicators of multiple foundation pit profiles;

[0019] The surrounding environment monitoring indicators include: cumulative building settlement, building settlement rate, cumulative surface settlement, surface settlement rate, cumulative pipeline settlement, and pipeline settlement rate.

[0020] The monitoring indicators for the foundation pit profile include: cumulative vertical displacement at the top of the wall, rate of change of vertical displacement at the top of the wall, cumulative horizontal displacement at the top of the wall, rate of change of horizontal displacement at the top of the wall, cumulative horizontal displacement at the deep layer, rate of change of horizontal displacement at the deep layer, cumulative internal force of the support, cumulative groundwater level, and rate of change of groundwater level.

[0021] Optionally, the sensors include: a displacement gauge, a stress gauge, and a water level gauge, which are used to measure the displacement of the structure or soil, monitor stress changes inside the structure, and observe fluctuations in the groundwater level, respectively.

[0022] The sensor is connected to the data acquisition unit to collect sensor data in real time and store it in the database for subsequent analysis and processing.

[0023] The data acquisition unit is equipped with an alarm. By setting early warning and alarm thresholds, when a certain monitoring indicator at a certain measuring point exceeds the set threshold, the system can automatically determine whether it is a "false alarm" or a real alarm hazard by integrating and analyzing the multi-source monitoring data of that measuring point and corroborating it with other monitoring indicators, thus ensuring that protective measures are taken in a timely manner.

[0024] The data preprocessing is used to remove noise, incomplete or inconsistent records from the data, improve data quality, and obtain the cumulative value and rate of change of monitoring indicators.

[0025] The cumulative value is the sum of data from a certain starting point to the current time point, calculated through an accumulation operation;

[0026] The rate of change is the ratio of the data difference between adjacent time points to the time interval.

[0027] Optionally, the framework for determining the safety level of the foundation pit includes:

[0028] S21. Define the identification framework: The focal elements in the identification framework are set according to the safety level requirements of the foundation pit. For example, according to safety evaluation standards and specifications, the general identification framework for the safety level of the foundation pit is set as Θ = {V1, V2, V3, V4} = {safe, relatively safe, relatively unsafe, unsafe}. Let m be the power set of the identification framework Θ. Θ A mapping on the interval [0,1], where V is a proposition in the recognition frame; if m(Φ) = 0 and , then m is called 2 Θ The basic probability assignment function; m(V) is the basic probability assignment value of proposition V, representing the degree of confidence of m in proposition V; if proposition V satisfies m(V)>0, then V is called a focal element on Θ.

[0029] S22. Normalization Processing: The raw data of different monitoring indicators are converted into dimensionless quantities. The collected multi-source time series monitoring data of the foundation pit are used as different pieces of evidence and normalized. The formula is as follows:

[0030] In the formula: Let be the value of the i-th monitoring indicator at time t; These are the maximum and minimum values ​​of the i-th monitoring indicator within the time interval 0-t, respectively. This is the normalized monitoring data.

[0031] Optionally, step S3 includes:

[0032] S31. Use cloud models to determine the basic probability allocation values ​​of different focal elements in the recognition framework;

[0033] The cloud model is a mathematical model based on probability theory. The core of the cloud model is to transform qualitative concepts into quantitative values ​​and to describe the uncertainty of a concept by generating clouds.

[0034] The generated cloud is a series of numerical points generated based on the digital features of the cloud model. The numerical points represent the specific form of a qualitative concept in a quantitative space.

[0035] Among them, cloud models are established for the safety level ranges of multi-source monitoring indicators, and the digital characteristics of the models can be calculated using a forward cloud generator.

[0036] The numerical characteristics include the expected E y Entropy E n and hyperentropy H e Let represent the central position of the qualitative proposition, the degree of uncertainty of the proposition, and the degree of uncertainty of the entropy, respectively. The expression is:

[0037] H e,ij =k

[0038] In the formula: This indicates the range of values ​​for a specific element within the identification framework corresponding to a certain monitoring indicator;

[0039] Where: k is a constant, with a value of 0.01;

[0040] i = 1, 2, ..., M;

[0041] j = 1, 2, ..., N;

[0042] M represents the number of multi-source monitoring indicators;

[0043] N represents the number of focal elements in the identification frame;

[0044] S32. Based on the time series monitoring data and the cloud model digital characteristics of different monitoring indicator grading intervals, calculate the certainty of the j-th focal element under the i-th monitoring indicator, and use it as the basic probability assignment value of that focal element:

[0045] In the formula: E y,ij Therefore, E n,ij The mean is H e,ij A normal random variable with standard deviation.

[0046] Optionally, step S4 includes:

[0047] S41. Weight Calculation: Based on the importance and reliability of the monitoring indicators, the weight of each indicator is calculated using the interquartile range method.

[0048] S42. Reliability Assessment: Assess the reliability of each monitoring data source, taking into account factors such as equipment accuracy and environmental factors, and assign a reliability value to each data source.

[0049] The weight represents the importance of the indicator, and its expression is:

[0050] Q d =Q3-Q1

[0051] In the formula: Q1 and Q3 are the first quartile and third quartile of the measured data of a certain monitoring indicator within the time interval 0-t, respectively;

[0052] Q d It is the interquartile range, used to describe the dispersion of data. The larger the value, the greater the variability and the more information it provides.

