Method and equipment for generating reinforcing scheme for unfavorable structural surface of arch dam abutment and storage medium
Through machine learning methods, the reinforcement solution generation model is constructed, and the adverse structural reinforcement solutions are automatically analyzed and optimized, which solves the problems of high cost and difficult to guarantee quality in traditional technologies, and achieves efficient and economical reinforcement solutions generation.
Patent Information
- Application Number
- CN202411114031.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-14
AI Technical Summary
The traditional adverse structural surface reinforcement scheme has high cost, difficult quality, and requires manual calculation and optimization, which is inefficient and prone to errors.
Using a machine learning-based method, by building sample data sets and reinforcement scheme generation models, using ResNet50 model for training, we generate automated adverse structural surface reinforcement schemes, including training stages and application stages, automatic analysis and optimization of reinforcement schemes.
The reinforcement plan is economical, high quality and high efficiency, reducing manual errors, and improving the applicability and engineering feasibility of the reinforcement plan.
Smart Images

Figure CN118965901A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of water conservancy and hydropower engineering, and in particular relates to a method, device and storage medium for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder based on machine learning. Background Art
[0002] Various geological structural surfaces in geotechnical engineering include: various fracture surfaces (such as cleavage, joints, fault surfaces, etc.), material differentiation surfaces (bedding, bedding, sedimentary discontinuities, foliation, etc.), as well as weak interlayers or weak zones, tectonic rocks, muddy central layers, filling interlayers, etc. (For details, please refer to the literature: Chen Gang. Combination of unfavorable structural surfaces and prevention and control measures of surrounding rock of Houyuntaishan Tunnel [J]. Theoretical Research on Urban Construction (Electronic Edition), 2012 (8).). The mechanical behavior of rock mass is largely affected by the mechanical properties of structural surfaces, and its destruction is often the result of structural surface expansion and compounding (For details, please refer to the literature: Wu Youlun, Wang Yuanhan, Gu Desheng. Experimental study on anchor rods to inhibit the expansion of air-side structural surfaces [J]. Yanshilixueyugongchengxuebao, 2006, 25 (z1): 3046-3050.). Unfavorable structural surfaces refer to weak parts in the structure, which are easily damaged by external forces. Hydropower development in southwest my country is mainly distributed in the surrounding areas of the Qinghai-Tibet Plateau (especially in the southwest and northwest regions). On the one hand, the surrounding river valleys such as the Lancang River and the Jinsha River are rapidly cutting down, forming the high mountain canyon landform landscape of the Hengduan Mountains on the east side of the plateau, and the slopes on both sides of the river valley are high and steep; on the other hand, the emergence of high-altitude stress environment and the frequent new activities of surrounding fault systems such as the Longmenshan Fault make the dynamic conditions in the crust abnormally active; at the same time, large-scale unloading shallow surface time-effect deformation structures caused by the strong downcutting of river valleys in high-altitude stress areas have also been discovered in large quantities (for details, please refer to the literature: Huang Runqiu. On the engineering geological problems and research countermeasures of hydropower development in southwest China [J]. Dizhizaihaiyuhuanjingbaohu, 2002, 13 (1): 1-5.). It can be seen that the influence of complex structural surfaces of slope rock masses cannot be ignored in the current and near future hydropower development in my country.
[0003] In order to deal with complex structural surfaces to ensure the safety of engineering construction and operation, it is necessary to identify and deal with unfavorable structural surfaces. The engineering community has a relatively mature process for dealing with unfavorable structural surfaces, which mainly includes four main links: "exploration-unfavorable structural surface identification-reinforcement scheme comparison-construction reinforcement". For the comparison of reinforcement schemes, the current mainstream method is to conduct it from four ways: mechanical mechanism, numerical simulation, indoor test and field test (for details, please refer to the literature: Wu Shangjie, Gao Xiaoming, Jiang Yujing. Research Status and Prospect of Rock Slope Anchoring Technology [J]. Roadbed Engineering, 2014 (4): 1-6.). The selected reinforcement schemes mainly include anchor rods, grouting, anchor rod-grouting combination and derivative schemes based on the above reinforcement methods. Due to the high cost of indoor and field tests, and the research on mechanical mechanisms is often more biased towards basic application scenarios, in the face of large-scale complex unfavorable structural surface simulation and reinforcement scheme comparison, numerical simulation is more economically feasible and the results are more intuitive.
