Development prediction method and system after crack repair of concrete gravity dam based on deep learning and finite element, electronic equipment and storage medium
By combining deep learning and finite element methods, using drone data and three-dimensional finite element models, the problem that traditional methods are difficult to accurately simulate and predict repaired cracks is solved, and accurate prediction of crack development trends is achieved.
Patent Information
- Application Number
- CN202510324451.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-19
- Publication Date
- 2025-06-13
AI Technical Summary
Traditional numerical calculation methods for cracking are difficult to accurately simulate and accurately predict the generation and development trend of ultra-fine cracks in the repaired crack areas.
Using deep learning and finite element methods, we use drones to collect apparent crack monitoring data before and after repair, build a three-dimensional finite element model and deep learning prediction model, and fuse multi-source data for accurate simulation and prediction.
Accurate simulation and accurate prediction of the repaired crack area can be achieved, and the crack generation location, morphology and expansion trend can be predicted, improving the accuracy and reliability of the prediction.
Smart Images
Figure CN120145769A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent monitoring and early warning for water conservancy and hydropower projects, and particularly relates to a method, system, electronic device and storage medium for predicting the development of cracks in a concrete gravity dam after crack repair based on deep learning and finite element methods. Background Art
[0002] Gravity dams are constructed with mass concrete. During the process from dam construction to operation, due to the comprehensive influence of external factors such as concrete pouring technology, temperature control measures, and water pressure, a certain number of cracks will inevitably occur. Once these cracks appear, they must be promptly treated through repair measures such as crack grouting to restore and ensure the integrity and stability of the gravity dam structure, and to ensure the safe and stable operation of the dam.
[0003] However, after the repair, due to the differences in material properties and mechanical properties between the grouting repair material and the original concrete structure, new cracks are very likely to occur or the original cracks may reopen under the continuous action of temperature changes and water pressure. When facing such complex situations, traditional numerical calculation methods for crack opening are difficult to accurately simulate the ultra-fine cracks in the repaired crack area, and even less able to accurately predict their generation and development trends. Summary of the Invention
[0004] In order to solve the above problems existing in the prior art, the purpose of the present invention is to provide a method, system, electronic device and storage medium for predicting the development of cracks in a concrete gravity dam after crack repair based on deep learning and finite element methods, so as to accurately simulate the ultra-fine cracks in the repaired crack area and accurately predict their generation and development trends.
[0005] The technical solution adopted by the present invention is as follows:
[0006] A method for predicting the development of cracks in a concrete gravity dam after crack repair based on deep learning and finite element methods, comprising the following steps:
[0007] S1: Using drones to collect the apparent monitoring data of the cracks in the gravity dam before and after repair;
[0008] S2: Constructing a three-dimensional finite element model of the cracks in the concrete gravity dam before and after repair, and inversely obtaining the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material;
[0009] S3: Constructing a gravity dam crack development prediction model based on deep learning, and performing model training and cross-validation;
[0010] S4: Using the gravity dam crack development prediction model after repair to predict the subsequent crack development trend.
[0011] With the development of artificial intelligence technology and monitoring technology in recent years, it is possible to comprehensively and completely monitor the generation and expansion process of cracks, which lays a solid foundation for accurately predicting the expansion trend of cracks after repair and brings new opportunities for overcoming this difficult problem.
[0012] The present invention has the ability of multi-source data fusion and can fuse multi-source information such as the characteristics of repair materials, the original condition of the gravity dam structure, water pressure, and temperature field. At the same time, relying on the accurate simulation of the mechanical behavior of the gravity dam structure by three-dimensional finite element, it deeply analyzes the intricate mechanical relationships such as stress transfer and strain distribution between the crack repair area and the original dam body structure. With the help of artificial intelligence technology, it accurately captures various complex factors generated by cracks after repair, and then optimizes the simulation parameters to achieve accurate prediction of the initiation and expansion path of cracks after repair. Taking the situation after the repair of the dam cracks as an example, the model can comprehensively consider key factors such as the elastic modulus of grouting repair materials, the quality of grouting construction, water pressure, and temperature field, and can accurately predict the specific location, length, width, direction and other key parameters of the reappearance of cracks after repair.
