Double-flood forecasting roughness limiting method and system based on hydrodynamic stability constraint

Abstract

This invention discloses a roughness limiting method and system for forecasting dual-super-flood systems based on hydrodynamic stability constraints, relating to the field of hydraulic control technology. The method includes the following steps: acquiring runoff boundary data output by the dual-super-flood generation model for the current time step; inputting the feedforward input and the hydrodynamic state variables obtained in the previous time step into the hydrodynamic model to generate the system state feedback deviation for the current time step; monitoring the asymptotic stability of the current system state space based on the system state feedback deviation; if the monitoring result shows that the asymptotic stability condition is met, converting the derived hydrodynamic state variables into a target control reference signal; if the monitoring result shows that the asymptotic stability condition is not met, restoring the current system state space to the hydrodynamic state variables obtained in the previous time step. This invention eliminates the risk of control command distortion.

Description

Technical Field

[0001] This invention relates to the field of water conservancy control technology, specifically to a method and system for roughness limiting in forecasting dual-super-flood based on hydrodynamic stability constraints. Background Technology

[0002] In existing automated watershed disaster prevention systems, the generation of control reference signals is highly dependent on the front-end flood forecasting model. The routine workflow is as follows: surface and interflow data are calculated using a dual-super-runoff model; the generated data is then used as boundary conditions to input into a hydrodynamic model for flow evolution calculations; and finally, water level and flow forecasts are output to downstream execution equipment.

[0003] The dual-overflow model calculates the surface overflow depth using an over-infiltration runoff calculation module and the subsurface and soil overflow depth using an over-storage runoff calculation module. The overflow depths and over-storage runoff depths are then superimposed and time-allocated to generate runoff boundary data. The hydrodynamic model performs inferences based on this runoff boundary data and the hydrodynamic state variables obtained from the previous time step, including water depth and velocity distribution data.

[0004] When a watershed experiences extreme rainfall, the runoff boundary data output by the dual super runoff generation model undergoes drastic changes in a short period of time. When such runoff boundary data is injected into the hydrodynamic model as a feedforward input, the convergence value of the residuals of the state equations or the numerical oscillation eigenvalues ​​within the hydrodynamic model are likely to exceed the preset threshold, causing the system to fail to generate system state feedback deviations that meet the asymptotic stability conditions. If such unstable projection results are directly converted into target control reference signals and sent to the water conservancy execution terminal, it may cause malfunctions of the execution terminal or delays in scheduling instructions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for roughness limiting in dual-super-flood forecasting based on hydrodynamic stability constraints.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints, applied to computing devices. The method includes the following steps: Obtain the runoff boundary data output by the dual super-runoff model for the current time step, and use the runoff boundary data as the feedforward input; The feedforward input and the hydrodynamic state variables obtained in the previous time step are input into the hydrodynamic model, and the hydrodynamic model performs the deduction to generate the system state feedback deviation for the current time step. The hydrodynamic state variables include water depth and flow velocity distribution data. Based on the system state feedback deviation, monitor the asymptotic stability of the current system state space; If the monitoring results meet the asymptotic stability condition, the hydrodynamic state variables obtained from the simulation will be converted into target control reference signals and sent to the lower-level water conservancy execution terminal to execute physical flood control scheduling. If the monitoring result does not meet the asymptotic stability condition, the current system state space is restored to the hydrodynamic state variables obtained in the previous time step, the local flow characteristics that lead to instability are extracted, and the roughness parameters of the current water area are corrected and physically limited by a preset physical constraint adaptive regulator based on the local flow characteristics. The hydrodynamic model is then restarted based on the corrected roughness parameters.

[0007] As a preferred embodiment of the present invention, the steps of generating the system state feedback deviation at the current time step and monitoring the asymptotic stability of the current system state space based on the system state feedback deviation include: At the end of each iteration of the hydrodynamic model, the residual convergence value of the state equation and the numerical oscillation characteristic value are extracted as the system state feedback deviation. When the convergence value of the residual of the state equation is less than a preset first safety threshold and the numerical oscillation characteristic value is less than a preset second safety threshold, it is determined that the current system state space satisfies the asymptotic stability condition. When the convergence value of the residual of the state equation is greater than or equal to the first safety threshold, or the numerical oscillation characteristic value is greater than or equal to the second safety threshold, it is determined that the current system state space does not meet the asymptotic stability condition.

[0008] As a preferred embodiment of the present invention, the step of restoring the current system state space to the hydrodynamic state variables obtained in the previous time step includes: When the asymptotic stability condition is not met, the register write instruction for the hydrodynamic state variable at the current time step is intercepted. Retrieve the hydrodynamic state variables obtained at the previous time step stored in the state snapshot database; The computational buffer is rewritten using the hydrodynamic state variables obtained in the previous time step to reset the initial integral boundary of the hydrodynamic model.

[0009] As a preferred embodiment of the present invention, the step of feedback correction and physical limiting of the roughness parameter of the current water area through a preset physical constraint adaptive regulator includes: Extract the hydrodynamic state variables of the grid cells that caused the instability obtained in the previous time step, and use the hydrodynamic state variables of adjacent stable grid cells for interpolation to supplement them, and analyze to obtain the hydrodynamic characteristic parameters and the relative submergence degree of the underlying surface; The physical constraint adaptive regulator uses the relative submergence of the underlying surface as a feedforward compensation amount and matches the corresponding roughness correction coefficient in a preset flow-resistance curve library. The initial roughness value of the current water area is updated using the roughness correction coefficient to calculate the equivalent roughness under the current flow state. The equivalent roughness is then limited to a physical confidence interval determined based on the land cover type after saturation limiting processing, and the output is the corrected roughness parameter.