[0004] The general process of numerical simulation research is: mainly carry out geological description and analysis of the unfavorable structural surface encountered in a certain project, establish a numerical model, propose several reinforcement schemes and conduct simulation analysis, and obtain the optimal scheme through evaluation indicators such as stress and displacement (for details, please refer to the literature: Zhang Yongchuang. Numerical simulation study on influencing factors of slope stability and reinforcement scheme selection [J]. Water Science and Engineering Technology, 2021(1):69-74. and Fang Zhichun, Zhu Zhengguo, Ma Chaoyi et al. Comparison of advanced reinforcement schemes for karst tunnels [J]. Railway Construction, 2020, 60(10):65-67.).
[0005] The above unfavorable structural surface reinforcement method has the following problems:
[0006] (1) In conventional numerical simulation studies on the selection of unfavorable structural surface reinforcement schemes, the results generally show that the combined structural surface reinforcement scheme is more effective, forming a fixed paradigm, which leads to increasingly complex reinforcement schemes and higher costs in modern engineering technology.
[0007] (2) In the selection of reinforcement methods, either different reinforcement methods are compared, or a single reinforcement measure is compared by controlling some variables (such as anchor length, layout density, etc.). There is a lack of unified scheme comparison principles, making it difficult to ensure the quality of the reinforcement scheme.
[0008] (3) Traditional reinforcement schemes often require structural engineers to perform manual calculations and optimization, which is time-consuming, labor-intensive and prone to errors.
[0009] Therefore, a method for automatically calculating the reinforcement scheme for unfavorable structural surfaces is needed. Summary of the invention
[0010] The purpose of the present invention is to provide a method, device and storage medium for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder, so as to solve the problems that traditional technologies have high costs, the quality of reinforcement schemes cannot be guaranteed, manual calculation and optimization are required, and the efficiency is low and errors are prone to occur.
[0011] The present invention solves the above technical problems through the following technical solutions: a method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment, the generation method comprising a training phase and an application phase; the training phase comprises the following steps:
[0012] Constructing a sample data set, wherein the sample data set includes a plurality of sample data, each sample data includes an input vector and a target output vector, wherein the input vector is a vector composed of lithology parameters, confining pressure, regularized stress distribution before reinforcement, and safety factor before reinforcement, and the target output vector is a vector composed of regularized stress distribution after reinforcement, and the type and quantity of reinforcement schemes;
[0013] Constructing a reinforcement scheme generation model, and training the reinforcement scheme generation model using the sample data set to obtain a target model; a loss function in the training process is constructed according to a target price and a predicted price, wherein the target price is calculated according to the type and quantity of reinforcement schemes in the target output vector, and the predicted price is calculated according to the type and quantity of reinforcement schemes output by the reinforcement scheme generation model;
[0014] The application phase includes the following steps:
[0015] Obtain the rock parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced;
[0016] The lithology parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced are input into the trained target model for analysis and processing to obtain the regularized stress distribution after reinforcement and reinforcement scheme of the area to be reinforced.
[0017] Furthermore, the constructing of the sample data set specifically includes the following steps:
[0018] Constructing a finite element basic unit including a rock mass structural surface, wherein the rock mass structural surface is filled with lithology parameters;
[0019] Apply stress to the finite element foundation unit and perform finite element calculation to obtain the stress distribution result of the finite element foundation unit before reinforcement; perform stability analysis on the finite element foundation unit to obtain the safety factor and the position of the unfavorable structural surface before reinforcement of the finite element foundation unit;
[0020] Applying a reinforcement scheme at the unfavorable structural surface position of the finite element foundation unit, and performing finite element calculation to obtain a stress distribution result after the finite element foundation unit is reinforced;
[0021] The stress distribution results before reinforcement and the stress distribution results after reinforcement are respectively regularized to obtain the regularized stress distribution before reinforcement and the regularized stress distribution after reinforcement;
[0022] A sample data set is constructed according to the lithological parameters, confining pressure, regularized stress distribution before reinforcement, regularized stress distribution after reinforcement, safety factor before reinforcement and reinforcement scheme of the finite element basic unit.
