Tunnel structure damage assessment method and system
Through the combination of real-time monitoring of data and environmental changes, dynamic window technology and integrated learning framework, the problem that real-time changes in tunnel structure damage assessment is difficult to reflect, achieving more accurate damage diagnosis and prediction, and improving the efficiency of tunnel management and Security.
Patent Information
- Application Number
- CN202411026633.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-12
AI Technical Summary
The existing tunnel structural damage assessment methods are difficult to reflect changes in internal and surrounding environmental conditions in real time, resulting in misjudgment of health conditions, increasing the risk of structural failure, lack of flexibility and depth, and affecting the formulation and safety of maintenance strategies.
Real-time monitoring data is used to combine temperature and humidity environment changes, and dynamic window technology is used to divide time series data, calculate posterior probability, quantify damage, integrate damage characteristics, input an integrated learning framework for probability prediction, and evaluate the future safety status of the tunnel.
Improve tunnel management efficiency and safety, reduce accident risks, optimize maintenance and repair decisions, and achieve more accurate damage diagnosis and prediction.
Smart Images

Figure CN118939970A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of civil engineering, and in particular to a tunnel structure damage assessment method and system. Background Art
[0002] The technical field of civil engineering is widely involved in infrastructure construction and maintenance, including the design, construction, monitoring and evaluation of important structures such as roads, bridges, tunnels, dams, etc. This field uses advanced technologies and methods to ensure structural safety, extend life and optimize performance. Civil engineering not only includes the physical construction of structures, but also involves regular assessment of existing structures to identify potential damage and destruction, and through various monitoring techniques and model analysis, early warning of possible structural failures, thereby protecting public safety and optimizing economic costs.
[0003] Among them, a tunnel structure damage assessment method refers to a technology that uses scientific and technological means to monitor and analyze the integrity and safety status of the tunnel. This method usually includes steps such as sensor data collection, data analysis and interpretation, and damage diagnosis, aiming to timely detect and evaluate possible cracks, deformations or other structural problems in the tunnel. The main purpose is to prevent accidents and structural failures, ensure the safe operation of the tunnel, and provide a scientific basis for maintenance and repair work, thereby significantly improving the efficiency and safety of tunnel management.
[0004] Existing technologies usually rely on traditional monitoring technology and regular inspections to assess tunnel structural damage, which makes it difficult to reflect the rapid changes in tunnel internal and surrounding environmental conditions in real time. This delay may lead to misjudgment of the tunnel's health status and increase the risk of tunnel structural failure. Existing methods lack flexibility and depth in the collection and analysis of damage indicators, resulting in incomplete damage diagnosis and affecting the effective formulation of maintenance strategies. The lack of dynamic and comprehensive evaluation methods limits the ability to fully evaluate tunnel safety, making it impossible to optimize tunnel operation and maintenance strategies, which may cause economic burdens and safety hazards. Summary of the invention
[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a tunnel structure damage assessment method and system.
[0006] In order to achieve the above object, the present invention adopts the following technical solution: a tunnel structure damage assessment method, comprising the following steps:
[0007] S1: Collect real-time monitoring data of tunnel structure, integrate temperature and humidity environment change data, and use dynamic window technology to segment time series data to obtain time series sets;
[0008] S2: updating the parameters of the time series set, calculating the posterior probability of data changes, quantifying the target damage of crack extension and concrete peeling, and establishing a health status assessment;
[0009] S3: Analyze the health status assessment, identify damage responses caused by the increase in tunnel service life and environmental changes, collect damage indicators, including internal corrosion and stress concentration, perform trend analysis and pattern recognition on the damage indicators, and integrate them into a damage feature set;
[0010] S4: Input the damage feature set into the integrated learning framework, combine the probability prediction and damage feature data, predict the development trend of tunnel structure damage, evaluate the safety status of the tunnel in the future time period, and obtain dynamic damage prediction analysis results.
