Radial intercepting, draining and flow guiding method for controlling reservoir entering flood of tailings pond

Abstract

The invention discloses a radial interception, drainage and diversion method for controlling reservoir flood of a tailing pond, which comprises the following steps of: S1, acquiring topographic and geological data of the tailing pond, and constructing a numerical model of interception, drainage and diversion outside the tailing pond; s2, simulating a rainfall flow process and a water level change, loading a simulation result into the numerical model as a boundary condition, and outputting a water level flow boundary condition; s3, calculating water flow and pressure distribution conditions under different interception and drainage positions and flow guide angles, and screening out an optimal combination of the interception and drainage positions and the flow guide angles; s4, according to the screened optimal interception and drainage position and the flow guide angle, an interception and drainage flow guide pipe is built; s5, water level and flow data are monitored in real time, and the opening degree of an inlet valve is adjusted; and S6, monitoring the water flow data after flow guide, feeding back the water flow data to a control center, and controlling radial interception and drainage flow guide of the flood. Effective control over the reservoir entering flood is achieved, and safe operation of the tailing pond is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of radial intercepting, draining and guiding of tailing ponds, and specifically to a radial intercepting, draining and guiding method for controlling the flood water entering a tailing pond. Background Art

[0002] As an important facility for mining enterprises to handle tailings, a tailing pond faces many potential safety hazards during operation, among which the control of the flood water entering the pond is of crucial importance. Once the flood is not properly controlled, leading to a serious accident of the tailing pond breach, it will not only cause great damage to the surrounding environment, threaten the ecological balance, but also endanger the life and property safety of the downstream residents. Therefore, developing an efficient and reliable technology for controlling the flood water entering a tailing pond has important practical significance for ensuring the safe operation of the tailing pond and reducing the disaster risk.

[0003] Traditional technologies for controlling the flood water entering a tailing pond mainly include setting simple intercepting ditches and drainage wells. By setting intercepting ditches around the tailing pond to intercept surface runoff and leading the flood water to the drainage wells and then discharging it outside the pond. Its advantages are relatively simple structure and low construction cost, and to a certain extent, it can relieve the pressure of the flood on the tailing pond. However, this traditional solution has obvious disadvantages. On the one hand, the layout of the intercepting ditches and drainage wells often lacks scientific planning and is difficult to be accurately designed according to the topographic and geological conditions of the tailing pond, resulting in poor flood intercepting and draining effects and being unable to effectively cope with large-scale floods. On the other hand, the traditional solution lacks the ability to monitor and adjust the water flow state in real time. When the flood flow rate and water level change, the intercepting and draining strategy cannot be adjusted in time, making the safety of the tailing pond during the flood period difficult to be fully guaranteed.

[0004] With the development of technology, the existing technologies for controlling the flood water entering a tailing pond have been improved to a certain extent. For example, some technologies introduce numerical simulation means to preliminarily analyze the water flow situation of the tailing pond, so as to optimize the layout of the intercepting and draining facilities. At the same time, some tailing ponds are equipped with simple monitoring devices that can obtain basic data such as water level and flow rate. However, the existing technologies still have deficiencies. In terms of numerical simulation, the accuracy and reliability of the model need to be improved, and its adaptability to complex topographic and geological conditions is poor, resulting in a deviation between the simulation results and the actual situation, affecting the design and optimization of the intercepting and draining facilities. In terms of monitoring devices, the real-time and accuracy of the data are difficult to guarantee, and the monitoring data is not fully combined with the regulation of the intercepting and draining facilities, unable to achieve dynamic and accurate control of the flood water entering the pond. In addition, the existing technologies lack effective measures to deal with the radial dispersion problem of the flood, and the concentrated discharge of the flood causes a greater impact on the downstream, threatening the downstream ecology and safety.

[0005] Therefore, in view of the above problems, the present application proposes a radial intercepting, draining and guiding method for controlling the flood water entering a tailing pond, which solves the above-mentioned existing technical problems. Summary of the Invention

[0006] Based on the above, the present application proposes a radial intercepting and diversion method for controlling the flood inflow into a tailings pond, including:

[0007] S1. Obtain the topographic and geological data of the tailings pond, construct a numerical model for external intercepting and diversion of the tailings pond, and set the structural parameters of the intercepting and diversion structures in the model;

[0008] S2. Collect the rainfall data of the tailings pond, combine it with the underlying surface data of the basin, simulate the rainfall runoff process and water level changes through a hydrological model, and load the simulation results into the numerical model as boundary conditions to output the water level and flow boundary conditions of the intercepting and diversion;

[0009] S3. According to the output water level and flow boundary conditions, based on the radial basis function neural network algorithm, calculate the water flow velocity distribution and pressure distribution under different intercepting positions and diversion angles, and screen out the optimal combination of intercepting positions and diversion angles that can evenly disperse the flood radially and minimize the pressure loss;

[0010] S4. Build an intercepting and diversion pipe according to the selected optimal intercepting position and diversion angle to ensure the stable transportation of water flow in the pipe;

[0011] S5. Install an intelligent flow regulating device at the inlet of the intercepting and diversion pipe, adjust the opening of the inlet valve according to the real-time monitored water level and flow data, and control the flood inflow;

[0012] S6. Set monitoring points downstream of the tailings pond, monitor the water flow data after diversion in real time, and feed back the water flow data to the control center to conduct radial intercepting and diversion of the flood control.