[0023] Furthermore, the construction of the finite element basic unit including the rock mass structural surface specifically comprises the following steps:
[0024] Construct two-dimensional finite element elements;
[0025] generating a rock mass structural surface within the two-dimensional finite element unit;
[0026] The lithology parameters of different types of rock masses are obtained, and the rock mass structural surfaces in the two-dimensional finite element unit are filled with the lithology parameters to obtain a finite element basic unit containing the rock mass structural surface.
[0027] Furthermore, generating a rock mass structural surface in the two-dimensional finite element unit specifically comprises the following steps:
[0028] Step a1: reading all boundaries of the two-dimensional finite element unit, forming a boundary set from the points on each boundary, and forming an endpoint set from the intersection points of the boundaries of the two-dimensional finite element unit;
[0029] Step a2: randomly selecting point A and point B from the boundary set, and point A and point B are located on two different boundaries respectively;
[0030] Step a3: Calculate the distance between point A and each intersection point in the endpoint set, and calculate the distance between point B and each intersection point in the endpoint set;
[0031] Step a4: Determine whether the distances obtained in step a3 are all greater than the set threshold value. If so, proceed to step a5; if not, proceed to step a2;
[0032] Step a5: remove point A and point B from the boundary set and add them to the endpoint set, and at the same time add the points on the line connecting point A and point B to the boundary set;
[0033] Step a6: Repeat steps a2 to a5 multiple times to obtain a two-dimensional finite element unit containing the rock mass structural surface.
[0034] Furthermore, the strength reduction method is used to perform stability analysis on the finite element foundation unit.
[0035] Furthermore, the specific formula of the regularization process is:
[0036]
[0037] in, represents the horizontal component of the regularized stress distribution, τ x represents the horizontal component of the unregularized stress distribution, σ x represents the horizontal component of the confining pressure, σ y represents the vertical component of the confining pressure, represents the vertical component of the regularized stress distribution, τ y represents the vertical component of the unregularized stress distribution.
[0038] Furthermore, the reinforcement scheme generation model adopts a ResNet50 model, and the activation function of the ResNet50 model is:
[0039]
[0040] Among them, x represents the input of the activation function.
[0041] Furthermore, the application stage also includes a safety evaluation step for the reinforcement scheme of the area to be reinforced, specifically including:
[0042] Perform stability analysis on the finite element foundation unit with reinforcement scheme in the specific steps of constructing the sample data set, and obtain the safety factor of the finite element foundation unit after reinforcement;
[0043] Calculate the difference between the stress distribution results before reinforcement and after reinforcement;
[0044] Fitting the confining pressure of the finite element foundation unit, the difference of the stress distribution results before and after reinforcement, the safety factor before reinforcement, and the safety factor after reinforcement to obtain a nonlinear relationship between the safety factor after reinforcement and the confining pressure, the difference of the stress distribution results before and after reinforcement, and the safety factor before reinforcement;
[0045] The safety factor of the reinforcement scheme of the area to be reinforced is calculated using the nonlinear relationship, and the safety of the reinforcement scheme of the area to be reinforced is evaluated based on the safety factor.
[0046] Furthermore, the nonlinear relationship is:
[0047]
[0048] Among them, S' represents the safety factor after reinforcement, S represents the safety factor before reinforcement, σ represents the confining pressure, △τ represents the difference in stress distribution results before and after reinforcement, and k1 and k2 represent fitting coefficients.
[0049] Based on the same concept, the present invention also provides a terminal device, the terminal device comprising:
[0050] Memory for storing computer programs;
[0051] A processor is used to implement the above-mentioned method for generating a reinforcement plan for an unfavorable structural surface of an arch dam abutment when executing the computer program.
[0052] Based on the same concept, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder as described above is implemented.
[0053] Beneficial Effects
[0054] Compared with the prior art, the advantages of the present invention are:
[0055] The method for generating a reinforcement scheme for an unfavorable structural surface provided by the present invention takes into account the actual construction cost of the reinforcement scheme, ensures the economy of the reinforcement scheme, and is more engineering feasible; the present invention fully considers different reinforcement schemes and adjustable variables of a single reinforcement scheme, as well as different load conditions, improves the quality of the reinforcement scheme, and has a wider applicability of the effect; the present invention can be implemented using computer software, realizes the automatic generation of the reinforcement scheme for the unfavorable structural surface, is convenient and practical, and is easy to promote and apply. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0057] Figure 1 is a flow chart of a method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment in an embodiment of the present invention;
[0058] Figure 2 It is the improved residual network unit architecture in the embodiment of the present invention. DETAILED DESCRIPTION
[0059] The following is a clear and complete description of the technical solution in the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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.