[0010] As a preferred embodiment of the present invention, the step of determining the physical confidence interval includes: Obtain the land use type data corresponding to the current computing grid; Based on the land use type data, query the standard roughness table to determine the benchmark roughness value; Based on the benchmark roughness value, an upper and lower floating threshold are set to form the physical confidence interval. The physical confidence interval is used to limit the corrected roughness parameter to a physically reasonable range under extreme hydrological conditions for this land use type.

[0011] As a preferred embodiment of the present invention, the method further includes the following steps: Compare the calculated equivalent roughness with the boundary value of the physical confidence interval; If the equivalent roughness exceeds the boundary value of the physical confidence interval, it is determined that the current mesh has a flow anomaly or the input data is distorted; In response to the determination, the current adaptive parameter adjustment process is interrupted, the spatial coordinates of the abnormal grid are recorded, and a fault interruption signal is generated.

[0012] As a preferred embodiment of the present invention, the step of restarting the hydrodynamic model derivation based on the corrected roughness parameters is performed using a cyclic trial-and-error method, including: Before each restart of the simulation, increment the trial count by one; The hydrodynamic model is invoked to perform a positive integral based on the reset hydrodynamic state variables and the corrected roughness parameters to generate a new system state feedback deviation. If the new system state feedback deviation satisfies the asymptotic stability condition, the loop exits and jumps to the step of generating the target control reference signal; If the trial count reaches the preset maximum number of trials and still does not meet the asymptotic stability condition, then the fault-tolerant interlocking mechanism is triggered.

[0013] As a preferred embodiment of the present invention, the step of triggering the fault-tolerant interlocking mechanism includes: Generate a model to calculate abnormal alarm information; Switch the control commands of the water conservancy execution terminal to a preset safe operation mode; Record the hydrodynamic state variables and trial parameters at the current time step and store them in the abnormal event log.

[0014] As a preferred embodiment of the present invention, the step of obtaining the runoff boundary data output by the dual super-runoff model for the current time step includes: Real-time rainfall monitoring data, anterior impact rainfall, and underlying surface characteristic parameters of the target watershed are obtained through a hydro-meteorological sensor network. Call the excess infiltration runoff calculation module to calculate the depth of surface excess infiltration runoff; Call the super-storage runoff calculation module to calculate the super-storage runoff depth in the ground and soil; The runoff boundary data is generated by superimposing the super-permeable runoff depth and the super-storage runoff depth over time.

[0015] This invention also provides a roughness limiting system for dual-super-flood forecasting based on hydrodynamic stability constraints, comprising: A communication interface device is used to acquire hydrological and meteorological input data of the target watershed and establish a two-way control link with the water conservancy execution terminal. The device includes a memory and a microprocessor electrically connected to the communication interface device and the memory, the microprocessor being configured to execute a dual superflood forecast roughness limiting method based on hydrodynamic stability constraints.

[0016] The beneficial effects of this invention are: 1. This invention determines whether the current system state space meets the asymptotic stability condition by monitoring the system state feedback deviation. If it does not meet the condition, state recovery and roughness correction are performed to ensure that only the deduction results that meet the asymptotic stability condition can be converted into target control reference signals and sent to the water conservancy execution terminal. This mechanism avoids the direct output of unstable deduction results caused by sudden changes in runoff boundary data to the execution terminal, thus eliminating the risk of control command distortion.

[0017] 2. This invention extracts the hydrodynamic state variables of the grid cells causing instability from the previous time step, analyzes the hydrodynamic characteristic parameters and the relative submergence of the underlying surface, matches the corresponding roughness correction coefficient in a pre-set flow-resistance curve library, calculates the equivalent roughness under the current flow state, and restricts the equivalent roughness within a physical confidence interval determined based on the land cover type. This mechanism allows the roughness parameter to automatically adjust with changes in flow state and always remains within a physically reasonable range, avoiding arbitrary changes to parameter values ​​in pursuit of convergence. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0019] Figure 1 This is a schematic diagram of the workflow of the dual-super-flood forecast roughness limiting method of the present invention. Detailed Implementation

[0020] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] like Figure 1 As shown, a roughness limiting method for forecasting dual-super-flood based on hydrodynamic stability constraints is applied to a computing device. The method includes the following steps: The control device first acquires the runoff boundary data output by the dual super-runoff model for the current time step, and uses the runoff boundary data as the feedforward input.

[0023] The dual-overflow model runs within the control equipment or on a server connected to it. The model acquires real-time rainfall monitoring data, anterior rainfall events, and underlying surface characteristic parameters of the target watershed through a network of hydro-meteorological sensors. The model includes both over-infiltration flow calculation modules and over-sustainment flow calculation modules.

[0024] The excess infiltration runoff calculation module is used to calculate the depth of surface excess infiltration runoff generated when rainfall intensity exceeds soil infiltration rate; the excess storage runoff calculation module is used to calculate the depth of underground and subsurface excess storage runoff generated when soil moisture content reaches saturation. The calculation results of the two modules are superimposed and time-allocated under a unified spatial grid division scale to generate the runoff boundary data. The control equipment uses the above runoff boundary data as feedforward input, preparing to send it to the hydrodynamic model for calculation.

[0025] The control device inputs the feedforward input and the hydrodynamic state variables obtained in the previous time step into the hydrodynamic model, which then performs a deduction to generate the system state feedback deviation for the current time step. The hydrodynamic state variables include water depth and flow velocity distribution data.

