A deformation monitoring method, device and system for a lock chamber wall

Abstract

The present application provides a method, device and system for monitoring the deformation of a lock chamber wall. The method includes: S1, obtaining a displacement data set of the lock chamber wall in a first preset period starting from the current moment; S2, based on the displacement data set, judging whether the deformation condition of the lock chamber wall meets a first alarm condition according to a preset judgment rule; when the first alarm condition is met, generating a first early warning message and sending it to a preset user device; when the first alarm condition is not met, executing S1; wherein, the preset judgment rule means generating a set of monitoring curves based on the displacement data set; obtaining a first similarity score between the set of monitoring curves and a set of sample curves; when the first similarity score is lower than a first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition. The method of using displacement-time curves for monitoring reduces the amount of data and thus improves the overall efficiency of monitoring.

Description

Technical Field

[0001] This application relates to the technical field of structural deformation monitoring methods, and particularly to a deformation monitoring method, device and system for a lock chamber wall. Background Art

[0002] The lock chamber wall of a ship lock is an important part of the ship lock, which is used to form the lock chamber. It not only plays a role in retaining water, but also bears the water pressure generated due to water level changes, etc. The lock chamber wall and the lock chamber floor can be an integral structure connected together, or a separated structure not connected together. The lock chamber wall bears the action of various forces and the deformation under construction conditions. The deformation monitoring of the lock chamber wall usually includes the measurement of indicators such as displacement. Monitoring methods can use instruments such as level gauges, total stations, GPS, fiber optic sensors, etc. The analysis and interpretation of monitoring data require combining professional knowledge and experience to ensure the accuracy and reliability of monitoring results. Therefore, the original manual monitoring cannot keep up with the existing requirements in terms of accuracy and monitoring efficiency.

[0003] Exemplarily, taking the deformation monitoring and status warning method for lock and pumping station hydraulic structures disclosed in the application number CN201510622076.2 as an example, it is used to judge the operation status of hydraulic structures and give a real-time warning of unsafe states to ensure the safe operation of water conservancy projects. It includes steps: the step of selecting engineering monitoring points; the step of collecting monitoring information; the step of analyzing and comparing monitoring information; the step of processing analysis results; the step of status warning. It can realize functions such as the release of monitoring information and data analysis results, and the status warning of the project, and conduct real-time analysis on the safety of lock and pumping station buildings, enabling staff to timely master the deformation amount of the buildings and analyze its change trend. However, in order to avoid inaccurate monitoring results caused by fluctuations in the collected monitoring information, the above method requires a large number of monitoring information for multiple engineering monitoring points and cannot achieve a balance between monitoring accuracy and monitoring efficiency.

[0004] Therefore, this application aims to provide a deformation monitoring method, device and system for a lock chamber wall to solve the above problems. Summary of the Invention

[0005] The purpose of this application is to provide a deformation monitoring method, device and system for a lock chamber wall to improve the overall monitoring efficiency on the premise of ensuring the accuracy of judgment. The purpose of this application is achieved by adopting the following technical solutions:

[0006] This application provides a deformation monitoring method for a lock chamber wall, which is applied to an Internet of Things network. The Internet of Things network includes a gateway module and a control module, as well as a plurality of displacement sensors arranged on the lock chamber wall. Each of the displacement sensors uploads displacement data that changes over time to the control module through the gateway module. The method includes:

[0007] S1. In a first preset period starting at the current moment, use each of the displacement sensors to obtain a displacement data set of the lock chamber wall; the displacement data set includes multiple sets of displacement data that change over time corresponding to each preset monitoring point, and the preset monitoring points are the monitoring points set at the joints of adjacent lock segments. The displacement data that changes over time is used to indicate the deformation displacement situation between the corresponding adjacent lock segments.

[0008] S2. Based on the displacement data set, determine whether the deformation situation of the lock chamber wall meets the first alarm condition according to a preset determination rule; when the first alarm condition is met, generate a first warning message and send it to a preset user device; when the first alarm condition is not met, execute S1.

[0009] Among them, the preset determination rule means that a monitoring curve set is generated based on the displacement data set, and the monitoring curve set includes the displacement-time curves of each preset monitoring point; obtain the first similarity score between the monitoring curve set and the sample curve set; when the first similarity score is lower than the first score threshold, it is considered that the deformation situation of the lock chamber wall meets the first alarm condition.

[0010] Further, each first preset period includes a second preset period. The method for obtaining the sample curve set includes:

[0011] Take the first preset period starting at the current moment as the target preset period, and in the second preset period within the target preset period, obtain the stress monitoring information of multiple stress acquisition areas of the lock chamber wall.

[0012] Take the stress information obtained in the second preset periods within multiple first preset periods before the target preset period as stress sample information.

[0013] Compare each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score.

[0014] Take the stress sample information that is not lower than the second similarity threshold and is closest to the current moment as the target stress information, and take the monitoring curve set corresponding to the target stress information as the sample curve set.

[0015] Further, the step of taking the stress information obtained in the second preset periods within multiple first preset periods before the target preset period as stress sample information includes:

[0016] When the deformation situation of the lock chamber wall in the first preset period before the target preset period does not meet the first alarm condition, input the stress information of the first preset period that does not meet the first alarm condition into an abnormal area detection model to obtain the abnormal detection result of each stress acquisition area; the abnormal detection result includes abnormal and normal.

[0017] Determine whether the number of abnormal stress acquisition regions is greater than a preset number; the stress information includes the identifier of each stress acquisition region and a set of stress data for the region corresponding to each identifier;

[0018] Use the stress information of the first preset period in which the number of abnormal stress acquisition regions is greater than the preset number as stress sample information for obtaining the second similarity score of the stress monitoring information.

[0019] Further, the abnormal region detection model is obtained by training a deep learning model, and the training process of the abnormal region detection model includes the following steps:

[0020] Obtain a training set, where the training set includes a plurality of training data, and each training data includes a data group of a set of sample stress data and annotation data of the abnormal region corresponding to the data group; for each training data in the training set, perform the following processing:

[0021] Input the data group in the training data into a preset deep learning model to obtain prediction data of the abnormal region corresponding to the data group;

[0022] Update the model parameters of the deep learning model based on the prediction data and the annotation data corresponding to the sample image data;

[0023] Detect whether a preset training end condition is satisfied; if so, use the trained deep learning model as the abnormal region detection model; if not, continue to train the deep learning model with the next training data.

[0024] Further, obtaining the first similarity score between the monitoring curve set and the sample curve set includes:

[0025] Obtain the similarity value of the displacement-time curves of the same preset monitoring points in the monitoring curve set and the sample curve set to obtain the similarity value of each preset monitoring point;

[0026] For each preset monitoring point, when the corresponding stress acquisition region is an abnormal region, use the product of the obtained similarity value and the abnormal coefficient of the stress acquisition region corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, use the obtained similarity value as the first similarity intermediate score;

[0027] Use the sum of the first similarity intermediate scores of each obtained preset monitoring point as the first similarity score between the monitoring curve set and the sample curve set.

[0028] Further, the obtaining method of the abnormal coefficient of the stress acquisition region includes:

[0029] Obtain the corresponding relationship of abnormal coefficients; the corresponding relationship of abnormal coefficients is used to indicate the corresponding relationship among the number, proportion, and abnormal values of the stress acquisition areas with abnormalities.

[0030] According to the number of stress acquisition areas indicating abnormalities in the target stress information, use the corresponding relationship of abnormal coefficients to obtain the corresponding abnormal value, and use the abnormal value as the abnormal coefficient of the stress acquisition area indicating abnormalities.

[0031] In a second aspect, the present application also provides a deformation monitoring device for a lock chamber wall. The deformation monitoring device for a lock chamber wall includes:

[0032] A first acquisition module, configured to obtain a displacement data set of the lock chamber wall by using each of the displacement sensors within a first preset period starting from the current moment; the displacement data set includes multiple groups of displacement data varying with time corresponding to each preset monitoring point, and the preset monitoring points are the monitoring points set at the joints of adjacent lock sections. The displacement data varying with time is used to indicate the deformation displacement situation between the corresponding adjacent lock sections.

[0033] A first judgment module, configured to judge whether the deformation situation of the lock chamber wall meets a first alarm condition based on the displacement data set according to a preset judgment rule; when the first alarm condition is met, generate a first early warning message and send it to a preset user device.

