A tunnel lining concrete damage risk control method and system

Abstract

The present invention discloses a tunnel lining concrete damage risk control method and system, which relates to tunnel risk control technology. The method includes: regularly collecting monitoring data of lining concrete in the control tunnel; identifying abnormal change parameters of the lining concrete monitoring data in the control tunnel; inputting the identified abnormal change parameters into a lining concrete damage risk prediction model, outputting the prediction result of the damage risk; and transmitting the output result of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge. The monitoring and diagnosis capabilities of tunnel lining concrete damage risks are improved, the safety and stability of the tunnel are improved, and the service life of the tunnel is extended.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel risk management and control, and in particular to a tunnel lining concrete damage risk management and control method and system. Background Art

[0002] Under different geological conditions, the loads and risks borne by tunnel linings vary greatly. For example, tunnel lining maintenance under special geological conditions such as soft soil, expansive soil, and karst has its own particularity and difficulty, and environmental factors such as groundwater level changes, acid rain, and salt damage will affect the durability of tunnel lining concrete, but these factors are relatively difficult to predict and control. In order to overcome these defects and challenges, it is necessary to research and develop a tunnel lining concrete damage risk management method with high monitoring and diagnostic capabilities to improve the safety and stability of the tunnel and extend the service life of the tunnel. Summary of the invention

[0003] The present invention provides a tunnel lining concrete damage risk management method, comprising:

[0004] Step 1: Regularly collect monitoring data of the lining concrete in the control tunnel;

[0005] Step 2: Identify and control abnormal change parameters of the monitoring data of the lining concrete in the tunnel;

[0006] Step 3, input the identified abnormal change parameters into the lining concrete damage risk prediction model, and output the prediction result of damage risk;

[0007] Step 4: Transmit the output results of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge.

[0008] The tunnel lining concrete damage risk control method described above, wherein the abnormal change parameters of the lining concrete monitoring data in the tunnel are identified and controlled, specifically includes the following sub-steps:

[0009] Select a tunnel with simple soil environment as the standard tunnel;

[0010] Regularly collect monitoring data of standard tunnel lining concrete;

[0011] Based on the monitoring data of standard tunnel lining concrete, abnormal change parameters of lining concrete in the controlled tunnel are detected.

[0012] In the above-mentioned tunnel lining concrete damage risk control method, the establishment of the lining concrete damage risk prediction model specifically includes the following sub-steps:

[0013] Create a training data set based on the historical damage records and historical monitoring records of lining concrete;

[0014] A lining concrete damage risk prediction model is trained based on the training data set;

[0015] Human intervention was introduced to test the performance of the lining concrete damage risk prediction model.

[0016] A tunnel lining concrete damage risk control method as described above, wherein a training data set is created based on historical damage assessment records and historical monitoring records of the lining concrete, specifically comprising the following sub-steps:

[0017] The damage type and damage degree in each historical damage assessment record are fused and encoded to form an output set;

[0018] The abnormal change parameters corresponding to each historical damage assessment record in the historical monitoring records are extracted as the input set.

[0019] A tunnel lining concrete damage risk control method as described above, wherein the content reflected in the risk report includes damage risk level, specific damage type, damage degree, and recent monitoring data; wherein the damage risk level, specific damage type, and damage degree are all derived from the interpretation of the output results of the lining concrete damage risk prediction model.

[0020] The present invention also provides a tunnel lining concrete damage risk management and control system, including: a monitoring data acquisition module, an abnormal change identification module, a damage risk prediction module, and a risk report creation module;

[0021] Monitoring data collection module, used to regularly collect monitoring data of the lining concrete in the control tunnel;

[0022] Abnormal change identification module, used to identify abnormal change parameters of lining concrete monitoring data in the control tunnel;

[0023] A damage risk prediction module is used to input the identified abnormal change parameters into the lining concrete damage risk prediction model and output the damage risk prediction result;

[0024] The risk report creation module is used to transmit the output results of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge.