[0026] The hydrodynamic model is constructed based on one-dimensional or two-dimensional shallow water equations and is solved numerically using the finite volume method or finite difference method. When the control equipment starts the calculation of the current time step, it retrieves the hydrodynamic state variables that converged in the previous time step from the state snapshot database. These hydrodynamic state variables include the water depth values ​​and velocity components along the x and y directions of each computational grid node in the watershed.

[0027] The control device uses the feedforward input as boundary conditions and the hydrodynamic state variables from the previous time step as initial conditions, inputting them together into the hydrodynamic model. The hydrodynamic model performs numerical derivation within a set number of iterations, progressively solving for the water depth and velocity distribution at each grid node.

[0028] At the end of each iteration, the control device extracts the convergence value of the state equation residual and the numerical oscillation characteristic value, and combines these two values ​​as the system state feedback deviation. The convergence value of the state equation residual characterizes the degree of difference between the solution vector of the current iteration step and the previous iteration step, while the numerical oscillation characteristic value characterizes whether abnormal fluctuations occur in the local gradient of flow velocity or water level during the calculation process.

[0029] The control device monitors the asymptotic stability of the current system state space based on the system state feedback deviation.

[0030] The control device has a first safety threshold and a second safety threshold preset internally. The first safety threshold is used to limit the acceptable upper limit of the convergence value of the state equation residual, and the second safety threshold is used to limit the acceptable upper limit of the numerical oscillation characteristic value. Both thresholds are pre-calibrated using historical flood calibration data and stored in the control device's configuration file.

[0031] The control device compares the convergence value of the state equation residual in the system state feedback deviation generated at the current time step with a first safety threshold, and compares the numerical oscillation characteristic value with a second safety threshold. When the convergence value of the state equation residual is less than the first safety threshold and the numerical oscillation characteristic value is less than the second safety threshold, the control device determines that the current system state space meets the asymptotic stability condition. Conversely, when the convergence value of the state equation residual is greater than or equal to the first safety threshold, or the numerical oscillation characteristic value is greater than or equal to the second safety threshold, the control device determines that the current system state space does not meet the asymptotic stability condition.

[0032] If the monitoring results meet the asymptotic stability conditions, the control equipment will convert the hydrodynamic state variables obtained from the simulation into target control reference signals and send them to the lower-level water conservancy execution terminal to perform physical flood control scheduling.

[0033] Specifically, the control equipment converts the water depth and flow velocity distribution data of each grid node, which has finally converged after numerical simulation, into a control command format recognizable by the hydraulic execution terminal according to a preset conversion rule. The hydraulic execution terminal includes a programmable logic controller for a flood discharge gate, a frequency converter for a pumping station unit, or a signal receiving module for an early warning broadcast system.

[0034] For example, when the water level at a certain gate section, obtained through simulation, reaches the preset gate opening threshold, the control equipment converts the calculated target opening value into a standard industrial control signal and sends it to the gate control cabinet via a communication interface device. Upon receiving this signal, the water conservancy execution terminal drives the gate motor to perform the opening operation, thus realizing physical flood control scheduling.

[0035] If the monitoring results indicate that the asymptotic stability condition is not met, the control device will perform the following adaptive adjustment operation: The control device restores the current system state space to the hydrodynamic state variables obtained in the previous time step.

[0036] The control device extracts the local flow characteristics that lead to instability.

[0037] The control device first identifies the computational grid cells that are causing the divergence. This identification process is achieved by scanning the state equation residual convergence values ​​and numerical oscillation eigenvalues ​​of all grid cells, and marking grid cells whose residuals or oscillation values ​​exceed the corresponding thresholds as unstable cells.

[0038] For a grid cell marked as unstable, the control device extracts the hydrodynamic state variables obtained by that grid cell in the previous time step. If the data of the unstable grid cell in the previous time step is missing or abnormal, the control device uses the hydrodynamic state variables of adjacent stable grid cells for interpolation to supplement the data. Specifically, the inverse distance weighted interpolation method or bilinear interpolation method is used to calculate the supplementary data of the grid cell.

[0039] Based on the aforementioned hydrodynamic state variables, the control equipment analyzes and obtains the hydrodynamic characteristic parameters and the relative submergence of the underlying surface. The hydrodynamic characteristic parameters include the Froude number, Reynolds number, or velocity gradient, while the relative submergence of the underlying surface is calculated based on the ratio of the water depth corresponding to the grid cell to the height of the surface vegetation or the height of the roughness cell.

[0040] Based on the local flow characteristics, the control device uses a preset physical constraint adaptive regulator to provide feedback correction and physical limitation on the roughness parameters of the current water area.

[0041] The control equipment has a pre-installed flow-resistance curve library. This curve library is constructed based on flume test data or field observation calibration results for different land use types (such as cultivated land, forest land, grassland, water area, and construction land). The relative submergence degree of the underlying surface is used as the input key, and the corresponding roughness correction coefficient is used as the output value to form a key-value pair mapping relationship. The curve library is stored in the non-volatile memory of the control equipment and organized in data table or JSON format.

[0042] The physical constraint adaptive regulator uses the analytically obtained relative submergence degree of the underlying surface as the feedforward compensation amount, and performs exact matching or linear interpolation query in the flow-resistance curve library to obtain the corresponding roughness correction coefficient.

[0043] The control equipment acquires the initial roughness value of the current water area. This initial roughness value is determined by consulting a standard roughness table based on the land use type data corresponding to the current calculation grid. The standard roughness table is compiled according to the water conservancy engineering industry specifications and includes the empirical range of Manning roughness values ​​for different land use types.

[0044] The control equipment updates the initial roughness value of the current water area using the roughness correction coefficient. Specifically, it uses multiplication correction: the initial roughness value is multiplied by the roughness correction coefficient to calculate the equivalent roughness under the current flow state.

