Reservoir dam safety intelligent inspection method and system

Abstract

The invention relates to the technical field of reservoir dam safety inspection, and discloses a reservoir dam safety intelligent inspection method and system, which comprises an inspection intelligent implementation module, a dam safety non-automatic monitoring data entry intelligent implementation module and a multi-source inspection information safety transmission management module, the inspection intelligent implementation module comprises an inspection planning unit, an inspection result input unit and an inspection result intelligent identification unit, and the dam safety non-automatic monitoring data input intelligent implementation module comprises measured value input task planning, measured value input task execution and safety state measured value intelligent judgment. And the multi-source inspection information security transmission management module has a multi-source data asynchronous transmission function and a multi-source data security management function. Compared with the prior art, scientific transmission and storage of unstructured recorded data of the engineering safety state are promoted, the working efficiency of reservoir dam safety inspection is effectively improved in combination with an artificial intelligence method, and the engineering digital management level is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of safe operation and maintenance of water conservancy projects, and in particular to a reservoir dam safety intelligent inspection method and system. Background Art

[0002] The complex construction and operation environment of water conservancy projects can easily lead to hidden dangers and damage to dam materials. At present, in the reservoir dams that have been put into operation, building a mature and reliable safety operation and maintenance system is an important guarantee for the project to achieve the expected benefits. In addition to the automated monitoring system, the entry of unstructured inspection information and non-automated monitoring data is also an important supplement to the safe operation and maintenance of reservoir dams. However, for a long time, the entry of inspection results and non-automated monitoring data has been difficult in terms of on-site situation perception, data retrieval, comprehensive analysis and comparison. The reason is that although there are standardized guidance for manual inspection work, the specific implementation effect depends more on the work habits and experience of the inspectors, and does not have the consistency of the automated monitoring system. At present, it has brought historic opportunities for the vigorous development and utilization of water resources. A large number of water conservancy and hydropower projects are under construction or have been included in the planning. In order to comprehensively improve the level of safe operation and maintenance of reservoir dams and provide valuable analysis data for the safety status analysis of major projects, it is urgent to establish a technical solution that can complete the inspection work in a standardized, comprehensive and safe manner. Summary of the invention

[0003] Purpose of the invention: In view of the problems in the background technology, the present invention provides a reservoir dam safety intelligent inspection method and system, which supports intelligent engineering inspection and intelligent entry of non-automatic monitoring data, and securely transmits and stores monitoring data and other unstructured inspection result information through a hierarchical database architecture, which can assist power station staff to complete daily inspection tasks in a standardized and standard manner, and meet the current needs of digital transformation of water conservancy projects. It can be promoted and applied to water-related structures such as port engineering and marine engineering, and provides a practical, effective, and multifunctional inspection intelligent implementation management platform for the integrated, real-time, global, and intelligent monitoring of large-volume water-related projects in complex environments, which has outstanding advantages in reducing monitoring costs, improving monitoring accuracy, and enhancing engineering practical capabilities.

[0004] Technical solution: The present invention discloses a reservoir dam safety intelligent inspection method, comprising the following steps:

[0005] Step 1: Determine the specific inspection items corresponding to the inspection tasks issued by the reservoir dam safety management agency, plan the order of execution of the inspection items, and use the dynamic programming algorithm to plan the route of the inspection tasks;

[0006] Step 2: Determine the project list according to the non-automatic monitoring data entry requirements corresponding to the reservoir dam inspection task, and use the dynamic programming algorithm to adaptively generate the optimal non-automatic safety monitoring entry route;

[0007] Step 3: Carry out inspections according to the planned inspection task routes, record the actual on-site conditions of specific inspection items of the reservoir dam, including text, image and video unstructured record data forms; carry out data monitoring according to the planned non-automatic monitoring data entry routes, read the readings of non-automatic monitoring project equipment and record them in the mobile inspection equipment;

[0008] Step 4: Apply the image feature intelligent recognition algorithm to process the unstructured recorded images of the actual situation at the inspection site, extract feature information and obtain the inspection results;

[0009] Step 5: Use extreme value theory to establish an early warning model for the input data of non-automated monitoring project equipment, adaptively determine whether the input data exceeds the upper and lower warning limits, and issue an early warning;

[0010] Step 6: Use multi-threading technology to cache the measured values, text, images, and video data obtained from the inspection in the mobile inspection device, and form a message queue for asynchronous pulling by the server of the reservoir dam inspection intelligent implementation management platform;

[0011] Step 7: The reservoir dam inspection intelligent implementation management platform completes access control, database hierarchical storage architecture, data transmission control, and data backup operations based on roles.