[0013] The present invention can carry out in-depth learning training on the deformation data of surface cracks of the dam body, has excellent generalization ability, and can also give reasonable and reliable predictions in the face of complex environments such as complex material mechanics and multiple working conditions after repair. The prediction results are presented in the form of intuitive three-dimensional graphics, clearly showing the possible positions, shapes and expansion trends of cracks after repair in the dam.
[0014] As a preferred embodiment of the present invention, step S1 is specifically as follows:
[0015] Using the micro-deformation photogrammetry technology of unmanned aerial vehicles, multi-period monitoring data of surface cracks of the gravity dam before and after crack repair are collected to obtain the true expansion process and distribution pattern of the gravity dam cracks.
[0016] As a preferred embodiment of the present invention, the true expansion process and distribution pattern of the gravity dam cracks include:
[0017] The apparent crack length L before repair Q , width W Q , extension direction D Q , cracking location LOC Q , denoted as CR Q =(L Q , W Q , D Q , LOC Q );
[0018] The apparent crack length L after repair H , width W H , extension direction D H , cracking location LOC H, denoted as CR H =(L H , W H , D H , LOC H ).
[0019] As a preferred embodiment of the present invention, step S2 is specifically as follows:
[0020] Based on the appearance dimensions, structural characteristics, mechanical parameters, environmental parameters, and crack distribution of the concrete gravity dam, a three-dimensional finite element model MODEL1 of the overall concrete gravity dam considering the combined action of the dam-foundation-load is constructed: For different load combination conditions that may cause surface cracking of the gravity dam, a crack propagation simulation method based on the maximum tensile stress criterion is used to accurately simulate the pre-repair crack apparent index CR Q =(L Q , W Q , D Q , LOC Q ) identified by the drone in step S1 in the three-dimensional finite element model, so as to back-calculate the comprehensive stress load LOAD Q of the crack area of the gravity dam;
[0021] Subsequently, considering adding a grouting repair material module after repair on the three-dimensional finite element model MODEL1 of the overall concrete gravity dam, a three-dimensional finite element model MODEL2 of the overall concrete gravity dam after repair is constructed: The comprehensive stress load LOAD Q back-calculated from model MODEL1 is brought into the three-dimensional finite element model MODEL2 of the overall concrete gravity dam after repair, and the post-repair apparent crack data index CR H =(L H , W H , D H , LOC H ) identified by the drone in step S1 is accurately simulated. Through system parameter inversion, the comprehensive performance index parameters CP G =(Rm G , Ib G , Me G , Qu G ) of the grouting repair material for repair are obtained; among them, Rm G represents the cured tensile strength of the grouting repair material, Me G represents the elastic modulus parameter of the grouting repair material, Ib G represents the interfacial bonding strength between the grouting repair material and the concrete, and Qu G represents the construction quality of the grouting repair material.
[0022] As a preferred embodiment of the present invention, step S3 is specifically as follows:
[0023] For the time series data of crack propagation after repair, a long short-term memory network of deep learning is used to build a prediction model MODEL for crack propagation of gravity dams pred , to predict the crack development in the repaired area:
[0024] The apparent crack index CR before repair obtained in step S1 Q =(L Q , W Q , D Q , LOC Q ), the comprehensive stress load LOAD obtained in step S2 Q and the comprehensive index parameters CP of the grouting repair material G =(Rm G , Ib G , Qu G ) are used as the input of the prediction model, and then the apparent crack index CR after repair obtained in S1 H =(L H , W H , D H , LOC H ) is used as the output of the prediction model for prediction model training;
[0025] Cross-validation is carried out using the test set and the prediction results of the artificial intelligence model for crack propagation of gravity dams; by reasonably adjusting the model parameters, the model performance is further optimized to improve the accuracy and generalization ability of the model.