[0025] A tunnel lining concrete damage risk management and control system as described above, wherein the abnormal change identification module includes a standard tunnel definition submodule and an abnormal monitoring submodule;

[0026] The standard tunnel definition submodule is used to select a tunnel with a simple soil environment as a standard tunnel;

[0027] The abnormal monitoring submodule is used to detect abnormal change parameters of the lining concrete in the controlled tunnel based on the monitoring data of the standard tunnel lining concrete.

[0028] A tunnel lining concrete damage risk management and control system as described above, wherein the damage risk prediction module includes a training data set creation submodule, a lining concrete damage risk prediction model training submodule, and a lining concrete damage risk prediction model testing submodule;

[0029] A training data set creation submodule is used to create a training data set based on the historical damage assessment records and historical monitoring records of lining concrete;

[0030] A lining concrete damage risk prediction model training submodule is used to train a lining concrete damage risk prediction model based on a training data set;

[0031] The lining concrete damage risk prediction model testing submodule is used to introduce human intervention to test the performance of the lining concrete damage risk prediction model.

[0032] The beneficial effects achieved by the present invention are as follows: the damage details of the lining concrete damage risk are obtained in advance, and the risk control plan is customized in time according to the damage details to prevent accidents in real time; the monitoring and diagnosis capabilities of the tunnel lining concrete damage risk are improved; the safety and stability of the tunnel are improved, and the service life of the tunnel is extended. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0034] Figure 1 It is a flow chart of a tunnel lining concrete damage risk management method provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0035] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0036] Embodiment 1

[0037] like Figure 1As shown, the first embodiment of the present invention provides a tunnel lining concrete damage risk control method, including:

[0038] Step S10: regularly collecting monitoring data of the lining concrete in the control tunnel;

[0039] The monitoring points to be set up need to cover key locations such as tunnel intersections, lining joints, and geological change areas; by combining automatic sensor collection with manual collection, key parameters such as tunnel lining integrity, stress state, crack development, displacement changes, sound wave velocity, carbonization depth, steel bar condition, groundwater level, and temperature changes at the monitoring points should be regularly collected.

[0040] Step S20: Identify abnormal change parameters of monitoring data of lining concrete in the control tunnel;

[0041] By comparing the changing trends of key parameters of the lining concrete of the standard tunnel and the control tunnel, the abnormally changing parameters are detected. These parameters are the key parameters that may cause damage to the tunnel lining concrete. Specifically:

[0042] Step S21: Select a tunnel with a simple soil environment as a standard tunnel;

[0043] The simple soil environment can ensure that the damage to the tunnel lining concrete comes mostly from natural aging rather than environmental influences. Using such a tunnel as a benchmark, we can observe the abnormal damage to the tunnel caused by the complex environment.

[0044] Step S22: regularly collecting monitoring data of the lining concrete in the standard tunnel;

[0045] The time interval for collecting monitoring data of lining concrete in standard tunnels must be consistent with the time interval for collecting data in control tunnels, and the collected monitoring data must be consistent with the data collected in control tunnels.

[0046] Step S23: using the monitoring data of the standard tunnel lining concrete as a benchmark, detecting abnormal change parameters of the lining concrete in the control tunnel;

[0047] This step is implemented through the anomaly detection function. The key parameters of the lining concrete of the control tunnel collected are used as the input parameters of the anomaly detection function. Then, according to the detection results returned by the function, it is judged whether the input key parameters are abnormal change parameters. The anomaly detection function is expressed as:

[0048]

[0049] where a i is the i-th key parameter for controlling tunnel lining concrete, a it+1 Indicates that in the t+1th data collection, a i The value of ait Indicates that in the tth data collection, a i The value of b it+1 represents the i-th key parameter b of the standard tunnel lining concrete in the t+1th acquisition data i The value of b it represents the i-th key parameter b of the standard tunnel lining concrete in the t-th acquisition data i The value of t ranges from 1 to T, T is the detection period, and λ represents the detection sensitivity. Returns 1, indicating a i For abnormal change parameters, 0 is returned in other cases, indicating a i It is a normal variation parameter.