[0013] Preferably, in S1, obtaining the topographic and geological data of the tailings pond includes obtaining the topographic and geomorphic features, stratigraphic lithology, and geological structure data of the tailings pond; obtaining the three-dimensional point cloud data of the tailings pond topography, detecting the geological structure information of the tailings pond, and combining the obtained macroscopic topographic and geological image data of the tailings pond area to obtain the topographic and geological data of the tailings pond.

[0014] Preferably, in S1, constructing the numerical model for external intercepting and diversion of the tailings pond specifically includes:

[0015] Converting the obtained topographic and geomorphic data into a digital elevation model through three-dimensional data processing, setting the flow velocities of the water flow in the x and y directions as v x 、v y , establishing the water flow motion equation, and the formula is: where q is the flow rate change rate per unit volume; for the flow velocities v x 、v y of the water flow in the x and y directions, the formula is: where k m is the permeability coefficient, h is the water head height. For the permeability coefficient k m is obtained through the formula where α m is the geological coefficient of the tailings pond, β m is the lithology coefficient of the tailings pond, γ m is the lithology pore parameter, p m is the pore pressure; According to the water flow motion equation, a numerical model for external intercepting drainage diversion of the tailings pond is constructed.

[0016] Preferably, in S2, rainfall data of the tailings pond is collected, combined with the underlying surface data of the basin, and the rainfall runoff process and water level changes are simulated through a hydrological model, specifically including:

[0017] According to the rainfall data of the tailings pond and the underlying surface data of the basin, the vegetation interception amount I is obtained. The formula is: I = α·C·P, where α is the interception coefficient, C is the vegetation coverage, and P is the rainfall intensity; According to the soil type and topographic slope data of the underlying surface data of the basin, the infiltration rate f of the tailings pond is obtained. The formula is: where K s is the saturated hydraulic conductivity, ψ is the soil suction, Δθ is the change in soil water content, and L is the wetting front advance distance; The rainfall runoff process and water level changes are obtained through the continuous equation of basin water storage. The formula is: where S is the basin water storage, R is the surface runoff, and t is the time.

[0018] Preferably, in S2, the simulation results are used as boundary conditions and loaded into the numerical model to output the water level - discharge boundary conditions for intercepting drainage diversion, specifically including the following steps:

[0019] S2.1. Perform multiple simulation operations on the numerical model after loading the boundary conditions, record the water flow state data at different positions inside the model during each operation, including flow velocity and pressure, and combine with the loaded boundary condition data to form a training data set;

[0020] S2.2. For the training data set, perform feature extraction through a convolutional neural network, mine the feature relationship between the boundary conditions and the water flow state inside the model, and extract the spatial and temporal features in the data through the convolutional layer, pooling layer, and fully - connected layer of the convolutional neural network;

[0021] S2.3. During the training process, use the simulated water flow state data as the input, and the water level - discharge data at the intercepting drainage diversion position obtained through actual observation or theoretical calculation as the output, and adjust the weight parameters of the convolutional neural network to minimize the error between the model output and the actual data;

[0022] S2.4. After training is completed, the new simulation results are used as inputs and fed into the trained convolutional neural network. The output of the model is the water level-discharge boundary condition for intercepting and diverting flow.

[0023] Preferably, for the output water level-discharge boundary condition in S3, the water flow velocity distribution and pressure distribution under different intercepting positions and diversion angles are calculated according to the radial basis function neural network, specifically including:

[0024] Normalize the output water level-discharge boundary condition data. The processed water level is h n , and the discharge is q n . Normalize the coordinates of different intercepting positions and the diversion angle θ n to form an input vector X. The formula is: Input it into the radial basis function neural network; the radial basis function neural network includes an input layer, a hidden layer, and an output layer. The radial basis function of the hidden layer selects the Gaussian function, and the formula is: where c i is the center vector of the i-th hidden layer neuron, σ i is the neuron width, and ||·|| is the Euclidean distance; the output of the hidden layer is passed to the output layer through weighted summation, and the water flow velocity distribution v b and the pressure distribution formula p b are output, where the water flow velocity distribution v b formula is: where M is the number of hidden layer neurons, w vi is the weight of the water flow velocity output layer, b v is the water flow velocity bias term, and the pressure distribution formula is: w pi is the weight of the pressure distribution output layer, b p is the pressure distribution bias term.

[0025] Preferably, in S3, according to the obtained water flow velocity distribution and pressure distribution, the optimal combination of intercepting position and diversion angle that enables the radial uniform dispersion of flood is screened, specifically including:

[0026] Using the intercepting position and diversion angle as variables, taking the radial uniform dispersion degree of flood as the optimization goal, the radial uniform dispersion degree of flood is measured by calculating the flow difference in different directions. Set polar coordinates with the intercepting point as the center, divide the circumference into r directions, and the flow in each direction is Q l . Calculate the average flow at the intercepting point The formula is: According to the average flow at the intercepting point and the flow Q lObtain the radial dispersion uniformity U of the flood, and the formula is: Through iterative calculation, obtain the minimum value of the radial dispersion uniformity U of the flood to find the optimal combination of the cut-off and drainage position and the diversion angle for the radial uniform dispersion of the flood.