[0011] As a further solution of the present invention, the step of acquiring the time series set is specifically:
[0012] S111: Collect real-time monitoring data of the tunnel structure, including temperature and humidity, and use a sliding average model to smooth the data using the formula:
[0013] E t =α·T t +(1-α)·H t
[0014] Calculate the sliding mean temperature and humidity index to obtain the environmental data index;
[0015] Among them, T t Represents the temperature data at time point t, H t represents the humidity data at time point t, and α is the weight parameter;
[0016] S112: Based on the environmental data index, the dynamic window technology is applied to segment the time series data, and the weighted moving average formula is adopted:
[0017]
[0018] Get the segmented time series data;
[0019] Among them, W k represents the dataset of the kth time window, E t-i is the environmental data index at time point ti, n is the time window size, β i is the weight of multiple time points in the time window;
[0020] S113: Summarize the time series data after segmentation in all time windows, using the formula:
[0021]
[0022] Calculate the time series set within the entire target time range to obtain the time series set;
[0023] Among them, S represents the time series set within the target time range, W j represents the dataset of the jth time window, and k is the number of time windows.
[0024] As a further solution of the present invention, the steps of obtaining the health status assessment are specifically as follows:
[0025] S211: extracting the monitoring data of a time point from the time series set parameters, and calculating the probability of the monitoring data at each time point in combination with historical trend analysis, using the formula:
[0026]
[0027] Calculate the generated conditional probability;
[0028] Among them, D t represents the monitoring data at time point t, P(D t |S) is the monitoring of D under the time series set S t The conditional probability, P(S|D t ) is in D t The probability of monitoring S when it occurs, P(D t ) is D t The prior probability of , λ is the adjustment factor, which adjusts the probability value;
[0029] S212: Using the conditional probability, apply the weighted summation method to combine the conditional probabilities of all time points, using the formula:
[0030] P(S)=∑ t P(D t ∣S)·W t
[0031] Calculate the overall posterior probability and obtain the updated posterior probability;
[0032] Among them, P(S) is the posterior probability of the updated data change, W t It is the weight assigned based on historical data;
[0033] S213: Using the updated posterior probability, evaluate the current health status of the tunnel, define a damage identification formula and a threshold, and use the formula:
[0034] H=max(0,κ·(P(S)-θ))
[0035] Generate health status assessments;
[0036] Among them, H is the health status assessment result, κ is the coefficient for adjusting the response sensitivity, and θ is the benchmark threshold for damage assessment.
[0037] As a further solution of the present invention, data related to tunnel service life and environmental changes are extracted from the health status assessment, and the composite impact formula is used:
[0038]
[0039] The damage response R is calculated;
[0040] Among them, γ and δ adjust the contribution of age increase and environmental change, a and b are exponential parameters, and λ is the normalization factor;
[0041] S312: Integrate the damage response with the damage indicators of the tunnel, including internal corrosion and stress concentration, using the weighted sum formula:
[0042]
[0043] Calculate and generate damage characteristic vector V;
[0044] Among them, α, β, and γ are weight parameters that adjust the influence of multiple damage indicators in the eigenvector, and c and d provide nonlinear responses to corrosion and stress reactions;
[0045] S313: Perform trend analysis and pattern recognition on the damage feature vector using a statistical pattern recognition formula:
[0046]
[0047] The damage feature set F is calculated;
[0048] Among them, κ is the regulatory factor.
[0049] As a further solution of the present invention, the steps of obtaining the dynamic damage prediction analysis results are specifically as follows:
[0050] S411: The damage feature set is input into an integrated learning framework, and multiple prediction models are combined to enhance the accuracy and reliability of the prediction, using a weighted average formula:
[0051]
[0052] Adjust and stabilize the output to generate a probability prediction result P;
[0053] Among them, w i is the weight of the ith model, M i (F) is the output of model i based on the damage feature set F, η and ξ are parameters for adjusting the smoothness of the output;
[0054] S412: Combining the probability prediction and historical damage data, perform trend analysis using the formula:
[0055] T=αP+βH t +γ
[0056] Increase the flexibility of the model and generate trend analysis results T;
[0057] Among them, α and β adjust the relative importance of current forecasts and historical data, and γ adjusts the baseline level;
[0058] S413: Based on the trend analysis results, a mathematical model is used to predict the damage development trend of the tunnel structure and evaluate the safety status in the future time period, using the formula:
[0059]
[0060] A variant of the sigmoid function is used to ensure that S is between 0 and 1, representing the damage probability, and to generate the dynamic damage prediction analysis result U;
[0061] Among them, e is the base of natural logarithm, which is used for sigmoid function transformation.