[0012] Furthermore, the dynamic programming algorithm in step 1 plans the route of the inspection task and the non-automatic monitoring data entry route, specifically:

[0013] Step 1.1: Use P = {p 1 ,p 2 ,...,p n} represents all inspection points, and the point coordinates correspond to the GPS coordinates of the inspection points;

[0014] Step 1.2: Construct the distance matrix D, D ij Indicates p i With p j If the distance between i With p j If the two are not directly reachable, then D ij =∞, construct the state space dp[mask][i];

[0015] Step 1.3: Define the initial state dp[1<<0][0], the state transition dp[mask|(1<<j)][j]=min(dp[mask|(1<<j)][j],dp[mask][i]+D[i][j]), and the termination condition result=min(dp[(1<<n)-1][i]+D[i][0]);

[0016] Step 1.4: Calculate the shortest path length to visit all points using the termination condition formula;

[0017] Step 1.5: Starting from the terminal state, trace the optimal path in reverse order to determine the order in which the inspection points are visited.

[0018] Furthermore, the image feature intelligent recognition algorithm in step 4 processes the unstructured recorded image of the actual situation at the inspection site, and the specific process is:

[0019] Step 4.1: Use the annotation tool to select the defective parts in the historical inspection image and assign labels;

[0020] Step 4.2: Divide the labeled image set into training set, validation set and test set;

[0021] Step 4.3: Integrate neural network and transfer learning to establish a reservoir dam defect identification network and determine the arrangement order and parameters of each type of layer. The type of layer includes input layer, convolution layer, activation layer, normalization layer, pooling layer, fully connected layer, and output layer;

[0022] Step 4.4: Start the training process of the network model, monitor the loss value and accuracy index, and avoid overfitting;

[0023] Step 4.5: Use the images in the validation set to evaluate the model performance, view key indicators, and use the test set to evaluate the final performance of the model;

[0024] Step 4.6: Use the trained model to process the unstructured recorded images of the actual situation at the inspection site.

[0025] Furthermore, in step 5, the extreme value theory is used to establish an early warning model for the input data, and adaptively judge whether the input data exceeds the upper and lower warning limits. The specific process is as follows:

[0026] Step 5.1: Input the non-automatic monitoring time series of the reservoir dam and use the distribution fitting method to determine the sequence threshold;

[0027] Step 5.2: Apply GPD distribution to fit the time series exceeding the threshold;

[0028] Step 5.3: Apply the maximum likelihood estimation method to calculate the specific parameters of the GPD distribution and output the super-threshold distribution function CEDF;

[0029] Step 5.4: Use the probability distribution conversion formula to obtain the overall distribution function of the super-threshold distribution function;

[0030] Step 5.5: Obtain the inverse function of the overall distribution function, and combine it with the probability of reservoir dam failure to obtain the warning limit of the non-automated monitoring time series.

[0031] Furthermore, in step 5, multi-threading technology is used to set up an acquisition thread and an upload thread. The acquisition thread is used to cache the measured values, text, image, and video data obtained during the inspection in the mobile inspection device, and a message queue is formed. The upload thread is used to upload the measured values, text, image, and video data obtained during the inspection to the water supply reservoir dam inspection intelligent implementation management platform server.

[0032] The present invention also discloses a system based on the above-mentioned reservoir dam safety intelligent inspection method, including an inspection intelligent implementation module and a dam safety non-automatic monitoring data entry intelligent implementation module;

[0033] The inspection intelligent implementation module includes an inspection planning unit, an inspection result input unit, and an inspection result intelligent identification unit;

[0034] The inspection planning unit is used to determine the specific inspection items corresponding to the inspection tasks issued by the reservoir dam safety management agency, and plan the execution sequence of the inspection items, and use a dynamic programming algorithm to plan the execution route of the inspection tasks. The specific inspection items are from the technical regulations for on-site inspection of hydropower station dam safety;

[0035] The inspection result input unit is used to record the actual situation of the specific inspection items of the reservoir dam during the inspection process, and the description method includes text, image and video unstructured recording data form;

[0036] The inspection result intelligent recognition unit is used to process the unstructured recorded images of the actual situation at the inspection site using an image feature intelligent recognition algorithm, extract feature information, extract feature information reflecting abnormal structural safety status, and issue a warning;