[0053] w i It represents the importance weights of multi-source monitoring indicators, i = 1, 2, ..., M.

[0054] Optionally, step S5 includes:

[0055] S61. Evidence synthesis: Using the synthesis rules of Dempster-Shafer (DS) evidence theory, evidence from different monitoring indicators and data sources is fused together;

[0056] S62. Safety Level Determination: Determine the final safety evaluation level of the foundation pit based on the fused Basic Probability Allocation (BPA) value.

[0057] In the DS evidence theory, the Basic Probability Assignment (BPA) represents the basic level of trust that each piece of evidence has for each proposition within the identification framework.

[0058] The focal element with the highest basic probability allocation value (BPA) is used as the final determination result.

[0059] Optionally, the evidence fusion step includes:

[0060] 1) Randomly select two pieces of evidence Y i and Y j ;

[0061] 2) Let m i and m jThese are the basic probability assignment values ​​in the corresponding recognition framework, i,j=1,2,…,M;

[0062] 3) Treat the basic probability allocation values ​​of different focal elements under the same evidence body as coordinate points;

[0063] 4) Calculate the Chebyshev distance between the two pieces of evidence;

[0064] d ij =max|m i -m j |,d ij ∈[0,1]

[0065] In the formula: d ij Indicates evidence Y i With evidence Y j The degree of similarity between them, d ij The larger the value, the greater the inconsistency between the two pieces of evidence; conversely, the smaller the value, the more similar the two pieces of evidence. Let's consider evidence Y... i With evidence Y j The similarity between them is expressed as follows:

[0066] s ij =1-d ij

[0067] Evidence Y i With evidence Y j The credibility between them is expressed as:

[0068] In the formula: credibility C rd (m i )satisfy

[0069] Credibility indicates the degree to which a piece of evidence is supported by other pieces of evidence; the higher the degree of support, the more credible the piece of evidence.

[0070] Optionally, in step S6, based on the importance weights of the multi-source monitoring indicators of the foundation pit and the credibility of the evidence, the following general fusion coefficient is proposed:

[0071] η i =α1w i +α2C rd (m i )

[0072] In the formula: α1 and α2 are constants with values ​​between 0 and 1, and satisfy α1 + α2 = 1;

[0073] If α1 > α2, then the importance weight of the multi-source monitoring indicators is considered more important than the credibility between the evidences;

[0074] Conversely, if α1 < α2, then the credibility of the evidence is considered more important than the importance weight of the multi-source monitoring indicators.

[0075] If α1 = 1, then the fusion coefficient η i Only consider the importance of the monitoring indicators;

[0076] Conversely, if α2 = 1, it indicates that the fusion coefficient η is... i The importance of multiple monitoring indicators is considered to be equal only based on the credibility of the evidence.

[0077] In summary, the values ​​of constants α1 and α2 can be determined according to actual needs.

[0078] Optionally, the fusion coefficient η can be used. i The basic probability assignment values ​​for different focal elements are corrected:

[0079] Where: m i (V j ) is the focal element V of the i-th evidence body. j The basic probability assignment values ​​are i = 1, 2, ..., M, j = 1, 2, ..., N;

[0080] According to the DS evidence theory composition rules, multiple pieces of evidence are combined.

[0081] In the formula: Φ represents the empty set; K is the conflict coefficient between pieces of evidence. Indicates the degree of conflict between pieces of evidence.

[0082] Based on the fusion results of multi-source monitoring data and the principle of maximum attribute, the safety level of the foundation pit is determined, and automated early warning is achieved.

[0083] The technical effects and advantages of this invention are as follows:

[0084] 1. By employing the interquartile range (ICM) method to calculate the weights of multi-source monitoring indicators and using the Chebyshev distance to assess the credibility of different pieces of evidence, the advantages of both methods can be combined to improve the overall accuracy and robustness of the assessment. Weight allocation ensures that the importance of different monitoring indicators is reasonably reflected in subsequent analyses; while credibility assessment helps to screen reliable sources of evidence, avoiding the impact of erroneous or misleading information on the analysis results.

[0085] 2. By adjusting the weights of multi-source monitoring indicators and the credibility of evidence, the basic probability allocation values ​​of different focal elements are corrected. Introducing the weights of multi-source monitoring indicators allows for a comprehensive consideration of the importance and reliability of different data sources, making the information fusion process more comprehensive and accurate. The fusion of multi-source data can reduce such biases and improve the overall accuracy of information. Through automated weight and credibility calculation methods, manual intervention and subjective judgment can be reduced, improving the efficiency and accuracy of information fusion. Furthermore, the corrected BPA can serve as an important input for intelligent decision-making systems, supporting the system's automated decision-making and intelligent reasoning processes.