[0034] Wherein, the preset judgment rule means that a monitoring curve set is generated based on the displacement data set, and the monitoring curve set includes the displacement-time curves of each preset monitoring point; obtain the first similarity score between the monitoring curve set and the sample curve set; when the first similarity score is lower than the first score threshold, it is considered that the deformation situation of the lock chamber wall meets the first alarm condition.

[0035] In a third aspect, the present application also provides a deformation monitoring system for a lock chamber wall. The deformation monitoring system for a lock chamber wall includes:

[0036] Multiple displacement sensors, respectively arranged at the preset monitoring points at the joints of adjacent lock sections of the lock chamber wall, for obtaining displacement data varying with time.

[0037] A control module, which is respectively communicatively connected to each of the displacement sensors.

[0038] A gateway module, which is arranged between the control module and the user device to realize the communication connection between the control module and the user device; and is arranged between the control module and each of the displacement sensors to realize the communication connection between the control module and each of the displacement sensors.

[0039] The control module is configured to:

[0040] S1, in a first preset cycle starting at the current moment, using each of the displacement sensors to obtain a displacement data set of the lock chamber wall; the displacement data set includes multiple groups of displacement data corresponding to each preset monitoring point that change with time, the preset monitoring point is a monitoring point set at the joint of adjacent lock sections, and the displacement data that change with time is used to indicate the deformation displacement between the corresponding adjacent lock sections;

[0041] S2, based on the displacement data set, judging whether the deformation of the lock chamber wall meets the first alarm condition according to a preset judgment rule; when the first alarm condition is met, generating a first warning message and sending it to a preset user device; when the first alarm condition is not met, executing S1;

[0042] Among them, the preset judgment rule refers to generating a monitoring curve set based on the displacement data set, the monitoring curve set including the displacement time curve of each preset monitoring point; obtaining a first similarity score between the monitoring curve set and the sample curve set; when the first similarity score is lower than a first score threshold, it is considered that the deformation of the gate chamber wall meets the first alarm condition.

[0043] Furthermore, it also includes: a plurality of stress monitoring sensors, which are respectively arranged in a plurality of stress collection areas of the lock chamber wall and are communicatively connected with the control module through the gateway module, so as to obtain stress data of each stress collection area;

[0044] Among them, the stress collection area is set at the intersection of the two concrete pouring stages of the gate chamber wall. The first concrete pouring stage is a pouring construction based on the bottom plate after construction and including chamfering construction. The second concrete pouring is a pouring construction after the first concrete pouring.

[0045] Furthermore, the construction elevation range of the first concrete pouring is 3 meters to 4 meters; the second concrete pouring is constructed by a mobile formwork construction method, and is constructed to the top based on the first concrete pouring construction.

[0046] The beneficial effects of the present application are as follows: Based on the collected displacement data set, a displacement time curve of each monitoring point is generated to form a monitoring curve set. The monitoring curve is then compared with the sample curve set to obtain a similarity score between the two. If the similarity score is lower than the preset score threshold, it is judged that the deformation of the lock chamber wall meets the first alarm condition, which means that the actual deformation of the lock chamber wall shows a significant difference from the sample curve, indicating that there is a risk of structural deformation. In this case, the first warning information is automatically generated and sent to the user device through the preset communication channel.

[0047] Thus, by predicting the status trend of the collected information of the lock chamber wall and directly comparing the deformation data with the historical data of the same period, the safe operation of the engineering equipment is ensured. Representing the continuous displacement data as a time curve, rather than relying solely on the original data at each moment, can capture the overall trend of the monitoring points over time. By generating the displacement time curve, abnormal conditions can be quickly identified by directly comparing the shape and trend of the curves, without the need to calculate the data at a large number of time points one by one. This reduces the amount of data comparison while ensuring the accuracy of the judgment, and improves the speed and efficiency of data processing. Through the real-time collection and analysis of displacement data, the deformation of the lock chamber wall can be continuously monitored. Once a deformation risk is detected, an early warning will be issued immediately to ensure that problems are discovered at an early stage and prevent the occurrence of major structural damage or failures. Ensuring the accuracy of the judgment means that by monitoring the displacement at the joint of adjacent lock segments, the deformation data of key parts can be accurately captured. At the same time, by using the similarity scoring method and comparing with the sample curve, subtle deformation differences can be effectively identified, thereby improving the accuracy of the judgment. Automatically collecting and analyzing data based on preset rules without manual intervention reduces the error rate of manual operations and improves the monitoring efficiency. By detecting abnormal conditions of structural deformation in advance, relevant personnel can take repair measures in time, thereby effectively reducing the safety hazards caused by the instability or failure of the lock chamber wall structure. Using the displacement time curve method for monitoring reduces the amount of data and thus improves the overall efficiency of monitoring. Description of the Drawings

[0048] The following further describes the present application in conjunction with the drawings and embodiments.

[0049] Figure 1 It is a flowchart showing the deformation monitoring method of the lock chamber wall provided by the embodiment of the present application.

[0050] Figure 2 It is a flowchart showing the process of obtaining the set of sample curves provided by the embodiment of the present application.

[0051] Figure 3 It is a flowchart showing the process of obtaining the first similarity score provided by the embodiment of the present application.

[0052] Figure 4 It is a block diagram showing the structure of the lock chamber wall deformation monitoring device provided by the embodiment of the present application. Detailed Embodiments

[0053] Next, in combination with the accompanying drawings and specific embodiments, the present application will be further described. It should be noted that, on the premise of non-conflict, any combination can be formed among the following-described embodiments or technical features to form new embodiments. The following will illustrate the implementation procedures of the present application with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. The present application can also be implemented or applied through other different specific implementation procedures. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for explaining the present application, rather than for limiting the protection scope of the present application.

[0054] Hereinafter, the application scenarios and some terms in the present application will be explained to facilitate the understanding of those skilled in the art.

[0055] Machine learning (ML) is a multi-disciplinary cross-cutting subject, involving multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. A computer program can learn experience E given a certain type of task T and performance metric P. If its performance in task T can be exactly measured by P, it will improve with experience E. Machine learning specifically studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve its own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications cover all fields of artificial intelligence.

[0056] Deep learning is a special type of machine learning that represents and realizes great functions and flexibility by learning to use nested concept hierarchies, where each concept is defined in relation to simple concepts, and more abstract representations are computed in a less abstract way. Machine learning and deep learning generally include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning from demonstration.

[0057] At present, although there are already some technologies for the deformation monitoring of the lock chamber wall, these technologies often cannot achieve a balance between monitoring accuracy and monitoring efficiency. Based on this, the present application provides a method and system for the deformation monitoring of the lock chamber wall, which is used to detect the deformation of the lock chamber wall in a timely and accurate manner. Through the collection and analysis of data of multiple groups of monitoring points and the combination of a similarity scoring mechanism, real-time intelligent monitoring and early warning are realized. Representing the continuous displacement data as a time curve instead of relying solely on the original data at each moment can capture the overall trend of the monitoring points changing over time. For the present application, since it reduces the difficulty of maintenance and management of the lock chamber wall and other components in the later stage of the project, reduces the possibility of its deformation, and improves the overall safety and stability of the project, the above-mentioned time serialization can extract key feature points or change trends. For example, the impact of local minor changes on the overall trend does not show up immediately, so there is no need to record all the original data at every moment, which can reduce the recording of redundant data and thus achieve a balance between monitoring accuracy and monitoring efficiency. The method will be described first below, and then the system and others will be described.

[0058] Method embodiment.

[0059] See Figure 1 , Figure 1 is a schematic flowchart of the method for monitoring the deformation of the lock chamber wall provided by the embodiment of the present application.

[0060] The present embodiment provides a method for monitoring the deformation of the lock chamber wall, which is applied to an Internet of Things network. The Internet of Things network includes a gateway module, a control module, and a plurality of displacement sensors arranged on the lock chamber wall. Each of the displacement sensors uploads the displacement data that changes over time to the control module through the gateway module. The method for monitoring the deformation of the lock chamber wall is executed by an electronic device, and this electronic device can be the control module.

[0061] The method includes S101 and S102.

[0062] S101, in a first preset period starting from the current moment, use each of the displacement sensors to obtain a set of displacement data of the lock chamber wall.

[0063] The set of displacement data includes multiple groups of displacement data that change over time corresponding to each preset monitoring point. The preset monitoring points are the monitoring points set at the joints of adjacent lock sections, and the displacement data that changes over time is used to indicate the deformation displacement situation between the corresponding adjacent lock sections.