[0037] The dam safety non-automatic monitoring data entry intelligent implementation module includes a measurement value entry task planning unit, a measurement value entry task execution unit and a safety status measurement value intelligent judgment unit;

[0038] The measurement value entry task planning unit is used to determine the project list according to the non-automatic monitoring data entry requirements corresponding to the reservoir dam inspection task, and after the task is issued, according to the distribution of non-automatic monitoring projects of the reservoir dam safety status, adaptively generate the optimal non-automatic safety monitoring entry route;

[0039] The measurement value entry task execution unit is used to read the readings and record them in the mobile inspection device when arriving at the target non-automated monitoring project equipment;

[0040] The safety status measurement intelligent judgment unit is used in the reservoir dam inspection intelligent implementation management platform to establish an early warning model for input data based on historical monitoring data using extreme value theory, adaptively judge whether the input data exceeds the upper and lower warning limits, and issue an early warning.

[0041] Preferably, it also includes a multi-source inspection information security transmission management module, and the multi-source inspection data security transmission management module includes a multi-source data asynchronous transmission unit and a multi-source data security management unit;

[0042] The multi-source data asynchronous transmission unit is used to process the measured values, text, image, and video data obtained from the inspection using multi-threading technology, cache the information in the mobile inspection device, and form a message queue for asynchronous pulling by the server of the reservoir dam inspection intelligent implementation management platform, so as to realize the offline execution of the inspection task;

[0043] The multi-source data security management unit is used in the reservoir dam inspection intelligent implementation management platform to realize role-based access control, database hierarchical storage, data transmission control and data backup.

[0044] Preferably, the inspection route planning unit, the inspection result entry unit, the measurement entry task planning unit, and the measurement entry task execution unit are integrated into the mobile inspection device, and the measurement value intelligent judgment unit and the inspection result intelligent identification unit are integrated into the server side of the reservoir dam inspection intelligent implementation management platform. The readings of various equipment for reservoir dam safety monitoring are obtained by the mobile inspection device and connected to the reservoir dam inspection intelligent implementation management platform server through the communication module, and multi-source data security management is implemented through the database hierarchical storage architecture.

[0045] The present invention also discloses a reservoir dam safety intelligent inspection device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of the above-mentioned reservoir dam safety intelligent inspection method when loaded into the processor.

[0046] The present invention also discloses a computer storage medium storing computer executable instructions, wherein the computer executable instructions are used to execute the steps of the above-mentioned reservoir dam safety intelligent inspection method.

[0047] Beneficial effects:

[0048] 1. The intelligent inspection method and system for reservoir dam safety provided by the present invention meet the requirements of the "Technical Regulations for On-site Inspection of Hydropower Station Dam Safety", have standardized recording methods, clear implementation process logic, support real-time uploading and synchronization of multi-source inspection data, and can also complete offline inspection tasks. The introduction of artificial intelligence methods can greatly improve the inspection efficiency of reservoir dams and meet the needs of digital management of water conservancy projects.

[0049] 2. The image feature intelligent recognition algorithm of the present invention processes the unstructured recorded images of the actual situation at the inspection site, and is trained based on the defective parts in the historical inspection images. It can more accurately display the actual situation at the inspection site, reduce the pressure on mobile inspection equipment, and save identifiable data.

[0050] 3. The dynamic programming algorithm used in the present invention can reduce the deviations and omissions that may occur when patrol personnel subjectively determine the route when planning non-automated monitoring data entry routes. In the case of many patrol points with complex distribution, a systematic optimal route determination plan can also be quickly given.

[0051] 4. The present invention uses extreme value theory to establish an early warning model for input data, which can focus on the modeling and early warning of extreme outliers, and provide accurate anomaly detection capabilities for non-automated monitoring projects of reservoir dam safety. It has the advantages of flexible modeling, quantitative prediction, and sensitivity to outliers. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the module composition of a reservoir dam safety intelligent inspection system;

[0053] Figure 2 It is a schematic diagram of the relationship between the hardware and software components of a reservoir dam safety intelligent inspection system;

[0054] Figure 3 This is a schematic diagram of the on-site inspection process of the reservoir dam;

[0055] Figure 4 This is a schematic diagram of the process of inputting non-automatic monitoring data for reservoir dams;

[0056] Figure 5 This is a schematic diagram of the generation and management process of reservoir dam inspection data;

[0057] Figure 6 This is a schematic diagram of the inspection image intelligent recognition results;

[0058] Figure 7 Provide intelligent input and discrimination diagram for monitoring data;

[0059] Figure 8 This is a schematic diagram showing inspection result statistics and routes. DETAILED DESCRIPTION

[0060] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0061] The present invention provides a reservoir dam safety intelligent inspection method and system, which supports engineering intelligent inspection and intelligent entry of non-automatic monitoring data, and securely transmits and stores monitoring data and other unstructured inspection result information through a hierarchical database architecture, which can assist power station staff to complete daily inspection tasks in a standardized and standard manner, meeting the current needs of digital transformation of water conservancy projects.