[0062] A tunnel structure damage assessment system, the tunnel structure damage assessment system is used to perform the above tunnel structure damage assessment method, the system comprises:
[0063] The data acquisition module collects real-time monitoring data of the tunnel from the monitoring equipment, and simultaneously obtains environmental data including temperature and humidity, updates time and environmental parameters, and generates a monitoring data set;
[0064] The time series analysis module inputs the monitoring data set, applies dynamic window technology to perform time series segmentation on the data, identifies key change points within the time range, and obtains a time series parameter set;
[0065] The health status assessment module calculates the posterior probability of data changes of the tunnel structure based on the time series parameter set, associates the data to a single structural problem, including cracks and peeling, and generates a health assessment result;
[0066] The damage signature analysis module analyzes the health assessment results, collects key damage indicators, including internal corrosion and stress concentration, performs trend analysis and pattern recognition on the damage indicators, evaluates the current damage state of the tunnel, and generates a damage signature set;
[0067] The prediction integration module inputs the damage feature set into the integrated learning framework, uses the probabilistic prediction method to predict the future development trend of tunnel structure damage, evaluates the safety status of the tunnel in the future time period, and obtains dynamic damage prediction analysis results.
[0068] Compared with the prior art, the advantages and positive effects of the present invention are:
[0069] In the present invention, by combining real-time monitoring data with environmental change data and applying dynamic window technology to finely segment time series data, the real-time changes in tunnel internal and environmental conditions are effectively mapped. By updating the parameters of the time series data to calculate the posterior probability of data changes, the current health status of the tunnel can be accurately quantified. The health status assessment is analyzed to identify damage responses caused by increased tunnel service life and environmental changes, damage indicators are collected, and trend analysis and pattern recognition are performed to effectively integrate damage characteristics. The damage characteristics are input into the integrated learning framework, combined with probabilistic prediction, to accurately predict the development trend of tunnel structure damage and evaluate the future safety status of the tunnel. This method improves the efficiency and safety of tunnel management, reduces the risk of accidents, and optimizes maintenance and repair decisions. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] Figure 1 It is a schematic diagram of the workflow of the present invention;
[0071] Figure 2 A flow chart of the steps for obtaining a time series set of the present invention;
[0072] Figure 3 A flowchart of the steps for obtaining the health status assessment of the present invention;
[0073] Figure 4 is a flow chart of the steps for obtaining the damage feature set of the present invention;
[0074] Figure 5 The present invention is a flowchart of the steps for obtaining the dynamic damage prediction analysis results. DETAILED DESCRIPTION
[0075] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0076] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.
[0077] Embodiment 1
[0078] See also Figure 1 The present invention provides a technical solution: a tunnel structure damage assessment method, comprising the following steps:
[0079] S1: Collect real-time monitoring data of tunnel structure, integrate temperature and humidity environment change data, and use dynamic window technology to segment time series data to obtain time series sets;
[0080] S2: Update the parameters of the time series set, calculate the posterior probability of data changes, quantify the target damage of crack extension and concrete peeling, and establish a health status assessment;
[0081] S3: Analyze health status assessment, identify damage responses caused by tunnel age and environmental changes, collect damage indicators, including internal corrosion and stress concentration, perform trend analysis and pattern recognition on damage indicators, and integrate them into a damage feature set;
[0082] S4: Input the damage feature set into the integrated learning framework, combine the probability prediction and damage feature data, predict the development trend of tunnel structure damage, evaluate the safety status of the tunnel in the future time period, and obtain the dynamic damage prediction analysis results.
[0083] The time series set includes a monitoring data subset and an environmental data subset. The health status assessment includes crack status and concrete surface condition. The damage indicators are specifically corrosion degree and stress level. The dynamic damage prediction analysis results include damage trend and safety rating.