[0050] Step S30: inputting the identified abnormal change parameters into the lining concrete damage risk prediction model, and outputting the prediction result of the damage risk;

[0051] The establishment of the lining concrete damage risk prediction model specifically includes the following sub-steps:

[0052] Step S31: creating a training data set based on historical damage assessment records and historical monitoring records of lining concrete;

[0053] The historical damage assessment records of lining concrete include damage type, damage degree, damage assessment time and location; the historical monitoring records include collection time, collection location, and various key parameters collected; the training data set includes two parts: input set and output set. The specific creation process includes the following sub-steps:

[0054] Step S311: fusing and encoding the damage type and damage degree in each historical damage assessment record to form an output set;

[0055] The output set is represented by D = {d1, d2, d3, ..., d N}, where d1, d2, d3, ..., d N They respectively represent the fusion code of damage type and damage degree in each historical damage assessment record, and N represents the total number of historical damage assessment records.

[0056] The fusion coding calculation formula is expressed as: k =μ k ×ε k , where d k is the fusion code of damage type and damage degree in the kth historical damage assessment record, μ k is the risk weight of the damage type in the kth historical damage record, ε k is the quantitative value of the damage degree in the kth historical damage assessment record, where e krepresents the damage degree in the kth historical damage record, e max Indicates the highest value of the damage degree in all historical damage records, e min Indicates the lowest value of the damage degree in all historical damage records. It represents the average of the damage extent in all historical damage assessment records.

[0057] After the fusion coding calculation is completed, it needs to be associated with the original damage type and damage degree, and these associations are stored in a mapping table.

[0058] Step S312: extracting the abnormal change parameters corresponding to each historical damage assessment record in the historical monitoring records as an input set;

[0059] The key parameters of the historical monitoring data of the same location (the collection location is the same as the damage location) within a detection period T before the damage assessment time of each historical damage assessment record are passed as input parameters to the anomaly detection function F(a i ), the key parameter passed in is determined according to the return result of the detection function to determine whether it is an abnormal change parameter. If so, it is stored in the input set, otherwise it is ignored. The input set is represented by: H = {{h1,h2,h3,...h L1},{h1,h2,h3,...h L2},...,{h1,h2,h3,...h LN}}, where {h1,h2,h3,...h L1} is the abnormal change parameter set corresponding to the first damage assessment record, {h1,h2,h3,...h L2} is the abnormal change parameter set corresponding to the second damage assessment record, {h1,h2,h3,...h LN} is the abnormal change parameter set corresponding to the Nth damage assessment record, h1,h2,h3,...h L1 are the abnormal change parameters corresponding to the first damage assessment record, h1, h2, h3, ...h L2 、h1,h2,h3,...h LN By analogy, L1, L2, and LN are the numbers of abnormal change parameters corresponding to the first, second, and Nth damage assessment records respectively.

[0060] Step S32: training a lining concrete damage risk prediction model based on the training data set;

[0061] Select any classification model as the base model, and then use the training data set to train and adjust the base model so that it has the ability to predict the damage risk of lining concrete. Specifically:

[0062] The abnormal change parameter sets {h1,h2,h3,...hL1},{h1,h2,h3,...h L2},...,{h1,h2,h3,...h LN} as the model input sample, and its corresponding fusion codes d1, d2, d3, ..., d N As the actual output, the base model is supervised and trained, and the loss function is designed as the basis for adjusting the model parameters. The loss function R() designed for the lining concrete damage risk prediction model is expressed as: R(d k ,p k )=-d k *log(p k ), where d k represents the kth input sample {h1,h2,h3,...,h Lk}, LK is the number of abnormal change parameters in the kth input sample, p k represents the kth input sample {h1,h2,h3,...,h Lk}Prediction output of lining concrete damage risk prediction model.