[0045] The control device, after saturation limiting processing, restricts the equivalent roughness to a physical confidence interval determined based on land cover type, and outputs the corrected roughness parameter. The physical confidence interval is determined as follows: Land use type data corresponding to the current computational grid is obtained; a benchmark roughness value is determined by querying a standard roughness table based on this land use type data; and upper and lower floating thresholds are set based on this benchmark roughness value to form the physical confidence interval. This physical confidence interval is used to limit the corrected roughness parameter to a physically reasonable range for this land use type under extreme hydrological conditions.

[0046] The specific processing method is as follows: the control equipment compares the calculated equivalent roughness with the boundary value of the physical confidence interval. If the equivalent roughness exceeds the boundary value of the physical confidence interval, an exception handling process is triggered; if the equivalent roughness does not exceed the boundary value, the calculated value is used directly.

[0047] The control equipment restarts the hydrodynamic model derivation based on the corrected roughness parameters. This restart operation is performed using a cyclic trial calculation method. The number of retries is recorded by the trial calculation count value, and the system decides to exit the loop or trigger the fault-tolerant interlocking mechanism based on whether the asymptotic stability condition is met.

[0048] Through the above steps, this embodiment can ensure that the target control reference signal finally output to the water conservancy execution terminal has reliable asymptotic stability when the runoff boundary data of the dual super-runoff model outputs drastic changes. This is achieved through mechanisms such as monitoring system state feedback deviation, execution state recovery, extraction of local flow characteristics, roughness physical correction and amplitude limiting, and cyclic trial calculation. This avoids the distortion of control commands or malfunction of the execution terminal caused by numerical divergence.

[0049] Furthermore, the steps of generating the system state feedback deviation at the current time step and monitoring the asymptotic stability of the current system state space based on the system state feedback deviation include: At the end of each iteration of the hydrodynamic model, the microprocessor extracts the state equation residual convergence value and numerical oscillation characteristic value from the computational cache as the system state feedback deviation.

[0050] The convergence value of the state equation residual is used to characterize the degree of approximation of the hydrodynamic state of the spatial grid system in two consecutive iterations. The microprocessor performs the calculation by traversing all spatial grids and reading the floating-point water depth data from memory. The calculation formula is as follows:

[0051] ; Among them, R c M represents the convergence value of the residuals of the state equation, and H represents the total number of spatial grids involved in the computation. a,i H represents the discrete water depth value obtained by spatial grid i in the current iteration calculation step. b,i The discrete water depth value of spatial grid i is represented by the value obtained in the previous iteration calculation step.

[0052] Numerical oscillation eigenvalues ​​are used to characterize the physical constraint relationship between the propagation velocity of characteristic water flow waves in a spatial grid and the computational grid size. Their calculation formula is as follows: ; Among them, C o V represents the numerical oscillation eigenvalue at the global maximum. i T represents the local water flow scalar velocity stored in spatial grid i. s L represents the current time step constant of the hydrodynamic model. i The step size represents the spatial discrete characteristic of the spatial grid i, and max(·) is the maximization function. The microprocessor iterates through all grids to calculate the local eigenvalues ​​and extracts the maximum value as the criterion for judgment.

[0053] After the microprocessor obtains the system status feedback deviation, it executes the conditional branch judgment logic.

[0054] When the convergence value of the residual of the state equation is less than the preset first safety threshold and the numerical oscillation characteristic value is less than the preset second safety threshold, the current system state space is determined to meet the asymptotic stability condition.

[0055] The first safety threshold is set to a value between 0.001 and 0.005. The basis for setting this range is that if the threshold is set too high, data that has not fully converged may be allowed to pass; if it is set too low, the solution process may take too long. Preferably, the first safety threshold is set to 0.003.

[0056] The second safety threshold is set to a value between 0.80 and 0.95. The basis for setting this range is that in the numerical calculation of water flow, the propagation distance of the characteristic wave within one time step should not exceed the size of one spatial grid. Setting the upper limit to 0.95 is to reserve a safety margin for extreme working conditions.

[0057] When the convergence value of the residual of the state equation is greater than or equal to the first safety threshold, or the numerical oscillation characteristic value is greater than or equal to the second safety threshold, it is determined that the current system state space does not meet the asymptotic stability condition.

[0058] Taking a flood forecasting and dispatching scenario in a certain river basin as an example, the control equipment is deployed in the industrial control computer of the flood control command center of the basin. During a certain simulation, the microprocessor calculates that the residual convergence value of the state equation is 0.002 and the numerical oscillation characteristic value is 0.85 at the end of the hydrodynamic model iteration. The preset first safety threshold is 0.005 and the second safety threshold is 0.90. Since 0.002 is less than 0.005 and 0.85 is less than 0.90, the microprocessor determines that the current system state space meets the asymptotic stability condition and allows the simulation results to be converted into target control reference signals and sent to the water conservancy execution terminal.

[0059] Through the above dual-indicator judgment logic, the system can accurately identify numerical divergence caused by boundary mutations, providing a reliable trigger basis for subsequent state recovery and roughness correction.

[0060] Furthermore, during the process of the control equipment performing hydrodynamic model deduction, when it is determined that the current system state space does not meet the asymptotic stability condition based on the system state feedback deviation, the control equipment immediately triggers the state recovery process. This trigger occurs after the deduction result of the current time step is determined to be unusable and before any result data is written into the system state area or output to external devices.

[0061] When the asymptotic stability condition is not met, the control device intercepts register write instructions for the hydrodynamic state variables at the current time step.