[0062] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] See also Figure 1 , Figure 1 A schematic diagram of the module composition of a reservoir dam safety intelligent inspection system provided by an embodiment of the present invention, with reference to Figure 1 As shown, the reservoir dam safety intelligent inspection system includes:

[0064] The inspection intelligent implementation module includes an inspection planning unit, an inspection result input unit, and an inspection result intelligent identification unit.

[0065] The intelligent implementation module for non-automatic monitoring data entry of dam safety includes a measurement entry task planning unit, a measurement entry task execution unit and a safety status measurement intelligent judgment unit.

[0066] The multi-source inspection data security transmission management module includes a multi-source data asynchronous transmission unit and a multi-source data security management unit.

[0067] The inspection planning unit is used to determine the specific inspection items corresponding to the inspection tasks issued by the reservoir dam safety management agency, and to plan the execution sequence of these inspection items, and to use the dynamic programming algorithm to plan the execution route of the inspection tasks, specifically:

[0068] Step 1: Use P = {p 1 ,p 2 ,...,p n} represents all inspection points, and the point coordinates correspond to the GPS coordinates of the inspection points.

[0069] Step 2: Construct a distance matrix D, where D ij Indicates p i With p j If the distance between i With p j If the two are not directly reachable, then D ij=∞, construct the state space dp[mask][i].

[0070] Step 3: Define the initial state dp[1<<0][0], the state transition dp[mask|(1<<j)][j]=min(dp[mask|(1<<j)][j],dp[mask][i]+D[i][j]), and the termination condition result=min(dp[(1<<n)-1][i]+D[i][0]).

[0071] Step 4: Calculate the shortest path length to visit all points using the termination condition formula.

[0072] Step 5: Starting from the termination state, trace the optimal path in reverse order to determine the visit order of the inspection points.

[0073] The specific inspection items are from the technical regulations for on-site inspection of hydropower station dam safety. If the concrete dam foundation is required to be inspected, the corresponding concrete dam foundation inspection items shall be checked according to the inspection parts and items specified in the regulations (depending on the specific project situation, when it comes to specific projects, the items specified in the regulations may not exist). The concrete dam foundation inspection items include the dam shoulder, the junction of the dam body and the slope, the hollow or wide-slit dam cavity, the dam toe, and the downstream near-dam area.

[0074] Furthermore, the detailed inspection contents of the dam shoulder project are as follows: ① landslide, collapse, falling blocks and water seepage of the dam shoulder rock mass; cracks, erosion, damage, dislocation, cracking and loosening of the supporting body concrete; ② dislocation, damage, cracks and water seepage of the lining concrete of the grouting hole or drainage hole; collapse, falling blocks, extrusion, loosening of the rock mass of the unlined cavern and water discharge, turbidity and precipitates from the drainage holes.

[0075] The detailed inspection contents of the joint between the dam body and the bank slope include: dislocation, detachment, water seepage, etc.

[0076] The detailed inspection contents of the foundation corridor are: ① Concrete structure joint displacement, water seepage, concrete damage, cracks, water seepage, calcium precipitation, etc.; ② Water outflow and precipitation from drainage holes.

[0077] The detailed inspection contents of the hollow or wide-slit dam cavity include: extrusion, dislocation, seepage, precipitation, etc. of the foundation rock and the joint between the foundation rock and the dam body.

[0078] The detailed inspection contents of the dam toe include: leakage, subsidence, scour, etc.

[0079] The detailed inspection contents of the downstream near-dam area include: leakage, turbidity, precipitates, water gushing from the riverbank slope of the deep overburden foundation, carryout of fine particles, collapse, etc.

[0080] The inspection result entry unit can be used to record the actual on-site conditions of specific inspection items of the reservoir dam during the inspection process, and the description method includes unstructured recording data forms such as text, images and videos. The text description should correspond to the detailed inspection content of the specific items in the above inspection area, that is, the detailed inspection content provided provides necessary prompts for the inspectors. During the inspection process, if there are multiple inspection items to be completed at the location where the inspectors are located, the inspectors can temporarily store part of the inspection items.