[0063] Step S33: introducing manual intervention to test the performance of the lining concrete damage risk prediction model;

[0064] The prediction results of the lining concrete damage risk prediction model are tested with new samples to see if they are accurate. The results are manually reviewed and judged. If the test results do not meet the standards, it is necessary to return to step S32 to increase the training weights of the new samples similar to the training samples and further optimize the model. Only the lining concrete damage risk prediction model that passes the test can be put into the production environment.

[0065] Step S40: The output result of the lining concrete damage risk prediction model together with the monitoring data is transmitted to the person in charge in the form of a risk report;

[0066] The risk report should reflect the risk level of future damage to the concrete lining in the controlled tunnel, the specific type and degree of damage, and include recent monitoring data to provide a reference for the person in charge to customize the control plan.

[0067] The type and degree of damage that the lining concrete in the control tunnel will suffer in the future can be obtained from the mapping table that stores the damage type, damage degree, and fusion coding association relationship, based on the output results of the lining concrete damage risk prediction model.

[0068] The risk level is defined based on the fusion coding of damage type and damage degree in historical damage assessment records. The higher the fusion coding value, the higher the risk level, and the lower the fusion coding value, the lower the risk level.

[0069] Embodiment 2

[0070] Embodiment 2 of the present invention provides a tunnel lining concrete damage risk management and control method system, including: a monitoring data acquisition module, an abnormal change identification module, a damage risk prediction module, and a risk report creation module;

[0071] (1) Monitoring data collection module, used to regularly collect monitoring data of the lining concrete in the control tunnel;

[0072] The monitoring points to be set up need to cover key locations such as tunnel intersections, lining joints, and geological change areas; by combining automatic sensor collection with manual collection, key parameters such as tunnel lining integrity, stress state, crack development, displacement changes, sound wave velocity, carbonization depth, steel bar condition, groundwater level, and temperature changes at the monitoring points should be regularly collected.

[0073] (2) An abnormal change identification module, which is used to identify abnormal change parameters of the lining concrete monitoring data in the control tunnel, including a standard tunnel definition submodule and an abnormal monitoring submodule;

[0074] 1. The standard tunnel definition submodule is used to select a tunnel with a simple soil environment as the standard tunnel;

[0075] The simple soil environment can ensure that the damage to the tunnel lining concrete comes mostly from natural aging rather than environmental influences. Using such a tunnel as a benchmark, we can observe the abnormal damage to the tunnel caused by the complex environment.

[0076] 2. The abnormal monitoring submodule is used to detect abnormal change parameters of the lining concrete in the controlled tunnel based on the monitoring data of the standard tunnel lining concrete;

[0077] The time interval for collecting monitoring data of lining concrete in standard tunnels must be consistent with the time interval for collecting data in control tunnels, and the collected monitoring data must be consistent with the data collected in control tunnels.

[0078] The key parameters of the tunnel lining concrete collected are used as input parameters of the anomaly detection function. Then, according to the detection results returned by the function, it is determined whether the input key parameters are abnormal change parameters. The anomaly detection function is expressed as:

[0079]

[0080] where a i is the i-th key parameter for controlling tunnel lining concrete, a it+1 Indicates that in the t+1th data collection, a i The value of a it Indicates that in the tth data collection, a i The value of b it+1represents the i-th key parameter b of the standard tunnel lining concrete in the t+1th acquisition data i The value of b it represents the i-th key parameter b of the standard tunnel lining concrete in the t-th acquisition data i The value of t ranges from 1 to T, T is the detection period, and λ represents the detection sensitivity. Returns 1, indicating a i For abnormal change parameters, 0 is returned in other cases, indicating a i It is a normal variation parameter.