[0064] Collect displacement data from each preset monitoring point in the first preset period. The monitoring points are located at the joints of adjacent lock sections and are used to monitor the deformation displacement situation between adjacent lock sections. Each monitoring point will collect a set of displacement data that changes over time.

[0065] S102. Based on the displacement data set, determine whether the deformation condition of the lock chamber wall meets the first alarm condition according to a preset determination rule; when the first alarm condition is met, generate a first warning message and send it to a preset user device; when the first alarm condition is not met, execute S101. That is to say, the above deformation monitoring method can continuously run through one first preset period after another until the first alarm condition is found to be met.

[0066] Among them, the preset determination rule means that a set of monitoring curves is generated based on the displacement data set, and the set of monitoring curves includes displacement-time curves of each preset monitoring point; obtain the first similarity score between the set of monitoring curves and the set of sample curves; when the first similarity score is lower than the first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition. The first similarity score can be obtained by using the cosine similarity method, representing the displacement-time curve as a vector. Calculate the cosine similarity between the monitoring curve and the sample curve, and the closer the value is to 1, the more similar the two are.

[0067] Based on the collected displacement data set, generate displacement-time curves for each monitoring point to form a set of monitoring curves. Then compare the monitoring curves with the set of sample curves to obtain the similarity score between the two. If the similarity score is lower than the preset score threshold, it is determined that the deformation condition of the lock chamber wall meets the first alarm condition, which means that the actual deformation condition of the lock chamber wall shows a significant difference from the sample curve, indicating a risk of structural deformation. In this case, a first warning message is automatically generated and sent to the user device through a preset communication channel, and the user device can be, for example, the control center or the mobile phone, tablet or control center of the maintenance personnel.

[0068] Thus, compared with the related art, in which the state trend of the collected information of the lock chamber wall is predicted and the deformation data is directly compared with the historical data of the same period to ensure the safe operation of engineering equipment, in the embodiment of the present application, the continuous displacement data is represented as a time curve, rather than relying solely on the original data at each moment, and the overall trend of the monitoring point changing with time can be captured. By generating the displacement time curve, since the shapes and trends of the curves are directly compared, abnormal situations can be quickly identified without calculating the data at a large number of time points one by one, reducing the amount of data comparison while ensuring the accuracy of judgment and improving the speed and efficiency of data processing. Through the real-time collection and analysis of the displacement data, the deformation of the lock chamber wall can be continuously monitored, and once a deformation risk is detected, an early warning will be issued immediately to ensure that problems are discovered at an early stage and prevent the occurrence of major structural damage or failures. Ensuring the accuracy of judgment means that by monitoring the displacement at the joint of adjacent lock segments, the deformation data of the key parts can be accurately captured. At the same time, by using the similarity scoring method and comparing with the sample curve, the subtle deformation differences can be effectively identified, thereby improving the accuracy of judgment. Automatically collecting and analyzing data based on preset rules without manual intervention reduces the error rate of manual operations and improves the monitoring efficiency. By detecting abnormal situations of structural deformation in advance, relevant personnel can take repair measures in time, thereby effectively reducing the safety hazards caused by the instability or failure of the lock chamber wall structure. Using the displacement time curve for monitoring reduces the amount of data and thus improves the overall efficiency of monitoring.

[0069] In specific applications, the displacement time curve can be further simplified by data compression algorithms. For example, by using methods such as Fourier transform or wavelet transform, the time series data can be transformed into frequency domain data, and the high-frequency noise data irrelevant to the overall monitoring result can be discarded, thereby reducing the storage and transmission burden.

[0070] In specific applications, for the use scenario of highly pervious sandy soil layers, a continuous and complete combined anti-seepage curtain can be formed in the highly pervious sandy soil layers. The combined anti-seepage curtain is obtained through a closed combination strategy to form a continuous and complete anti-seepage curtain in the highly pervious sandy soil layers. It can be understood that the anti-seepage curtain obtained through the closed combination strategy can not only effectively prevent the seepage of groundwater, prevent soil erosion and structural damage, but also reduce the difficulty of maintenance and management of the lock chamber wall and other components in the later stage of the project, reduce the possibility of its deformation, and improve the overall safety and stability of the project.

[0071] The closed combination strategy includes: using the combination of diaphragm walls, cement mixing pile walls, geomembranes and high-pressure jet grouting piles to form a multi-layer anti-seepage curtain; through high-pressure jetting and mud solidification technologies, a uniform anti-seepage barrier is formed in the sandy soil layer to control the occurrence of local concentrated seepage and improve the overall anti-seepage performance; an anti-seepage layer is set in the depth direction of the sandy soil layer, and anti-seepage isolation belts are arranged in the horizontal direction to form a three-dimensional protection system to ensure that water seepage can be effectively blocked both vertically and horizontally; an independent mud circulation system is adopted to improve the construction speed. The lock chamber wall mentioned in this application is applicable to the lock use scenario of highly permeable sandy soil layers, that is, the case of a combined anti-seepage curtain.

[0072] Taking the deformation monitoring and status warning method for sluice and pumping station hydraulic structures disclosed in the application number CN201510622076.2 as an example, it does not consider the situation that the combined anti-seepage curtain can effectively improve and solve the anti-seepage of highly permeable sandy soil layers under the condition of the combined anti-seepage curtain. When applied to the deformation monitoring of the lock chamber wall improved by the combined anti-seepage curtain, in order to avoid inaccurate monitoring results caused by fluctuations in the collected monitoring information, a large number of monitoring information needs to be collected at multiple engineering monitoring points. Similarly, a balance cannot be achieved between monitoring accuracy and monitoring efficiency.

[0073] The technical solution provided in this embodiment is for the application scenario of forming a continuous and complete anti-seepage curtain in highly permeable sandy soil layers (that is, under the condition of the combined anti-seepage curtain). Since the combined anti-seepage curtain reduces the difficulty of maintenance and management of the lock chamber wall and other structures in the later stage of the project in highly permeable sandy soil layers, reduces the possibility of its deformation, and improves the overall safety and stability of the project, the above time series can extract key feature points or change trends. For example, local minor changes do not affect the overall trend, and it is not necessary to record all the original data at every moment. Therefore, the recording of redundant data can be reduced. Thus, the impact on the detection efficiency is reduced, and the detection accuracy is ensured on the premise that local minor changes do not affect the overall trend, achieving a balance between accuracy and efficiency.

[0074] See Figure 2 , Figure 2 is a schematic flow chart for obtaining the sample curve set provided by the embodiment of the present application.

[0075] In an exemplary embodiment, each first preset period includes a second preset period. The method for obtaining the sample curve set includes S201 to S204.

[0076] S201, take the first preset period starting from the current moment as the target preset period, and obtain the stress monitoring information of multiple stress acquisition areas of the lock chamber wall in the second preset period within the target preset period;

[0077] S202. Take the stress information obtained in the second preset cycle within multiple first preset cycles before the target preset cycle as stress sample information.

[0078] S203. Compare each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score.

[0079] S204. Take the stress sample information that is not lower than the second similarity threshold and is closest to the current time as the target stress information, and take the set of monitoring curves corresponding to the target stress information as the sample curve set.

[0080] It can be considered that the displacement data set of the lock chamber wall is collected within each first preset cycle. In the second preset cycle within the target preset cycle, the stress information of multiple stress acquisition areas of the lock chamber wall is obtained to form the stress monitoring information for reflecting the stress conditions borne by the lock chamber wall at different positions. At the same time, the stress information collected within multiple historical preset cycles before the target preset cycle is called as stress sample information. Compare the stress sample information of each second preset cycle with the current stress monitoring information to obtain the corresponding second similarity score. Select the stress sample information that is not lower than the second similarity threshold and is closest to the current time as the target stress information, and take the set of monitoring curves within the first preset cycle corresponding to the target stress information as the sample curve set. Generate a set of monitoring curves based on the displacement data set and compare it with the sample curve set to calculate the first similarity score. When the first similarity score is lower than the first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition. Once the alarm condition is met, a warning message is automatically generated and sent to the user device through a preset communication channel.