[0086] Other features and advantages of the invention will become clear from the following detailed description of exemplary embodiments of the invention with reference to the accompanying drawings. Attached Figure Description

[0087] Figure 1 is a schematic diagram of the structure provided by the present invention;

[0088] Figure 2 is a schematic diagram of the cloud model of foundation pit safety level provided in an embodiment of the present invention;

[0089] Figure 3 is a schematic diagram showing the change in the probability allocation value of the proposition after fusion of different amounts of evidence provided in the embodiments of the present invention;

[0090] Figure 4 is a schematic diagram of the foundation pit safety evaluation index system provided in the embodiment of the present invention. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0092] As shown in Figures 1 to 4, this invention provides a method for early warning of foundation pit safety through multi-source monitoring data fusion.

[0093] S1. Based on the characteristics of foundation pit engineering and the features of safety monitoring, construct a foundation pit safety evaluation index system and obtain the cumulative value and rate of change of multi-source time series monitoring indicators;

[0094] S2. Determine the framework for identifying the safety level of the foundation pit, and use multi-source time series monitoring data as different evidence and perform normalization processing;

[0095] S3. Based on multi-source time series monitoring data, determine the basic probability allocation values ​​of different focal elements in the identification framework;

[0096] S4. Calculate the weights of multi-source monitoring indicators using the interquartile range method, and obtain the credibility of different evidence using the Chebyshev distance.

[0097] S5. Based on the weights of the multi-source monitoring indicators and the credibility of the evidence, adjust the basic probability allocation values ​​of different focal elements.

[0098] S6. The multi-source monitoring data of the foundation pit are fused using the DS evidence theory synthesis rules to obtain the foundation pit safety evaluation level.

[0099] Furthermore, this invention constructs a foundation pit safety evaluation index system that covers the characteristics of foundation pit engineering and safety monitoring. This system can more accurately reflect the safety status of foundation pit engineering and provide a solid foundation for safety assessment. When determining the foundation pit safety level identification framework, normalization processing is performed to ensure the comparability of multi-source time series monitoring data in different evidence, avoiding errors caused by different data units or magnitudes, and providing a reliable data foundation for subsequent weight calculation and credibility assessment.

[0100] Using the interquartile range method to calculate the weights of multi-source monitoring indicators can effectively reflect the importance of each indicator in the overall evaluation system. Simultaneously, obtaining the credibility of different pieces of evidence through Chebyshev distance further improves the accuracy of the evaluation. The rationality of the allocation and credibility assessment makes the evaluation results more consistent with reality. Furthermore, the DS evidence theory synthesis rules are used to fuse multi-source monitoring data for the foundation pit. This theory has significant advantages in handling uncertainties and conflicting information. By fusing multi-source monitoring data, the impact of various factors on foundation pit safety can be comprehensively considered, resulting in a more comprehensive and accurate safety evaluation level.

[0101] The entire evaluation process is based on cumulative values ​​and rates of change over time, enabling real-time reflection of changes in the safety status of the foundation pit project. This real-time and dynamic nature allows for timely adjustments and optimization of construction plans, ensuring the safe progress of the foundation pit project. The final foundation pit safety rating provides a scientific basis for project management and decision-making. Through comprehensive evaluation results, managers can promptly understand the safety status of the foundation pit project, take corresponding preventative measures and emergency responses, and ensure project quality and personnel safety.

[0102] The application of the above steps in foundation pit safety evaluation improves the accuracy and reliability of foundation pit safety evaluation by constructing a comprehensive evaluation index system, scientific data processing, reasonable weight allocation and credibility assessment, the integration advantages of evidence theory, and the characteristics of real-time and dynamic nature, thus providing strong support for the safety management of foundation pit projects.

[0103] Specifically, the construction pit safety evaluation index system includes:

[0104] S11. Determine the evaluation index system. Based on the characteristics of the foundation pit project, such as geological conditions, support structure type, excavation depth, and surrounding environmental impact, determine the key safety evaluation indicators.

[0105] S12. Data collection: Real-time monitoring of various indicators through sensors to obtain time series data;

[0106] S13. Data preprocessing: Cleaning, denoising, and interpolating the data to obtain cumulative values ​​and rates of change.

[0107] The determination index system consists of monitoring indicators of the surrounding environment and monitoring indicators of multiple foundation pit profiles;

[0108] The surrounding environment monitoring indicators include: cumulative building settlement, building settlement rate, cumulative surface settlement, surface settlement rate, cumulative pipeline settlement, and pipeline settlement rate.

[0109] The monitoring indicators for the foundation pit profile include: cumulative vertical displacement at the top of the wall, rate of change of vertical displacement at the top of the wall, cumulative horizontal displacement at the top of the wall, rate of change of horizontal displacement at the top of the wall, cumulative horizontal displacement at the deep layer, rate of change of horizontal displacement at the deep layer, cumulative internal force of the support, cumulative groundwater level, and rate of change of groundwater level.

[0110] The sensors include: displacement gauges, stress gauges, and water level gauges, which are used to measure the displacement of the structure or soil, monitor stress changes inside the structure, and observe fluctuations in the groundwater level, respectively.

[0111] The sensor is connected to the data acquisition unit to collect sensor data in real time and store it in the database for subsequent analysis and processing.

[0112] Furthermore, the sensor features high precision and high reliability, which ensures the accuracy of the monitoring data. This accuracy guarantees the accuracy of the evaluation results and avoids evaluation bias caused by data errors.

