Method for predicting valley amplitude deformation of water storage dam based on rock mass mechanical parameters or rock strata

Abstract

The invention discloses a method for predicting valley amplitude deformation of a water storage dam based on rock mass mechanical parameters, and the method comprises the following steps: S100, dividing a river valley into a plurality of sub-regions, endowing the sub-regions with a first number, and obtaining a first coordinate result; s200, rock mass samples are obtained through multiple times of rock mass sampling, and the rock mass mechanical parameter of each rock mass sample is a first parameter; s300, forming a first parameter set by each group of first parameters; s400, a plurality of sub-regions of the position where any valley amplitude observation section in the river valley is located are selected; s500A, reading the plurality of sub-regions selected in the step S400 to establish a first vector set; s600, obtaining a total water head distribution field in the river valley; and S700A, inputting the first vector set and the local water head value into a pre-trained first prediction model, and outputting a second coordinate result. The valley amplitude deformation of the dam is predicted through a plurality of models, and a prediction result obtained based on factors such as rock mass mechanical parameters and coordinate positions is more reliable.

Description

Technical Field

[0001] The invention belongs to the field of predicting valley deformation based on computer technology, and particularly relates to a method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters or rock strata. Background Art

[0002] Valley deformation refers to the horizontal deformation (i.e. horizontal displacement) of the reservoir bank (here, the reservoir bank refers to the river valley slope) in a large range upstream and downstream of the dam caused by reservoir water storage under specific geological conditions. It is the manifestation of regional deformation in the horizontal direction, which is equal to the change in the shortest horizontal distance between the same elevation points on the left and right banks of the reservoir (river valley). Usually, the observer will select a valley observation section to measure the valley deformation.

[0003] Valley deformation is different from the common valley slope erosion that causes slope collapse, resulting in sediment or rock and soil settling at the bottom of the valley and causing valley deformation. It is mainly determined by geological conditions and reservoir water storage, and the stress causes the internal valley width structure to change.

[0004] At present, direct measurement is usually used to measure valley deformation. The valley width at the time of measurement is measured, and then whether more severe deformation will occur is estimated based on experience or calculation. The prediction results obtained in this way are not accurate enough and have insufficient reference value. If the estimation of valley deformation is wrongly judged, it may lead to a series of serious consequences such as dam cracking.

[0005] Therefore, a method is needed to assist observers in predicting valley deformation at future moments. Summary of the invention

[0006] In view of the above problems, the purpose of the present invention is to disclose a method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters.

[0007] Another object of the present invention is to disclose a method for predicting valley deformation of a water storage dam based on rock formations.

[0008] According to a method for predicting valley deformation of a water storage dam using rock mechanical parameters of the present invention, the water storage dam is constructed on any cross section of a river valley, and comprises the following steps: S100, dividing the river valley into a plurality of sub-areas, assigning a first number to each of the sub-areas, and obtaining the coordinates of each of the sub-areas at an initial moment as a first coordinate result; S200, performing multiple rock sampling on each of the sub-areas to obtain a group of rock samples, obtaining the rock mechanical parameters of each of the rock samples as a first parameter, and obtaining the coordinates of each of the rock samples as a sub-coordinate; S300, a group of the rock samples collected in each of the sub-areas corresponds to a group of the first parameters, and each group of the first parameters is formed into a first parameter set; S400, selecting a plurality of the sub-areas at the location of any valley width observation section in the river valley, and reading the selected plurality of the sub-areas for the first parameter set. The first parameters corresponding to the plurality of sub-areas selected in S400 are read; S500A, the first coordinate results corresponding to the plurality of sub-areas selected in S400 are read, and a first vector set is established according to the first number, wherein the first vector set includes a plurality of first sub-vectors, and each first sub-vector includes: the first number corresponding to the sub-area; the first parameters and sub-coordinates corresponding to a rock sample; S600, the total head distribution field in the river valley is obtained, and the total head at the coordinates of each sub-area is retrieved as the local head value according to the total head distribution field; S700A, the first vector set and the local head value are input into the pre-trained first prediction model; the first prediction model outputs the coordinates of each sub-area at the prediction moment as the second coordinate results, and the second coordinate results include a plurality of new sub-coordinates corresponding to the coordinates of each rock sample at the prediction moment.