[0081] Thus, the technical solution provided in this embodiment uses historical data as samples, improving the prediction ability of the deformation trend. By introducing stress monitoring information, the stress information can more directly reflect the stress state of the lock chamber wall. Through the similarity comparison between the historical stress sample information and the current stress information, the sample curve set closest to the current working condition can be selected. The dynamic selection mechanism ensures the reliability of the monitoring results, can adapt to different environments and load changes, avoids using irrelevant sample data for comparison, and thus reduces the risk of misjudgment. By dynamically analyzing the similarity between the current stress situation and the historical stress data, the application of stress information can be flexibly adjusted, especially in different environmental conditions, the most suitable sample curve can be automatically selected, enhancing the adaptability of the monitoring method, and being applicable to the complex and changeable application of the lock chamber wall. Using the similarity scoring algorithm for stress information comparison and curve screening facilitates the realization of a fully automated monitoring process. The stress monitoring information can not only be used to judge whether there is a current deformation risk, but also be used to optimize the selection of sample data to ensure the accuracy and efficiency of the monitoring.

[0082] In summary, by introducing a multi-level stress information monitoring and similarity analysis mechanism, the intelligence and accuracy of the lock chamber wall deformation monitoring system are enhanced, enabling it to dynamically adjust the selection of sample curves and ensuring reliable judgment results under different working conditions.

[0083] In an exemplary embodiment, taking the stress information obtained in the second preset period within multiple first preset periods before the target preset period as stress sample information includes:

[0084] When the deformation condition of the lock chamber wall in the first preset period before the target preset period does not meet the first alarm condition, input the stress information of the first preset period that does not meet the first alarm condition into the abnormal area detection model to determine whether the number of abnormal stress acquisition areas is greater than the preset number; the stress information includes the identifier of each stress acquisition area and a set of stress data for the area corresponding to each identifier;

[0085] Take the stress information of the first preset period in which the number of abnormal stress acquisition areas is greater than the preset number as stress sample information for obtaining the second similarity score of the stress monitoring information.

[0086] The technical solution provided in this embodiment screens the stress information of historical periods through the abnormal area detection model to determine stress data with reference value, thereby improving the accuracy and reliability of the lock chamber wall deformation monitoring. Specifically, when the deformation condition of the lock chamber wall within multiple first preset periods before the target preset period does not meet the first alarm condition, it is considered that there is no obvious deformation risk in the structure during these periods, but there may still be potential stress abnormalities. Input the stress information of the non-alarm periods into the abnormal area detection model, and based on the data of each stress acquisition area, determine whether there is a possibility of stress abnormality in that area. The abnormal detection result of each area will be marked as "abnormal" or "normal". Count the number of stress acquisition areas determined to be "abnormal" and determine whether this number is greater than the preset number. The preset number is a set threshold used to distinguish the severity of abnormal situations, such as 3, 5, 8. If the number of abnormal areas exceeds this set threshold, it is considered that the stress information of this period has potential representativeness, indicating that these stress data may be related to future structural deformation risks. When the detected number of abnormal areas is greater than the preset number, the stress information of this first preset period will be selected as stress sample information. The stress sample information will be used for similarity score analysis (second similarity score) with the stress monitoring information of the current period to determine whether there is a deformation risk in the lock chamber wall at the current moment.

[0087] Thus, the technical solution provided in this embodiment screens historical stress data through an abnormal area detection model, only retaining the cycles with stress anomalies in multiple areas. Compared with the traditional direct data comparison method, the detection based on abnormal areas can better capture the potential precursors of structural deformation. The screening process not only simplifies the complexity of large-scale data analysis but also ensures that the stress sample data finally used for comparison is more representative and accurate. Only when a large range of stress anomalies is detected will the data of that cycle be used as sample information, reducing the inaccurate selection of stress sample information caused by local stress fluctuations. Automatically adjusting the selection of stress samples according to the stress conditions of historical cycles makes the monitoring mechanism highly adaptable. Especially in engineering scenarios with complex environmental conditions and frequent stress changes, the system can flexibly respond to different stress change patterns and provide efficient deformation monitoring. By setting a preset quantity threshold, historical data with less severe or scattered stress anomalies can be effectively filtered out, and only the cycles with a relatively large proportion of abnormal areas are selected as stress samples; when the lock chamber wall is relatively stable, the preset quantity threshold can be artificially reduced to ensure that the stress sample information can be updated in a timely manner, making the sample data input into the similarity analysis have high reference value. The dynamic adaptive ability of screening helps to avoid over-reliance on specific historical data while ensuring the stability and continuity of monitoring.

[0088] In an exemplary embodiment, the abnormal area detection model is obtained by training a deep learning model, and the training process of the abnormal area detection model includes the following steps:

[0089] Obtain a training set, where the training set includes multiple training data, and each training data includes a data group of a set of sample stress data and annotation data of the corresponding abnormal area of the data group; for each training data in the training set, perform the following processing:

[0090] Input the data group in the training data into a preset deep learning model to obtain prediction data of the corresponding abnormal area of the data group;

[0091] Update the model parameters of the deep learning model based on the prediction data and annotation data corresponding to the sample image data;

[0092] Detect whether a preset training end condition is met; if so, use the trained deep learning model as the abnormal area detection model; if not, continue to train the deep learning model with the next training data.

[0093] Thus, through design, by establishing an appropriate number of neuron computing nodes and a multi-layer operation hierarchy, and selecting appropriate input and output layers, a preset deep learning model can be obtained. Through the learning and optimization of this deep learning model, a functional relationship from input to output can be established. Although the functional relationship between input and output cannot be found 100%, it can approximate the real correlation relationship as much as possible. The anomaly region detection model trained thereby can obtain an anomaly judgment corresponding to the stress acquisition region based on the stress information in the first preset period, with high calculation accuracy and high reliability.

[0094] Among them, during the training process, a large amount of training data can be used, including a data set of a group of sample stress data and the annotation data of the corresponding anomaly regions of the data set. Using these data to train a preset deep learning model, based on the prediction data output by the deep learning model and the stability scoring data of the actual annotation, the backpropagation algorithm is used to adjust and update the parameters of the deep learning model so that it can more accurately predict the influence of the data set of sample stress data on the classification of anomaly regions.

[0095] In a specific application, the architecture of the anomaly region detection model may include an input layer, a feature extraction layer, a fusion layer, a fully connected layer, and an output layer. The input layer is used to receive multiple time series data sets, such as the stress data of different stress acquisition regions, to reflect the stress changes in the regions. It can also receive the time series data of a single stress monitoring data point. The feature extraction layer, including the LSTM layer, is used to process the time series data and capture the temporal features of stress changes so that the model can identify anomaly patterns and trends; the CNN layer is used to extract the local features of the stress data and enhance the sensitivity to local anomalies, and identify local fluctuations or anomalies in the time series data through convolution operations. The fusion layer is used to integrate the features extracted from the LSTM layer and the CNN layer to facilitate capturing the mutual relationships between different features. The fusion layer can use concatenation operations or weighted summation to integrate the information from different data sources. The fully connected layer is used to process a single stress monitoring data point and combine it with the fused features to ensure that the model can consider both global and local features. The output layer is used to output an anomaly score or classification label based on the fused features to indicate whether there is an anomaly region. The score can be a continuous value (indicating the degree of anomaly) or binary classification (indicating normal or abnormal).

[0096] By combining LSTM and CNN, the model can effectively capture the complex non-linear relationships in the time series data, which helps to better identify potential anomalies. By fusing data from different sources (time series and single data points), the accuracy of anomaly detection is improved, and the anomaly risk can be considered from multiple dimensions.

[0097] Embodiments of the present application can train an abnormal area detection model. In some other alternative embodiments, the present application can use a pre-trained abnormal area detection model. The preset deep learning model can be a convolutional neural network model or a recurrent neural network model. The present application does not limit the implementation manner of the preset deep learning model. The present application does not limit the training process of the abnormal detection model. For example, it can adopt the above-mentioned supervised learning training method, or can adopt the semi-supervised learning training method, or can adopt the unsupervised learning training method. The present application does not limit the preset training end condition. For example, it can be that the number of training times reaches a preset number (the preset number is, for example, 1 time, 3 times, 10 times, 100 times, 1000 times, 10000 times, etc.), or it can be that all the training data in the training set have completed one or more trainings, or it can be that the total loss value obtained in this training is not greater than the preset loss value.

[0098] Among them, the training data in the training set can be historical data and monitoring records, including the stress monitoring information of each stress acquisition area obtained by multiple field data collections, and the evaluation of abnormal conditions by technicians; it can also be simulated data generated after modeling and simulation according to known physical models and assumed conditions to simulate the evaluation of abnormal conditions of the stress monitoring information in different stress acquisition areas.