[0060] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.
[0061] like Figure 1 As shown, a method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder provided by an embodiment of the present invention includes a training phase and an application phase. The training phase mainly trains a reinforcement scheme generation model to obtain a target model; the application phase is the specific application of the target model, that is, using the target model to generate an actual reinforcement scheme.
[0062] The training phase specifically includes the following steps:
[0063] Step S11: construct a sample data set;
[0064] Step S12: construct a reinforcement solution generation model, and use the sample data set in step S11 to train the reinforcement solution generation model to obtain a target model.
[0065] In a specific embodiment of the present invention, the specific steps of constructing a sample data set include:
[0066] Step S11.1: constructing a finite element basic unit including a rock mass structural surface, wherein the rock mass structural surface is filled with lithology parameters;
[0067] Step S11.2: applying stress to the finite element foundation unit in step S11.1, and performing finite element calculation to obtain the stress distribution result of the finite element foundation unit before reinforcement; performing stability analysis on the finite element foundation unit to obtain the safety factor and the position of the unfavorable structural surface before reinforcement of the finite element foundation unit;
[0068] Step S11.3: applying a reinforcement scheme to the unfavorable structural surface position of the finite element foundation unit, and performing finite element calculation to obtain the stress distribution result of the finite element foundation unit after reinforcement; performing stability analysis on the finite element foundation unit to which the reinforcement scheme is applied to obtain the safety factor of the finite element foundation unit after reinforcement;
[0069] Step S11.4: Regularizing the stress distribution results before reinforcement and the stress distribution results after reinforcement respectively to obtain regularized stress distribution before reinforcement and regularized stress distribution after reinforcement;
[0070] Step S11.5: Construct a sample data set based on the lithological parameters of the finite element basic unit (including Poisson's ratio, elastic modulus, density, cohesion, friction angle and the coordinate position of the group of lithological parameters in the finite element basic unit), confining pressure, regularized stress distribution before reinforcement, regularized stress distribution after reinforcement, safety factor before reinforcement and reinforcement scheme.
[0071] The sample data set includes multiple sample data, each sample data includes an input vector and a target output vector. The input vector is a vector composed of lithology parameters, confining pressure, regularized stress distribution before reinforcement, and safety factor before reinforcement. The target output vector is a vector composed of regularized stress distribution after reinforcement, the type and quantity of reinforcement schemes.
[0072] In step S11.1, a finite element basic unit including a rock mass structural surface is constructed, which specifically includes the following steps:
[0073] Use finite element analysis software to construct a two-dimensional finite element unit (for example, a 30m*30m rectangular unit); generate a rock structure surface in the two-dimensional finite element unit; obtain the lithology parameters of different types of rock masses, and use the lithology parameters to fill the rock structure surface in the two-dimensional finite element unit to obtain a finite element basic unit containing the rock structure surface.
[0074] In this embodiment, generating a rock mass structural surface in a two-dimensional finite element unit specifically includes the following steps:
[0075] Step a1: Read all boundaries of the two-dimensional finite element unit, use a linear function to record the boundary coordinates, form a boundary set from the points on each boundary, and form an endpoint set from the intersection of the boundaries of the two-dimensional finite element unit. For a 30m*30m rectangular unit, the boundary set includes the points on the four sides of the rectangular unit, and the endpoint set includes the four vertices of the rectangular unit.
[0076] Step a2: randomly select point A and point B from the boundary set, and point A and point B are located on two different boundaries (for example, on two different sides of a rectangular unit).
[0077] Step a3: Calculate the distance between point A and each intersection point in the endpoint set, and calculate the distance between point B and each intersection point in the endpoint set.
[0078] Step a4: Determine whether the distances obtained in step a3 are all greater than the set threshold. If so, proceed to step a5; if not, proceed to step a2.
[0079] Step a5: Remove point A and point B from the boundary set and add them to the endpoint set (i.e., the endpoint set is changed from the original four vertices to 6 points). At the same time, a linear function is used to record the coordinates of the line connecting point A and point B (i.e., the new boundary), and the points on the line connecting point A and point B are added to the boundary set.