[0009] Optionally, S200 includes: S210, collecting multiple rock samples in sequence along the depth direction of the valley slope of each sub-area, and assigning a second number to each rock sample in sequence according to the depth of each rock sample; S220, obtaining the rock mechanical parameters of each rock sample as the first parameter.

[0010] Furthermore, the first vector set obtained in S500A is stored, and the first vector set in the stored data is read when S600 to S700 are executed.

[0011] Furthermore, the method further includes S710 performed after S700A: S710, obtaining the second coordinate result, selecting the second coordinate result corresponding to the valley width observation section; and obtaining the valley width deformation result by comparing the vector difference between the first coordinate result and the second coordinate result.

[0012] Furthermore, the S710 includes: S711, matching the first coordinate result and the second coordinate result according to the first number, and matching the sub-coordinate with the new sub-coordinate; S712, obtaining the vector difference between each pair of the sub-coordinates and the new sub-coordinates as a single vector; S713, reading the rock layer division in each of the sub-areas and the position in the total head distribution field, and assigning a first weight coefficient to a number of the single vectors corresponding to each of the sub-areas; S714, performing vector summation on each of the single vectors to obtain a first change; S715, assigning the first weight coefficient to each of the first change, weighting each of the first change, and performing mathematical synthesis to obtain the overall displacement result of the valley width observation section; S716, obtaining the valley width deformation result at each valley width observation section based on the overall displacement result.

[0013] Optionally, the method further includes S800 performed after S710: S800, updating the first coordinate result, and completing the update after the second coordinate result covers the first coordinate result; and performing S100 again after the update is completed.

[0014] Optionally, the method further includes S610 performed after S600; S610, inputting the first vector set and the local water head value into a pre-trained third prediction model; the third prediction model outputs a corrected permeability coefficient value, and re-performing S310 based on the corrected permeability coefficient.

[0015] Optionally, S310 performed after S300 is also included:

[0016] S310, reading the second number, assigning a permeability coefficient value to each of the rock samples according to the arrangement order of the second number; dividing the sub-area into rock layers according to the permeability coefficient; the rock layer division includes permeable rock layers and general rock layers divided along the depth direction of the valley slope of the river valley.

[0017] Further, in S310: the permeable rock layer is a rock mass with a high permeability coefficient value, and the permeable rock layer includes: a fractured rock layer and a broken rock layer; the general rock layer is a rock mass with a low permeability coefficient value, and the general rock layer includes: granite and mudstone.

[0018] According to a method for predicting valley deformation of a water storage dam based on rock strata according to an embodiment of the present invention, according to the method for predicting valley deformation of a water storage dam based on rock mechanical parameters described in an embodiment of the present invention, S500A is replaced by S500B; S500B, reads the first coordinate results corresponding to the several sub-areas selected in S400, and establishes a second vector set according to the first number, wherein the second vector set includes several second sub-vectors, each second sub-vector includes: the first number corresponding to the sub-area; the first parameters and sub-coordinates corresponding to a rock sample and the result of rock strata division; S700 is replaced by S700B; S700B, inputs the second vector set and the local head value into a pre-trained second prediction model; the second prediction model outputs the coordinates of each of the sub-areas at the prediction moment as the second coordinate results, and the second coordinate results include multiple new sub-coordinates corresponding to the coordinates of each rock sample at the prediction moment.