[0084] See also Figure 2 , the specific steps for obtaining the time series set are:
[0085] S111: Collect real-time monitoring data of the tunnel structure, including temperature and humidity, and use a sliding average model to smooth the data using the formula:
[0086] E t =α·T t +(1-α)·H t
[0087] Calculate the sliding mean temperature and humidity index to obtain the environmental data index;
[0088] Among them, T t Represents the temperature data at time point t, H t represents the humidity data at time point t, and α is the weight parameter;
[0089] S112: Based on the environmental data index, dynamic window technology is applied to segment time series data, and the weighted moving average formula is adopted:
[0090]
[0091] Get the segmented time series data;
[0092] Among them, W k represents the dataset of the kth time window, E t-i is the environmental data index at time point ti, n is the time window size, β i is the weight of multiple time points in the time window;
[0093] S113: Summarize the time series data after all time windows are segmented, using the formula:
[0094]
[0095] Calculate the time series set within the entire target time range to obtain the time series set;
[0096] Among them, S represents the time series set within the target time range, W j represents the dataset of the jth time window, and k is the number of time windows.
[0097] Detailed explanation of formulas and parameters
[0098] Environmental data index calculation formula
[0099] formula:
[0100] E t =α·T t +(1-α)·H t
[0101] Parameter explanation and source:
[0102] T t : Temperature data at time point t, assuming the measurement value at a specific time point.
[0103] H t : Humidity data at time point t, also assumed to be the measured value at time point.
[0104] α: Weight coefficient of temperature data, which is set based on historical data analysis and expert opinions, and the value range is [0,1].
[0105] Calculation example:
[0106] Assume that at a certain point in time, the measured temperature T t =20℃ and humidity H t =80%. Set the weight α = 0.7. Then the environmental data index is calculated as:
[0107] E t=0.7·20+(1-0.7)·80=14+24=38
[0108] Results E t =38 represents the environmental index after referring to the temperature and humidity at the time point.
[0109] Time window data set calculation formula
[0110] formula:
[0111]
[0112] Parameter explanation and source:
[0113] E t-i : Represents the environmental data index at time point ti.
[0114] β i : Weight coefficients for multiple time points within a time window, set according to data importance and time proximity.
[0115] n: Time window size, determined based on the frequency of the monitoring system and the required data coverage.
[0116] Calculation example:
[0117] Assume that time window n = 3, time window weight β 1 =0.5,β 2 =0.3,β 3 =0.2, and assume that the environmental data index E at three time points t-1 =38, E t-2 =35, E t-3 =40.
[0118] The calculation is as follows:
[0119]
[0120] Results k =37.5 represents the average environmental data index within the time window.
[0121] Time series set calculation formula
[0122] formula:
[0123]
[0124] Parameter explanation and source:
[0125] W j : The dataset of the j-th time window.
[0126] k: The number of time windows, which depends on the monitoring period length and analysis requirements.
[0127] Calculation example:
[0128] Assume there are 3 time windows and the data sets are W 1 =37.5,W 2 =36,W 3 =38. The calculation is as follows:
[0129]
[0130] The result S=37.17 represents the average environmental data index during the entire monitoring period.
[0131] See also Figure 3 , the specific steps for obtaining health status assessment are:
[0132] S211: Extract the monitoring data of the time point from the time series set parameters, combine the historical trend analysis, and calculate the probability of the monitoring data at each time point using the formula:
[0133]
[0134] Calculate the generated conditional probability;
[0135] Among them, D t represents the monitoring data at time point t, P(D t |S) is the monitoring of D under the time series set S t The conditional probability, P(S|D t ) is in D t The probability of monitoring S when it occurs, P(D t ) is D t The prior probability of , λ is the adjustment factor, which adjusts the probability value;
[0136] S212: Using conditional probability, apply weighted summation method to combine the conditional probabilities of all time points, using the formula:
[0137] P(S)=∑ t P(D t ∣S)·W t
[0138] Calculate the overall posterior probability and obtain the updated posterior probability;
[0139] Among them, P(S) is the posterior probability of the updated data change, W t It is the weight assigned based on historical data;
[0140] S213: Use the updated posterior probability to evaluate the current health status of the tunnel and define the damage identification formula and threshold using the formula:
[0141] H=max(0,κ·(P(S)-θ))
[0142] Generate health status assessments;
[0143] Among them, H is the health status assessment result, κ is the coefficient for adjusting the response sensitivity, and θ is the benchmark threshold for damage assessment.
[0144] Step 1: Calculate the conditional probability P(D t ∣S)
[0145] Formula analysis:
[0146]
[0147] parameter:
[0148] P(S|D t ): In the event of D t The probability of observing S under the condition .