[0026] As a preferred solution of the present invention, in step S3,
[0027] The objective function of the deep learning prediction model is:
[0028]
[0029] Among them, N is the number of samples used for training and validation of the prediction model; is the predicted crack index after repair obtained by the prediction model through the fusion model of the three-dimensional finite element model and the LSTM network; CR Hi is the true data index of the crack development after repair obtained in step S1; is the penalty term considering the influence of the comprehensive load of the gravity dam on the prediction result of the crack after repair, λ 1 is the hyperparameter used to adjust the weight of the penalty term related to the gravity dam load, and m is the number of parameters included in the gravity dam comprehensive load LOAD Q ; is the penalty term considering the influence of the original crack before repair on the prediction result of the crack after repair, λ 2is a hyperparameter for adjusting the weight of the penalty term related to the original cracks before repair, and n is the number of parameters included in the original crack index CRQ before repair; is a penalty term for considering the influence of the performance parameters of the grouting repair material used for repair on the predicted results of the cracks after repair, λ 3 is a hyperparameter for adjusting the weight of the penalty term related to the performance of the grouting repair material used for repair, and l is the number of parameters included in the performance index CPG of the grouting repair material.
[0030] As a preferred embodiment of the present invention, step S4 is specifically:
[0031] Input the multi-stage expansion data of the cracks and multi-condition parameters collected by the unmanned aerial vehicle into the latest three-dimensional finite element model and the gravity dam crack development prediction model based on artificial intelligence, and predict the future development trend of the concrete cracks in the gravity dam. The prediction content covers key parameters such as the generation location, expansion length, expansion width, and expansion direction of the cracks.
[0032] A development prediction system for concrete gravity dam cracks after repair based on deep learning and finite element
[0033] including a data acquisition unit: the unmanned aerial vehicle collects the apparent monitoring data of the gravity dam cracks before and after repair;
[0034] a model construction unit: construct a three-dimensional finite element model of the concrete gravity dam cracks before and after repair, and inversely obtain the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material;
[0035] a model training unit: construct a gravity dam crack development prediction model based on deep learning, and perform model training and cross-validation;
[0036] a trend prediction unit: perform subsequent crack development trend prediction through the gravity dam crack development prediction model after repair.
[0037] An electronic device includes at least one processor and at least one memory communicatively connected to the processor; wherein, the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the method described in any one of the above.
[0038] A non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the method described in any one of the above.
[0039] The beneficial effects of the present invention are:
[0040] 1. The present invention has the ability of multi-source data fusion and can fuse and repair multi-source information such as material properties, the original condition of the gravity dam structure, water pressure, and temperature field. At the same time, relying on the accurate simulation of the mechanical behavior of the gravity dam structure by three-dimensional finite element, it deeply analyzes the intricate mechanical relationships such as stress transfer and strain distribution between the crack repair area and the original dam body structure. With the help of artificial intelligence technology, it accurately captures various complex factors generated by the cracks after repair, and then optimizes the simulation parameters to achieve accurate prediction of the initiation and propagation paths of the cracks after repair. Taking the situation after the dam cracks are repaired as an example, this model can comprehensively consider key factors such as the elastic modulus of the grouting repair material, the quality of grouting construction, water pressure, and temperature field, and can accurately predict the specific location, length, width, direction, and other key parameters of the reappearance of the cracks after repair.
[0041] 2. The present invention can carry out in-depth learning training on the deformation data of the cracks on the dam surface and has excellent generalization ability. In the face of complex material mechanics and multi-condition complex environmental situations after repair, it can also give reasonable and reliable predictions, and present the prediction results in the form of intuitive three-dimensional graphics, clearly showing the possible positions, shapes, and expansion trends of the cracks after repair on the dam. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 is the flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions 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 some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown here can be arranged and designed in various different configurations.
[0044] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the present invention claimed, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.
[0045] Embodiment 1:
[0046] As Figure 1 shown, the method for predicting the development of cracks after repair of a concrete gravity dam based on deep learning and finite element in this embodiment includes the following steps:
[0047] S1: The drone collects the apparent monitoring data of the gravity dam cracks before and after repair.