[0027] Preferably, the cut-off and diversion pipe in S4 adopts a gradually changing aperture structure. The aperture gradually increases from the cut-off end to the diversion end, and the increasing ratio is designed according to the law of water flow velocity distribution; the law of water flow velocity distribution is obtained through the water flow velocity change curve, and the aperture increase value is Δd, and the formula is: where Δv is the water flow velocity increase value, d b is the initial reference aperture, and λ is the aperture correction coefficient.

[0028] Preferably, the opening degree of the inlet valve in S5 is adjusted by constructing a valve function to control the inflow flood discharge. The valve function formula is: V = V 0 +k H (H - H 0 ) + k Q (Q - Q 0 ), where H and Q are the real-time monitored water level and flow rate respectively, V is the valve opening degree, V 0 is the initial valve opening degree, and k H and k Q are the water level and flow rate adjustment coefficients respectively.

[0029] Preferably, the radial cut-off and diversion of the flood in S6 is controlled by remotely adjusting the parameters of the cut-off and diversion pipe or starting the standby cut-off and diversion device; adjusting the parameters of the cut-off and diversion pipe includes adjusting the opening degree of the pipeline valve and the angle of the internal diversion blade of the pipeline; the standby cut-off and diversion device includes an independent main cut-off and diversion pipeline and an auxiliary cut-off and diversion pipeline. The diversion pipeline and the auxiliary diversion pipeline have different pipe diameter specifications to adapt to different flow rates, and the opening and opening degree are remotely controlled to cooperate with the main cut-off and diversion pipe to achieve effective control of the flood.

[0030] Compared with the prior art, the technical solution of the present application has the following technical effects:

[0031] The present invention constructs a numerical model by obtaining the topographic and geological data of the tailings pond, combines rainfall and underlying surface data of the basin to simulate the rainfall flow process and water level change, and outputs the water level and flow boundary conditions of the cut-off and diversion, solves the problem of poor flood cut-off effect caused by lack of precise design and real-time monitoring and adjustment in the traditional scheme, and obtains the technical effect of being able to scientifically design the cut-off and diversion facilities according to the actual situation and effectively respond to the change of water level and flow rate during the flood period, greatly improving the safety of the tailings pond in dealing with floods.

[0032] The present invention calculates the water flow velocity and pressure distribution at different drainage positions and diversion angles by using a radial basis function neural network algorithm, screens the optimal combination, solves the problems of uneven radial dispersion of flood water and large pressure loss, and obtains the technical effect of making the flood water radially evenly dispersed and minimizing the pressure loss, reducing the damage of flood water to the drainage and diversion facilities, ensuring the high efficiency of drainage and diversion, and reducing the impact on the downstream.

[0033] The present invention constructs a drainage and diversion pipe with a gradually changing aperture structure according to the optimal drainage position and diversion angle, and installs an intelligent flow regulating device at the inlet. By constructing a valve function to adjust the valve opening to control the flood water flow into the reservoir, it solves the problems that traditional drainage and diversion facilities cannot be adjusted in real time according to the water flow state and are difficult to adapt to different flow rates, and obtains the technical effect of being able to stably transport water according to real-time water flow data and accurately control the flood water flow into the reservoir, further enhancing the flood control ability of the tailings pond.

[0034] The present invention sets monitoring points downstream of the tailings pond, feeds the water flow data back to the control center, and remotely adjusts the parameters of the drainage and diversion pipe or starts the standby drainage and diversion device, solving the problems that traditional technologies cannot achieve dynamic and accurate flood control and have insufficient ability to cope with emergencies, and obtaining the technical effect of being able to master the flood dynamics in real time, timely and flexibly adjust the drainage and diversion strategy, and effectively ensure the safety of the tailings pond under different flood conditions.

[0035] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, so as to be implemented in accordance with the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following describes the preferred embodiments of the present application in detail in conjunction with the accompanying drawings.

[0036] Those skilled in the art will understand the above and other purposes, advantages and features of the present application more clearly according to the following detailed description of the specific embodiments of the present application in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings. In all the drawings, similar elements or parts are generally denoted by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to actual scale.

[0038] Figure 1 It is a flowchart of a radial drainage and diversion method for controlling the flood water entering the tailings pond of the present invention;

[0039] Figure 2 Boundary condition output flowchart for a radial cut-off and diversion method for controlling the flood entering a tailings pond in the present invention;

[0040] Figure 3 Curve correlation diagram of flood and dispersion uniformity for a radial cut-off and diversion method for controlling the flood entering a tailings pond in the present invention. Detailed implementation manners

[0041] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. In the following description, specific details such as specific configurations and components are provided only to assist in a comprehensive understanding of the embodiments of the present application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, descriptions of known functions and structures are omitted for clarity and conciseness in the embodiments.

[0042] It should be understood that the term "one embodiment" or "the present embodiment" mentioned throughout the specification means that a specific feature, structure or characteristic related to the embodiment is included in at least one embodiment of the present application. Therefore, the appearances of the term "one embodiment" or "the present embodiment" throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in any suitable manner in one or more embodiments.