[0019] The beneficial effects of the present invention are:

[0020] The valley deformation of the dam is predicted by integrating multiple models, and the prediction results are obtained based on factors such as rock mass mechanical parameters and coordinate positions, so that the final prediction results are more reliable. The model parameters involve multiple core parameters, so that the present invention can cover different types of data and various factors that may affect the valley deformation, so that it can be applied to the prediction of valley deformation of dams under different conditions, and has strong versatility.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 An exemplary valley width observation section structure according to an embodiment of the present invention is shown;

[0024] Figure 2 A schematic diagram for assisting understanding of a first vector set according to an embodiment of the present invention is shown;

[0025] Figure 3A flow chart of a method for predicting valley deformation of a water storage dam based on rock mass mechanics according to an embodiment of the present invention is shown;

[0026] Figure 4 A flow chart of a method for predicting valley deformation of a water storage dam based on rock formations according to an embodiment of the present invention is shown.

[0027] In the attached figure:

[0028] 1-sub-area, 2-river valley, 3-valley slope, 4-water level, 5-reservoir plate, 6-reservoir edge; DETAILED DESCRIPTION

[0029] 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.

[0030] The professional terms such as "river valley", "rock sampling", "valley width observation section", "total water head distribution field" described in the embodiments of the present application are well known to ordinary technicians in this field, and therefore will not be repeated in the embodiments of the present application.

[0031] Next, combine Figure 1-4 Understand the first aspect embodiment and the second aspect embodiment disclosed in this embodiment.

[0032] First aspect embodiment;

[0033] A method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters, wherein the water storage dam is constructed on any cross section of a river valley 2, comprises the following steps:

[0034] S100, dividing the river valley 2 into a plurality of sub-regions 1, assigning a first number to each of the sub-regions 1, obtaining the coordinates of each of the sub-regions 1 at the initial moment as the first coordinate result, and combining Figure 1 In a certain valley observation section shown in the figure, the area between the left and right reservoir edges 6 of the valley 2 and the area extending to both sides for a certain distance is the reservoir plate 5, and the valley 2 can be divided into several sub-areas 1 along the valley slope 3. Of course, there are also several sub-areas 1 in other observation sections; its purpose is to predict the deformation of different areas of the valley 2, and the corresponding first coordinate result and the first number can be used to locate different sub-areas 1.

[0035] S200, performing multiple rock sampling on each of the sub-areas 1 to obtain a group of rock samples, obtaining the rock mechanical parameters of each of the rock samples as the first parameters, and obtaining the coordinates of each of the rock samples as sub-coordinates; it can be understood that since the valley deformation mainly depends on the mechanical parameters of the rock sample and the head distribution at its location, it is necessary to perform multiple rock sampling on each of the sub-areas 1 to obtain the corresponding rock mechanical parameters.

[0036] S300, each group of the rock samples collected in the sub-area 1 corresponds to a group of the first parameters, and each group of the first parameters is combined into a first parameter set; in this embodiment, the first parameters include but are not limited to: elastic modulus, compressive strength, and the specific selected parameters can be selected by technicians in the field based on existing common knowledge.

[0037] S400, selecting a plurality of sub-areas 1 at the location of any valley width observation section in the valley 2, for example Figure 1 The valley width observation section shown reads the first parameters corresponding to the selected sub-areas 1.

[0038] It is understandable that valley deformation is affected by many factors, the main influencing factors include rock mechanical parameters and head distribution. By measuring these parameters at the actual section position, direct data of the rock mass at a specific location can be obtained. In this example, the river valley 2 is divided into several sub-areas 1, and the rock mechanical parameters are obtained for each sub-area 1 in order to capture the changes in the valley width at different positions. In the above example, step S400 selects any valley width observation section and reads the first parameter set of the relevant sub-area 1, then each observation section can be regarded as a representative area to obtain the rock mass response at this specific section.

[0039] And when the valley width is deformed, the deformation phenomenon of the mountain near the dam toward the riverbed in the water storage dam is called valley width deformation. Therefore, by selecting one or several specific valley width observation sections for analysis, the changes of the dam body in one or more valley width observation sections can be obtained to accurately obtain the result of valley width deformation.