[0099] See Figure 3 , Figure 3 which is a schematic flowchart of obtaining the first similarity score provided by the embodiments of the present application.

[0100] In an exemplary embodiment, obtaining the first similarity score between the monitoring curve set and the sample curve set includes S301 to S303.

[0101] S301, obtain the similarity value of the displacement-time curves of the same preset monitoring points in the monitoring curve set and the sample curve set, and obtain the similarity value of each preset monitoring point;

[0102] S301, for each preset monitoring point, when the corresponding stress acquisition area is an abnormal area, take the product of the obtained similarity value and the abnormal coefficient of the stress acquisition area corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, take the obtained similarity value as the first similarity intermediate score;

[0103] S303, take the sum of the obtained first similarity intermediate scores of each preset monitoring point as the first similarity score between the monitoring curve set and the sample curve set.

[0104] The technical solution provided by this embodiment analyzes the abnormal conditions of different monitoring points more precisely by weighting the similarity values with abnormal coefficients. First, a set of monitoring curves and a set of sample curves are obtained. For each preset monitoring point, the displacement-time curves at the same position (i.e., the same monitoring point) in the monitoring curve and the sample curve are compared, and the similarity value between the two is calculated. This similarity value is used to evaluate the similarity between the displacement change of the current monitoring point and the displacement change of the historical sample. It can be considered that each preset monitoring point will have a separate similarity value for subsequent scoring calculations. Then, for each preset monitoring point, it is checked whether the corresponding stress acquisition area (in the set of sample curves) is marked as an abnormal area. If it is an abnormal area, the similarity value of this monitoring point is multiplied by the abnormal coefficient corresponding to this area to obtain the first intermediate similarity score of this monitoring point. The abnormal coefficient is used to weight the similarity value of this monitoring point, reflecting the importance and potential risk of this area. If it is a normal area, the similarity value is directly used as the first intermediate similarity score of this monitoring point. The weighting mechanism for abnormal areas ensures that in potentially high-risk areas, the deviation of similarity is more sensitive. The first intermediate similarity scores of all preset monitoring points are added together to obtain the first similarity score between the entire set of monitoring curves and the set of sample curves. This first similarity score represents the similarity between the overall monitoring situation at the current moment and the historical sample situation. If the score is low, it indicates that there are significant differences between the current monitoring data and the historical sample, which may indicate a deformation risk of the lock chamber wall.

[0105] Thus, each preset monitoring point adopts different scoring mechanisms (weighted or unweighted) according to whether the area where it is located is an abnormal area, enabling dynamic adaptation to changes in the stress state, focusing on monitoring high-risk areas, and maintaining routine analysis for low-risk areas. By summarizing the first intermediate similarity scores to obtain the first similarity score, the overall deformation situation of the lock chamber wall can be reflected more accurately. Compared with the simple average similarity value, it can better consider the importance of each monitoring point and the differences in stress states, thereby improving the representativeness and accuracy of the judgment results. The similarity values of each monitoring point are analyzed point by point, and the high-risk areas are weighted, which can improve the overall reliability on the basis of a low false alarm rate.

[0106] In specific applications, the value of the abnormal coefficient is less than 1 and greater than 0. The value of the abnormal coefficient is set to be less than 1 and greater than 0, so that the similarity value of the abnormal area is weakened, thereby highlighting the possible potential problems in the abnormal area. In the final first similarity score, the influence of this area is more significantly reflected.

[0107] In an exemplary embodiment, the method for obtaining the abnormal coefficient of the stress acquisition area includes:

[0108] Obtain the corresponding relationship of abnormal coefficients; the corresponding relationship of abnormal coefficients is used to indicate the corresponding relationship among the number, proportion, and abnormal values of abnormal stress acquisition regions;

[0109] According to the number of stress acquisition regions indicating abnormality in the target stress information, use the corresponding relationship of abnormal coefficients to obtain the corresponding abnormal value, and use the abnormal value as the abnormal coefficient of the stress acquisition region indicating abnormality.

[0110] The technical solution provided in this embodiment dynamically adjusts the abnormal coefficient of the abnormal region according to the number of abnormal regions through the corresponding relationship of abnormal coefficients. In specific applications, a corresponding relationship of abnormal coefficients is defined first before obtaining the corresponding relationship of abnormal coefficients. The corresponding relationship of abnormal coefficients is used to indicate the corresponding relationship between the number of abnormal stress acquisition regions and the abnormal value. That is, different numbers of abnormal regions correspond to different abnormal values. In specific applications, according to the target stress information, obtain the number of abnormal stress acquisition regions at the current moment. Then, use the preset corresponding relationship of abnormal coefficients to calculate the corresponding abnormal value in combination with the number of abnormal regions. The abnormal value is used to further reflect the potential risk or importance of these abnormal regions. Use the obtained abnormal value as the abnormal coefficient of the abnormal stress acquisition region.

[0111] Thus, by introducing the corresponding relationship of abnormal coefficients, the abnormal coefficient of the abnormal region corresponding to the new sample curve set can be dynamically adjusted. The introduction of the corresponding relationship of abnormal coefficients enables configuration and expansion according to the actual needs of the engineering environment.

[0112] As an example, a method for monitoring the deformation of a lock chamber wall is provided, including:

[0113] R1. In a first preset period starting from the current moment, use each of the displacement sensors to obtain a displacement data set of the lock chamber wall; the displacement data set includes multiple groups of displacement data changing with time corresponding to each preset monitoring point. The preset monitoring points are monitoring points set at the joints of adjacent lock sections. The displacement data changing with time is used to indicate the deformation displacement situation between the corresponding adjacent lock sections;

[0114] R2. Use the first preset period starting from the current moment as the target preset period, and in a second preset period within the target preset period, obtain the stress monitoring information of multiple stress acquisition regions of the lock chamber wall;

[0115] R3. When the deformation situation of the lock chamber wall in the first preset period before the target preset period does not meet the first alarm condition, input the stress information of the first preset period that does not meet the first alarm condition into the abnormal region detection model to obtain the abnormal detection result of each stress acquisition region; the abnormal detection result includes abnormal and normal;

[0116] R4, determine whether the number of abnormal stress acquisition regions is greater than a preset number; the stress information includes the identifier of each stress acquisition region and a set of stress data for the region corresponding to each identifier.

[0117] R5, use the stress information of the first preset period in which the number of abnormal stress acquisition regions is greater than the preset number as stress sample information for obtaining the second similarity score of the stress monitoring information.

[0118] R6, compare each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score.

[0119] R7, use the stress sample information that is not lower than the second similarity threshold and is closest to the current time as the target stress information, and use the set of monitoring curves corresponding to the target stress information as the sample curve set.

[0120] R8, generate a set of monitoring curves based on the displacement data set, where the set of monitoring curves includes the displacement-time curves of each preset monitoring point; obtain the similarity values of the displacement-time curves of the same preset monitoring points in the set of monitoring curves and the sample curve set to obtain the similarity values of each preset monitoring point.

[0121] R9, for each preset monitoring point, when the corresponding stress acquisition region is an abnormal region, use the product of the obtained similarity value and the abnormal coefficient of the stress acquisition region corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, use the obtained similarity value as the first similarity intermediate score.

[0122] R10, use the sum of the obtained first similarity intermediate scores of each preset monitoring point as the first similarity score between the set of monitoring curves and the sample curve set. When the first similarity score is lower than the first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition; when the first alarm condition is met, generate a first warning message and send it to the preset user device; when the first alarm condition is not met, execute R1 to start the deformation monitoring of the lock chamber wall in the next first preset period.

[0123] Among them, the method for obtaining the abnormal coefficient of the stress acquisition region includes: obtaining the abnormal coefficient correspondence relationship; the abnormal coefficient correspondence relationship is used to indicate the correspondence relationship between the number, proportion, and abnormal value of the abnormal stress acquisition regions; according to the number of abnormal stress acquisition regions indicated in the target stress information, use the abnormal coefficient correspondence relationship to obtain the corresponding abnormal value, and use the abnormal value as the abnormal coefficient of the abnormal stress acquisition region.