[0043] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity, and does not itself indicate the relationship between the various embodiments and / or arrangements discussed.

[0044] The term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, B exists alone, and both A and B exist simultaneously. The term " / and" in this article is a description of another association object relationship, indicating that two relationships can exist. For example, A / and B can represent: A exists alone, and both A and B exist. In addition, the character " / " in this article generally represents that the associated objects before and after are in an "or" relationship.

[0045] The term "at least one" in this article is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, at least one of A and B can represent: A exists alone, both A and B exist simultaneously, and B exists alone.

[0046] It should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion.

[0047] Example 1

[0048] This example mainly describes a radial cut-off and diversion method for controlling the flood inflow into the tailings pond, as Figure 1 shown, including the following steps:

[0049] S1. Obtain the topographic and geological data of the tailings pond, construct a numerical model for external cut-off and diversion of the tailings pond, and set the structural parameters of the cut-off and diversion structures in the model;

[0050] S2. Collect the rainfall data of the tailings pond, combine with the underlying surface data of the basin, simulate the rainfall runoff process and water level changes through a hydrological model, and load the simulation results as boundary conditions into the numerical model to output the water level and flow boundary conditions of the cut-off and diversion;

[0051] S3. According to the output water level and flow boundary conditions, based on the radial basis function neural network algorithm, calculate the water flow velocity distribution and pressure distribution under different cut-off positions and diversion angles, and screen out the optimal combination of cut-off positions and diversion angles that can make the flood disperse radially uniformly and have the minimum pressure loss;

[0052] S4. Build cut-off and diversion pipes according to the selected optimal cut-off positions and diversion angles to ensure the stable transportation of water flow in the pipes;

[0053] S5. Install an intelligent flow regulating device at the inlet of the cut-off and diversion pipe, adjust the opening of the inlet valve according to the real-time monitored water level and flow data, and control the inflow flood discharge;

[0054] S6. Set monitoring points downstream of the tailings pond, monitor the water flow data after diversion in real time, and feedback the water flow data to the control center for radial cut-off and diversion of flood control.

[0055] Preferably, in S1, obtaining the topographic and geological data of the tailings pond includes obtaining the topographic and geomorphic features, stratigraphic lithology and geological structure data of the tailings pond; by obtaining the three-dimensional point cloud data of the tailings pond topography, detecting the geological structure information of the tailings pond, and combining with the obtained macroscopic topographic and geological image data of the tailings pond area, the topographic and geological data of the tailings pond are obtained.

[0056] Preferably, in S1, constructing the numerical model for external cut-off and diversion of the tailings pond specifically includes:

[0057] The obtained topographic and geomorphic data is transformed into a digital elevation model through three-dimensional data processing, and the flow velocities of the water flow in the x and y directions are set as v x and v y , and a water flow motion equation is established. The formula is: where q is the flow rate change rate per unit volume; for the flow velocities v x and v y in the x and y directions of the water flow, the formula is: where k m is the permeability coefficient, h is the water head height, and for the permeability coefficient k m it is obtained through the formula where α m is the tailings pond geological coefficient, β m is the tailings pond lithology coefficient, γ m is the lithology pore parameter, p m is the pore pressure; according to the water flow motion equation, a numerical model for external interception and diversion of the tailings pond is constructed.

[0058] Preferably, in S2, the rainfall data of the tailings pond is collected, combined with the underlying surface data of the basin, and the rainfall flow process and water level change are simulated through a hydrological model, specifically including:

[0059] According to the rainfall data of the tailings pond and the underlying surface data of the basin, the vegetation interception amount I is obtained. The formula is: I = α·C·P, where α is the interception coefficient, C is the vegetation coverage, and P is the rainfall intensity; according to the soil type and topographic slope data of the underlying surface data of the basin, the infiltration rate f of the tailings pond is obtained. The formula is: where K s is the saturated hydraulic conductivity, ψ is the soil suction, Δθ is the change in soil water content, and L is the wetting front advance distance; the rainfall flow process and water level change are obtained through the continuous equation of basin water storage. The formula is: where S is the basin water storage, R is the surface runoff, and t is the time.

[0060] Preferably, as Figure 2 shown, in S2, the simulation result is used as a boundary condition and loaded into the numerical model to output the water level and flow boundary conditions for interception and diversion, specifically including the following steps:

[0061] S2.1. Perform multiple simulation operations on the numerical model after loading the boundary conditions, record the water flow state data at different positions inside the model during each operation, including flow velocity and pressure, and combine the loaded boundary condition data to form a training data set;

[0062] S2.2. For the training data set, feature extraction is performed through a convolutional neural network to explore the characteristic relationship between boundary conditions and the water flow state inside the model. The spatial and temporal features in the data are extracted through the convolutional layer, pooling layer and fully connected layer of the convolutional neural network.

[0063] S2.3. During the training process, the simulated water flow state data is used as input, and the water level flow data at the interception and diversion location obtained by actual observation or theoretical calculation is used as output, and the weight parameters of the convolutional neural network are adjusted to minimize the error between the model output and the actual data;

[0064] S2.4. After the training is completed, the new simulation results are passed as input into the trained convolutional neural network, and the model outputs the water level and flow boundary conditions for interception, drainage and diversion.