[0149] P(D t ): D t The prior probability of occurrence.
[0150] P(S): D not observed t The probability of S under the condition .
[0151] λ: Adjustment factor, used to balance the denominator, set to a small positive number.
[0152] Assumptions and examples:
[0153] Suppose you are evaluating a surveillance system where:
[0154] P(S|D t )=0.8(very high probability S is in D t Occurs when)
[0155] P(D t ) = 0.1 (lower prior probability)
[0156] P(S) = 0.3 (S is observed but D t May not happen)
[0157] λ=0.05
[0158] Calculate P(D t |S):
[0159]
[0160] This means that given S, D t The posterior probability of occurrence is about 22.9%.
[0161] Step 2: Calculate the overall posterior probability P(S)
[0162] Formula analysis:
[0163]
[0164] parameter:
[0165] P(D t |S): The conditional probability obtained from step 1.
[0166] W t : The weight at time point t.
[0167] Assumptions and examples:
[0168] Using the above P(D t The result is ∣S)≈0.229.
[0169] Assume there are three time points t, with weights W 1 =0.5,W 2 =0.3,W 3 =0.2.
[0170] Calculate P(S):
[0171] P(S)=0.229×0.5+0.229×0.3+0.229×0.2=0.1145+0.0687+0.0458
[0172] =0.229
[0173] This shows that the weighted posterior probability combined across all time points remains at approximately 22.9%.
[0174] Step 3: Health status assessment H
[0175] Formula analysis:
[0176] H=max(0,κ·(P(S)-θ))
[0177] parameter:
[0178] P(S): posterior probability obtained from step .
[0179] κ: sensitivity coefficient.
[0180] θ: damage assessment threshold.
[0181] Assumptions and examples:
[0182] The result of P(S)=0.229 was used.
[0183] Assume κ=2 and θ=0.2.
[0184] Calculate H:
[0185] H=max(0,2·(0.229-0.2))=max(0,2·0.029)=0.058
[0186] This means that the health status assessment result is 0.058, indicating a mild potential injury state.
[0187] See also Figure 4 , the steps for obtaining the damage feature set are as follows:
[0188] S311: Extract data related to tunnel age and environmental changes from health status assessment, and apply the composite impact formula:
[0189]
[0190] The damage response R is calculated;
[0191] Among them, γ and δ adjust the contribution of age increase and environmental change, a and b are exponential parameters, and λ is the normalization factor;
[0192] S312: Integrate the damage response with the tunnel damage indicators, including internal corrosion and stress concentration, using the weighted sum formula:
[0193]
[0194] Calculate and generate damage characteristic vector V;
[0195] Among them, α, β, and γ are weight parameters that adjust the influence of multiple damage indicators in the eigenvector, and c and d provide nonlinear responses to corrosion and stress reactions;
[0196] S313: Perform trend analysis and pattern recognition on the damage feature vector using the statistical pattern recognition formula:
[0197]
[0198] The damage feature set F is calculated;
[0199] Among them, κ is the regulatory factor.
[0200] Formula analysis and derivation of step 1
[0201] formula:
[0202]
[0203] in:
[0204] γ, δ: These are the weights of the influencing factors, adjusting the contribution of increased service life (represented by H) and environmental changes (represented by E) to the damage response.
[0205] a, b: These are nonlinear adjustment indices used to enhance the model’s ability to match data complexity.
[0206] λ: Normalization factor, used to stabilize the calculation results and avoid the influence of extreme values.
[0207] Assumed parameter values:
[0208] Assume that γ = 0.8 and δ = 0.2, which reflects the critical impact of increasing service life on damage.
[0209] Assume a=2, b=2 to increase the nonlinear effect.
[0210] Assume λ = 0.05 and adjust the denominator of the formula slightly to ensure that the denominator is not zero.
[0211] Calculation example:
[0212] Assume that H=10 (degree of impact of years) and E=3 (degree of environmental change).
[0213] Calculation process:
[0214]
[0215] Result analysis:
[0216] The calculated result R≈77.90 represents the damage response index, and a higher value indicates that the tunnel faces serious structural problems.