[0040] S500A, reading the first coordinate results corresponding to the plurality of sub-areas 1 selected in S400, wherein the first sub-coordinate results include a plurality of sub-coordinates corresponding to each rock sample, and establishing a first vector set for each sub-area 1 according to the first serial number. The first vector set can be combined with Figure 2 It is understood that the first vector set includes a plurality of first sub-vectors, each of which includes:

[0041] The first number corresponding to the sub-area 1; the first parameter and sub-coordinate corresponding to a rock sample.

[0042] In a specific example, the first sub-vector V i for:.

[0043] V ia =(i,P ia1 , P ia2 ,…,P ian , X ia1 , Y ia1 , Y ia1 );

[0044] Wherein, i is the first number corresponding to the i-th sub-region 1, and a is the a-th rock sample corresponding to the i-th sub-region 1; (P ia1 , P ia2 ,…,P ian ) are the mechanical parameters in the first parameter set corresponding to the a-th rock sample in the i-th sub-region 1, a total of n rock mechanical parameters; (X ia1 , Y ia1 , Y ia1 ) is the sub-coordinate corresponding to the a-th rock sample in the i-th sub-region 1.

[0045] Similarly, the first vector set corresponding to the same sub-area 1 also includes:

[0046] V ib =(i,P ib1 , P ib2 ,…,P ibn , X ib1 , Y ib1 , Y ib1 );

[0047] Among them, b represents the bth rock sample corresponding to the i-th sub-region 1, and so on.

[0048] The first set of vectors is:

[0049] V i =(V ia， V ib，… );

[0050] In an optional example, the first vector set obtained in S500A is stored, and the first vector set in the stored data is read when executing S600-S700. It is worth mentioning that the conventional means of obtaining rock mechanical parameters all require sampling and analysis of the rock mass, and this method is generally only performed once or periodically. Therefore, in this example, when a rock sampling task is performed, the parameters obtained by rock sampling are stored, that is, the first vector obtained in S500A is stored after acquisition and retrieved in subsequent multiple prediction processes. Each time a prediction is made, the total head distribution field continues to change and needs to be measured in real time.

[0051] S600, obtaining a total water head distribution field in the river valley 2, and retrieving the total water head at the coordinates of each sub-area 1 as a local water head value according to the total water head distribution field.

[0052] In this example, the local water head value of sub-area 1 is obtained, and the first coordinate result and the first number obtained in steps S100-S500A can be used to locate the corresponding coordinates of different sub-areas 1.

[0053] When making predictions, a data set is obtained by measuring inside the river valley 2. Specifically, hydrogeological methods and tools can be used to measure the water pressure at each point in the river valley 2, usually by installing a water pressure sensor or using other hydrological measurement techniques. The specific detection process for obtaining the total water head distribution field is well known to those skilled in the art and will not be described in detail here.

[0054] Then extract the local head value, specifically, extract and calculate the head value at the coordinate of the specific sub-area 1. In this example, the coordinates of each sub-area 1 can be matched with the total head distribution field to obtain the local head value of each specific sub-area 1. The total head distribution field data is simulated in the form of a graph or table, and the local head value can be obtained by reading the total head value corresponding to the coordinate of the sub-area 1, or directly collecting the local head value of the corresponding sub-area 1.

[0055] S700A. Input the first vector set and the local head value into the pre-trained first prediction model; the first prediction model outputs the coordinates of each of the sub-areas 1 at the prediction moment as the second coordinate result, and the second coordinate result includes multiple new sub-coordinates corresponding to the coordinates of each of the rock samples at the prediction moment.

[0056] The new sub-coordinates are the coordinates predicted by the first prediction model after the valley deformation occurs. In this example, since the valley deformation is affected by multiple factors and the scene is rich and complex in geological data, the first model uses a gradient boosting model, which is a supervised learning model with regression function. The first model can map the relationship between the input parameters and the coordinate changes well, and is not easily affected by outliers, and does not require a large amount of data like a neural network to avoid overfitting.

[0057] Specifically, we selected various data corresponding to other water storage dams that have produced valley deformation in the past, and sorted them out based on the historical data obtained: the mechanical parameters and corresponding coordinates of each rock sample, as well as the head distribution data. We then used the historical data to train the model to learn the mapping relationship between the input parameters and the coordinate changes.