[0065] Preferably, in S3, for the output water level flow boundary conditions, the water flow velocity distribution and pressure distribution at different interception positions and diversion angles are calculated according to a radial basis function neural network, specifically including:

[0066] The output water level and flow boundary condition data are normalized, and the processed water level is t n , the flow rate is q n , the coordinates of different cut positions and the flow angle θ n Normalize to form the input vector X, the formula is: Input into the radial basis function neural network; the radial basis function neural network includes an input layer, a hidden layer and an output layer. The radial basis function of the hidden layer uses a Gaussian function, and the formula is: where c i is the center vector of the i-th hidden layer neuron, σ i is the neuron width, ||·|| is the Euclidean distance; the output of the hidden layer is transmitted to the output layer through weighted summation, and the output water flow velocity distribution v b And the pressure distribution formula p b , where the water velocity distribution v b The formula is: Where M is the number of neurons in the hidden layer, w vi is the water velocity output layer weight, b v is the water velocity bias term, and the pressure distribution formula is: w pi is the pressure distribution output layer weight, b p is the pressure distribution bias term.

[0067] Preferably, in step S3, according to the obtained water flow velocity distribution and pressure distribution, the optimal combination of the intercepting and draining position and the diversion angle that enables the radial uniform dispersion of the flood is screened out, which specifically includes:

[0068] Taking the intercepting and draining position and the diversion angle as variables, and taking the radial dispersion uniformity of the flood as the optimization objective, the radial dispersion uniformity of the flood is measured by calculating the flow rate differences in different directions. Taking the intercepting and draining point as the center, polar coordinates are set, the circumference is divided into r directions, and the flow rate in each direction is Q l , and calculate the average flow rate at the intercepting and draining point The formula is: According to the average flow rate at the intercepting and draining point and the flow rate Q in each direction l Obtain the radial dispersion uniformity U of the flood, and the formula is: Through iterative calculation, obtain the minimum value of the radial dispersion uniformity U of the flood, and find out the optimal combination of the intercepting and draining position and the diversion angle for the radial uniform dispersion of the flood.

[0069] Preferably, the intercepting and draining diversion pipe in step S4 adopts a gradually changing aperture structure, and the aperture gradually increases from the intercepting end to the diversion end, and the increasing ratio is designed according to the water flow velocity distribution law; the water flow velocity distribution law is obtained through the water flow velocity change curve, and the aperture increase value is Δd, and the formula is: Where Δv is the water flow velocity increase value, and d b is the initial reference aperture, and λ is the aperture correction coefficient.

[0070] Preferably, the opening degree of the inlet valve in step S5 is adjusted by constructing a valve function to control the inflow flood discharge, and the valve function formula is: V = V 0 +k H (H - H 0 ) + k Q (Q - Q 0 ), where H and Q are the water level and flow rate monitored in real time, V is the valve opening degree, V 0 is the initial valve opening degree, and k H and k Q are the water level and flow rate adjustment coefficients respectively.

[0071] Preferably, in step S6, the radial intercepting and draining diversion of the flood is controlled by remotely adjusting the parameters of the intercepting and draining diversion pipe or starting the standby intercepting and draining diversion device; adjusting the parameters of the intercepting and draining diversion pipe includes adjusting the opening degree of the pipeline valve and the angle of the internal diversion blade of the pipeline; the standby intercepting and draining diversion device includes an independent main intercepting and draining diversion pipeline and an auxiliary intercepting and draining diversion pipeline, and the diversion pipeline and the auxiliary diversion pipeline have different pipe diameter specifications to adapt to different flow rates, and the opening and opening degree are remotely controlled to cooperate with the main intercepting and draining diversion pipe to effectively control the flood.

[0072] This embodiment describes in detail that the present application obtains the topographic geology, rainfall and underlying surface data of the tailings pond, constructs a numerical model to simulate the water flow state, uses a neural network algorithm to select the optimal interception and drainage position and diversion angle, constructs a gradient aperture interception and diversion pipe and installs an intelligent flow regulating device, and cooperates with the downstream monitoring point feedback control to effectively solve the problem of flood control entering the tailings pond, achieve radially uniform dispersion of floods and minimize pressure losses, and can accurately control the flood flow entering the pond in real time, greatly improving the safety of the tailings pond during floods.

[0073] Based on Example 1, this example describes a preferred solution for loading the simulation results as boundary conditions into the numerical model to output the water level and flow boundary conditions for interception, drainage and diversion, specifically:

[0074] The convolutional neural network uses a multi-scale convolutional neural network. By setting convolution kernels of different sizes, the multi-scale convolutional neural network can simultaneously extract features from the numerical model simulation operation data after loading boundary conditions at multiple scales. Smaller convolution kernels can capture detailed features in water flow data, such as slight changes in local water flow velocity, subtle fluctuations in pressure, etc., while larger convolution kernels can obtain overall macroscopic features, such as large-scale water flow trends, major pressure distribution areas, etc. When processing tailings pond water flow state data, the multi-scale convolutional neural network can explore the characteristic relationship between boundary conditions and the water flow state inside the model from different angles. Through the convolution layer, convolution operations are performed at different scales, the pooling layer screens and reduces the dimensions of the features, and the fully connected layer fuses the multi-scale features, thereby more comprehensively and accurately extracting the spatial and temporal features in the data.