[0081] (3) a damage risk prediction module, which is used to input the identified abnormal change parameters into the lining concrete damage risk prediction model and output the damage risk prediction result, including a training data set creation submodule, a lining concrete damage risk prediction model training submodule, and a lining concrete damage risk prediction model testing submodule;

[0082] 1. The training data set creation submodule is used to create a training data set based on the historical damage assessment records and historical monitoring records of lining concrete;

[0083] The historical damage assessment records of lining concrete include damage type, damage degree, damage assessment time and location; the historical monitoring records include collection time, collection location, and various key parameters collected; the training data set includes two parts: input set and output set. The specific creation process includes:

[0084] I. Fusion code the damage type and damage degree in each historical damage assessment record to form an output set;

[0085] The output set is represented by D = {d1, d2, d3, ..., d N}, where d1, d2, d3, ..., d N They respectively represent the fusion code of damage type and damage degree in each historical damage assessment record, and N represents the total number of historical damage assessment records.

[0086] The fusion coding calculation formula is expressed as: k =μ k ×ε k , where d k is the fusion code of damage type and damage degree in the kth historical damage assessment record, μ k is the risk weight of the damage type in the kth historical damage record, ε k is the quantitative value of the damage degree in the kth historical damage assessment record; where e k represents the damage degree in the kth historical damage record, e max Indicates the highest value of the damage degree in all historical damage records, emin Indicates the lowest value of the damage degree in all historical damage records. It represents the average of the damage extent in all historical damage assessment records.

[0087] After the fusion coding calculation is completed, it needs to be associated with the original damage type and damage degree, and these associations are stored in a mapping table.

[0088] II. Extract the abnormal change parameters corresponding to each historical damage assessment record in the historical monitoring records as the input set;

[0089] The key parameters of the historical monitoring data of the same location (the collection location is the same as the damage location) within a detection period T before the damage assessment time of each historical damage assessment record are passed as input parameters to the anomaly detection function F(a i ), the key parameter passed in is determined according to the return result of the detection function to determine whether it is an abnormal change parameter. If so, it is stored in the input set, otherwise it is ignored. The input set is represented by: H = {{h1,h2,h3,...h L1},{h1,h2,h3,...h L2},...,{h1,h2,h3,...h LN}}, where {h1,h2,h3,...h L1} is the abnormal change parameter set corresponding to the first damage assessment record, {h1,h2,h3,...h L2} is the abnormal change parameter set corresponding to the second damage assessment record, {h1,h2,h3,...h LN} is the abnormal change parameter set corresponding to the Nth damage assessment record, h1,h2,h3,...h L1 are the abnormal change parameters corresponding to the first damage assessment record, h1, h2, h3, ...h L2 、h1,h2,h3,…h LN By analogy, L1, L2, and LN are the numbers of abnormal change parameters corresponding to the first, second, and Nth damage assessment records respectively.

[0090] 2. Lining concrete damage risk prediction model training submodule, used to train a lining concrete damage risk prediction model based on the training data set;

[0091] Select any classification model as the base model, and then use the training data set to train and adjust the base model so that it has the ability to predict the damage risk of lining concrete. Specifically:

[0092] The abnormal change parameter sets {h1,h2,h3,...h L1},{h1,h2,h3,...h L2},...,{h1,h2,h3,...h LN} as the model input sample, and its corresponding fusion codes d1, d2, d3, ..., d N As the actual output, the base model is supervised and trained, and the loss function is designed as the basis for adjusting the model parameters. The loss function R() designed for the lining concrete damage risk prediction model is expressed as:

[0093] R(d k ,p k )=-d k *log(p k ), where d k represents the kth input sample {h1,h2,h3,...,h Lk}, LK is the number of abnormal change parameters in the kth input sample, p k represents the kth input sample {h1,h2,h3,...,h Lk}Prediction output of lining concrete damage risk prediction model.

[0094] 3. Lining concrete damage risk prediction model test submodule, used to introduce manual intervention to test the performance of the lining concrete damage risk prediction model;

[0095] The prediction results of the lining concrete damage risk prediction model are tested with new samples to see if they are accurate. The results are manually reviewed and judged. If the test results do not meet the standards, it is necessary to return to step S32 to increase the training weights of the new samples similar to the training samples and further optimize the model. Only the lining concrete damage risk prediction model that passes the test can be put into the production environment.