[0058] In one example, the first vector set and the local hydraulic head value are subjected to feature processing before being input into the model, specifically including scaling, normalization or encoding. Specifically, the rock mass mechanical parameters need to be normalized so as to have the same scale.

[0059] When training the model, the gradient enhancement model used is G(x), where x is a feature vector containing rock mass mechanical parameters and local head values. The model training process can be expressed as:

[0060]

[0061] Among them, J min (G) is the total loss function to be minimized, L is the loss function that calculates the difference between the model prediction and the actual observation, and y i It is the true coordinate corresponding to the i-th rock sample in the training parameters. It can be understood that the aforementioned true coordinate is the true coordinate selected in the parameters used to train the model, rather than the sub-coordinate in the aforementioned or subsequent examples; N is the number of training samples, and Ω is a regularization term used to penalize model complexity to prevent overfitting.

[0062] In order to evaluate the generalization ability of the model, the cross-validation method is used. The data is divided into K parts, and K-1 parts are used for training in turn, and the remaining part is used for validation. Model evaluation can select the best performing model parameters as the final model based on the results of cross-validation. The evaluation indicators include the determination coefficient R2, mean square error (MSE) or root mean square error (RMSE), and the specific process is not described here.

[0063] The trained model G is finally used to predict valley deformation under future or new conditions. That is, new rock mass mechanical parameters and local head values ​​are input to predict new coordinate points:

[0064]

[0065] in, is the new sub-coordinate output by prediction, V new is the first vector set at the current moment (the moment of prediction). It can be understood that V new As the first vector set at the prediction time, it can be unobserved and continuously input into the subsequent model update as an unobservable parameter. The user only needs to observe That is, the new sub-coordinates can obtain the corresponding valley amplitude change prediction results. For specific content, please refer to the subsequent examples.

[0066] Finally, the performance of the model is evaluated on an independent validation set, and necessary adjustments are made to obtain the first prediction model that has been trained.

[0067] According to the foregoing embodiment, in an optional example, the S200 includes:

[0068] S210, collecting rock samples for each sub-area 1 in sequence along the depth direction of the valley slope 3 of the river valley 2 for multiple times. In an example, reference may be made to Figure 1 It is understood that the direction of the valley slope 3 of the rock valley 2 is Figure 1 The arrow direction shown in . Each rock sample is assigned a second number in turn according to the depth of each rock sample; it can be understood that another factor affecting the valley deformation is the rock layer of the valley 2.

[0069] S220: Obtain a rock mass mechanical parameter of each of the rock mass samples as a first parameter.

[0070] It should be understood that the rock layer division also has an impact on the valley deformation. Therefore, in order to improve the accuracy of the valley deformation prediction, the rock layer division should be considered. The rock sample collection along the depth direction of the valley slope 3 of the river valley 2 in step S210 can refer to Figure 1 It is understood that in the embodiment of the present application, the pre-stored permeability coefficient value is read according to the first parameter, and the specific value corresponding to the numerical value can be known by technicians in the field based on the existing public content, and will not be described here.

[0071] In a further example, the rock formation is divided in S310 , and S310 is performed after S300 .

[0072] S310, reading the second number, and assigning a permeability coefficient value to each of the rock samples according to the arrangement order of the second number. It is not difficult to understand that the arrangement order of the second number actually means assigning a permeability coefficient to the rock in sub-area 1 in sequence according to the sampling depth of the corresponding sub-area 1.

[0073] The rock formations of the sub-area 1 are divided according to the permeability coefficients. As mentioned above, after the permeability coefficients are assigned according to the depth sequence indicated by the second number, the rock formations can be divided according to the permeability coefficients at different depths.

[0074] In a further example, the rock layer division includes permeable rock layers and general rock layers divided along the depth direction of the valley slope 3 of the valley 2 .