[0080] Step a6: Repeat steps a2 to a5 multiple times to obtain a two-dimensional finite element unit containing the rock mass structural surface.
[0081] In this embodiment, for a 30m*30m rectangular unit, the threshold is set to 1m. In order to avoid the generated rock mass structural surface being a deformed unit, which causes the problem of difficult convergence, the distance between the endpoints of the boundary (such as point A and point B) and the initial endpoint (i.e., the four vertices of the rectangular unit) needs to be greater than the set threshold to avoid the distance between the generated boundary endpoint and the initial endpoint being too close.
[0082] Based on the rock mass types and lithology parameters known in the project, the lithology parameters of different types of rock masses are obtained, including Poisson's ratio, elastic modulus, density, cohesion and friction angle. Based on the boundary of the rock mass structural surface, the rock mass structural surface in the two-dimensional finite element unit is filled with lithology parameters to obtain the finite element basic unit containing the rock mass structural surface.
[0083] In step S11.2, a stress combination of 0.1MPa to 100MPa is applied to the finite element foundation unit in the vertical and horizontal directions. For example, a stress combination scheme is composed of a 0.1MPa stress applied vertically and a 100MPa stress applied horizontally, and a stress combination scheme is composed of a 1MPa stress applied vertically and a 90MPa stress applied horizontally, ... Different stress combinations can be applied to the finite element foundation unit, that is, different load conditions are considered to improve the quality of the reinforcement scheme.
[0084] In both step S11.2 and step S11.3, the strength reduction method is used for stability analysis, that is, the safety assessment of the unfavorable structural surface is calculated using the strength reduction method. The specific calculation formula is:
[0085]
[0086] Among them, c, They represent the cohesion and friction angle of the unfavorable structural surface, c', They represent the cohesion and friction angle of the unfavorable structural surface after strength reduction, F s Represents the reduction factor.
[0087] In order to evaluate the safety of the reinforcement scheme output by the target model in the application phase, it is also necessary to obtain the calculation formula of the safety factor after reinforcement. In this embodiment, the confining pressure in step S11.2 and step S11.3 (i.e., the combination of vertical and horizontal stresses in step S11.2), the stress distribution results before reinforcement, the stress distribution results after reinforcement, the safety factor before reinforcement, and the safety factor after reinforcement are fitted to obtain the calculation formula of the safety factor after reinforcement. The specific process is:
[0088] The difference between the stress distribution results before reinforcement and the stress distribution results after reinforcement is calculated; the confining pressure of the finite element foundation unit, the difference between the stress distribution results before and after reinforcement, the safety factor before reinforcement and the safety factor after reinforcement are fitted to obtain the nonlinear relationship between the safety factor after reinforcement and the confining pressure, the difference between the stress distribution results before and after reinforcement and the safety factor before reinforcement, that is, the calculation formula of the safety factor after reinforcement, which is specifically:
[0089]
[0090] Among them, S' represents the safety factor after reinforcement, S represents the safety factor before reinforcement, σ represents the confining pressure, △τ represents the difference in stress distribution results before and after reinforcement, and k1 and k2 represent fitting coefficients.
[0091] Since some values of the stress distribution results are large and some are small, it is difficult to converge. Therefore, in step S11.4, the stress distribution results before reinforcement and the stress distribution results after reinforcement are respectively regularized to ensure that the data are in the same order of magnitude. In this embodiment, the regularization processing formula is:
[0092]
[0093] in, represents the horizontal component of the regularized stress distribution, τ x represents the horizontal component of the unregularized stress distribution, σ x represents the horizontal component of the confining pressure, σ y represents the vertical component of the confining pressure, represents the vertical component of the regularized stress distribution, τ y represents the vertical component of the unregularized stress distribution.
[0094] In order to construct rich sample data, different reinforcement schemes are applied to the unfavorable structural surface positions of the finite element basic unit in step S11.3, that is, sample data corresponding to each reinforcement scheme is obtained. Different reinforcement schemes and adjustable variables of a single reinforcement scheme (such as the type and quantity of reinforcement means) are fully considered to improve the quality of the reinforcement scheme.
[0095] In order to ensure the economy and safety of the reinforcement scheme, the economic calculation of the reinforcement scheme in the sample data set is performed, that is, the price of the reinforcement scheme is calculated, and the data corresponding to the uneconomical reinforcement scheme is eliminated; the data corresponding to the safety factor after reinforcement of the finite element foundation unit that does not meet the safety requirements is eliminated.