[0075] Integrating the attention mechanism, the attention mechanism can automatically learn the importance of different features in the data, assign higher weights to important features, so that the model pays more attention to features that have an important impact on the output results. In the process of processing tailings pond water flow data, not all water flow state data features have an equally important contribution to the water level and flow boundary conditions of interception and diversion. For example, during the peak of the flood, the changes in water flow velocity and pressure in key areas have a more significant impact on the water level and flow boundary conditions. Combined with the convolutional neural network with the attention mechanism, after extracting features in the convolution layer and pooling layer, the importance weight of each feature is calculated through the attention module. This weight will play a role in the subsequent fully connected layer calculation, so that important features are enhanced, while secondary features are relatively weakened. In this way, the model can focus on key features more accurately, dig out more valuable feature relationships, adjust weight parameters more effectively during the training process, reduce the error between the model output and the actual data, and thus obtain interception and diversion water level and flow boundary conditions that are more in line with the actual situation, and improve the reliability of tailings pond flood control.

[0076] As the number of convolutional neural network layers increases, the problem of gradient vanishing or gradient exploding will occur, making model training difficult and performance degraded. Residual connections are introduced. Residual connections allow the network to directly learn residual mappings. In the convolutional neural network for tailings pond water flow data processing, residual connections establish shortcut connection channels between different convolutional layers; when the network is forward propagating, in addition to normal convolution calculations, input data can be directly passed to subsequent layers through residual connections. In this way, during the back propagation process, the gradient can be more smoothly transmitted back through the residual connection, avoiding the situation of gradient vanishing or gradient exploding, so that the network can be trained more stably; residual connections help the network learn richer features, and even if the depth of the network is increased, the performance of the model can be continuously improved. For the complex water flow data of the tailings pond, a deeper and more stably trained convolutional neural network can dig out more detailed and comprehensive feature relationships, and perform better in the process of adjusting weight parameters to minimize the error between the model output and the actual data, and then more accurately output the water level and flow boundary conditions for interception and diversion, providing more reliable data support for tailings pond flood control.

[0077] This embodiment describes in detail a multi-scale convolutional neural network based on convolutional neural network optimization. By incorporating attention mechanism and residual connection, the problems of output and actual data error and model training difficulty are solved, thereby obtaining interception, drainage and diversion water level and flow boundary conditions that are more in line with the actual situation, improving the reliability of tailings pond flood control, and outputting interception, drainage and diversion water level and flow boundary conditions to provide more reliable data support for tailings pond flood control.

[0078] Based on Example 1, this example describes a detailed example of a radial interception and drainage method for controlling floods entering a tailings pond, specifically:

[0079] A tailings pond is selected. The tailings pond is located in a mountainous area with a catchment area of ​​50km 2 , total storage capacity 8 million m 3 , and carried out data collection work. The three-dimensional point cloud data of the tailings pond terrain was obtained by using three-dimensional laser scanning technology. Combined with the geological exploration drilling data, detailed topography, lithology and geological structure data were obtained; nearly 20 years of tailings pond rainfall data were collected, including information such as rainfall intensity and rainfall duration, as well as basin underlying surface data, such as vegetation coverage of 30%, soil type mainly loam, saturated hydraulic conductivity of 5×10 -5 m / s.

[0080] The acquired topographic data is converted into a digital elevation model, and the tailings pond geological coefficient α m is 0.5, the tailings pond lithology coefficient β m is 0.2, and the lithology pore parameter γ m is 0.3, the pore pressure p mis 100 kPa, and the calculated permeability coefficient k m is 0.5003. The water flow motion equation is established to construct a numerical model for the external intercepting drainage diversion of the tailings pond.

[0081] Simulate the rainfall flow rate and water level changes. According to the collected data, calculate the vegetation interception amount I = 1.5 mm (assuming a certain rainfall intensity P is 50 mm), and the infiltration rate f = 6×10 -5 m / s (assuming the soil suction ψ is 10 kPa, the change in soil water content Δθ is 0.1, and the advancing distance of the wetting front L is 0.5 m). Simulate the rainfall flow rate process and water level changes under different rainfall conditions through the continuous equation of basin water storage. For example, during a rainfall process lasting 3 hours with a total rainfall of 60 mm, the rising situation of the water level at different times is simulated.

[0082] Determine the optimal intercepting and drainage position and diversion angle. According to the simulation results, output the water level - flow boundary condition data. After normalizing these data, input them into a radial basis function neural network with the number of neurons M in the hidden layer being 50. After multiple calculations, obtain the water flow velocity distribution and pressure distribution under different intercepting and drainage positions and diversion angles. Set polar coordinates with the intercepting and drainage point as the center, divide the circumference into 8 directions, and calculate the radial flood dispersion uniformity U. After a large number of iterative calculations, finally determine that at the coordinate (100, 200) (unit: meter) and the diversion angle of 45°, U obtains the minimum value of 0.1, and this is the optimal combination of the intercepting and drainage position and the diversion angle at this time.