[0075] According to the above embodiment, after obtaining a number of new sub-coordinates, in a further embodiment, S710 is performed after S700A:

[0076] S710, obtaining the second coordinate result, selecting the second coordinate result corresponding to the valley width observation section; and obtaining the valley width deformation result by comparing the vector difference between the first coordinate result and the second coordinate result.

[0077] Specifically, the S710 includes:

[0078] S711. Match the first coordinate result and the second coordinate result according to the first number, and match the sub-coordinates with the new sub-coordinates; it can be understood that each of the sub-areas 1 corresponds to a set of sub-coordinates and new sub-coordinates, and the position change of the sub-area 1 can be obtained by comparing the new sub-coordinates with the sub-coordinates.

[0079] Continue to refer Figure 1 Understand S712.

[0080] S712, obtaining the vector difference between each pair of the sub-coordinates and the new sub-coordinates as a separate vector;

[0081] S713, reading the rock formation division in each of the sub-areas 1 and the position in the total hydraulic head distribution field, and assigning a first weight coefficient to a plurality of the individual vectors corresponding to each of the sub-areas 1;

[0082] S714, performing vector summation on each of the individual vectors to obtain a first variation;

[0083] S715, assigning the first weight coefficient to each of the first changes, weighting each of the first changes, and performing mathematical synthesis to obtain an overall displacement result of the valley width observation section;

[0084] S716. Obtain valley deformation results at each valley observation section according to the overall displacement results.

[0085] For example, in a specific scenario, take a sub-area 1 as an example:

[0086] First, perform S711 sub-region 1 coordinate matching to obtain two corresponding sub-coordinates:

[0087] The sub-coordinates at the initial moment are:

[0088] (X1, Y1, Z1);

[0089] The sub-coordinates at the prediction time are:

[0090] (X 1new , Y new , Z 1new );

[0091] Then, the vector difference calculation of S712 is performed. For each rock sample i in sub-area 1j, the calculated difference is:

[0092] Δr i，j =(X 1new，ij -X 1,ij , Y 1new，ij -Y 1，ij , Z 1new，ij -Z 1，ij );

[0093] Then, weight coefficients are assigned in S713. Each individual vector r ij A weight coefficient w is assigned according to its rock type and location ij Then, S714 is performed to sum all the individual vectors to obtain the first change ΔR j :

[0094]

[0095] Then, S715 is performed to weight the changes in all regions to obtain the overall displacement ΔR of the valley width observation section:

[0096]

[0097] Finally, S716 is performed to analyze the obtained overall displacement ΔR to obtain the final valley deformation result.

[0098] In a further example, the method further includes S800 performed after S710:

[0099] S800, updating the first coordinate result, and completing the update after the second coordinate result covers the first coordinate result; after the update is completed, repeat S100.

[0100] It is not difficult to understand that valley deformation will continue to occur over time, and the mechanical parameters of the rock mass will basically not change, but valley deformation will continue to occur. Therefore, in order to improve the accuracy of the prediction results, it is necessary to continuously update the coordinates of each sub-area 1. Specifically, as in S800, the result of the prediction at the current moment is used as the input value of the next prediction, that is, the update is completed after the second coordinate result covers the first coordinate result.

[0101] Second aspect embodiment;

[0102] This embodiment discloses a method for predicting valley deformation of a water storage dam based on rock strata, which is different from the method disclosed in the first aspect embodiment in that S500A is replaced by S500B, and S700A is replaced by S700B.

[0103] In this embodiment, the prediction of the new sub-coordinates also takes into account the rock stratum division, and therefore, the input content for predicting the valley deformation is the second vector including the rock stratum division.

[0104] In a specific example, S500B includes: reading the first coordinate results corresponding to the plurality of sub-areas 1 selected in S400, and establishing a second vector set for each of the sub-areas according to the first number, wherein the second vector set includes a plurality of second sub-vectors, and each second sub-vector includes:

[0105] The sub-region|corresponds to the first number; a first parameter and sub-coordinate corresponding to a rock sample and the result of rock layer division.