[0217] Formula analysis and derivation of step 2
[0218] formula:
[0219]
[0220] in:
[0221] α, β, γ: weight parameters, which adjust the influence of damage response R, internal corrosion C, and stress concentration S in the damage eigenvector respectively.
[0222] c, d: Nonlinear adjustment index, which increases the adaptability of the model to complex damage.
[0223] Assumed parameter values:
[0224] α=0.5, β=0.3, γ=0.2.
[0225] c=2, d=2.
[0226] Calculation example:
[0227] Assume that R=77.90 (calculated in the previous step), C=4, and S=5.
[0228] Calculation process:
[0229]
[0230] Result analysis:
[0231] V = 48.75 represents the comprehensive score of the damage feature vector, and a higher score indicates a significant risk of structural damage.
[0232] Formula analysis and derivation of step 3
[0233] formula:
[0234]
[0235] in:
[0236] κ: Sensitivity adjustment factor, used to adjust the input sensitivity during pattern recognition.
[0237] Assumed parameter values:
[0238] κ=0.1.
[0239] Calculation example:
[0240] V = 48.75 (calculated in the previous step).
[0241] Calculation process:
[0242]
[0243] Result analysis:
[0244] F≈0.424 represents the characteristic set score of structural damage, and the value reflects the significance of damage characteristics and the estimated risk level.
[0245] See also Figure 5 , the specific steps for obtaining dynamic damage prediction analysis results are as follows:
[0246] S411: The damage feature set is input into the ensemble learning framework, combining multiple prediction models to enhance the accuracy and reliability of the prediction, using a weighted average formula:
[0247]
[0248] Adjust and stabilize the output to generate a probability prediction result P;
[0249] Among them, w i is the weight of the ith model, M i (F) is the output of model i based on the damage feature set F, η and ξ are parameters for adjusting the smoothness of the output;
[0250] S412: Combine probability predictions with historical damage data to perform trend analysis using the formula:
[0251] T=αP+βH t +γ
[0252] Increase the flexibility of the model and generate trend analysis results T;
[0253] Among them, α and β adjust the relative importance of current forecasts and historical data, and γ adjusts the baseline level;
[0254] S413: Based on the trend analysis results, a mathematical model is used to predict the damage development trend of the tunnel structure and evaluate the safety status in the future time period, using the formula:
[0255]
[0256] A variant of the sigmoid function is used to ensure that S is between 0 and 1, representing the damage probability, and to generate the dynamic damage prediction analysis result U;
[0257] Among them, e is the base of natural logarithm, which is used for sigmoid function transformation.
[0258] Detailed explanation and derivation of the formula for step 1
[0259] formula:
[0260]
[0261] in:
[0262] w i is the weight of the ith model, indicating the relative importance of the model in the overall prediction.
[0263] M i (F) is the output of model i based on the damage feature set F.
[0264] η and ξ are adjustment parameters added to increase the robustness of the model, to avoid the denominator being zero and to smooth the output.
[0265] Assumed parameter values:
[0266] Assume there are three models (n=3), where w 1 =0.5, w 2 =0.3, w 3 =0.2.
[0267] The model output is assumed to be M 1 (F) = 0.8, M 2 (F) = 0.6, M 3 (F) = 0.4.
[0268] Set η=0.1 and ξ=1.
[0269] Calculation process:
[0270]
[0271] Result analysis:
[0272] The calculated result P = 0.19 represents the average damage probability predicted by the model, which is a low value, indicating that the current structural damage risk of the tunnel is low.
[0273] Detailed explanation and derivation of the formula for step 2
[0274] formula:
[0275] T=αP+βH t +γ
[0276] in:
[0277] α and β adjust the relative importance of current forecasts and historical data.
[0278] γ is the intercept term, which provides a baseline adjustment for the model.
[0279] Assumed parameter values:
[0280] α=0.7, β=0.3.
[0281] Historical injury data t =0.15.
[0282] γ=0.05.
[0283] Calculation process:
[0284] T=0.7×0.19+0.3×0.15+0.05=0.133+0.045+0.05=0.228
[0285] Result analysis:
[0286] T = 0.228 indicates the trend analysis result after referring to the current and historical damage data. The lower value reflects the good safety risk management of the tunnel.
[0287] Detailed explanation and derivation of the formula for step 3
[0288] formula:
[0289]
[0290] in:
[0291] T is the trend analysis result obtained from step 2.