[0081] The inspection result intelligent recognition unit is used to intelligently process the unstructured inspection record data (including text, images and videos). Mobile inspection equipment usually has limited performance, so such unstructured record data is uploaded to the server. The server's built-in intelligent recognition algorithm automatically extracts feature information reflecting abnormal structural safety status and issues warnings. See Figure 6 , Figure 6 This is a schematic diagram of the inspection image intelligent recognition results, as follows:

[0082] Step 1: Use the annotation tool to select the defective parts in the historical inspection image and assign labels (such as crack or seepage).

[0083] Step 2: Divide the labeled image set into training set, validation set and test set.

[0084] Step 3: Integrate neural networks and transfer learning to establish a reservoir dam defect identification network and determine the arrangement order and parameters of each type of layer (input layer, convolution layer, activation layer, normalization layer, pooling layer, fully connected layer, output layer, etc.).

[0085] Step 4: Start the training process of the network model, monitor the loss value, accuracy and other indicators to avoid overfitting.

[0086] Step 5: Use the images in the validation set to evaluate the model performance, check key indicators such as accuracy, etc. Use the test set to evaluate the final performance of the model, and measure the model's generalization ability and actual performance.

[0087] Step 6: Deploy the trained model on the cloud, establish a communication link between the mobile inspection device and the cloud, set the human-computer interaction mode of the alarm, and reserve a model update channel.

[0088] The measurement input task planning unit is used by the inspection personnel to determine the input project and plan the input task execution route according to the distribution of non-automatic monitoring projects of the reservoir dam safety status after the task is issued. For example, a dam has seepage monitoring projects and stress monitoring projects, among which the seepage monitoring project is distributed on the slope above the left bank dam road, the upstream and downstream sides of the right bank grouting hole, the curtain grouting corridor, the left bank grouting hole, and the dam foundation. The stress monitoring project is distributed in the two dam sections a and b. If the input project this time is a stress monitoring project, all the stress gauge measurement point positions corresponding to the two dam sections a and b are selected. And the optimal input route is adaptively planned according to the location of each measuring point.

[0089] The measurement value entry task execution unit is used to read the equipment readings and record them in the mobile inspection device when the inspection personnel arrive at the target non-automated monitoring project equipment. At this time, the mobile inspection device will adaptively display the historical input data process line of the non-automated measuring point.

[0090] The measurement value intelligent judgment unit refers to the intelligent implementation management platform of the reservoir dam inspection, which adaptively establishes a monitoring model for non-automatic monitoring data based on rich monitoring data, judges whether the input data exceeds the upper and lower warning limits, and adaptively issues an early warning. Figure 7 , Figure 7 The figure is a schematic diagram for intelligent input and discrimination of monitoring data. For example, when applying the method for formulating dam safety monitoring indicators based on the POT model, for the non-automatic measured data sequence of dam effect quantity, the POT model first determines a threshold value according to certain rules, and then uses the measured values ​​exceeding the threshold value as the sub-sample sequence, and uses the generalized Pareto distribution to fit the sub-sample sequence to obtain its distribution function, and then obtains the distribution function of the parent sample according to the statistical conditional probability formula, so as to estimate the dam safety monitoring indicators based on the probability of dam failure, as follows:

[0091] Step 1: Input the non-automatic monitoring time series of the reservoir dam and use the distribution fitting method to determine the sequence threshold.

[0092] Step 2: Apply GPD distribution to fit the time series exceeding the threshold.

[0093] Step 3: Apply the maximum likelihood estimation method to calculate the specific parameters of the GPD distribution and output the excess threshold distribution function (CEDF).

[0094] Step 4: Use the probability distribution conversion formula to obtain the overall distribution function of the super-threshold distribution function.

[0095] Step 5: Obtain the inverse function of the overall distribution function, and combine it with the probability of reservoir dam failure to obtain the warning limit of the non-automated monitoring time series.

[0096] The present invention integrates the functions of inspection route planning unit, inspection result input unit, measurement input task planning unit, and measurement input task execution unit into the mobile inspection terminal, and integrates the functions of measurement intelligent discrimination unit and inspection result intelligent identification unit into the server side. The readings of various equipment for reservoir dam safety monitoring are obtained by the mobile inspection terminal, and are connected to the reservoir dam inspection intelligent implementation management platform server through a communication device, and multi-source data security management is implemented through a database hierarchical storage architecture. The reservoir dam inspection intelligent implementation management platform can promote the scientific transmission and storage of unstructured record data of engineering safety status, and combined with artificial intelligence methods, it can effectively improve the efficiency of reservoir dam safety inspection work and enhance the level of engineering digital management.