[0062] When the status flag bit maintained internally by the control equipment fails the asymptotic stability check, the memory management unit of the control equipment refuses to perform any operation to write hydrodynamic state variables into the system status register. Only after the asymptotic stability check passes will the control equipment allow the data in the calculation buffer to be submitted to the system status register.

[0063] The control equipment retrieves the hydrodynamic state variables obtained from the previous time step from the state snapshot database.

[0064] The state snapshot database is a data structure stored in the non-volatile memory or dedicated memory partition of the control device. After the asymptotic stability check is successfully passed at each time step, the database stores the hydrodynamic state variables that have converged at the current time step as a snapshot record.

[0065] When the control device triggers a state recovery, it calculates the identifier of the previous time step by subtracting one from the current time step number. Using this identifier as the retrieval key, it performs an exact match query in the state snapshot database to extract the corresponding hydrodynamic state variable record.

[0066] The control device rewrites the temporary data in the computation buffer using the hydrodynamic state variables obtained from the previous time step, in order to reset the initial integral boundary of the hydrodynamic model.

[0067] Specifically, the control device writes the retrieved hydrodynamic state variables from the previous time step into the computational buffer node by node, overwriting the original temporary data. After the buffer is rewritten, the computational state of the hydrodynamic model is reset to the physically stable equilibrium point of the previous time step. When the control device subsequently restarts the simulation based on the corrected roughness parameters, the hydrodynamic model uses the hydrodynamic state variables from the previous time step as the initial integration boundary and the feedforward input as the boundary condition, and re-executes the numerical integration calculation from this stable baseline point, thereby avoiding divergence propagation caused by initial state instability.

[0068] Furthermore, the step of feedback correction and physical limiting of the roughness parameter of the current water area through a preset physical constraint adaptive regulator includes: The microprocessor extracts the hydrodynamic state variables of the mesh cells that caused the instability at the previous time step.

[0069] Since the calculation results of the grid cells that caused the instability may have diverged at the current time step, directly extracting the data from the current time step is not meaningful. Therefore, the microprocessor reads the basic water depth and flow velocity data obtained from the previous time step. For grid cells with missing data, the microprocessor uses the hydrodynamic state variables of adjacent stable grid cells for interpolation to supplement the data, for example, using the inverse distance weighted interpolation method to calculate smooth transition values.

[0070] Based on the above data, the microprocessor analyzes and obtains the hydrodynamic characteristic parameters and the relative submergence of the underlying surface. The formulas for calculating the hydrodynamic characteristic parameters are as follows:

[0071] ; Among them, F r Characterizing the hydrodynamic parameters, V s The interpolated flow velocity is represented by G, which represents the gravitational acceleration constant, and H represents the interpolated flow velocity within the discrete mesh. s The interpolated water depth is represented by the discrete mesh depth, where ε1 is a preset minimum zero-prevention bias constant, such as 10. -8 .

[0072] The formula for calculating the relative submergence of the underlying surface is: ; Among them, S r H represents the relative submergence of the underlying surface. s The definition is the same as above, H v The average reference height of water-blocking obstacles corresponding to the current grid surface cover type is used. This value is read from the system's preset static geographic information database. ε2 is a preset minimum zero-bias constant, such as 10. -6 .

[0073] The physical constraint adaptive regulator uses the relative submergence of the underlying surface as a feedforward compensation amount and matches the corresponding roughness correction coefficient in a preset flow-resistance curve library.

[0074] It should be noted that the physical constraint adaptive regulator is a software function module pre-installed in the control device, used to perform feedback correction and physical limiting of the roughness parameter based on local flow characteristics. This regulator is not an independent physical hardware, but a set of program code executed by a microprocessor. Its input parameter is the relative submergence degree of the underlying surface, and the output result is the corrected roughness parameter.

[0075] The microprocessor performs interpolation calculations in the curve library based on the input relative submergence of the underlying surface to generate the corresponding roughness correction coefficient.

[0076] The microprocessor updates the initial roughness value of the current water area using a roughness correction coefficient, and calculates the equivalent roughness under the current flow regime. The calculation formula is as follows: ; Where, N e The equivalent roughness is represented by N0, which represents the initial roughness value assigned to the current grid based on static geomorphological survey data, and K. c Characterizes the roughness correction coefficient.

[0077] After obtaining the equivalent roughness, the microprocessor performs saturation limiting processing, confining it within a physical confidence interval determined based on land cover type, and outputs the corrected roughness parameter. The saturation limiting processing employs the following piecewise function logic:

[0078] ; Where, N m Characterizing the corrected roughness parameter, N min With N max Representing the lower and upper limits of the physical confidence interval, respectively, when N e Less than N min When, the value is assigned to N. min When N e Greater than N max When, the value is assigned to N. max .

[0079] Taking a flood forecasting and scheduling scenario in a certain watershed as an example, the land use type of a certain grid cell is forest land, and the initial roughness value is 0.08. During an extreme rainstorm, the water depth of this grid cell rises sharply. The microprocessor calculates that the relative inundation degree of the underlying surface is 1.5. The physical constraint adaptive regulator matches the roughness correction coefficient of 0.85 in the flow-resistance curve library based on this inundation degree, and calculates the equivalent roughness to be 0.068. This value is within the forest land physical confidence interval of 0.06 to 0.12. The microprocessor outputs 0.068 as the corrected roughness parameter.

[0080] Furthermore, the steps for determining the physical confidence interval include: The microprocessor obtains the land use type data corresponding to the current computing grid.

[0081] A static geographic information database is pre-configured in the system's memory. This database uses the coordinate identifiers of spatial grids as the primary key and stores the land cover attribute identifiers of the corresponding grids. The microprocessor performs a query operation in the geographic information database based on the coordinate identifiers of the current computational grid that has experienced numerical instability, and extracts the corresponding land use type data. This land use type data is manifested as classification and coding data representing different landform features such as forest land, grassland, urban construction land, or exposed riverbed.