[0096] (4) Risk report creation module, used to transmit the output results of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge;

[0097] The risk report should reflect the risk level of future damage to the concrete lining in the controlled tunnel, the specific type and degree of damage, and include recent monitoring data to provide a reference for the person in charge to customize the control plan.

[0098] The type and degree of damage that the lining concrete in the control tunnel will suffer in the future can be obtained from the mapping table that stores the damage type, damage degree, and fusion coding association relationship, based on the output results of the lining concrete damage risk prediction model.

[0099] The risk level is defined based on the fusion coding of damage type and damage degree in historical damage assessment records. The higher the fusion coding value, the higher the risk level, and the lower the fusion coding value, the lower the risk level.

[0100] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A tunnel lining concrete damage risk management method, characterized in that: The method comprises: Step 1: Regularly collect monitoring data of the lining concrete in the control tunnel; Step 2: Identify and control abnormal change parameters of the monitoring data of the lining concrete in the tunnel; Step 3, input the identified abnormal change parameters into the lining concrete damage risk prediction model, and output the prediction result of damage risk; Step 4: Transmit the output results of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge; Identify and control abnormal change parameters of the monitoring data of the lining concrete in the tunnel, including the following sub-steps: Select a tunnel with simple soil environment as the standard tunnel; Regularly collect monitoring data of standard tunnel lining concrete; Based on the monitoring data of standard tunnel lining concrete, the abnormal change parameters of lining concrete in the controlled tunnel are detected; The establishment of the lining concrete damage risk prediction model specifically includes the following sub-steps: Create a training data set based on the historical damage records and historical monitoring records of lining concrete; A lining concrete damage risk prediction model is trained based on the training data set; Human intervention was introduced to test the performance of the lining concrete damage risk prediction model; The training data set is created based on the historical damage records and historical monitoring records of lining concrete, which includes the following sub-steps: The damage type and damage degree in each historical damage assessment record are fused and encoded to form an output set; The abnormal change parameters corresponding to each historical damage assessment record in the historical monitoring records are extracted as the input set.

2. A tunnel lining concrete damage risk management method according to claim 1, characterized in that: The risk report includes the injury risk level, specific injury type, injury degree, and recent monitoring data; The damage risk level, specific damage type and damage degree are all derived from the interpretation of the output results of the lining concrete damage risk prediction model.

3. A tunnel lining concrete damage risk management and control system, characterized in that: The system includes: a monitoring data acquisition module, an abnormal change identification module, a damage risk prediction module, and a risk report creation module; Monitoring data collection module, used to regularly collect monitoring data of the lining concrete in the control tunnel; Abnormal change identification module, used to identify abnormal change parameters of lining concrete monitoring data in the control tunnel; A damage risk prediction module is used to input the identified abnormal change parameters into the lining concrete damage risk prediction model and output the damage risk prediction result; The risk report creation module is used to transmit the output results of the lining concrete damage risk prediction model together with the monitoring data into a risk report to the person in charge; The abnormal change identification module includes a standard tunnel definition submodule and an abnormal monitoring submodule; The standard tunnel definition submodule is used to select a tunnel with a simple soil environment as a standard tunnel; The abnormal monitoring submodule is used to detect abnormal change parameters of the lining concrete in the controlled tunnel based on the monitoring data of the standard tunnel lining concrete; The damage risk prediction module includes a training data set creation submodule, a lining concrete damage risk prediction model training submodule, and a lining concrete damage risk prediction model testing submodule; The training data set creation submodule is used to create a training data set based on the historical damage assessment records and historical monitoring records of the lining concrete; the training data set is created based on the historical damage assessment records and historical monitoring records of the lining concrete, specifically: the damage type and damage degree in each historical damage assessment record are fused and encoded to form an output set; the abnormal change parameters corresponding to each historical damage assessment record in the historical monitoring record are extracted as the input set; A lining concrete damage risk prediction model training submodule is used to train a lining concrete damage risk prediction model based on a training data set; The lining concrete damage risk prediction model testing submodule is used to introduce human intervention to test the performance of the lining concrete damage risk prediction model.

Images