[0097] The multi-source data asynchronous transmission unit refers to the measured values, texts, images, videos and other data obtained from the inspection. It is processed using multi-threading technology and can be cached in the mobile inspection device through the acquisition thread. It also forms a message queue and is asynchronously pulled by the server of the reservoir dam inspection intelligent implementation management platform through the upload thread. This can realize the offline execution of the inspection task and ensure the real-time transmission of the inspection results to the greatest extent.

[0098] The multi-source data security management unit refers to the intelligent implementation management platform for reservoir dam inspection, which ensures the integrity, confidentiality and consistency of multi-source data through role-based access control, database hierarchical storage architecture, data transmission control, data backup and other technical methods. Figure 5 shown.

[0099] The specific units of the patrol inspection intelligent implementation module are as follows, and its specific implementation process is as follows: Figure 3 As shown:

[0100] The inspection task receiving unit is used to receive the inspection tasks issued by the reservoir dam safety management agency.

[0101] The multi-level list unit of inspection items is used to display the inspection locations, items and detailed contents specified in the technical regulations for on-site inspection of dam safety of hydropower stations in a classified and graded manner.

[0102] The inspection item check-in unit is used to confirm the specific inspection item combination.

[0103] The inspection project route planning unit is used to adaptively generate the optimal inspection project route using a dynamic programming algorithm.

[0104] Contains a sequential inspection item combination import unit, providing a human-computer interaction interface for executing inspections.

[0105] The inspection information entry unit is used to provide an upload interface for the text, images, and videos recorded during the inspection, and to provide raw data for intelligent identification of inspection information.

[0106] The inspection information intelligent identification unit is used to automatically display the time, route and related information statistics of the inspection, and is connected to the intelligent algorithm running in the reservoir dam inspection intelligent implementation management platform server to realize the intelligent identification of inspection information.

[0107] The specific units of the intelligent implementation module for non-automatic monitoring data entry of dam safety are as follows: Figure 4 As shown:

[0108] The measurement value entry task receiving unit is used to receive non-automatic monitoring data entry tasks issued by the reservoir dam safety management agency.

[0109] Multi-level list unit for measurement entry items: includes all non-automated safety monitoring items that need to be entered in reservoir dam projects, and is displayed in levels by location, category, number, etc.

[0110] The measurement entry item check-in unit is used to confirm the specific non-automated safety monitoring entry item combination.

[0111] The measurement entry project route planning unit is used to adaptively generate the optimal non-automated safety monitoring entry project route.

[0112] A combination import unit containing non-automated safety monitoring entry items in the optimal sequence provides a human-computer interaction interface for executing entry.

[0113] The measurement value collection unit is used to provide an interface control for measurement value input and can display the historical non-automated monitoring data process line.

[0114] The input data security intelligent judgment unit is used to intelligently judge the legality of the input data format, and combines with the cloud-based intelligent monitoring model to adaptively issue abnormal warnings for input data.

[0115] The specific units of the multi-source inspection data security transmission management module are as follows, and its specific implementation process is as follows Figure 5 As shown:

[0116] Multi-source inspection data cache unit: temporarily stores the measured values, text images, videos and other data in different formats collected during the inspection task in the mobile inspection device, so that the inspection task can be executed offline.

[0117] Multi-source inspection data multi-threaded processing unit: can process the recording, transmission and analysis of multi-source data in parallel to improve data processing efficiency.

[0118] Multi-source inspection data message queue generation unit: Multi-source inspection data can be generated into a message queue for asynchronous pulling by the server, realizing asynchronous transmission of data in batches, and ensuring the real-time stability of inspection task execution.

[0119] Multi-source inspection data offline synchronization unit: the inspection data can be automatically synchronized and cached to the server to ensure that the inspection results are uploaded and updated in real time.

[0120] Multi-source inspection data hierarchical storage unit: Inspection data is stored and backed up at different levels according to hardware facilities to ensure that data can be stored and accessed efficiently and securely.

[0121] Multi-source inspection data access control unit: divide permissions based on user roles to ensure data is protected from unauthorized access.