[0082] The microprocessor queries the standard roughness table based on land use type data to determine the baseline roughness value.

[0083] The microprocessor sets an upper and lower floating threshold based on a reference roughness value to form a physical confidence interval.

[0084] When encountering extreme rainstorms and flood peaks, the resistance of surface water flow will undergo certain physical changes. For example, high flow velocity can cause flexible vegetation to collapse, thereby reducing resistance, or floods can carry silt and floating objects, thereby increasing the local water-blocking effect. Therefore, the roughness value will fluctuate around the baseline value, but this fluctuation is limited by the physical boundaries of the surface material itself.

[0085] The microprocessor calculates the upper and lower floating thresholds using the following formulas: ; ; Where, N b Characterizing the reference roughness value, N min Characterizing the floating threshold, N max Characterizing the floating threshold, C down Characterized by the floating ratio coefficient, C up Characterizes the floating ratio coefficient.

[0086] Optionally, the downward floating ratio C down The numerical range is set to 0.1 to 0.3. This range is determined based on hydraulic experimental data. There is a limit to the reduction of surface roughness by high flow velocity. Even after flexible vegetation is felled, there is still friction from the underlying soil or root system. The resistance attenuation usually does not exceed 30% of the benchmark value.

[0087] Optionally, the upward floating ratio C up The numerical range is set to 0.2 to 0.5. This range is determined based on observational data of debris accumulation and local topographic damage under extreme flood conditions. The increase in resistance is constrained by macro-topography, and usually, a maximum upward fluctuation of 50% can cover the physical resistance increase effect under most extreme hydrological conditions.

[0088] Furthermore, the method also includes the following steps: The control equipment compares the boundary values ​​of the equivalent roughness calculated with the physical confidence interval.

[0089] The physical confidence interval consists of an upper limit and a lower limit. The upper limit is the sum of the baseline roughness value and the upward floating threshold, and the lower limit is the difference between the baseline roughness value and the downward floating threshold. The baseline roughness value is obtained by querying the standard roughness table, and the upward and downward floating thresholds are preset according to the physical variation range of land use type.

[0090] When performing the comparison, the control device compares the equivalent roughness with the upper and lower limits of the physical confidence interval. The specific conditional judgment logic is as follows: first, it checks whether the equivalent roughness is greater than the upper limit; if not, it checks whether the equivalent roughness is less than the lower limit.

[0091] If the equivalent roughness exceeds the boundary value of the physical confidence interval, it is determined that the current grid has a flow anomaly or the input data is distorted.

[0092] This indicates that the equivalent roughness has deviated from the physically reasonable range for this land use type under extreme hydrological conditions. The control equipment records the anomaly determination result in the status flag bit in memory.

[0093] In response to this determination, the control device interrupts the current adaptive parameter adjustment process: skips the operation of outputting the equivalent roughness as the corrected roughness parameter, and prevents the equivalent roughness from being written into the parameter area of ​​the hydrodynamic model; at the same time, it terminates the cyclic trial calculation process of the current time step and does not restart the hydrodynamic model deduction based on the equivalent roughness.

[0094] The control device records the spatial coordinates of abnormal grids. Each computational grid is assigned a unique grid identifier and corresponding spatial coordinate information during watershed spatial discretization, including the longitude and latitude coordinates of the grid center point, as well as the row and column indices in the computational grid matrix. The control device extracts the above information from the metadata of the current grid, associates it with the current time step identifier and the anomaly type identifier, stores it, and appends it to the anomaly event log file.

[0095] The control equipment generates a fault interruption signal and sends an alarm message to the upper-level monitoring system via the fieldbus communication interface. The alarm message includes information such as the timestamp of the anomaly, the spatial coordinates of the abnormal grid, the calculated value of the equivalent roughness, and the upper and lower limits of the physical confidence interval.

[0096] Furthermore, the step of restarting the hydrodynamic model derivation based on the corrected roughness parameters is performed using a cyclic trial-and-error method, including: Before each restart of the simulation, the control device increments the trial count value in the memory register by one.

[0097] The trial count is stored in the control device's memory register to record the number of restart simulations performed within the current time step after a failure to meet the asymptotic stability condition. At the start of each new time step, the control device initializes the trial count to zero. When the failure to meet the asymptotic stability condition is first triggered and the adaptive adjustment process begins, the control device increments the trial count from zero to one before performing a restart simulation. If the simulation result after this restart still does not meet the asymptotic stability condition, the control device performs state recovery and roughness correction again, and increments the trial count from one to two before the next restart simulation, and so on.

[0098] The control equipment calls the hydrodynamic model to perform positive integration based on the reset hydrodynamic state variables and the corrected roughness parameters, generating a new system state feedback deviation.

[0099] The hydrodynamic model uses the reset hydrodynamic state variables as initial conditions, the corrected roughness parameters as Manning roughness coefficients, and the feedforward input as boundary conditions to perform numerical derivation. At the end of each iteration, the control equipment extracts the convergence value of the residuals of the state equations and the numerical oscillation eigenvalues, and combines them as the new system state feedback deviation.

[0100] If the new system state feedback deviation meets the asymptotic stability condition, the control device exits the loop and jumps to the step of generating the target control reference signal.

[0101] If the trial count reaches the preset maximum number of trials and still does not meet the asymptotic stability condition, the control device triggers a fault-tolerant interlocking mechanism. The preset maximum number of trials is stored in a configuration file, with a typical value range of 3 to 10.