[0048] The drone has the characteristics of flexibility and high efficiency. By using the micro-deformation photogrammetry technology of the drone, multi-period monitoring data of the surface cracks of the gravity dam before and after crack repair are collected, and the true expansion process and distribution pattern of the gravity dam cracks are obtained. The true expansion process and distribution pattern of the gravity dam cracks include:
[0049] The apparent crack length L before repair Q , width W Q , extension direction D Q , cracking location LOC Q , denoted as CR Q =(L Q , W Q , D Q , LOC Q );
[0050] The apparent crack length L after repair H , width W H , extension direction D H , cracking location LOC H , denoted as CR H =(L H , W H , D H , LOC H ).
[0051] S2: Construct a three-dimensional finite element model of the concrete gravity dam cracks before and after repair, and inversely obtain the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material.
[0052] According to the appearance size, structural characteristics, mechanical parameters, environmental parameters and crack distribution of the concrete gravity dam, a three-dimensional finite element model MODEL1 of the concrete gravity dam considering the combined action of the dam-foundation-load is constructed: For different load combination conditions that may cause surface cracking of the gravity dam, a crack propagation simulation method based on the maximum tensile stress criterion is adopted to accurately simulate the apparent crack index CR Q =(L Q , W Q , D Q , LOC Q ) identified by the drone in step S1 in the three-dimensional finite element model, so as to inversely obtain the comprehensive stress load LOAD Q of the gravity dam crack area.
[0053] Subsequently, considering adding the repaired grouting repair material module to the overall three-dimensional finite element model MODEL1 of the concrete gravity dam, the overall three-dimensional finite element model MODEL2 of the repaired concrete gravity dam is constructed: the comprehensive stress load LOAD in the crack area of the gravity dam is inversely calculated from model MODEL1 Q and input into the overall three-dimensional finite element model MODEL2 of the repaired concrete gravity dam, and the repaired apparent crack data index CR identified by the UAV in step S1 is accurately simulated H =(L H , W H , D H , LOC H ). Through system parameter inversion, the comprehensive performance index parameters CP of the grouting repair material for repair are obtained G =(Rm G , Ib G , Me G , Qu G ); among them, Rm G represents the cured tensile strength of the grouting repair material, Me G represents the elastic modulus parameter of the grouting repair material, Ib G represents the interfacial bonding strength between the grouting repair material and concrete, and Qu G represents the construction quality of the grouting repair material, that is, whether it is full without cavities, etc.
[0054] S3: Construct a gravity dam crack development prediction model based on deep learning, and conduct model training and cross-validation
[0055] For the repaired crack propagation time series data, use the long short-term memory network of deep learning to build a prediction model MODEL pred for the gravity dam crack propagation to predict the crack development in the repaired area
[0056] Take the pre-repair crack apparent index CR Q =(L Q , W Q , D Q , LOC Q ) obtained in step S1, the comprehensive stress load LOAD Q obtained in step S2, and the comprehensive index parameters CP of the grouting repair material G =(Rm G , Ib G , Qu G ) as the input of the prediction model, and then take the post-repair crack apparent index CR H =(L H , W H , D H , LOCH ) As the output of the prediction model, the prediction model is trained.
[0057] Among them, the objective function of the deep learning prediction model is:
[0058]
[0059] Among them, N is the number of samples used for training and validation of the prediction model; is the predicted index of the repaired crack obtained by the prediction model through the fusion model of the three-dimensional finite element model and the LSTM network; CR Hi is the true data index of the repaired crack development obtained in step S1; is the penalty term considering the influence of the comprehensive load of the gravity dam on the prediction result of the repaired crack, λ 1 is the hyperparameter used to adjust the weight of the penalty term related to the gravity dam load, m is the number of parameters included in the comprehensive load LOAD of the gravity dam Q included; is the penalty term considering the influence of the original crack before repair on the prediction result of the repaired crack, λ 2 is the hyperparameter used to adjust the weight of the penalty term related to the original crack before repair, n is the number of parameters included in the original crack index CRQ before repair; is the penalty term considering the influence of the performance parameters of the grouting repair material used for repair on the prediction result of the repaired crack, λ 3 is the hyperparameter used to adjust the weight of the penalty term related to the performance of the grouting repair material, l is the number of parameters included in the performance index CPG of the grouting repair material.