[0096] Since the ResNet50 model performs well in image classification and object detection tasks, the residual learning of the ResNet50 model can effectively solve the gradient vanishing problem, making the model easier to train. Therefore, in step S12, the ResNet50 model is selected as the reinforcement solution generation model. The commonly used activation function in the traditional ResNet50 model is ReLU, and the ReLU function is defined as:
[0097] F(x)=max(0,x) (6)
[0098] Among them, max() means setting the part of the input value x less than 0 to 0, and the part greater than or equal to 0 remains unchanged. Since the gradient of the ReLU function is 0 when x<0, some neurons of the ResNet50 model will no longer be updated, which is not conducive to model training. Therefore, in order to apply the ResNet50 model to the reinforcement of unfavorable structural surfaces, the activation function of the ResNet50 model is changed to the ELU function, which is specifically defined as:
[0099]
[0100] The residual network units in the improved ResNet50 model are as follows: Figure 2 shown.
[0101] Before training the ResNet50 model, the ResNet50 model was initialized with an input vector of 128*128, a learning rate of 0.1, a batch size of 16, a regularization parameter of 0.01, and a number of training rounds of 150.
[0102] In this embodiment, the sample data set is divided into a training set and a test set at a ratio of 9:1, and the "lithology parameters-confining pressure-stress distribution before regularization reinforcement-safety factor before reinforcement" in the training set is used as the input of the ResNet50 model, and the "stress distribution after regularization reinforcement-type and quantity of reinforcement schemes" is used as the target output of the ResNet50 model. During the training process, the loss function is constructed according to the target price and the predicted price, the target price is calculated according to the type and number of reinforcement schemes in the target output vector, and the predicted price is calculated according to the type and number of reinforcement schemes output by the reinforcement scheme generation model. That is, the predicted price (i.e., the predicted reinforcement scheme price) is calculated according to the type and number of reinforcement schemes output by the ResNet50 model, the target price (i.e., the reinforcement scheme price in the training set) is calculated according to the type and number of corresponding reinforcement schemes in the training set, and the predicted price and target price are substituted into the loss function to calculate the loss value to adjust the ResNet50 model and realize the training of the ResNet50 model.
[0103] The price of the reinforcement plan is equal to the sum of the costs of various reinforcement methods. The cost calculation formula for a single reinforcement method is: cost = (unit price (factors affecting the unit price) + labor costs) × volume. The commonly used reinforcement methods for unfavorable structural surfaces are anchor rods and grouting. The cost calculation formula for anchor rods is:
[0104] C1=[L*g1(D,T)+M C1 ]*N (8)
[0105] Among them, C1 is the cost of anchor reinforcement, g1(D,T) is the unit price of anchor, D is the diameter of anchor, L is the length of anchor, T is the special anchor technical fee, N is the number of anchors, M is the C1 Labor costs for anchor reinforcement.
[0106] The cost calculation formula for grouting is:
[0107] C2=[g2(K,H)+M C2 ]*V (9)
[0108] Among them, C2 is the cost of grouting reinforcement, g2(K,H) is the unit price of grouting, K is the type of grouting, H is the grouting depth, V is the grouting volume, M C2 Labor costs for grouting reinforcement.
[0109] After completing the learning and training of the ResNet50 model on the training set, the ResNet50 model is evaluated using the test set. If the accuracy of the ResNet50 model reaches more than 95%, the ResNet50 model training is considered complete and the target model is obtained.
[0110] The application phase includes the following steps:
[0111] Step S21: obtaining the lithology parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced;
[0112] Step S22: inputting the lithology parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced into the trained target model for analysis and processing to obtain the regularized stress distribution after reinforcement and reinforcement scheme of the area to be reinforced.
[0113] According to formula (3), the safety factor of the area to be reinforced after reinforcement can be calculated. According to formulas (8) and (9), the price of the reinforcement plan for the area to be reinforced output by the target model can be calculated. Thus, the regularized stress distribution after reinforcement of the area to be reinforced - the safety factor after reinforcement - the price of the reinforcement plan can be obtained, which is helpful to intuitively judge the advantages of the reinforcement plan.