[0102] After generating a new system state feedback deviation, if the control device determines that the stability condition is not met, it first checks whether the trial count value has reached the maximum number of trials; if it is less than the maximum number of trials, it returns to the beginning of the loop to continue execution; if it is equal to the maximum number of trials, it triggers the fault-tolerant interlocking mechanism.

[0103] The fault-tolerant interlocking mechanism specifically includes: The control equipment generates a model to calculate abnormal alarm information.

[0104] The timestamp of the anomaly, the current time step identifier, the preset maximum number of trials, the convergence value of the state equation residual generated in the last trial, and the numerical oscillation characteristic value are encapsulated in a preset message format and sent to the upper monitoring system through the fieldbus communication interface.

[0105] The control equipment switches the control commands of the water conservancy execution terminal to the preset safe operation mode.

[0106] The safe operation mode is a set of operating status parameters pre-stored in the configuration file, which is determined according to the type of water conservancy execution terminal and flood control dispatching procedures.

[0107] For floodgate control terminals, the safe operation mode is set to maintain the current opening or close to the designed flood control safety opening; for pump station control terminals, the mode is set to stop operation or maintain minimum safe flow operation. The control equipment reads the control command parameters corresponding to the safe operation mode and sends them to the water conservancy execution terminal via the fieldbus communication interface.

[0108] The control equipment records the hydrodynamic state variables and trial parameters at the current time step and stores them in the abnormal event log.

[0109] The recorded data includes: the current time step identifier; the feedforward input data, i.e., the runoff boundary data; the hydrodynamic state variables obtained in the previous time step; the corrected roughness parameters generated during each retry in the cyclic trial calculation; the convergence value of the state equation residuals and the numerical oscillation eigenvalues ​​generated during each retry; and the judgment result of the last trial calculation, etc.

[0110] Furthermore, the step of obtaining the runoff boundary data output by the dual super-runoff model for the current time step includes: The control equipment acquires real-time rainfall monitoring data, antecedent rainfall, and underlying surface characteristic parameters of the target watershed through a network of hydrological and meteorological sensors.

[0111] The hydro-meteorological sensor network includes rain gauges, water level stations, soil moisture monitoring stations, and satellite remote sensing data receiving terminals deployed within the target watershed. Real-time rainfall monitoring data is collected by rain gauges at a frequency of minutes and uploaded to the data receiving queue of the control equipment via a wireless communication network. Anterior impact rainfall is a cumulative indicator reflecting the anterior soil moisture level in the watershed, calculated by the control equipment by reading historical rainfall and evaporation data from the previous several days. Underlying surface characteristic parameters include spatial distribution data of land use type, soil type, vegetation cover, and topographic slope for each spatial grid unit within the watershed. This data is pre-stored in the control equipment's memory and organized in a geographic information system grid file format.

[0112] When the control equipment starts the calculation of the current time step, it reads the latest real-time rainfall monitoring data from the data receiving queue and reads the previous impact rainfall and underlying surface characteristic parameters from the memory as inputs for subsequent runoff calculation.

[0113] The control equipment calls the excess infiltration runoff calculation module to calculate the depth of surface excess infiltration runoff.

[0114] For each grid cell, the rainfall intensity value is compared with the soil infiltration rate. If the rainfall intensity is greater than the soil infiltration rate, the excess rainfall is accumulated as the surface excess runoff depth; otherwise, it is zero.

[0115] The control equipment combines the excess infiltration runoff depth and the excess storage runoff depth, and adds the two runoff depths of the same grid unit to obtain the total runoff depth.

[0116] The control equipment allocates the total runoff depth over time. Based on the runoff time parameters of the watershed, it distributes the runoff depth to multiple calculation sub-time steps of the hydrodynamic model according to the triangular unit curve or exponential decay curve. It generates the inflow boundary value for each sub-time step and organizes the inflow boundary value of each sub-time step into runoff boundary data, which is stored in memory as a structure array.

[0117] This invention also provides a roughness limiting system for dual-super-flood forecasting based on hydrodynamic stability constraints, comprising: The communication interface device is used to acquire hydrological and meteorological input data of the target watershed and establish a two-way control link with the water conservancy execution terminal.

[0118] The communication interface device and the water conservancy execution terminal adopt a master-slave communication architecture. The control device, as the master station, initiates communication requests, and the water conservancy execution terminal, as the slave station, responds to the instructions and returns the execution status.

[0119] The memory is used to store the state-space control instruction set, the flow-resistance curve library, and the state snapshot database.

[0120] The memory includes non-volatile and volatile storage media. Non-volatile storage media can be solid-state drives, embedded multimedia cards, or NAND flash memory, used for long-term storage of system configuration parameters, control instruction sets, curve library data, and status snapshot databases. Volatile storage media can be double data rate synchronous dynamic random access memory, used for temporary storage of intermediate computational data during system operation.

[0121] The system includes a microprocessor electrically connected to the communication interface device and the memory. The microprocessor is configured to execute a roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints. The main control microprocessor can be an industrial-grade embedded processor, a programmable logic controller (PLC) central processing unit, or a high-performance microcontroller unit. The main control microprocessor integrates an arithmetic logic unit, a control unit, and a register set to execute the state-space control instruction set stored in the memory.