[0113] By monitoring various indicators in real time through sensors, real-time time-series data can be obtained, enabling the evaluation process to reflect changes in the safety status of the foundation pit project in a timely manner, and providing strong support for subsequent decision-making.

[0114] The adoption of automated monitoring reduces human intervention and errors, improves monitoring efficiency and data quality. At the same time, automated monitoring can also achieve all-weather, uninterrupted monitoring, providing strong support for the safety management of foundation pit projects.

[0115] The data preprocessing is used to remove noise, incomplete or inconsistent records from the data, improve data quality, and calculate cumulative values ​​and rates of change.

[0116] The cumulative value is the sum of data from a certain starting point to the current time point, calculated through an accumulation operation;

[0117] The rate of change is the ratio of the data difference between adjacent time points to the time interval.

[0118] Furthermore, cumulative values ​​and rates of change are used in data analysis and engineering monitoring to provide critical information about changes in the state of a system or structure;

[0119] The cumulative value represents the total change from the start of monitoring to the current time point, and is used to provide information on the overall deformation of the system throughout the monitoring period;

[0120] By observing the trend of the cumulative value, it is possible to determine whether the system or structure is in a stable state or whether there are abnormalities such as accelerated deformation.

[0121] The rate of change represents the amount of change of a monitored indicator per unit time, providing direct information about the speed of change;

[0122] The magnitude of the rate of change is directly related to the stability of the system; a faster rate of change may mean that the system or structure is undergoing rapid deformation or instability.

[0123] By setting a reasonable threshold for the rate of change, real-time early warning and monitoring of system or structure deformation can be achieved. When the rate of change exceeds the threshold, an alarm mechanism can be triggered through the fusion and analysis of multi-source monitoring data, reminding relevant personnel to take timely measures to prevent the situation from worsening.

[0124] In risk assessment, the rate of change is an important reference indicator. By comparing the rate of change at different time periods or locations, the risk level of a system or structure under different conditions can be assessed, and corresponding risk management strategies can be formulated.

[0125] Specifically, the framework for determining the safety level of a foundation pit includes:

[0126] S21. Define the identification framework: The focal elements in the identification framework are set according to the safety level requirements of the foundation pit. For example, according to safety evaluation standards and specifications, the general identification framework for the safety level of the foundation pit is set as Θ = {V1, V2, V3, V4} = {safe, relatively safe, relatively unsafe, unsafe}. Let m be the power set of the identification framework Θ. Θ A mapping on the interval [0,1], where V is a proposition in the recognition frame; if m(Φ) = 0 and , then m is called 2 Θ The basic probability assignment function; m(V) is the basic probability assignment value of proposition V, representing the degree of confidence of m in proposition V; if proposition V satisfies m(V)>0, then V is called a focal element on Θ.

[0127] S22. Normalization Processing: The raw data of different monitoring indicators are converted into dimensionless quantities. The collected multi-source time series monitoring data of the foundation pit are used as different pieces of evidence and normalized. The formula is as follows:

[0128] In the formula: Let be the value of the i-th monitoring indicator at time t; These are the maximum and minimum values ​​of the i-th monitoring indicator within the time interval 0-t, respectively. This is the normalized monitoring data.

[0129] Furthermore, after obtaining the normalized multi-source time-series monitoring data, it can be input as different pieces of evidence into the foundation pit safety level identification model. This model can calculate the probability allocation value for each safety level based on the input evidence and a predefined identification framework, and output the final foundation pit safety level.

[0130] It is important to note that building and training a foundation pit safety level identification model is a complex process that requires extensive historical data and expert experience. In practical applications, appropriate adjustments and optimizations are also necessary based on specific engineering conditions.

[0131] By establishing a framework for identifying the safety level of foundation pits, collecting multi-source time-series monitoring data, and performing normalization processing, strong data support is provided for the identification and early warning of foundation pit safety levels.

[0132] Specifically, step S3 includes:

[0133] S31. Use cloud models to determine the basic probability allocation values ​​of different focal elements in the recognition framework;

[0134] The cloud model is a mathematical model based on probability theory. The core of the cloud model is to transform qualitative concepts into quantitative values ​​and to describe the uncertainty of a concept by generating clouds.

[0135] The generated cloud is a series of numerical points generated based on the digital features of the cloud model. The numerical points represent the specific form of a qualitative concept in a quantitative space.

[0136] Among them, cloud models are established for the safety level ranges of multi-source monitoring indicators, and the digital characteristics of the models can be calculated using a forward cloud generator.

[0137] It should be noted that in foundation pit safety early warning, many factors such as geological conditions, excavation depth, and support structure performance exhibit varying degrees of randomness and ambiguity. Cloud models, through their unique digital characteristics (expected value E...), address these issues. y Entropy E n and hyperentropy H eIt can simultaneously express the randomness and ambiguity of these factors, making the early warning results more consistent with the actual situation;

[0138] Cloud models transform qualitative concepts into concrete numerical representations through quantitative methods, thereby reducing the subjectivity and randomness of human judgment. In foundation pit safety early warning, cloud models can be used to determine the basic probability allocation values ​​of different focal elements (such as different risk levels), and these values ​​are more objective and accurate. By comprehensively considering the probability allocation values ​​of multiple focal elements, more precise early warning results can be obtained.