[0292] e is the base of the natural logarithm and is often used in the sigmoid function.
[0293] Calculation process:
[0294]
[0295] Result analysis:
[0296] U≈0.746 represents the damage probability after sigmoid adjustment. This value is closer to 1, indicating that there is a higher risk of structural damage in the future.
[0297] A tunnel structure damage assessment system, which is used to execute the above tunnel structure damage assessment method, and includes:
[0298] The data acquisition module collects real-time monitoring data of the tunnel from the monitoring equipment, and simultaneously obtains environmental data including temperature and humidity, updates time and environmental parameters, and generates a monitoring data set;
[0299] The time series analysis module inputs the monitoring data set, applies the dynamic window technology to segment the data into time series, identifies the key change points within the time range, and obtains the time series parameter set;
[0300] The health status assessment module calculates the posterior probability of data changes of the tunnel structure based on the time series parameter set, associates the data to single structural problems, including cracks and peeling, and generates health assessment results;
[0301] The damage signature analysis module analyzes the health assessment results, collects key damage indicators, including internal corrosion and stress concentration, performs trend analysis and pattern recognition on the damage indicators, evaluates the current damage state of the tunnel, and generates a damage signature set;
[0302] The prediction integration module inputs the damage feature set into the integrated learning framework, uses the probabilistic prediction method to predict the future development trend of tunnel structure damage, evaluates the safety status of the tunnel in the future time period, and obtains dynamic damage prediction analysis results.
[0303] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.
Claims
1. A tunnel structure damage assessment method, characterized in that: The following steps are involved: Collect real-time monitoring data of tunnel structures, integrate temperature and humidity environment change data, and use dynamic window technology to segment time series data to obtain time series sets; updating the parameters of the time series set, calculating the posterior probability of data changes, quantifying the target damage of crack extension and concrete peeling, and establishing a health status assessment; Analyze the health status assessment, identify damage responses caused by increased tunnel age and environmental changes, collect damage indicators, including internal corrosion and stress concentration, perform trend analysis and pattern recognition on the damage indicators, and integrate them into a damage feature set; The damage feature set is input into the integrated learning framework, and the probability prediction and damage feature data are combined to predict the development trend of tunnel structure damage, evaluate the safety status of the tunnel in the future time period, and obtain dynamic damage prediction analysis results.
2. The tunnel structure damage assessment method according to claim 1, characterized in that: The steps for obtaining the time series set are specifically as follows: Real-time monitoring data of the tunnel structure, including temperature and humidity, are collected and smoothed using a sliding average model using the formula: E t =α·T t +(1-α)·H t Calculate the sliding mean temperature and humidity index to obtain the environmental data index; Among them, T t Represents the temperature data at time point t, H t represents the humidity data at time point t, and α is the weight parameter; Based on the environmental data index, dynamic window technology is applied to segment the time series data, and the weighted moving average formula is adopted: Get the segmented time series data; Among them, W k represents the dataset of the kth time window, E t-i is the environmental data index at time point ti, n is the time window size, β i is the weight of multiple time points in the time window; Summarize the split time series data in all time windows using the formula: Calculate the time series set within the entire target time range to obtain the time series set; Among them, S represents the time series set within the target time range, W j represents the dataset of the jth time window, and k is the number of time windows.
3. The tunnel structure damage assessment method according to claim 2, characterized in that: The steps for obtaining the health status assessment are specifically as follows: The monitoring data of the time point is extracted from the time series set parameters, and the probability of the monitoring data at each time point is calculated by combining the historical trend analysis, using the formula: Calculate the generated conditional probability; Among them, D t represents the monitoring data at time point t, P(D t |S) is the monitoring of D under the time series set S t The conditional probability, P(S|D t ) is in D t The probability of monitoring S when it occurs, P(D t ) is D t The prior probability of , λ is the adjustment factor, which adjusts the probability value; Using the conditional probabilities, the weighted summation method is applied to synthesize the conditional probabilities of all time points, using the formula: P(S)=∑ t P(D t ∣S)·W t Calculate the overall posterior probability and obtain the updated posterior probability; Among them, P(S) is the posterior probability of the updated data change, W t It is the weight assigned based on historical data; Using the updated posterior probability, the current health status of the tunnel is evaluated, and the damage identification formula and threshold are defined, using the formula: H=max(0,κ·(P(S)-θ)) Generate health status assessments; Among them, H is the health status assessment result, κ is the coefficient for adjusting the response sensitivity, and θ is the benchmark threshold for damage assessment.