[0124] The abnormal area detection model is obtained by training a deep learning model. The training process of the abnormal area detection model includes the following steps: obtaining a training set, where the training set includes a plurality of training data, and each training data includes a data set of a group of sample stress data and annotation data of the abnormal area corresponding to the data set; for each training data in the training set, perform the following processing: input the data set in the training data into a preset deep learning model to obtain prediction data of the abnormal area corresponding to the data set; update the model parameters of the deep learning model based on the prediction data and annotation data corresponding to the sample image data; detect whether a preset training end condition is satisfied; if so, use the trained deep learning model as the abnormal area detection model; if not, continue to train the deep learning model with the next training data.

[0125] The gate chamber wall deformation monitoring method provided in this example can more comprehensively evaluate the deformation of the gate chamber wall and improve the accuracy of the monitoring results by combining displacement data and stress monitoring information. Generating early warning information in a timely manner when deformation risks are detected helps to take preventive measures and avoid structural damage. By introducing an abnormal area detection model and a similarity scoring mechanism, false alarms caused by local fluctuations or noise are reduced. It can dynamically select sample curves according to the actual stress state, improving the adaptability and flexibility of the monitoring method. By representing continuous displacement data as a time curve, the data volume is reduced, the data processing process is simplified, and the efficiency is improved. Through continuous monitoring and timely early warning, it helps to achieve the structural health monitoring and preventive maintenance of the gate chamber wall. The automated monitoring and early warning process reduces manual intervention, reduces the possibility of human errors, and improves the reliability of the monitoring.

[0126] Based on the content described in the foregoing method embodiments, those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the devices and systems described below can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0127] Device embodiment.

[0128] See Figure 4 , Figure 4 which is a structural block diagram of a gate chamber wall deformation monitoring device provided in an embodiment of the present application.

[0129] This embodiment provides a gate chamber wall deformation monitoring device, and the gate chamber wall deformation monitoring device includes:

[0130] A first acquisition module is configured to obtain a displacement data set of the lock chamber wall by using each of the displacement sensors during a first preset period starting from the current moment; the displacement data set includes multiple sets of displacement data varying with time corresponding to each preset monitoring point, and the preset monitoring points are monitoring points set at the joints of adjacent lock sections, and the displacement data varying with time is used to indicate the deformation displacement condition between the corresponding adjacent lock sections;

[0131] A first judgment module is configured to judge whether the deformation condition of the lock chamber wall meets a first alarm condition according to a preset judgment rule based on the displacement data set; when the first alarm condition is met, a first early warning message is generated and sent to a preset user device;

[0132] Wherein, the preset judgment rule means that a monitoring curve set is generated based on the displacement data set, and the monitoring curve set includes displacement-time curves of each preset monitoring point; a first similarity score between the monitoring curve set and a sample curve set is obtained; when the first similarity score is lower than a first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition.

[0133] In an exemplary embodiment, each first preset period includes a second preset period, and the method for obtaining the sample curve set includes:

[0134] Taking the first preset period starting from the current moment as a target preset period, and obtaining stress monitoring information of multiple stress acquisition areas of the lock chamber wall during the second preset period within the target preset period;

[0135] Taking the stress information obtained during the second preset period within multiple first preset periods before the target preset period as stress sample information;

[0136] Comparing each stress sample information with the stress monitoring information respectively to obtain a corresponding second similarity score;

[0137] Taking the stress sample information that is not lower than the second similarity threshold and is closest to the current moment as target stress information, and taking the monitoring curve set corresponding to the target stress information as the sample curve set.

[0138] In an exemplary embodiment, the step of taking the stress information obtained during the second preset period within multiple first preset periods before the target preset period as stress sample information includes:

[0139] When the deformation condition of the lock chamber wall in the first preset period before the target preset period does not meet the first alarm condition, inputting the stress information of the first preset period that does not meet the first alarm condition into an abnormal area detection model to obtain an abnormal detection result of each stress acquisition area; the abnormal detection result includes abnormal and normal;

[0140] Determine whether the number of abnormal stress acquisition regions is greater than a preset number; the stress information includes the identifier of each stress acquisition region and a set of stress data for the region corresponding to each identifier.

[0141] Use the stress information of the first preset period in which the number of abnormal stress acquisition regions is greater than the preset number as stress sample information for obtaining the second similarity score of the stress monitoring information.

[0142] In an exemplary embodiment, the abnormal region detection model is obtained by training a deep learning model, and the training process of the abnormal region detection model includes the following steps:

[0143] Obtain a training set, where the training set includes a plurality of training data, and each training data includes a data group of a set of sample stress data and annotation data of the abnormal region corresponding to the data group; for each training data in the training set, perform the following processing:

[0144] Input the data group in the training data into a preset deep learning model to obtain prediction data of the abnormal region corresponding to the data group.

[0145] Update the model parameters of the deep learning model based on the prediction data and annotation data corresponding to the sample image data.

[0146] Detect whether a preset training end condition is satisfied; if so, use the trained deep learning model as the abnormal region detection model; if not, continue to train the deep learning model with the next training data.

[0147] In an exemplary embodiment, obtaining the first similarity score between the monitoring curve set and the sample curve set includes:

[0148] Obtain the similarity value of the displacement-time curve of the same preset monitoring points in the monitoring curve set and the sample curve set to obtain the similarity value of each preset monitoring point.

[0149] For each preset monitoring point, when the corresponding stress acquisition region is an abnormal region, use the product of the obtained similarity value and the abnormal coefficient of the stress acquisition region corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, use the obtained similarity value as the first similarity intermediate score.

[0150] Use the sum of the first similarity intermediate scores of each obtained preset monitoring point as the first similarity score between the monitoring curve set and the sample curve set.

[0151] In an exemplary embodiment, the method for obtaining the anomaly coefficient of the stress acquisition area includes:

[0152] Obtain the anomaly coefficient correspondence; the anomaly coefficient correspondence is used to indicate the correspondence between the number, proportion, and anomaly value of the abnormally stressed acquisition areas;

[0153] According to the number of abnormally stressed acquisition areas indicated in the target stress information, use the anomaly coefficient correspondence to obtain the corresponding anomaly value, and use the anomaly value as the anomaly coefficient of the abnormally stressed acquisition area.

[0154] System embodiment.

[0155] This embodiment provides a deformation monitoring system for a lock chamber wall, which is applicable to the deformation monitoring method for a lock chamber wall in the method embodiment. The deformation monitoring system for the lock chamber wall includes:

[0156] A plurality of displacement sensors are respectively arranged at preset monitoring points at the joints of adjacent lock segments of the lock chamber wall, and are used to obtain displacement data that changes over time;

[0157] A control module, and the control module is respectively communicatively connected to each of the displacement sensors;

[0158] A gateway module, which is arranged between the control module and the user device to realize the communication connection between the control module and the user device; and is arranged between the control module and each of the displacement sensors to realize the communication connection between the control module and each of the displacement sensors;

[0159] The control module is configured to:

[0160] S1. In a first preset period starting from the current moment, use each of the displacement sensors to obtain a displacement data set of the lock chamber wall; the displacement data set includes multiple sets of displacement data that change over time corresponding to each preset monitoring point, and the preset monitoring points are the monitoring points arranged at the joints of adjacent lock segments, and the displacement data that changes over time is used to indicate the deformation displacement situation between the corresponding adjacent lock segments;

[0161] S2. Based on the displacement data set, judge whether the deformation situation of the lock chamber wall meets the first alarm condition according to a preset determination rule; when the first alarm condition is met, generate a first warning message and send it to a preset user device; when the first alarm condition is not met, execute S1;

[0162] Among them, the preset determination rule means that a set of monitoring curves is generated based on the displacement data set, and the set of monitoring curves includes displacement-time curves of each preset monitoring point; obtaining a first similarity score between the set of monitoring curves and the set of sample curves; when the first similarity score is lower than the first score threshold, it is considered that the deformation condition of the lock chamber wall meets the first alarm condition.

[0163] By setting multiple displacement sensors and transmitting data to the control module in real time, real-time deformation monitoring of the lock chamber wall can be achieved, which helps to detect potential structural problems at an early stage and prevent serious accidents. Displacement sensors can detect tiny displacement changes and have higher accuracy and sensitivity compared to traditional manual detection methods, which helps to more accurately judge the deformation condition of the lock chamber wall. Among them, the displacement sensor is, for example, a laser displacement sensor, which measures the distance change between two points by emitting a laser beam and has the advantage of high precision. It can also be, for example, a fiber optic displacement sensor, an LVDT (linear variable differential transformer) displacement sensor, a MEMS accelerometer and displacement sensor, etc. The control module can be an industrial computer or a cloud server, which includes multiple microprocessors, storage devices and input / output interfaces, etc.