[0122] Multi-source inspection data transmission optimization unit: dynamically adjusts the data transmission rate based on changes in network bandwidth, sets data transmission priority according to data importance, and automatically executes a restart mechanism when data transmission fails to ensure reliable data transmission.

[0123] The present invention also discloses a reservoir dam safety intelligent inspection device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the computer program is loaded into the processor, the inspection process of the reservoir dam safety intelligent inspection system is realized. A computer storage medium is also disclosed, which stores computer executable instructions, and the computer executable instructions are used to execute the inspection process of the reservoir dam safety intelligent inspection system.

[0124] The process of using the above-mentioned reservoir dam safety intelligent inspection method and system to conduct inspection is shown in Figure 8 , Figure 8 This is a schematic diagram showing inspection result statistics and routes of inspection items in an embodiment of the present invention.

[0125] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A reservoir dam safety intelligent inspection method, characterized in that: The following steps are involved: Step 1: Determine the specific inspection items corresponding to the inspection tasks issued by the reservoir dam safety management agency, plan the execution sequence of the inspection items, and use the dynamic programming algorithm to plan the execution route of the inspection tasks; Step 2: Determine the project list according to the non-automatic monitoring data entry requirements corresponding to the reservoir dam inspection task, and use the dynamic programming algorithm to adaptively generate the optimal non-automatic safety monitoring entry route; Step 3: Carry out inspections according to the planned inspection task routes, record the actual on-site conditions of specific inspection items of the reservoir dam, including text, image and video unstructured record data forms; carry out data monitoring according to the planned non-automatic monitoring data entry routes, read the readings of non-automatic monitoring project equipment and record them in the mobile inspection equipment; Step 4: Apply the image feature intelligent recognition algorithm to process the unstructured recorded images of the actual situation at the inspection site, extract feature information and obtain the inspection results; Step 5: Use extreme value theory to establish an early warning model for the input data of non-automated monitoring project equipment, adaptively determine whether the input data exceeds the upper and lower warning limits, and issue an early warning; Step 6: Use multi-threading technology to cache the measured values, text, images, and video data obtained from the inspection in the mobile inspection device, and form a message queue for asynchronous pulling by the server of the reservoir dam inspection intelligent implementation management platform; Step 7: The reservoir dam inspection intelligent implementation management platform completes access control, database hierarchical storage architecture, data transmission control, and data backup operations based on roles.

2. A reservoir dam safety intelligent inspection method according to claim 1, characterized in that: The dynamic programming algorithm in step 1 and step 2 plans the route of the inspection task and the non-automatic monitoring data entry route, specifically: Step 1.1: Use P = {p1, p2, ..., p n } represents all inspection points, and the point coordinates correspond to the GPS coordinates of the inspection points; Step 1.2: Construct the distance matrix D, D ij Indicates p i With p j If the distance between i With p j If the two are not directly reachable, then D ij =∞, construct the state space dp[mask][i]; Step 1.3: Define the initial state dp[1<<0][0], the state transition dp[mask|(1<<j)][j]=min(dp[mask|(1<<j)][j],dp[mask][i]+D[i][j]), and the termination condition result=min(dp[(1<<n)-1][i]+D[i][0]); Step 1.4: Calculate the shortest path length to visit all points using the termination condition formula; Step 1.5: Starting from the terminal state, trace the optimal path in reverse order to determine the order in which the inspection points are visited.

3. A reservoir dam safety intelligent inspection method according to claim 1, characterized in that: In step 4, the image feature intelligent recognition algorithm processes the unstructured recorded image of the actual situation at the inspection site. The specific process is: Step 4.1: Use the annotation tool to select the defective parts in the historical inspection image and assign labels; Step 4.2: Divide the labeled image set into training set, validation set and test set; Step 4.3: Integrate neural network and transfer learning to establish a reservoir dam defect identification network and determine the arrangement order and parameters of each type of layer. The type of layer includes input layer, convolution layer, activation layer, normalization layer, pooling layer, fully connected layer, and output layer; Step 4.4: Start the training process of the network model, monitor the loss value and accuracy index, and avoid overfitting; Step 4.5: Use the images in the validation set to evaluate the model performance, view key indicators, and use the test set to evaluate the final performance of the model; Step 4.6: Use the trained model to process the unstructured recorded images of the actual situation at the inspection site.