[0060] At the same time, to further verify the reliability of the prediction model, cross-validation is carried out using the test set and the prediction result of the artificial intelligence model for the crack propagation of the gravity dam. By reasonably adjusting the model parameters, the model performance is further optimized, and the accuracy and generalization ability of the model are improved.
[0061] S4: Through the development prediction model of the gravity dam after crack repair, the subsequent crack development trend is predicted.
[0062] The multi-stage expansion data of the crack and multi-condition parameters collected by the unmanned aerial vehicle are input into the latest three-dimensional finite element model and the artificial intelligence-based gravity dam crack development prediction model to predict the future development trend of the concrete crack of the gravity dam. The prediction content covers key parameters such as the generation location, expansion length, expansion width, and expansion direction of the crack.
[0063] In summary, the present invention needs to collect the apparent data before and after crack repair by using an unmanned aerial vehicle, and the data includes crack length, width, extension direction, cracking position, etc.
[0064] The present invention needs to construct three-dimensional finite element models of the cracks in the concrete gravity dam before and after repair, and inversely obtain the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material. A whole three-dimensional finite element model MODEL1 of the concrete gravity dam that comprehensively considers the co-action of "dam - foundation - load" is constructed. On MODEL1, a module of the grouting repair material after repair is considered to be added to construct a whole three-dimensional finite element model MODEL2 of the concrete gravity dam after repair. The comprehensive stress load LOAD in the crack area of the gravity dam inversely obtained from model MODEL1 Q is brought into the whole three-dimensional finite element model MODEL2 of the concrete gravity dam after repair, and the apparent crack data index CR identified by the unmanned aerial vehicle in S1 is accurately simulated H =(L H , W H , D H , LOC H ). Through the inversion of system parameters, the comprehensive performance index parameters of the grouting repair material for repair are obtained.
[0065] The present invention needs to construct a prediction model for the development of cracks in the gravity dam based on deep learning, and conduct model training and cross-validation. For the time series data of the crack propagation after repair, a long short-term memory network (LSTM) of deep learning is used to build a prediction model MODELpred suitable for the crack propagation in the gravity dam to predict the crack development in the repaired area. At the same time, to further verify the reliability of the prediction model, a test set is used for cross-validation with the prediction results of the artificial intelligence model for the crack propagation in the gravity dam. By reasonably adjusting the model parameters, the model performance is further optimized to improve the accuracy and generalization ability of the model.
[0066] The present invention conducts subsequent prediction of the crack development trend through the prediction model for the development of cracks in the gravity dam after repair. The multi-period expansion data of the cracks collected by the unmanned aerial vehicle and multi-condition parameters are input into the latest three-dimensional finite element model and the artificial intelligence-based prediction model for the crack development in the gravity dam to predict the future development trend of the concrete cracks in the gravity dam. The prediction content covers key parameters such as the generation location, expansion length, expansion width, and expansion direction of the cracks.
[0067] Example 2:
[0068] The development prediction system for the cracks in the concrete gravity dam after repair based on deep learning and finite element in this example includes a data acquisition unit: the unmanned aerial vehicle collects the apparent monitoring data of the cracks in the gravity dam before and after repair;
[0069] a model construction unit: constructs three-dimensional finite element models of the cracks in the concrete gravity dam before and after repair, and inversely obtains the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material;
[0070] Model training unit: Construct a prediction model for the crack development of gravity dams based on deep learning, and perform model training and cross-validation;
[0071] Trend prediction unit: Through the prediction model for the development after the repair of gravity dam cracks, predict the subsequent crack development trend.
[0072] Example 3:
[0073] The electronic device of this embodiment includes at least one processor and at least one memory communicatively connected to the processor; wherein, the memory stores program instructions executable by the processor, and the processor can execute the method described in Example 1 by invoking the program instructions.