[0164] In a specific application, the monitoring system is integrated into the cloud server, and a mobile application is set up so that users can remotely access the detection system of the cloud server to obtain monitoring data and alarm information in real time.

[0165] In an exemplary embodiment, it further includes: multiple stress monitoring sensors, which are respectively arranged in multiple stress acquisition areas of the lock chamber wall and are connected to the control module for obtaining stress data of each stress acquisition area;

[0166] Among them, the stress acquisition area is set at the junction area of two concrete pouring stages of the lock chamber wall. The first concrete pouring stage is a pouring construction based on the post-construction floor slab and including chamfer construction, and the second concrete pouring is the pouring construction after the first concrete pouring.

[0167] The stress monitoring sensors are arranged in the junction area of the two concrete pourings. The first concrete pouring completes the floor slab and chamfer construction, and the second pouring forms a junction with the first one, which may cause stress concentration at the junction. The stress monitoring sensors collect stress data in real time and transmit it to the control module. By analyzing the stress data, the stress distribution of the lock chamber wall during operation can be judged, and combined with the displacement data, the structural health status of the wall can be more comprehensively evaluated. The stress monitoring sensor is, for example, a strain gauge sensor: it can also be a fiber Bragg grating sensor (FBG sensor), which is based on the light wave reflection characteristics in the optical fiber and measures stress through the strain change of the optical fiber caused by stress.

[0168] In an exemplary embodiment, the construction elevation range for the first concrete pouring is from 3 meters (inclusive) to 4 meters (inclusive); the construction of the second concrete pouring is in the form of a movable formwork construction method, and the construction reaches the top based on the first concrete pouring construction.

[0169] The construction of the second concrete pouring is in the form of a movable formwork construction method. The large-area steel formwork is selected for the movable formwork to cooperate with the second concrete pouring, reducing the possibility of deformation of the lock chamber wall during construction. During the pouring process of the lock chamber wall, the method of two (or multiple) concrete pourings is gradually adopted, and the movable formwork construction method is adopted during the second concrete pouring. Through the large-area steel formwork and the formwork system with no or few tie rods of the movable formwork, the stability of the lock chamber wall is better (compared with the previous traditional construction process).

[0170] In an exemplary embodiment, each first preset period includes a second preset period. The method for obtaining the sample curve set includes:

[0171] Taking the first preset period starting from the current moment as the target preset period, and during the second preset period within the target preset period, obtaining the stress monitoring information of multiple stress acquisition areas of the lock chamber wall;

[0172] Taking the stress information obtained during the second preset periods within multiple first preset periods before the target preset period as stress sample information;

[0173] Comparing each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score;

[0174] Taking the stress sample information that is not lower than the second similarity threshold and is closest to the current moment as the target stress information, and taking the monitoring curve set corresponding to the target stress information as the sample curve set.

[0175] In a specific application, the method of comparing each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score can be implemented by a similarity measurement method. As an example, for each stress acquisition area, the stress sample information and the stress monitoring information are represented as two vectors, and the Euclidean distance between the two vectors is calculated. The smaller the distance, the more similar the stress states are. The distance is converted into a similarity score, for example, similarity = 1 / (1 + Euclidean distance). Suppose in a certain stress acquisition area, the stress sample information is [5, 7, 9], and the current stress monitoring information is [4, 6, 10]. The obtained Euclidean distance is 1.73, and the similarity score is 0.366.

[0176] In stress monitoring, the specific value of the stress usually depends on the properties of the material, loading conditions, environmental factors, etc. Therefore, the stress (sample and monitoring) information in the above specific applications is obtained through actual measurement or simulation calculation. Specifically, a stress monitoring sensor is used to measure in the actual environment, and the stress values (such as normal stress, shear stress) under specific conditions are recorded. Among them, since the stress values may vary greatly in different regions, before calculating the similarity, the data is normalized or standardized. Then, by extracting specific features of the stress change, it is converted into a comparable vector form.

[0177] In an exemplary embodiment, the control module takes the stress information obtained in a second preset period within a plurality of first preset periods before the target preset period as stress sample information in the following manner:

[0178] When the deformation condition of the lock chamber wall in the first preset period before the target preset period does not meet the first alarm condition, the stress information of the first preset period that does not meet the first alarm condition is input into the abnormal area detection model to obtain the abnormal detection result of each stress acquisition area; the abnormal detection result includes abnormal and normal;

[0179] Judge whether the number of abnormal stress acquisition areas is greater than the preset number; the stress information includes the identifier of each stress acquisition area and a set of stress data corresponding to the area of each identifier;

[0180] Take the stress information of the first preset period in which the number of abnormal stress acquisition areas is greater than the preset number as stress sample information for obtaining the second similarity score of the stress monitoring information.

[0181] In an exemplary embodiment, the abnormal area detection model is obtained by training a deep learning model, and the training process of the abnormal area detection model includes the following steps:

[0182] Obtain a training set, the training set includes a plurality of training data, each training data includes a data set of a group of sample stress data and the annotation data of the abnormal area corresponding to the data set; for each training data in the training set, perform the following processing:

[0183] Input the data set in the training data into a preset deep learning model to obtain the prediction data of the abnormal area corresponding to the data set;

[0184] Based on the prediction data and annotation data corresponding to the sample image data, update the model parameters of the deep learning model;

[0185] Check whether the preset training end condition is satisfied; if so, use the trained deep learning model as the abnormal area detection model; if not, continue to train the deep learning model with the next training data.

[0186] In an exemplary embodiment, the control module obtains the first similarity score between the monitoring curve set and the sample curve set in the following manner:

[0187] Obtain the similarity value of the displacement-time curves of the same preset monitoring points in the monitoring curve set and the sample curve set to obtain the similarity value of each preset monitoring point;

[0188] For each preset monitoring point, when the corresponding stress acquisition area is an abnormal area, use the product of the obtained similarity value and the abnormal coefficient of the stress acquisition area corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, use the obtained similarity value as the first similarity intermediate score;

[0189] Use the sum of the first similarity intermediate scores of each preset monitoring point obtained as the first similarity score between the monitoring curve set and the sample curve set.

[0190] In an exemplary embodiment, the control module obtains the abnormal coefficient of the stress acquisition area in the following manner:

[0191] Obtain the abnormal coefficient correspondence; the abnormal coefficient correspondence is used to indicate the correspondence between the number, proportion, and abnormal value of the abnormal stress acquisition areas;

[0192] According to the number of stress acquisition areas indicated as abnormal in the target stress information, use the abnormal coefficient correspondence to obtain the corresponding abnormal value, and use the abnormal value as the abnormal coefficient of the stress acquisition area indicated as abnormal.

[0193] In various embodiments of the specification of this application, the magnitudes of the sequence numbers of the various processes do not mean the order of execution. The order of execution of the various processes should be determined by their functions and internal logics, and should not constitute any limitation to the implementation process of this application.

[0194] In the description, claims and drawings of this application, the terms "first", "second", "third", etc. (if any) are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising", "corresponding to" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0195] This application is described from the perspectives of purpose of use, efficacy, progress and novelty, and has met the requirements of function enhancement and use emphasized by the patent law. The above description and accompanying drawings of this application are only preferred embodiments of this application and do not limit this application. Therefore, all those that are similar or identical to the structure, device, features, etc. of this application, that is, all equivalent replacements or modifications made according to the scope of the patent application of this application, shall fall within the scope of protection of the patent application of this application.