4. A reservoir dam safety intelligent inspection method according to claim 1, characterized in that: In step 5, extreme value theory is used to establish an early warning model for input data, and adaptively judge whether the input data exceeds the upper and lower warning limits. The specific process is as follows: Step 5.1: Input the non-automatic monitoring time series of the reservoir dam and use the distribution fitting method to determine the sequence threshold; Step 5.2: Apply GPD distribution to fit the time series exceeding the threshold; Step 5.3: Apply the maximum likelihood estimation method to calculate the specific parameters of the GPD distribution and output the super-threshold distribution function CEDF; Step 5.4: Use the probability distribution conversion formula to obtain the overall distribution function of the super-threshold distribution function; Step 5.5: Obtain the inverse function of the overall distribution function, and combine it with the probability of reservoir dam failure to obtain the warning limit of the non-automated monitoring time series.

5. The intelligent inspection method for reservoir dam safety according to claim 1 is characterized in that: In step 5, a collection thread and an upload thread are set up by multi-threading technology. The measurement values, text, image, and video data obtained from the inspection are cached in the mobile inspection device through the collection thread, and a message queue is formed. The measurement values, text, image, and video data obtained from the inspection are uploaded to the water supply reservoir dam inspection intelligent implementation management platform server through the upload thread.

6. A system based on the reservoir dam safety intelligent inspection method according to any one of claims 1 to 5, characterized in that: It includes inspection intelligent implementation module and dam safety non-automatic monitoring data entry intelligent implementation module; The inspection intelligent implementation module includes an inspection planning unit, an inspection result input unit, and an inspection result intelligent identification unit; The inspection planning unit is used to determine the specific inspection items corresponding to the inspection tasks issued by the reservoir dam safety management agency, and plan the execution sequence of the inspection items, and use a dynamic programming algorithm to plan the execution route of the inspection tasks. The specific inspection items are from the technical regulations for on-site inspection of hydropower station dam safety; The inspection result input unit is used to record the actual situation of the specific inspection items of the reservoir dam during the inspection process, and the description method includes text, image and video unstructured recording data form; The inspection result intelligent recognition unit is used to process the unstructured recorded images of the actual situation at the inspection site using an image feature intelligent recognition algorithm, extract feature information, extract feature information reflecting abnormal structural safety status, and issue a warning; The dam safety non-automatic monitoring data entry intelligent implementation module includes a measurement value entry task planning unit, a measurement value entry task execution unit and a measurement value intelligent judgment unit; The measurement value entry task planning unit is used to determine the project list according to the non-automatic monitoring data entry requirements corresponding to the reservoir dam inspection task, and after the task is issued, according to the distribution of non-automatic monitoring projects of the reservoir dam safety status, adaptively generate the optimal non-automatic safety monitoring entry route; The measurement value entry task execution unit is used to read the readings and record them in the mobile inspection device when arriving at the target non-automated monitoring project equipment; The measured value intelligent judgment unit is used for the reservoir dam inspection intelligent implementation management platform to establish an early warning model for input data based on historical monitoring data using extreme value theory, adaptively judge whether the input data exceeds the upper and lower warning limits, and issue an early warning.

7. The system according to claim 6, characterized in that It also includes a multi-source inspection information security transmission management module, the multi-source inspection data security transmission management module includes a multi-source data asynchronous transmission unit and a multi-source data security management unit; The multi-source data asynchronous transmission unit is used to process the measured values, text, image, and video data obtained from the inspection using multi-threading technology, cache the information in the mobile inspection device, and form a message queue for asynchronous pulling by the server of the reservoir dam inspection intelligent implementation management platform, so as to realize the offline execution of the inspection task; The multi-source data security management unit is used in the reservoir dam inspection intelligent implementation management platform to realize role-based access control, database hierarchical storage, data transmission control and data backup.

8. The system according to claim 6 or 7, characterized in that: The inspection route planning unit, inspection result entry unit, measurement entry task planning unit and measurement entry task execution unit are integrated into the mobile inspection equipment. The measurement value intelligent judgment unit and the inspection result intelligent identification unit are integrated into the server side of the reservoir dam inspection intelligent implementation management platform. The readings of various equipment for reservoir dam safety monitoring are obtained by the mobile inspection equipment and connected to the reservoir dam inspection intelligent implementation management platform server through the communication module. Multi-source data security management is implemented through the database hierarchical storage architecture.

9. A reservoir dam safety intelligent inspection device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is loaded into the processor, the steps of the reservoir dam safety intelligent inspection method according to any one of claims 1 to 5 are implemented.

10. A computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions are used to execute the steps of the reservoir dam safety intelligent inspection method according to any one of claims 1 to 5.

Images