[0106] Exemplarily, in a specific example, the second sub-vector Vi′ is:

[0107] V′ ia =(i, P′ ia1 , P′ ia2 , ..., P′ ian , X′ ia , Y′ ia , Z′ ia , K′ i )

[0108] Wherein, i is the first number corresponding to the ith sub-region 1, and a is the ath rock sample corresponding to the ith sub-region 1; (P' ia1 , P' ia2 , …, P' ian ) are the mechanical parameters in the first parameter set corresponding to the a-th rock sample in the i-th sub-region 1, a total of n rock mechanical parameters; (X' ia , Y' ia , Z' ia′) is the sub-coordinate corresponding to the a-th rock sample in the i-th sub-region 1; K i ′ represents the corresponding rock stratum division result.

[0109] As described in the aforementioned first aspect embodiment, it is also necessary to perform S600 to obtain the total water head distribution field in the valley 2, and the specific content will not be repeated here.

[0110] As mentioned above, the training parameters of the prediction model in this embodiment are changed compared to the first embodiment, which leads to a fundamental change in the training model. It is worth mentioning that although the same model and training steps can be used, the first prediction model in this example is different from the second prediction model, and the specific training process will not be repeated.

[0111] Specifically, S700B includes: inputting the second vector set and the local head value into a pre-trained second prediction model; the second prediction model outputs the coordinates of each of the sub-areas 1 at the prediction moment as a second coordinate result, and the second coordinate result includes multiple new sub-coordinates corresponding to the coordinates of each of the rock samples at the prediction moment.

[0112] In this example, the gradient enhancement model is also used, but now the second input vector contains information related to rock strata division in addition to rock mass mechanical parameters and head values.

[0113] According to the method described in the first aspect embodiment or the second aspect embodiment, in an optional example, the method further includes S610 (not shown in the flowchart) performed after S600. In fact, although the valley deformation will not affect the mechanical parameters of the rock mass, it will change the permeability of the rock layer in the valley 2, so the permeability coefficient also needs to be updated, as described in S610.

[0114] S610, input the first vector set and the local water head value into the pre-trained third prediction model; the third prediction model is used to calculate or predict the new permeability after valley deformation or its influence on valley deformation prediction, specifically, the third prediction model uses machine learning methods to capture and predict the complex relationship between mechanical parameters, sub-coordinates and water head values ​​of rock samples. The third prediction model can introduce the influence of valley deformation on permeability, and the output of the third prediction model includes the updated permeability of each sub-area 1 at the prediction time.

[0115] In S310: the permeable rock layer is a rock mass with a high permeability coefficient value, and the permeable rock layer includes: a fractured rock layer and a broken rock layer; the general rock layer is a rock mass with a low permeability coefficient value, and the general rock layer includes: granite and mudstone.

[0116] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters, wherein the water storage dam is constructed on any cross section of a river valley, characterized in that: The steps include: S100, dividing the river valley into a plurality of sub-regions, assigning a first number to each of the sub-regions, and obtaining the coordinates of each of the sub-regions at an initial moment as a first coordinate result; S200, performing multiple rock sampling on each of the sub-areas to obtain a group of rock samples, obtaining a rock mechanical parameter of each of the rock samples as a first parameter, and obtaining a coordinate of each of the rock samples as a sub-coordinate; S300, a group of the rock samples collected in each sub-area corresponds to a group of the first parameters, and each group of the first parameters is combined into a first parameter set; S400, selecting a plurality of sub-areas at the location of any valley width observation section in the river valley, and reading a plurality of the first parameters corresponding to the selected plurality of sub-areas; S500A, reading the first coordinate results corresponding to the plurality of sub-regions selected in S400, and establishing a first vector set according to the first serial number, wherein the first vector set includes a plurality of first sub-vectors, each of which includes: The first number corresponding to the sub-region; The first parameter and sub-coordinate corresponding to a rock sample; A plurality of the first sub-vectors having the same first number constitute the first vector set; S600, obtaining a total water head distribution field in the river valley, and retrieving the total water head at the coordinates of each sub-area as a local water head value according to the total water head distribution field; S700A, inputting the first vector set and the local water head value into a pre-trained first prediction model; The first prediction model outputs the coordinates of each of the sub-areas at the prediction moment as a second coordinate result, and the second coordinate result includes a plurality of new sub-coordinates corresponding to the coordinates of each of the rock mass samples at the prediction moment.

2. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 1, characterized in that: The S200 includes: S210, collecting rock samples for multiple times in sequence along the depth direction of the valley slope of each sub-area, and assigning a second number to each rock sample in sequence according to the depth of each rock sample; S220: Obtain a rock mass mechanical parameter of each of the rock mass samples as a first parameter.

3. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 2, characterized in that: The first vector set obtained in S500A is stored, and the first vector set in the stored data is read when executing S600 to S700.

4. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 3 is characterized in that: Also included is the S710 performed after the S700A: S710, obtaining the second coordinate result, and selecting the second coordinate result corresponding to the valley width observation section; The valley deformation result is obtained by comparing the vector difference between the first coordinate result and the second coordinate result.

5. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 4, characterized in that: The S710 includes: S711, matching the first coordinate result and the second coordinate result according to the first number, and matching the sub-coordinate with the new sub-coordinate; S712, obtaining the vector difference between each pair of the sub-coordinates and the new sub-coordinates as a separate vector; S713, reading the rock formation division in each of the sub-areas and the position in the total hydraulic head distribution field, and assigning a first weight coefficient to a plurality of the individual vectors corresponding to each of the sub-areas; S714, performing vector summation on each of the individual vectors to obtain a first variation; S715, assigning the first weight coefficient to each of the first changes, weighting each of the first changes, and performing mathematical synthesis to obtain an overall displacement result of the valley width observation section; S716. Obtain valley deformation results at each valley observation section according to the overall displacement results.

6. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 4, characterized in that: Also included is the S800 performed after the S710: S800, updating the first coordinate result, and completing the update after the second coordinate result covers the first coordinate result; after the update is completed, repeat S100.

7. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 4, characterized in that: It also includes S610 which is performed after S600; S610, inputting the first vector set and the local water head value into a pre-trained third prediction model; The third prediction model outputs a corrected permeability coefficient value, and S310 is performed again according to the corrected permeability coefficient.

8. The method for predicting valley deformation of a water storage dam based on rock mass mechanical parameters according to claim 4, characterized in that: Also included is S310 performed after S300: S310, reading the second numbers, and assigning a permeability coefficient value to each of the rock samples according to the arrangement order of the second numbers; Dividing the sub-areas into rock formations according to the permeability coefficient; The rock layer division includes permeable rock layers and general rock layers divided along the depth direction of the valley slope of the river valley.

9. A method for predicting valley deformation of a water storage dam based on rock strata, according to any one of claims 5 to 8, characterized in that: Replace S500A with S500B; S500B, reading the first coordinate results corresponding to the plurality of sub-regions selected in S400, and establishing a second vector set according to the first serial number, wherein the second vector set includes a plurality of second sub-vectors, each of which includes: The first number corresponding to the sub-region; The first parameter and sub-coordinates corresponding to a rock sample and the result of rock layer division; A plurality of the first sub-vectors having the same first number constitute the first vector set; S700 was replaced by S700B; S700B, inputting the second vector set and the local water head value into a pre-trained second prediction model; The second prediction model outputs the coordinates of each of the sub-areas at the prediction moment as a second coordinate result, and the second coordinate result includes a plurality of new sub-coordinates corresponding to the coordinates of each of the rock mass samples at the prediction moment.

10. The method for predicting valley deformation of a water storage dam based on rock formation according to claim 8, characterized in that: In S310: The permeable rock layer is a rock mass with a high permeability coefficient value, and the permeable rock layer includes: a fractured rock layer and a broken rock layer; The general rock formation is a rock mass with a low permeability coefficient value, and the general rock formation includes: granite and mudstone.

Images