Claims

1. A deformation monitoring method for a lock chamber wall, characterized in that Applied to the Internet of Things network, the Internet of Things network includes a gateway module and a control module, as well as a plurality of displacement sensors disposed on the gate chamber wall. Each of the displacement sensors uploads displacement data that changes over time to the control module through the gateway module. The method includes: S1. In a first preset period starting from the current moment, use each of the displacement sensors to obtain a set of displacement data of the gate chamber wall; the set of displacement data includes multiple sets of displacement data that change over time corresponding to each preset monitoring point. The preset monitoring point is a monitoring point set at the joint of adjacent gate segments. The displacement data that changes over time is used to indicate the deformation displacement situation between the corresponding adjacent gate segments; S2. Based on the set of displacement data, judge whether the deformation situation of the gate chamber wall meets the first alarm condition according to a preset judgment rule; when the first alarm condition is met, generate a first early warning message and send it to a preset user device; when the first alarm condition is not met, execute S1; Among them, the preset judgment rule means that a set of monitoring curves is generated based on the set of displacement data. The set of monitoring curves includes displacement-time curves of each preset monitoring point; obtain a first similarity score between the set of monitoring curves and a set of sample curves; when the first similarity score is lower than a first score threshold, it is considered that the deformation situation of the gate chamber wall meets the first alarm condition; Each first preset period includes a second preset period. The method for obtaining the set of sample curves includes: Take the first preset period starting from the current moment as the target preset period. In the second preset period within the target preset period, obtain stress monitoring information of multiple stress acquisition areas of the gate chamber wall; Take the stress information obtained in the second preset period within multiple first preset periods before the target preset period as stress sample information; Compare each stress sample information with the stress monitoring information respectively to obtain a corresponding second similarity score; Take the stress sample information that is not lower than the second similarity threshold and is the closest to the current moment as the target stress information, and take the set of monitoring curves corresponding to the target stress information as the set of sample curves.

2. The deformation monitoring method of the lock chamber wall according to claim 1, characterized in that The taking the stress information obtained in the second preset period within multiple first preset periods before the target preset period as stress sample information includes: When the deformation situation of the gate chamber wall in the first preset period before the target preset period does not meet the first alarm condition, input the stress information of the first preset period that does not meet the first alarm condition into an abnormal area detection model to obtain an abnormal detection result for each stress acquisition area; the abnormal detection result includes abnormal and normal; Judge whether the number of abnormal stress acquisition areas is greater than a preset number; the stress information includes the identifier of each stress acquisition area and a set of stress data of the area corresponding to each identifier; Take the stress information of the first preset period in which the number of abnormal stress acquisition areas is greater than the preset number as stress sample information for obtaining the second similarity score with the stress monitoring information.

3. The deformation monitoring method of the lock chamber wall according to claim 2, characterized in that, The abnormal area detection model is obtained by training a deep learning model. The training process of the abnormal area detection model includes the following steps: Obtain a training set, where the training set includes a plurality of training data, and each training data includes a data group of a set of sample stress data and annotation data of the abnormal area corresponding to the data group; for each training data in the training set, perform the following processing: Input the data group in the training data into a preset deep learning model to obtain prediction data of the abnormal area corresponding to the data group; Update the model parameters of the deep learning model based on the prediction data and annotation data corresponding to the sample image data; Detect whether a preset training end condition is satisfied; if so, use the trained deep learning model as the abnormal area detection model; if not, continue to train the deep learning model with the next training data.

4. The deformation monitoring method of the lock chamber wall according to claim 2, characterized in that, The obtaining of the first similarity score between the monitoring curve set and the sample curve set includes: Obtain the similarity values of the displacement-time curves of the same preset monitoring points in the monitoring curve set and the sample curve set to obtain the similarity values of each preset monitoring point; For each preset monitoring point, when the corresponding stress acquisition area is an abnormal area, use the product of the obtained similarity value and the abnormal coefficient of the stress acquisition area corresponding to the preset monitoring point as the first similarity intermediate score; otherwise, use the obtained similarity value as the first similarity intermediate score; Use the sum of the obtained first similarity intermediate scores of each preset monitoring point as the first similarity score between the monitoring curve set and the sample curve set.

5. The deformation monitoring method of the lock chamber wall according to claim 4, characterized in that, The obtaining method of the abnormal coefficient of the stress acquisition area includes: Obtain an abnormal coefficient correspondence; the abnormal coefficient correspondence is used to indicate the correspondence between the number, proportion, and abnormal value of the abnormal stress acquisition area; According to the number of stress acquisition areas indicated as abnormal in the target stress information, use the abnormal coefficient correspondence to obtain the corresponding abnormal value, and use the abnormal value as the abnormal coefficient of the stress acquisition area indicated as abnormal.

6. A deformation monitoring device for a lock chamber wall, characterized in that, The lock chamber wall deformation monitoring device includes: A first acquisition module, configured to obtain a displacement data set of the lock chamber wall by using each displacement sensor in a first preset period starting from the current moment; the displacement data set includes multiple groups of displacement data varying with time corresponding to each preset monitoring point, and the preset monitoring points are the monitoring points set at the joints of adjacent lock segments, and the displacement data varying with time is used to indicate the deformation displacement situation between the corresponding adjacent lock segments; A first judgment module, configured to judge whether the deformation situation of the lock chamber wall satisfies a first alarm condition based on the displacement data set according to a preset judgment rule; when the first alarm condition is satisfied, generate a first early warning message and send it to a preset user device; Wherein, the preset judgment rule means that a monitoring curve set is generated based on the displacement data set, and the monitoring curve set includes the displacement-time curves of each preset monitoring point; obtain the first similarity score between the monitoring curve set and the sample curve set; when the first similarity score is lower than a first score threshold, it is considered that the deformation situation of the lock chamber wall satisfies the first alarm condition; Each first preset period includes a second preset period. The method for obtaining the set of sample curves includes: Taking the first preset period starting from the current moment as the target preset period, and obtaining the stress monitoring information of multiple stress acquisition areas of the lock wall during the second preset period within the target preset period; Taking the stress information obtained during the second preset period within multiple first preset periods before the target preset period as stress sample information; Comparing each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score; Taking the stress sample information that is not lower than the second similarity threshold and is closest to the current moment as the target stress information, and taking the set of monitoring curves corresponding to the target stress information as the set of sample curves.

7. A deformation monitoring system for a lock chamber wall, characterized in that, The lock wall deformation monitoring system includes: A plurality of displacement sensors, which are respectively arranged at preset monitoring points at the joints of adjacent lock sections of the lock wall, and are used to obtain displacement data that changes over time; A control module, which is communicatively connected to each of the displacement sensors respectively; A gateway module, which is arranged between the control module and the user equipment to realize the communication connection between the control module and the user equipment; and is arranged between the control module and each of the displacement sensors to realize the communication connection between the control module and each of the displacement sensors; The control module is configured to: S1, within a first preset period starting from the current moment, use each of the displacement sensors to obtain a set of displacement data of the lock wall; the set of displacement data includes multiple sets of displacement data that change over time corresponding to each preset monitoring point, and the preset monitoring point is a monitoring point arranged at the joint of adjacent lock sections, and the displacement data that changes over time is used to indicate the deformation displacement situation between the corresponding adjacent lock sections; S2, based on the set of displacement data, judging whether the deformation situation of the lock wall meets the first alarm condition according to a preset judgment rule; when the first alarm condition is met, generating a first warning message and sending it to a preset user equipment; when the first alarm condition is not met, execute S1; Wherein, the preset judgment rule means that a set of monitoring curves is generated based on the set of displacement data, and the set of monitoring curves includes displacement time curves of each preset monitoring point; obtaining a first similarity score between the set of monitoring curves and the set of sample curves; when the first similarity score is lower than the first score threshold, it is considered that the deformation situation of the lock wall meets the first alarm condition; Each first preset period includes a second preset period. The method for obtaining the set of sample curves includes: Taking the first preset period starting from the current moment as the target preset period, and obtaining the stress monitoring information of multiple stress acquisition areas of the lock wall during the second preset period within the target preset period; Taking the stress information obtained during the second preset period within multiple first preset periods before the target preset period as stress sample information; Comparing each stress sample information with the stress monitoring information respectively to obtain the corresponding second similarity score; The stress sample information that is not lower than the second similarity threshold and is closest to the current moment is used as the target stress information, and the set of monitoring curves corresponding to the target stress information is used as the sample curve set.

8. The deformation monitoring system for the lock chamber wall according to claim 7, wherein, It further includes: Multiple stress monitoring sensors, which are respectively arranged in multiple stress acquisition areas of the lock chamber wall and are communicatively connected through the gateway module and the control module, and are used to acquire the stress data of each stress acquisition area; Wherein, the stress acquisition area is arranged at the junction area of two concrete pouring stages of the lock chamber wall. The first concrete pouring stage is a pouring construction based on the post-construction floor slab and including chamfer construction, and the second concrete pouring is the pouring construction after the first concrete pouring.

9. The deformation monitoring system for a lock chamber wall according to claim 8, wherein The construction elevation range of the first concrete pouring is 3 meters to 4 meters; the construction of the second concrete pouring is in the form of a moving formwork construction method, and it is constructed to the top based on the first concrete pouring construction.

Images