[0139] The numerical characteristics include the expected E y Entropy E n and hyperentropy H e Let represent the central position of the qualitative proposition, the degree of uncertainty of the proposition, and the degree of uncertainty of the entropy, respectively. The expression is:

[0140] H e,ij =k

[0141] In the formula: This indicates the range of values ​​for a specific element within the identification framework corresponding to a certain monitoring indicator;

[0142] Where: k is a constant, with a value of 0.01;

[0143] i = 1, 2, ..., M;

[0144] j = 1, 2, ..., N;

[0145] M represents the number of multi-source monitoring indicators;

[0146] N represents the number of focal elements in the identification frame;

[0147] S32. Based on the time series monitoring data and the cloud model digital characteristics of different monitoring indicator grading intervals, calculate the certainty of the j-th focal element under the i-th monitoring indicator, and use it as the basic probability assignment value of that focal element:

[0148] In the formula: E y,ij Therefore, E n,ij The mean is H e,ij A normal random variable with standard deviation.

[0149] Specifically, step S4 includes:

[0150] S41. Weight Calculation: Based on the importance and reliability of the monitoring indicators, the weight of each indicator is calculated using the interquartile range method.

[0151] S42. Reliability Assessment: Assess the reliability of each monitoring data source, taking into account factors such as equipment accuracy and environmental factors, and assign a reliability value to each data source.

[0152] The weight represents the importance of the indicator, and its expression is:

[0153] Q d =Q3-Q1

[0154] In the formula: Q1 and Q3 are the first quartile and third quartile of the measured data of a certain monitoring indicator within the time interval 0-t, respectively;

[0155] Q d It is the interquartile range, used to describe the dispersion of data. The larger the value, the greater the variability and the more information it provides.

[0156] w i It represents the importance weights of multi-source monitoring indicators, i = 1, 2, ..., M.

[0157] Furthermore, the interquartile range (ICM) not only provides a quantitative indicator of data dispersion but also implicitly reveals the data's distribution. By analyzing the ICM, we can understand how the data is distributed across different intervals, thereby determining the importance of each monitoring indicator within the dataset and assigning it appropriate weights. It's important to note that the ICM determines data dispersion by calculating the difference between the third and first quartiles; this method is less sensitive to outliers. Compared to statistics such as the mean or standard deviation, the ICM better reflects the overall distribution of the data and is less susceptible to significant bias due to a few outliers.

[0158] Furthermore, the Chebyshev distance focuses on the maximum coordinate difference between two points in vector space, which can capture the largest deviation in the data and ensure that any extreme cases that could lead to significant errors are not ignored when assessing the credibility of evidence.

[0159] Specifically, step S5 includes:

[0160] S61. Evidence synthesis: Using the synthesis rules of Dempster-Shafer (DS) evidence theory, evidence from different monitoring indicators and data sources is fused together;

[0161] S62. Safety Level Determination: Determine the final safety evaluation level of the foundation pit based on the fused Basic Probability Allocation (BPA) value.

[0162] In the DS evidence theory, the Basic Probability Assignment (BPA) represents the basic level of trust that each piece of evidence has for each proposition within the identification framework.

[0163] The focal element with the highest basic probability allocation value (BPA) is used as the final determination result.

[0164] Specifically, the evidence fusion steps include:

[0165] 1) Randomly select two pieces of evidence Y i and Y j ;

[0166] 2) Let m i and m j These are the basic probability assignment values ​​in the corresponding recognition framework, i,j=1,2,…,M;

[0167] 3) Treat the basic probability allocation values ​​of different focal elements under the same evidence body as coordinate points;

[0168] 4) Calculate the Chebyshev distance between the two pieces of evidence;

[0169] d ij =max|m i -m j |,d ij ∈[0,1]

[0170] In the formula: d ij Indicates evidence Y i With evidence Y j The degree of similarity between them, d ij The larger the value, the greater the inconsistency between the two pieces of evidence; conversely, the smaller the value, the more similar the two pieces of evidence. Let's consider evidence Y... i With evidence Y j The similarity between them is expressed as follows:

[0171] s ij =1-d ij

[0172] Evidence Y i With evidence Y j The credibility between them is expressed as:

[0173] In the formula: credibility C rd (m i )satisfy

[0174] Credibility indicates the degree to which a piece of evidence is supported by other pieces of evidence; the higher the degree of support, the more credible the piece of evidence.

[0175] It should be noted that in practical applications, there may be certain correlations or dependencies between different monitoring indicators and data sources. Therefore, it is necessary to assess the independence of evidence before fusion and take appropriate measures to handle non-independent evidence.

[0176] Furthermore, the DS evidence theory can directly handle unknown states. By fusing evidence from different monitoring indicators and data sources, it can provide a more comprehensive understanding of the safety status of the foundation pit and reduce the uncertainty caused by a single data source. The fused BPA provides a more accurate and comprehensive basis for the final safety assessment level of the foundation pit. By fusing evidence from multiple data sources, the dependence on a single data source can be reduced, thereby improving the robustness of the system. The DS evidence theory can handle evidence from different monitoring indicators and data sources, including heterogeneous, incomplete, or conflicting evidence. Through synthesis rules, these heterogeneous data can be fused into a unified assessment result, providing more comprehensive information support for the foundation pit safety assessment.

