Model prediction based channel-tunnel clear flow process control method and system

Abstract

The application discloses a kind of based on model prediction's channel-tunnel clear flow process control method and system, comprising: determining state constraint condition and dividing accident dispatch section range;Adopt hydraulic fluctuation evaluation index to quantify system fluctuation state, establish the optimization problem considering safety and water saving, dynamically determine safety weight and water-saving weight, obtain optimal control target by solving;Establish clear flow prediction model;With optimal control target as input, predict the hydrodynamic state quantity under the given gate action sequence, and with the action sequence of each gate as decision variable, establish prediction optimization control problem, and obtain each gate future control action sequence by solving;Each gate executes the action in future control action sequence, and after waiting control interval, the initial state of prediction model is updated and repeated.The application takes into account the different safety limits and water-saving needs of open channel section and tunnel section, reduces water abandonment under the premise of safety, and meets the requirements of rapidity, safety, flexibility and economy of emergency dispatch.

Description

Technical Field

[0001] This invention relates to the field of water conservancy technology, specifically to a method and system for controlling the full-flow process of a canal-tunnel based on model prediction. Background Technology

[0002] Open channels are used as the main water conveyance carrier in water diversion projects due to their advantages such as low cost, simple construction and convenient maintenance. However, they have poor terrain adaptability. When the project route needs to cross rivers, mountains, areas with poor geological conditions or important buildings, tunnels are often used for water conveyance. Therefore, water conveyance projects often use a combination of open channels and tunnels along the route.

[0003] Compared to pure open channel systems, canal-tunnel systems operate under more complex and variable conditions. Under normal conditions, the flow pattern in a canal-tunnel system is the same as that in a pure open channel system, exhibiting a gradually changing flow pattern. However, under conditions such as sudden changes in water load, rapid opening and closing of gates, and emergency dispatching, the tunnel section may experience alternating open and full flow, characterized by rapid transitions between open and pressurized flow. This alternating open and full flow is a special form of transient flow, often accompanied by significant reconfiguration of local velocity, pressure, and water surface morphology. Specifically, it manifests as strong transient pressure fluctuations in the tunnel section, causing severe water level fluctuations in adjacent open channel sections, significantly reducing the stability of system operation, and even leading to overpressure damage in the tunnel section or overflow of the open channel section.

[0004] Emergency control of water conveyance systems is a crucial issue in ensuring the safe operation of engineering projects. A common approach is to develop emergency dispatch plans, which involve assuming various possible accident scenarios, determining the optimal gate action scheme through numerical simulation, and compiling this into an emergency dispatch plan library. In the event of an accident, an appropriate gate dispatch scheme is selected based on experience for emergency dispatch management. However, this offline dispatch plan development method has significant drawbacks: the pre-set accident scenarios cannot cover all real-world accident scenarios, making it difficult to find a completely corresponding emergency plan when an accident occurs, requiring managers to determine a suitable dispatch scheme based on experience. In canal-tunnel systems, once open full flow is formed during dispatch, the relationship between gate action and system hydrodynamic response exhibits strong nonlinearity. Even small differences in gate action can lead to significant changes in the hydraulic response, resulting in the risk of system instability and damage. Without experience, this dispatch method carries substantial risks. Therefore, introducing automated control technology to customize an emergency control framework for canal-tunnel systems is of great significance.

[0005] In summary, given the complexity of canal-tunnel systems and the transient and highly nonlinear characteristics of the open-flow process during scheduling, a pressing technical problem that needs to be addressed by those skilled in the art is to propose a model predictive control framework suitable for emergency scheduling of canal-tunnel systems. This framework should meet the following requirements: 1) The controller design fully considers the different operating characteristics and safe operating ranges of open channels and tunnels, enabling unified scheduling and control of both to ensure the structural safety of the entire system; 2) The predictive model can quickly predict the open-flow transition process during the emergency scheduling period, with prediction accuracy and efficiency sufficient to meet the requirements of real-time online control; 3) The control objective can be flexibly adjusted according to the transient state of the system, balancing the dual objectives of structural safety and water conservation in real time, ensuring that the system does not suffer structural damage while minimizing drainage losses. Based on this, this invention proposes a model predictive control method and system for the open-flow process of canals and tunnels. Summary of the Invention

[0006] The technical problem this invention aims to solve is to address the shortcomings of existing emergency control systems for water conveyance systems by proposing a model-predictive control method for open-flow processes in canals and tunnels. This control method, based on model predictive control, can be used for emergency scheduling of canal-tunnel water conveyance systems. Through a customized model predictive control scheme, it reduces the impact of complex open-flow processes on system safety and minimizes unnecessary waste of runoff.

[0007] To address the aforementioned technical problems, the present invention adopts the following technical solution: A model-predictive-based method for controlling full-flow processes in a channel-tunnel includes the following steps: Determine the safe operation limits for each open channel and tunnel within the accident dispatch section, as state constraints. Based on the aforementioned state constraints, the number of channels and pools required for dispatching is estimated according to the location and scale of the sudden accident, as well as the available storage capacity of each upstream channel and pool, and the scope of the accident dispatching segment is divided. Read real-time monitoring data of each canal and pool within the accident dispatch section, quantify the system fluctuation state using hydraulic fluctuation evaluation index, establish an optimization problem considering safety and water conservation, determine the safety weight and water conservation weight in the optimization problem based on the fluctuation index, and solve to obtain the optimal control objective; A full-flow prediction model is established based on the channel parameters of the accident dispatch section; Using the optimal control objective as input, the Mingman Flow Prediction Model is used to predict the hydrodynamic state variables under a given gate action sequence. The action sequence of each gate in the accident scheduling section is used as the decision variable. At the same time, gate action constraints are set to establish a predictive optimization control problem and solve it to obtain the future control action sequence of each gate. Each gate executes the first control action in the future control action sequence, waits for the control interval, updates the initial state of the full flow prediction model, and repeats the above steps.