[0074] The electronic device includes a processor, a memory, and a bus; wherein, the processor and the memory complete communication with each other through the bus. The processor is used to invoke the program instructions in the memory to execute the methods provided in the above method embodiments. For example, it includes: the unmanned aerial vehicle collects the apparent monitoring data of the gravity dam cracks before and after repair; constructs a three-dimensional finite element model of the concrete gravity dam cracks before and after repair, and inversely obtains the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material; constructs a prediction model for the crack development of gravity dams based on deep learning, and performs model training and cross-validation; through the prediction model for the development after the repair of gravity dam cracks, predict the subsequent crack development trend.
[0075] In addition, when the logical instructions in the above-mentioned memory can be implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. And the aforementioned storage medium includes: USB flash drive, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other various media that can store program codes.
[0076] Example 4:
[0077] The non-transitory computer-readable storage medium of this embodiment stores computer instructions, and the computer instructions cause the computer to execute the method described in Embodiment 1. The non-transitory computer-readable storage medium of the present invention can store computer instructions, and the computer instructions cause the computer to execute the method of the present invention, which is convenient for automatic control. For example, it includes: the drone collects the apparent monitoring data of the cracks in the gravity dam before and after repair; constructs a three-dimensional finite element model of the cracks in the concrete gravity dam before and after repair, and inversely obtains the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material; constructs a prediction model for the development of the cracks in the gravity dam based on deep learning, and conducts model training and cross-validation; through the prediction model for the development of the cracks in the gravity dam after repair, predicts the subsequent crack development trend.
[0078] The present invention is not limited to the above optional implementation manners. Anyone can obtain other various forms of products under the inspiration of the present invention. However, no matter what changes are made in its shape or structure, as long as the technical solutions fall within the scope defined by the claims of the present invention, they all fall within the protection scope of the present invention.
Claims
1. A method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite elements, characterized by: The following steps are involved: S1: UAV collects apparent monitoring data of cracks in gravity dams before and after repair; S2: Construct a three-dimensional finite element model of the cracks in the concrete gravity dam before and after repair, and inversely obtain the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material; S3: Build a deep learning-based prediction model for the development of gravity dam cracks, and perform model training and cross-validation; S4: The subsequent crack development trend is predicted through the development prediction model of gravity dam cracks after repair.
2. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 1, characterized in that: Step S1 is specifically as follows: Using drone micro-variation photogrammetry technology, multi-period monitoring data of cracks on the surface of the gravity dam before and after crack repair were collected to obtain the actual expansion process and distribution pattern of cracks in the gravity dam.
3. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 2, characterized in that: The actual expansion process and distribution patterns of gravity dam cracks include: Apparent crack length before repair L Q , Width W Q , extension direction D Q 、Crack location LOC Q , denoted as CR Q =(L Q , W Q , D Q , LOC Q ); Apparent crack length after repair L H , Width W H , extension direction D H 、Crack location LOC H , denoted as CR H =(L H , W H , D H , LOC H ).
4. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 3 is characterized in that: Step S2 is specifically as follows: According to the appearance size, structural characteristics, mechanical parameters, environmental parameters and crack distribution of the concrete gravity dam, a three-dimensional finite element model of the concrete gravity dam is constructed which comprehensively considers the combined effects of dam-foundation-load. MODEL1: According to different load combination conditions that may cause surface cracking of the gravity dam, a crack propagation simulation method based on the maximum tensile stress criterion is adopted to accurately simulate the crack surface index CR before repair identified by the drone in step S1 in the three-dimensional finite element model. Q =(L Q , W Q , D Q , LOC Q ), thereby inverting the comprehensive load LOAD in the crack area of the gravity dam Q ; Subsequently, the repaired grouting repair material module was added to the overall three-dimensional finite element model MODEL1 of the concrete gravity dam to construct the overall three-dimensional finite element model MODEL2 of the repaired concrete gravity dam: The comprehensive force load LOAD of the crack area of the gravity dam was inverted by model MODEL1 Q Bring in the overall three-dimensional finite element model MODEL2 of the repaired concrete gravity dam, and accurately simulate the repaired apparent crack data index CR identified by the drone in step S1 H =(L H , W H , D H , LOC H ), through the inversion of system parameters, the comprehensive performance index parameter CP of the grouting repair material for repair is obtained G =(Rm G , Ib G , Me G , Qu G ), where Rm G Indicates the cured tensile strength of the grouting repair material, Me G Represents the elastic modulus parameter of the grouting repair material, Ib G Indicates the interfacial bonding strength between the grouting repair material and concrete, Qu G Indicates the construction quality of grouting repair materials.
5. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 4 is characterized in that: Step S3 is specifically as follows: Based on the time series data of crack expansion after repair, a long short-term memory network based on deep learning is used to build a prediction model for crack expansion in gravity dams. pred , used to predict crack development in the repaired area: The crack surface index CR before repair obtained in step S1 Q =(L Q , W Q , D Q , LOC Q ), and the comprehensive load LOAD obtained in step S2 Q And the comprehensive index parameter CP of grouting repair materials G =(Rm G , Ib G , Qu G ), as the input of the prediction model, and then the repaired crack apparent index CR obtained by S1 H =(L H , W H , D H , LOC H ) as the output of the prediction model, and perform prediction model training; The test set was used for cross-validation with the prediction results of the gravity dam crack expansion artificial intelligence model; by reasonably adjusting the model parameters, the model performance was further optimized to improve the accuracy and generalization ability of the model.
6. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 5, characterized in that: In step S3, The objective function of the deep learning prediction model is: Where N is the number of samples used for training and validation of the prediction model; It is the prediction index of the repaired crack obtained by the prediction model through the fusion model of the three-dimensional finite element model and the LSTM network; CR Hi is the real data indicator of crack development after repair obtained in step S1; is a penalty term that considers the impact of the gravity dam's comprehensive load on the crack prediction results after repair. λ1 is a hyperparameter used to adjust the weight of the penalty term related to the gravity dam load. m is the gravity dam's comprehensive load LOAD Q The number of parameters contained in; is a penalty term that considers the impact of the original cracks before repair on the crack prediction results after repair, λ2 is a hyperparameter used to adjust the weight of the penalty term related to the original cracks before repair, and n is the number of parameters included in the original crack index CRQ before repair; In order to consider the penalty term of the influence of the performance parameters of the grouting repair material on the prediction results of the cracks after repair, λ3 is a hyperparameter used to adjust the weight of the penalty term related to the performance of the grouting repair material, and l is the number of parameters included in the performance index CPG of the grouting repair material.
7. The method for predicting the development of cracks in concrete gravity dams after repair based on deep learning and finite element analysis according to claim 6, characterized in that: Step S4 is specifically as follows: The multi-period crack expansion data and multi-operating condition parameters collected by drones are input into the latest three-dimensional finite element model and artificial intelligence-based gravity dam crack development prediction model to predict the future development trend of gravity dam concrete cracks. The prediction content covers the crack location, expansion length, expansion width and expansion direction.
8. A system for predicting the development of cracks in a concrete gravity dam after repair based on deep learning and finite elements, used in a method for predicting the development of cracks in a concrete gravity dam after repair based on deep learning and finite elements as claimed in any one of claims 1 to 7, characterized in that: It includes a data collection unit: UAV collects apparent monitoring data of gravity dam cracks before and after repair; Model building unit: construct three-dimensional finite element models of concrete gravity dam cracks before and after repair, and inversely obtain the comprehensive load of the gravity dam and the comprehensive index parameters of the grouting repair material; Model training unit: build a deep learning-based prediction model for the development of gravity dam cracks, and conduct model training and cross-validation; Trend prediction unit: predict the subsequent crack development trend through the development prediction model after the gravity dam crack is repaired.
9. An electronic device, characterized in that: The method comprises at least one processor and at least one memory in communication with the processor; wherein the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to execute the method according to any one of claims 1 to 7.