[0177] Specifically, in step S6, based on the importance weights of the multi-source monitoring indicators of the foundation pit and the credibility of the evidence, the following general fusion coefficients are proposed:

[0178] η i =α1w i +α2C rd (m i )

[0179] In the formula: α1 and α2 are constants with values ​​between 0 and 1, and satisfy α1 + α2 = 1;

[0180] If α1 > α2, then the importance weight of the multi-source monitoring indicators is considered more important than the credibility between the evidences;

[0181] Conversely, if α1 < α2, then the credibility of the evidence is considered more important than the importance weight of the multi-source monitoring indicators.

[0182] If α1 = 1, then the fusion coefficient η i Only consider the importance of the monitoring indicators;

[0183] Conversely, if α2 = 1, it indicates that the fusion coefficient η is... i The importance of multiple monitoring indicators is considered to be equal only based on the credibility of the evidence.

[0184] In summary, the values ​​of constants α1 and α2 can be determined according to actual needs.

[0185] Specifically, using the fusion coefficient η i The basic probability assignment values ​​for different focal elements are corrected:

[0186] Where: mi (V j ) is the focal element V of the i-th evidence body. j The basic probability assignment values ​​are i = 1, 2, ..., M, j = 1, 2, ..., N;

[0187] According to the DS evidence theory composition rules, multiple pieces of evidence are combined.

[0188] In the formula: Φ represents the empty set; K is the conflict coefficient between pieces of evidence. Indicates the degree of conflict between pieces of evidence.

[0189] Based on the fusion results of multi-source monitoring data and the principle of maximum attribute, the safety level of the foundation pit is determined, and automated early warning is achieved.

[0190] Furthermore, in order to verify the effectiveness of the improved evidence theory in this patent, evidence bodies under three different working conditions were selected for fusion analysis. The basic probability distribution values ​​of different evidence bodies are shown in Table 1. Among them, evidence body Y2 is the interference item, which has a large conflict with other evidence bodies.

[0191] Table 1. Basic probability distribution values ​​of evidence under different working conditions.

[0192] The data fusion analysis of Table 1 was performed using traditional evidence theory, the Murphy method, the Deng method, and the method of this invention, respectively. Assuming that the constants α1 and α2 are both 0.5, the fusion results obtained by different methods are shown in Table 2.

[0193] Due to its inherent limitations, traditional evidence theory struggles to achieve a fusion result that accurately reflects the actual situation under all three conditions. Furthermore, the conflict coefficient K gradually approaches 1, leading to severe conflicts between pieces of evidence. Consequently, the results calculated by traditional evidence theory have significant errors.

[0194] Furthermore, since m(V1) = 0 in evidence Y2, the "veto power" of traditional evidence theory means that no matter how many pieces of evidence are added during the subsequent fusion process, the support for proposition V1 will not increase.

[0195] The "veto power" specifically means that when the basic probability assignment (BPA) of a certain piece of evidence for a certain proposition (such as V1) is 0, that is, m(V1) = 0, then no matter how many pieces of evidence are added later, and these new pieces of evidence show that m(V1) ≠ 0, the support for proposition V1 will not increase.

[0196] In contrast, the fusion results obtained by the Murphy method, the Deng method, and the method proposed in this invention are all consistent with the actual results, and the support of proposition V1 gradually increases with the increase of the number of evidence bodies.

[0197] Table 2 Comparison of fusion results obtained by different methods

[0198] Figure 3 shows the changes in the probability distribution of proposition V1 after fusion of different amounts of evidence. Under the same conditions, the conflict coefficient K obtained by the three methods gradually decreases, ensuring the accuracy of the fusion results. These results demonstrate that the method proposed in this invention is superior to other methods in terms of fusion effect, fast convergence speed, and anti-interference ability.

[0199] The main features of this invention are: based on fully utilizing the measured monitoring data of the foundation pit, it simultaneously considers the importance of different monitoring indicators and the credibility of the evidence, ensuring the objectivity of foundation pit safety status identification and avoiding interference from subjective human factors. It effectively overcomes the following shortcomings of traditional DS evidence theory: inability to resolve cases of severe and complete evidence conflict; evidence having a veto power over the proposition; and being sensitive to basic probability allocation with poor anti-interference ability. Therefore, it can not only obtain the overall assessment result of the foundation pit safety status but also obtain the local safety level of different monitoring sections of the foundation pit and the surrounding environment, providing a more reasonable approach and more accurate assessment results for foundation pit safety early warning.

[0200] In summary, this invention uses the interquartile range method to calculate the weights of multi-source monitoring indicators and uses the Chebyshev distance to assess the credibility of different pieces of evidence. This approach combines the advantages of both methods and improves the overall accuracy and robustness of the assessment.

[0201] By assigning weights, we can ensure that the importance of different monitoring indicators is reasonably reflected in subsequent analysis; and by conducting credibility assessment, we can screen out reliable sources of evidence and avoid the impact of erroneous or misleading information on the analysis results.