[0114] In step S21, obtaining the lithology parameters, confining pressure, regularized stress distribution before reinforcement, and safety factor before reinforcement of the area to be reinforced specifically includes:
[0115] Conduct detailed surveys on the geological structure, lithology, and structural morphology of the area to be reinforced to obtain basic geological information on the slope. Integrate and process the data from field surveys and geological surveys to summarize modeling information such as the topography, geological structure, etc. of the area to be reinforced. Construct a finite element model of the area to be reinforced based on the modeling information.
[0116] The finite element model of the area to be reinforced is numerically simulated to obtain the stress field distribution results of the area to be reinforced before reinforcement, and then the confining pressure is obtained; the stress field distribution results before reinforcement are regularized (such as formulas (4) and (5)) to obtain the regularized stress distribution before reinforcement;
[0117] The strength reduction method is used to perform stability analysis on the finite element model of the area to be reinforced (such as formulas (1) and (2)), and the safety factor and the position of the unfavorable structural surface of the area to be reinforced before reinforcement are obtained.
[0118] The lithology parameters of the area to be reinforced, confining pressure, regularized stress distribution before reinforcement, safety factor before reinforcement, and position of the unfavorable structural surface are input into the target model to obtain the regularized stress distribution after reinforcement of the area to be reinforced, the types and number of reinforcement schemes, and thus the optimal reinforcement scheme for the unfavorable structural surface is obtained.
[0119] The present invention aims at different reinforcement schemes and adjustable variables in the reinforcement schemes to obtain the most economical scheme under the premise of ensuring safety.
[0120] Example 2
[0121] An embodiment of the present invention further provides a terminal device, which includes: a processor and a memory storing a computer program, wherein the processor is configured to implement the above-mentioned method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder when executing the computer program.
[0122] Although not shown, the terminal device includes a processor, which can perform various appropriate operations and processes according to the program and / or data stored in a read-only memory (ROM) or the program and / or data loaded from the storage part into a random access memory (RAM). The processor can be a multi-core processor or can include multiple processors. In some embodiments, the processor can include a general main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. In RAM, various programs and data required for the operation of the terminal device are also stored. The processor, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.
[0123] The processor and the memory are used together to execute the program stored in the memory. When the program is executed by the computer, the methods, steps or functions described in the above embodiments can be implemented.
[0124] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam shoulder as described above is implemented.
[0125] Storage media in embodiments of the present invention include permanent and non-permanent, removable and non-removable items that can be implemented by any method or technology to store information. Examples of storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.
[0126] What is disclosed above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, which should be covered within the protection scope of the present invention.
Claims
1. A method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment, characterized in that: The generation method comprises a training phase and an application phase; the training phase comprises the following steps: Constructing a sample data set, wherein the sample data set includes a plurality of sample data, each sample data includes an input vector and a target output vector, wherein the input vector is a vector composed of lithology parameters, confining pressure, regularized stress distribution before reinforcement, and safety factor before reinforcement, and the target output vector is a vector composed of regularized stress distribution after reinforcement, and the type and quantity of reinforcement schemes; Constructing a reinforcement scheme generation model, and training the reinforcement scheme generation model using the sample data set to obtain a target model; a loss function in the training process is constructed according to a target price and a predicted price, wherein the target price is calculated according to the type and quantity of reinforcement schemes in the target output vector, and the predicted price is calculated according to the type and quantity of reinforcement schemes output by the reinforcement scheme generation model; The application phase includes the following steps: Obtain the rock parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced; The lithology parameters, confining pressure, regularized stress distribution before reinforcement and safety factor before reinforcement of the area to be reinforced are input into the trained target model for analysis and processing to obtain the regularized stress distribution after reinforcement and reinforcement scheme of the area to be reinforced.
2. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to claim 1, characterized in that: The construction of the sample data set specifically includes the following steps: Constructing a finite element basic unit including a rock mass structural surface, wherein the rock mass structural surface is filled with lithology parameters; Apply stress to the finite element foundation unit and perform finite element calculation to obtain the stress distribution result of the finite element foundation unit before reinforcement; perform stability analysis on the finite element foundation unit to obtain the safety factor and the position of the unfavorable structural surface before reinforcement of the finite element foundation unit; Applying a reinforcement scheme at the unfavorable structural surface position of the finite element foundation unit, and performing finite element calculation to obtain a stress distribution result after the finite element foundation unit is reinforced; The stress distribution results before reinforcement and the stress distribution results after reinforcement are respectively regularized to obtain the regularized stress distribution before reinforcement and the regularized stress distribution after reinforcement; A sample data set is constructed according to the lithological parameters, confining pressure, regularized stress distribution before reinforcement, regularized stress distribution after reinforcement, safety factor before reinforcement and reinforcement scheme of the finite element basic unit.
3. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to claim 2, characterized in that: The construction of the finite element basic unit including the rock mass structural surface specifically comprises the following steps: Construct two-dimensional finite element elements; generating a rock mass structural surface in the two-dimensional finite element unit; The lithology parameters of different types of rock masses are obtained, and the rock mass structural surfaces in the two-dimensional finite element unit are filled with the lithology parameters to obtain a finite element basic unit containing the rock mass structural surface.
4. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to claim 3 is characterized in that: Generating a rock mass structural surface in the two-dimensional finite element unit specifically comprises the following steps: Step a1: reading all boundaries of the two-dimensional finite element unit, forming a boundary set from the points on each boundary, and forming an endpoint set from the intersection points of the boundaries of the two-dimensional finite element unit; Step a2: randomly selecting point A and point B from the boundary set, and point A and point B are located on two different boundaries respectively; Step a3: Calculate the distance between point A and each intersection point in the endpoint set, and calculate the distance between point B and each intersection point in the endpoint set; Step a4: Determine whether the distances obtained in step a3 are all greater than the set threshold value. If so, proceed to step a5; if not, proceed to step a2; Step a5: remove point A and point B from the boundary set and add them to the endpoint set, and at the same time add the points on the line connecting point A and point B to the boundary set; Step a6: Repeat steps a2 to a5 multiple times to obtain a two-dimensional finite element unit containing the rock mass structural surface.
5. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to claim 2, characterized in that: The specific formula of the regularization process is: in, represents the horizontal component of the regularized stress distribution, τ x represents the horizontal component of the unregularized stress distribution, σ x represents the horizontal component of the confining pressure, σ y represents the vertical component of the confining pressure, represents the vertical component of the regularized stress distribution, τ y represents the vertical component of the unregularized stress distribution.
6. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to any one of claims 1 to 5, characterized in that: The reinforcement scheme generation model adopts the ResNet50 model, and the activation function of the ResNet50 model is: Among them, x represents the input of the activation function.
7. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to any one of claims 1 to 5, characterized in that: The application stage also includes a safety evaluation step of the reinforcement scheme for the area to be reinforced, specifically including: Perform stability analysis on the finite element foundation unit with reinforcement scheme in the specific steps of constructing the sample data set, and obtain the safety factor of the finite element foundation unit after reinforcement; Calculate the difference between the stress distribution results before reinforcement and after reinforcement; Fitting the confining pressure of the finite element foundation unit, the difference of the stress distribution results before and after reinforcement, the safety factor before reinforcement, and the safety factor after reinforcement to obtain a nonlinear relationship between the safety factor after reinforcement and the confining pressure, the difference of the stress distribution results before and after reinforcement, and the safety factor before reinforcement; The safety factor of the reinforcement scheme of the area to be reinforced is calculated using the nonlinear relationship, and the safety of the reinforcement scheme of the area to be reinforced is evaluated based on the safety factor.
8. The method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to claim 7, characterized in that: The nonlinear relationship is: Among them, S' represents the safety factor after reinforcement, S represents the safety factor before reinforcement, σ represents the confining pressure, △τ represents the difference in stress distribution results before and after reinforcement, and k1 and k2 represent fitting coefficients.
9. A terminal device, characterized in that: The terminal device comprises: Memory for storing computer programs; A processor is used to implement the method for generating a reinforcement plan for an unfavorable structural surface of an arch dam abutment as described in any one of claims 1 to 8 when executing the computer program.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating a reinforcement scheme for an unfavorable structural surface of an arch dam abutment according to any one of claims 1 to 8 is implemented.
Citation Information
Patent Citations
Stress-strain digital data processing method for reinforcing tunnel face of broken and weak surrounding rock
CN112800530A
Safety assessment method for danger-removing reinforced concrete gravity dam
CN116136941A
Pre-disintegration soft rock embankment structure based on space function partition and design method thereof
CN117349922A
Fiber grating sensor, strain monitoring method and system for a surrounding rock of a deep roadway
US20230296458A1