4. The tunnel structure damage assessment method according to claim 3, characterized in that: Data related to tunnel age and environmental changes are extracted from the health status assessment, using the composite impact formula: The damage response R is calculated; Among them, γ and δ adjust the contribution of age increase and environmental change, a and b are exponential parameters, and λ is the normalization factor; Integrating the damage response with the tunnel damage indicators, including internal corrosion and stress concentration, the weighted sum formula is applied: Calculate and generate damage characteristic vector V; Among them, α, β, and γ are weight parameters that adjust the influence of multiple damage indicators in the eigenvector, and c and d provide nonlinear responses to corrosion and stress reactions; The damage feature vector is subjected to trend analysis and pattern recognition using the statistical pattern recognition formula: The damage feature set F is calculated; Among them, κ is the regulatory factor.
5. The tunnel structure damage assessment method according to claim 4, characterized in that: The steps for obtaining the dynamic damage prediction analysis results are specifically as follows: The damage feature set is input into an integrated learning framework, combining multiple prediction models to enhance the accuracy and reliability of prediction, using a weighted average formula: Adjust and stabilize the output to generate a probability prediction result P; Among them, w i is the weight of the ith model, M i (F) is the output of model i based on the damage feature set F, η and ξ are parameters for adjusting the smoothness of the output; Combining the probabilistic forecasts with historical damage data, a trend analysis was performed using the formula: T=αP+βH t +g Increase the flexibility of the model and generate trend analysis results T; Among them, α and β adjust the relative importance of current forecasts and historical data, and γ adjusts the baseline level; Based on the trend analysis results, a mathematical model is used to predict the damage development trend of the tunnel structure and evaluate the safety status in the future time period, using the formula: A variant of the sigmoid function is used to ensure that S is between 0 and 1, representing the damage probability, and to generate the dynamic damage prediction analysis result U; Among them, e is the base of natural logarithm, which is used for sigmoid function transformation.
6. A tunnel structure damage assessment system, characterized in that: According to the tunnel structure damage assessment method according to any one of claims 1 to 5, the system comprises: The data acquisition module collects real-time monitoring data of the tunnel from the monitoring equipment, and simultaneously obtains environmental data including temperature and humidity, updates time and environmental parameters, and generates a monitoring data set; The time series analysis module inputs the monitoring data set, applies dynamic window technology to perform time series segmentation on the data, identifies key change points within the time range, and obtains a time series parameter set; The health status assessment module calculates the posterior probability of data changes of the tunnel structure based on the time series parameter set, associates the data to a single structural problem, including cracks and peeling, and generates a health assessment result; The damage signature analysis module analyzes the health assessment results, collects key damage indicators, including internal corrosion and stress concentration, performs trend analysis and pattern recognition on the damage indicators, evaluates the current damage state of the tunnel, and generates a damage signature set; The prediction integration module inputs the damage feature set into the integrated learning framework, uses the probabilistic prediction method to predict the future development trend of tunnel structure damage, evaluates the safety status of the tunnel in the future time period, and obtains dynamic damage prediction analysis results.
Citation Information
Patent Citations
Method for evaluating residual life of in-service digging rope with broken outer sheath based on defect equivalent processing
CN106769823A
Corrosion fatigue life prediction method and system for large-span bridge cable system
CN111191318A
Structure early warning analysis method based on Bayesian model
CN111291481A
Bridge structure damage identification method based on sparse Bayesian learning
CN112528564A
Power transmission line iron tower safety evaluation system and evaluation method thereof
CN117745077A
Cited By
Method and system for evaluating reliability of petroleum pipeline sealing structure
CN119129441A
Intelligent monitoring system for wharf breakwater
CN119475190A
Tunnel structure full life cycle health condition evaluation method and system
CN119989501A
Intelligent leakage current real-time monitoring and diagnosis system and method
CN120214633A
Intelligent leakage current real-time monitoring and diagnosis system and method
CN120214633B