[0202] By adjusting the weights of multi-source monitoring indicators and the credibility of evidence, the basic probability allocation values ​​of different focal elements can be corrected. Introducing the weights of multi-source monitoring indicators allows for a comprehensive consideration of the importance and reliability of different data sources, making the information fusion process more comprehensive and accurate. The fusion of multi-source data can reduce this bias and improve the overall accuracy of information.

[0203] By using automated weight and credibility calculation methods, human intervention and subjective judgment can be reduced, improving the efficiency and accuracy of information fusion. Furthermore, the corrected BPA can serve as an important input for intelligent decision-making systems, supporting the system's automated decision-making and intelligent reasoning processes.

[0204] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for early warning of foundation pit safety through multi-source monitoring data fusion, characterized in that: Includes the following steps: S1. Based on the characteristics of foundation pit engineering and the features of safety monitoring, construct a foundation pit safety evaluation index system and obtain the cumulative value and rate of change of multi-source time series monitoring indicators; S2. Determine the framework for identifying the safety level of the foundation pit, and use multi-source time series monitoring data as different evidence and perform normalization processing; S3. Based on multi-source time series monitoring data, determine the basic probability allocation values ​​of different focal elements in the identification framework; S4. Calculate the weights of multi-source monitoring indicators using the interquartile range method, and obtain the credibility of different evidence using the Chebyshev distance. S5. Based on the weights of the multi-source monitoring indicators and the credibility of the evidence, adjust the basic probability allocation values ​​of different focal elements. S6. The multi-source monitoring data of the foundation pit are fused using the DS evidence theory synthesis rules to obtain the foundation pit safety evaluation level.

2. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The safety evaluation index system for the construction pit includes: S11. Determine the evaluation index system. Based on the characteristics of the foundation pit project, such as geological conditions, support structure type, excavation depth, and surrounding environmental impact, determine the key safety evaluation indicators. S12. Data collection: Real-time monitoring of various indicators through sensors to obtain time series data; S13. Data preprocessing: Cleaning, denoising, and interpolating the data to obtain cumulative values ​​and rates of change. The determination index system consists of monitoring indicators of the surrounding environment and monitoring indicators of multiple foundation pit profiles; The surrounding environment monitoring indicators include: cumulative building settlement, building settlement rate, cumulative surface settlement, surface settlement rate, cumulative pipeline settlement, and pipeline settlement rate. The monitoring indicators for the foundation pit profile include: cumulative vertical displacement at the top of the wall, rate of change of vertical displacement at the top of the wall, cumulative horizontal displacement at the top of the wall, rate of change of horizontal displacement at the top of the wall, cumulative horizontal displacement at the deep layer, rate of change of horizontal displacement at the deep layer, cumulative internal force of the support, cumulative groundwater level, and rate of change of groundwater level.

3. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 2, characterized in that: The sensors include: displacement gauges, stress gauges, and water level gauges, which are used to measure the displacement of the structure or soil, monitor stress changes inside the structure, and observe fluctuations in the groundwater level, respectively. The sensor is connected to the data acquisition unit to collect sensor data in real time and store it in the database for subsequent analysis and processing. The data acquisition unit is equipped with an alarm. By setting early warning and alarm thresholds, when a certain monitoring indicator at a certain measuring point exceeds the set threshold, the system integrates and analyzes the multi-source monitoring data of that measuring point. That is, with the corroboration of other monitoring indicators, the system can automatically determine whether it is a "false alarm" or a real alarm danger, ensuring that protective measures are taken in a timely manner. The data preprocessing is used to remove noise, incomplete or inconsistent records from the data, improve data quality, and obtain the cumulative value and rate of change of monitoring indicators. The cumulative value is the sum of data from a certain starting point to the current time point, calculated through an accumulation operation; The rate of change is the ratio of the data difference between adjacent time points to the time interval.

4. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The framework for determining the safety level of a foundation pit includes: S21. Define the identification framework: The focal elements in the identification framework are set according to the safety level requirements of the foundation pit. For example, according to safety evaluation standards and specifications, the general identification framework for the safety level of the foundation pit is set as Θ = {V1, V2, V3, V4} = {safe, relatively safe, relatively unsafe, unsafe}. Let m be the power set of the identification framework Θ. Θ A mapping on the interval [0,1], where V is a proposition in the recognition frame; if m(Φ) = 0 and , then m is called 2 Θ The basic probability assignment function; m(V) is the basic probability assignment value of proposition V, representing the degree of confidence of m in proposition V; if proposition V satisfies m(V)>0, then V is called a focal element on Θ; S22. Normalization Processing: The raw data of different monitoring indicators are converted into dimensionless quantities. The collected multi-source time series monitoring data of the foundation pit are used as different pieces of evidence and normalized. The formula is as follows: In the formula: Let be the value of the i-th monitoring indicator at time t; These are the maximum and minimum values ​​of the i-th monitoring indicator within the time interval 0-t, respectively. This is the normalized monitoring data.

5. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The steps in S3 include: S31. Use cloud models to determine the basic probability allocation values ​​of different focal elements in the recognition framework; The cloud model is a mathematical model based on probability theory. The core of the cloud model is to transform qualitative concepts into quantitative values ​​and to describe the uncertainty of a concept by generating clouds. The generated cloud is a series of numerical points generated based on the digital features of the cloud model. The numerical points represent the specific form of a qualitative concept in a quantitative space. Among them, cloud models are established for the safety level ranges of multi-source monitoring indicators, and the digital characteristics of the models can be calculated using a forward cloud generator. The numerical characteristics include the expected E y Entropy E n and hyperentropy H e Let represent the central position of the qualitative proposition, the degree of uncertainty of the proposition, and the degree of uncertainty of the entropy, respectively. The expression is: H e,ij =k In the formula: This indicates the range of values ​​for a specific element within the identification framework corresponding to a certain monitoring indicator; Where: k is a constant, with a value of 0.01; i = 1, 2, ..., M; j = 1, 2, ..., N; M represents the number of multi-source monitoring indicators; N represents the number of focal elements in the identification frame; S32. Based on the time series monitoring data and the cloud model digital characteristics of different monitoring indicator grading intervals, calculate the certainty of the j-th focal element under the i-th monitoring indicator, and use it as the basic probability assignment value of that focal element: In the formula: E y,ij Therefore, E n,ij The mean is H e,ij A normal random variable with standard deviation.

6. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The steps in S4 include: S41. Weight Calculation: Based on the importance and reliability of the monitoring indicators, the weight of each indicator is calculated using the interquartile range method. S42. Reliability Assessment: Assess the reliability of each monitoring data source, taking into account factors such as equipment accuracy and environmental factors, and assign a reliability value to each data source. The weight represents the importance of the indicator, and its expression is: Q d =Q3-Q1 In the formula: Q1 and Q3 are the first quartile and third quartile of the measured data of a certain monitoring indicator within the time interval 0-t, respectively; Q d It is the interquartile range, used to describe the dispersion of data. The larger the value, the greater the variability and the more information it provides. w i It represents the importance weights of multi-source monitoring indicators, i = 1, 2, ..., M.

7. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The steps in S5 include: S61. Evidence Composition: Utilizing the composition rules of the Dempster-Shafer (DS) evidence theory, in the future... Evidence from different monitoring indicators and data sources is integrated; S62. Safety Level Determination: Determine the final safety evaluation level of the foundation pit based on the fused Basic Probability Allocation (BPA) value. In the DS evidence theory, the Basic Probability Assignment (BPA) represents the basic level of trust that each piece of evidence has for each proposition within the identification framework. The focal element with the highest basic probability allocation value (BPA) is used as the final determination result.

8. The method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: The evidence fusion steps include: 1) Randomly select two pieces of evidence Y i and Y j ; 2) Let m i and m j These are the basic probability assignment values ​​in the corresponding recognition framework, i,j=1,2,…,M; 3) Treat the basic probability allocation values ​​of different focal elements under the same evidence body as coordinate points; 4) Calculate the Chebyshev distance between the two pieces of evidence; d ij =max|m i -m j |,d ij ∈[0,1] In the formula: d ij Indicates evidence Y i With evidence Y j The degree of similarity between them, d ij The larger the value, the greater the inconsistency between the two pieces of evidence; conversely, the smaller the value, the more similar the two pieces of evidence. Let's consider evidence Y... i With evidence Y j The similarity between them is expressed as follows: s ij =1-d ij Evidence Y i With evidence Y j The credibility between them is expressed as: In the formula: credibility C rd (m i )satisfy Credibility indicates the degree to which a piece of evidence is supported by other pieces of evidence; the higher the degree of support, the more credible the piece of evidence.

9. A method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 7, characterized in that: In step S6, based on the importance weights of the multi-source monitoring indicators for the foundation pit and the credibility of the evidence, the following general fusion coefficient is proposed: η i =α1w i +α2C rd (m i ) In the formula: α1 and α2 are constants with values ​​between 0 and 1, and satisfy α1 + α2 = 1; If α1 > α2, then the importance weight of the multi-source monitoring indicators is considered more important than the credibility between the evidences; Conversely, if α1 < α2, then the credibility of the evidence is considered more important than the importance weight of the multi-source monitoring indicators. If α1 = 1, then the fusion coefficient η i Only consider the importance of the monitoring indicators; Conversely, if α2 = 1, it indicates that the fusion coefficient η is... i The importance of multiple monitoring indicators is considered to be equal only based on the credibility of the evidence. In summary, the values ​​of constants α1 and α2 can be determined according to actual needs.

10. A method for early warning of foundation pit safety based on multi-source monitoring data fusion according to claim 1, characterized in that: Using the fusion coefficient η i The basic probability assignment values ​​for different focal elements are corrected: Where: m i (V j ) is the focal element V of the i-th evidence body. j The basic probability assignment values ​​are i = 1, 2, ..., M, j = 1, 2, ..., N; According to the DS evidence theory composition rules, multiple pieces of evidence are combined. In the formula: Φ represents the empty set; K is the conflict coefficient between pieces of evidence. Indicate the degree of conflict between pieces of evidence; Based on the fusion results of multi-source monitoring data and the principle of maximum attribute, the safety level of the foundation pit is determined, and automated early warning is achieved.

Images