[0122] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints, applied to computing equipment, characterized in that... The method includes the following steps: Obtain the runoff boundary data output by the dual super-runoff model for the current time step, and use the runoff boundary data as the feedforward input; The feedforward input and the hydrodynamic state variables obtained in the previous time step are input into the hydrodynamic model, and the hydrodynamic model performs the deduction to generate the system state feedback deviation for the current time step. The hydrodynamic state variables include water depth and flow velocity distribution data. Based on the system state feedback deviation, monitor the asymptotic stability of the current system state space; If the monitoring results meet the asymptotic stability condition, the hydrodynamic state variables obtained from the simulation will be converted into target control reference signals and sent to the lower-level water conservancy execution terminal to execute physical flood control scheduling. If the monitoring result does not meet the asymptotic stability condition, the current system state space is restored to the hydrodynamic state variables obtained in the previous time step, the local flow characteristics that lead to instability are extracted, and the roughness parameters of the current water area are corrected and physically limited by a preset physical constraint adaptive regulator based on the local flow characteristics. The hydrodynamic model is then restarted based on the corrected roughness parameters.

2. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 1, characterized in that, The steps of generating the system state feedback deviation at the current time step and monitoring the asymptotic stability of the current system state space based on the system state feedback deviation include: At the end of each iteration of the hydrodynamic model, the residual convergence value of the state equation and the numerical oscillation characteristic value are extracted as the system state feedback deviation. When the convergence value of the residual of the state equation is less than a preset first safety threshold and the numerical oscillation characteristic value is less than a preset second safety threshold, it is determined that the current system state space satisfies the asymptotic stability condition. When the convergence value of the residual of the state equation is greater than or equal to the first safety threshold, or the numerical oscillation characteristic value is greater than or equal to the second safety threshold, it is determined that the current system state space does not meet the asymptotic stability condition.

3. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 1, characterized in that, The step of restoring the current system state space to the hydrodynamic state variables obtained in the previous time step includes: When the asymptotic stability condition is not met, the register write instruction for the hydrodynamic state variable at the current time step is intercepted. Retrieve the hydrodynamic state variables obtained at the previous time step stored in the state snapshot database; The computational buffer is rewritten using the hydrodynamic state variables obtained in the previous time step to reset the initial integral boundary of the hydrodynamic model.

4. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 1, characterized in that, The step of feedback correction and physical limiting of the roughness parameter of the current water area through a preset physical constraint adaptive regulator includes: Extract the hydrodynamic state variables of the grid cells that caused the instability obtained in the previous time step, and use the hydrodynamic state variables of adjacent stable grid cells for interpolation to supplement them, and analyze to obtain the hydrodynamic characteristic parameters and the relative submergence degree of the underlying surface; The physical constraint adaptive regulator uses the relative submergence of the underlying surface as a feedforward compensation amount and matches the corresponding roughness correction coefficient in a preset flow-resistance curve library. The initial roughness value of the current water area is updated using the roughness correction coefficient to calculate the equivalent roughness under the current flow state. The equivalent roughness is then limited to a physical confidence interval determined based on the land cover type after saturation limiting processing, and the output is the corrected roughness parameter.

5. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 4, characterized in that, The steps for determining the physical confidence interval include: Obtain the land use type data corresponding to the current computing grid; Based on the land use type data, query the standard roughness table to determine the benchmark roughness value; Based on the benchmark roughness value, an upper and lower floating threshold are set to form the physical confidence interval. The physical confidence interval is used to limit the corrected roughness parameter to a physically reasonable range under extreme hydrological conditions for this land use type.

6. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 4, characterized in that, The method further includes the following steps: Compare the calculated equivalent roughness with the boundary value of the physical confidence interval; If the equivalent roughness exceeds the boundary value of the physical confidence interval, it is determined that the current mesh has a flow anomaly or the input data is distorted; In response to the determination, the current adaptive parameter adjustment process is interrupted, the spatial coordinates of the abnormal grid are recorded, and a fault interruption signal is generated.

7. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 1, characterized in that, The step of restarting the hydrodynamic model derivation based on the corrected roughness parameters is performed using a cyclic trial-and-error method, including: Before each restart of the simulation, increment the trial count by one; The hydrodynamic model is invoked to perform a positive integral based on the reset hydrodynamic state variables and the corrected roughness parameters to generate a new system state feedback deviation. If the new system state feedback deviation satisfies the asymptotic stability condition, the loop exits and jumps to the step of generating the target control reference signal; If the trial count reaches the preset maximum number of trials and still does not meet the asymptotic stability condition, then the fault-tolerant interlocking mechanism is triggered.

8. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 7, characterized in that, The steps for triggering the fault-tolerant interlocking mechanism include: Generate a model to calculate abnormal alarm information; Switch the control commands of the water conservancy execution terminal to a preset safe operation mode; Record the hydrodynamic state variables and trial parameters at the current time step and store them in the abnormal event log.

9. The roughness limiting method for dual-super-flood forecasting based on hydrodynamic stability constraints according to claim 1, characterized in that, The steps for obtaining the runoff boundary data output by the dual super-runoff model for the current time step include: Real-time rainfall monitoring data, anterior impact rainfall, and underlying surface characteristic parameters of the target watershed are obtained through a hydro-meteorological sensor network. Call the excess infiltration runoff calculation module to calculate the depth of surface excess infiltration runoff; Call the super-storage runoff calculation module to calculate the super-storage runoff depth in the ground and soil; The runoff boundary data is generated by superimposing the super-permeable runoff depth and the super-storage runoff depth over time.

10. A roughness limiting system for dual-super-flood forecasting based on hydrodynamic stability constraints, characterized in that, include: A communication interface device is used to acquire hydrological and meteorological input data of the target watershed and establish a two-way control link with the water conservancy execution terminal. The memory, and a microprocessor electrically connected to the communication interface device and the memory, the microprocessor being configured to perform the hydrodynamic stability constraint-based roughness limiting method for dual superflood forecasting as described in any one of claims 1-9.

Images