[0083] Construct the intercepting and drainage diversion pipe and install the regulating device. Based on the determined optimal parameters, construct the intercepting and drainage diversion pipe. The intercepting and drainage diversion pipe adopts a gradually changing pore size structure, and the pore size gradually increases from the intercepting end to the diversion end. According to the water flow velocity distribution law, the initial reference pore size d b is 0.5 m, the pore size correction coefficient λ is 0.8. When the increase in water flow velocity in a certain section Δv is 0.2 m / s and the water flow velocity v b is 1 m / s, calculate the increase in pore size Δd = 0.08 m. Install an intelligent flow regulating device at the entrance of the intercepting and drainage diversion pipe to adjust the opening degree of the inlet valve according to the real - time monitored water level and flow rate data; the initial valve opening degree V 0 is 0.3, the water level adjustment coefficient k H is 0.1, the flow rate adjustment coefficient k Q is 0.05. When the real - time monitored water level H is 2 m higher than the initial water level H 0 and the flow rate Q is 10 m 0 / s larger than the initial flow rate Q 3 , the valve opening degree V = 0.3 + 0.1×2 + 0.05×10 = 1.0.

[0084] Through continuous iteration, such as Figure 3As shown, in this application, the optimal cut-off and drainage position and diversion angle combination are obtained by acquiring the dispersion uniformity U. When the flood changes, the dispersion uniformity U will respond in real time, and the dynamic change of its value intuitively reflects that the technical solution of this application can accurately adjust the cut-off and drainage diversion strategy according to the actual situation of the flood, achieve radial uniform dispersion of the flood, effectively reduce the impact of the flood on the tailings pond and downstream, and ensure the safe and stable operation of the tailings pond.

[0085] This embodiment details specific examples, which can scientifically design the cut-off and drainage diversion facilities according to the actual situation, and have the technical effect of effectively coping with the changes in water level and flow during floods, greatly improving the flood resistance safety of the tailings pond.

[0086] The above are only the preferred embodiments of the present invention, and it does not limit the protection scope of the present invention. For those skilled in the art, the present invention can have various changes and modifications; within the spirit and principle of the present invention, any changes, modifications, substitutions, integrations, and parameter changes made to these embodiments by conventional substitutions or by achieving the same functions without departing from the principle and spirit of the present invention all fall within the protection scope of the present invention.

Claims

1. A radial interception and drainage method for controlling flooding entering a tailings pond, characterized in that: include: S1. Obtain the topographic and geological data of the tailings pond and construct a numerical model of interception, drainage and diversion outside the tailings pond. In the model, the structural parameters of the interception, drainage and diversion structure are set; S2. Collect rainfall data of tailings pond, combine with watershed underlying surface data, simulate rainfall flow process and water level change through hydrological model, load the simulation results into numerical model as boundary conditions, and output water level flow boundary conditions for interception, drainage and diversion; S3. According to the output water level and flow boundary conditions, based on the radial basis function neural network algorithm, the water velocity distribution and pressure distribution under different interception and drainage positions and diversion angles are calculated, and the optimal interception and drainage position and diversion angle combination that can make the flood radially uniformly dispersed and minimize the pressure loss is selected; S4. According to the selected optimal interception and drainage position and diversion angle, an interception and drainage diversion pipe is constructed to ensure stable transportation of water flow in the pipe; S5. Install an intelligent flow regulating device at the entrance of the interception and drainage diversion pipe to adjust the opening of the inlet valve according to the real-time monitored water level and flow data to control the flood flow into the reservoir; S6. Set up monitoring points downstream of the tailings pond to monitor the water flow data after diversion in real time, and feed the water flow data back to the control center to control radial interception and diversion of floods.

2. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: The topographic and geological data of the tailings pond are obtained in S1, including obtaining the topographic and geomorphic, stratigraphic lithology and geological structure data of the tailings pond; the topographic and geological data of the tailings pond are obtained by obtaining the three-dimensional point cloud data of the tailings pond topography, detecting the geological structure information of the tailings pond, and combining the obtained macroscopic topographic and geological image data of the tailings pond area.

3. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: The numerical model for interception, drainage and diversion outside the tailings pond is constructed in S1, which specifically includes: The acquired topographic data is converted into a digital elevation model through three-dimensional data processing, and the flow velocity of the water in the x and y directions is set to v x 、v y , establish the water flow equation, the formula is: Where q is the rate of change of flow rate per unit volume; for the flow velocity v in the x and y directions x 、v y , the formula is: where k m is the permeability coefficient, h is the water head height, for the permeability coefficient k m By formula Get, where α m is the geological coefficient of the tailings pond, β m is the lithology coefficient of the tailings pond, γ m is the lithological pore parameter, p m is the pore pressure; based on the water flow equation, a numerical model for interception, drainage and diversion outside the tailings pond is constructed.

4. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: In S2, rainfall data of the tailings pond is collected, combined with the underlying surface data of the watershed, and the rainfall flow process and water level changes are simulated through a hydrological model, specifically including: According to the rainfall data of the tailings pond and the underlying surface data of the watershed, the vegetation interception amount I is obtained, and the formula is: I = α·C·P, where α is the interception coefficient, C is the vegetation coverage, and P is the rainfall intensity; according to the soil type and terrain slope data of the underlying surface data of the watershed, the infiltration rate f of the tailings pond is obtained, and the formula is: Where K s is the saturated hydraulic conductivity, ψ is the soil suction, Δθ is the change in soil moisture content, and L is the distance of the wetting front advancement. The rainfall flow process and water level change are obtained through the basin water storage continuity equation, and the formula is: Where S is the water storage capacity of the basin, R is the surface runoff, and t is the time.

5. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: In S2, the simulation results are loaded into the numerical model as boundary conditions, and the water level and flow boundary conditions for interception, drainage and diversion are output, which specifically includes the following steps: S2.

1. Perform multiple simulation operations on the numerical model after loading boundary conditions, record the water flow state data at different positions inside the model in each operation, including flow velocity and pressure, and combine them with the loaded boundary condition data to form a training data set; S2.

2. For the training data set, feature extraction is performed through a convolutional neural network to explore the characteristic relationship between boundary conditions and the water flow state inside the model. The spatial and temporal features in the data are extracted through the convolutional layer, pooling layer and fully connected layer of the convolutional neural network. S2.

3. During the training process, the simulated water flow state data is used as input, and the water level flow data at the interception and diversion location obtained by actual observation or theoretical calculation is used as output, and the weight parameters of the convolutional neural network are adjusted to minimize the error between the model output and the actual data; S2.

4. After the training is completed, the new simulation results are passed as input into the trained convolutional neural network, and the model outputs the water level and flow boundary conditions for interception, drainage and diversion.

6. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: In S3, for the output water level flow boundary conditions, the water flow velocity distribution and pressure distribution at different interception and drainage positions and diversion angles are calculated according to the radial basis function neural network, specifically including: The output water level and flow boundary condition data are normalized, and the processed water level is h n , the flow rate is q n , the coordinates of different cut positions and the flow angle θ n Normalize to form the input vector X, the formula is: Input into the radial basis function neural network; the radial basis function neural network includes an input layer, a hidden layer and an output layer. The radial basis function of the hidden layer uses a Gaussian function, and the formula is: where c i is the center vector of the i-th hidden layer neuron, σ i is the neuron width, ∥·∥ is the Euclidean distance; the output of the hidden layer is transmitted to the output layer through weighted summation, and the output water flow velocity distribution v b And the pressure distribution formula p b , where the water velocity distribution v b The formula is: Where M is the number of neurons in the hidden layer, w vi is the water velocity output layer weight, b v is the water velocity bias term, and the pressure distribution formula is: w pi is the pressure distribution output layer weight, b p is the pressure distribution bias term.

7. A radial interception and drainage method for controlling flood water entering a tailings pond according to claim 1 or 6, characterized in that: In S3, based on the acquired water velocity distribution and pressure distribution, the optimal interception and drainage position and diversion angle combination for uniform radial dispersion of the flood are selected, which specifically includes: The interception position and diversion angle are used as variables, and the uniformity of radial dispersion of flood is used as the optimization target. The uniformity of radial dispersion of flood is measured by calculating the difference in flow in different directions. The polar coordinates are set with the interception point as the center, and the circumference is divided into r directions. The flow in each direction is Q l , calculate the average flow rate at the cut-off point The formula is: According to the average flow rate at the cut-off point and the flow rate in each direction Q l The radial dispersion uniformity of flood water, U, is obtained by the following formula: Through iterative calculation, the minimum value of the radial dispersion uniformity U of the flood is obtained, and the optimal interception and drainage position and diversion angle combination for radial uniform dispersion of the flood is obtained.

8. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: The interception and drainage pipe in S4 adopts a gradual aperture structure, the aperture of the gradual aperture structure gradually increases from the interception end to the diversion end, and the increase ratio is designed according to the water flow velocity distribution law; the water flow velocity distribution law is obtained through the water flow velocity change curve, and the aperture increase value is Δd, and the formula is: Where Δv is the increase in water velocity, d b is the initial reference aperture, and λ is the aperture correction coefficient.

9. A radial interception and drainage method for controlling flooding entering a tailings pond according to claim 1, characterized in that: The opening of the S5 inlet valve is adjusted by constructing a valve function to control the flood flow into the reservoir. The valve function formula is: V = V0 + k H (H-H0)+k Q (Q-Q0), where H and Q are the real-time monitored water level and flow rate, V is the valve opening, V0 is the initial valve opening, k H and k Q are the water level and flow adjustment coefficients respectively.

10. A radial interception and drainage method for controlling flood water entering a tailings pond according to claim 1, characterized in that: The radial interception and drainage diversion for controlling flood in S6 is controlled by remotely adjusting the parameters of the interception and drainage diversion pipe or starting the spare interception and drainage diversion device; adjusting the parameters of the interception and drainage diversion pipe includes adjusting the opening of the pipeline valve and the angle of the guide blades inside the pipeline; the spare interception and drainage diversion device includes an independent main interception and drainage diversion pipe and an auxiliary interception and drainage diversion pipe, the diversion pipe and the auxiliary diversion pipe have different pipe diameter specifications to adapt to different flow rates, and the opening and closing and opening are remotely controlled to cooperate with the main interception and drainage diversion pipe to achieve effective control of the flood.

Images