[0008] Furthermore, the state constraint condition is:

[0009] in, , Representing canals and pools respectively i Maximum safe storage capacity and current storage capacity, in cubic meters (m³) 3 ; , These represent the upstream inflow and downstream outflow of the accident dispatch segment at the moment the accident occurred, respectively, in meters (m). 3 / s; , These represent the closing times of the upstream and downstream gates in the accident dispatching section, respectively, in seconds. Indicates the various water distribution points during the accident dispatch period. r Flow rate, unit m 3 / s; Indicates the duration of the accident dispatch period, in seconds (s). Indicates the number of canals and pools in the accident dispatch section; Indicates the number of water distribution points in the accident dispatch section; This indicates the water outlet number of the accident dispatch section.

[0010] Furthermore, the hydraulic fluctuation evaluation index is as follows:

[0011] in, Indicates the current time Canal Pond i Hydraulic fluctuation evaluation index, unit: m; Indicates the canal pool number; Indicates the current time, in seconds (s); This indicates the width of the sliding window, in seconds (s). Represents the time integral variable, in seconds (s). Indicates canal pool i Water level at the downstream control point, in meters (m). Indicates the current time The average water level at the downstream control point of the canal pool during the previous sliding window period, in meters.

[0012] Furthermore, the optimization problem considering safety and water conservation is as follows:

[0013] in, Indicates canal pooli The objective function used to address optimization problems that consider both safety and water conservation; Indicates canal pool i The safety weight varies with the time-varying hydrodynamic fluctuations and meets the following requirements. ; Indicates the canal pool i Water conservation weight; Indicates canal pool i The target water level at the downstream control point, in meters (m). Indicates the canal / pool in still water condition i The water level at the downstream control point is The storage capacity of the canal at that time, m 3 ; Indicates canal pool i The safety margin that the target water level setting must meet, in meters; Indicates canal pool i Safety limits on water head.

[0014] Furthermore, the open channel flow prediction model is an implicit volume method discrete Saint-Venant prediction model. The implicit volume method discrete Saint-Venant prediction model is based on the Pressman slit method to unify the control equations of open channel flow and pressure flow. The control equations are numerically discretized using the finite volume method, and Taylor expansion is performed on the flux and source terms of the equations.

[0015] Furthermore, the Mingman flow prediction model predicts the hydrodynamic processes of each canal and pool within the prediction domain, and within a single prediction step, the canal and pool... Medium grid The state variables are updated as follows:

[0016] in, , These represent the channel pool number and the discrete grid number of the open flow prediction model, respectively. This represents the change in hydrodynamic state parameters within a single-step prediction time interval. These represent the coefficient matrices for updating the state variables of the Mingman flow prediction model. The initial conditions of the Mingman flow prediction model are the water level and flow sensor measurements of each canal and pool in the scheduling section at the start of the prediction; the inner and outer boundary conditions are the flow rates through each gate and the flow rates at the water distribution outlets in the scheduling section.

[0017] Furthermore, the predictive optimization control problem is:

[0018] in, This represents the objective function used for gate action optimization; Indicates the length of the prediction domain, in seconds; This represents the water level sequence at the downstream control point of each canal and pool within the prediction domain, in meters (m). Indicates each canal and pool i The target water level at the downstream control point, in meters (m). This indicates the penalty for exceeding the peak value of each canal / pool within the prediction domain; , , These represent the penalty weights for the water level deviation path term, terminal term, and peak term, respectively.

[0019] Furthermore, the penalty for exceeding the limit for each canal and pool within the predicted domain is as follows:

[0020] in, This indicates the safety head limit for each canal and pool.

[0021] Furthermore, a nonlinear optimization solver is used to solve the predictive optimization control problem.

[0022] On the other hand, the present invention provides a model-predictive channel-tunnel full-flow process control system, comprising: The constraint determination module is used to determine the safe operation limits of each open channel and tunnel within the accident dispatch section, as state constraint conditions. The scheduling segment division module is used to estimate the number of channels and pools required to participate in the scheduling based on the state constraints, the location and scale of the sudden accident, and the available storage capacity of each upstream channel and pool, and to divide the scope of the accident scheduling segment. The target optimization module is used to read real-time monitoring data of each canal and pool within the accident dispatch section, quantify the system fluctuation state using hydraulic fluctuation evaluation index, establish an optimization problem considering safety and water conservation, determine the safety weight and water conservation weight in the optimization problem based on the fluctuation index, and solve for the optimal control target. The prediction model building module is used to establish a full-flow prediction model based on the channel parameters of the accident scheduling section. The predictive optimization control module is used to predict the hydrodynamic state variables under a given gate action sequence using the optimal control objective as input and the Mingman flow prediction model. It also uses the action sequence of each gate in the accident scheduling section as decision variables, sets gate action constraints, establishes a predictive optimization control problem, and solves the future control action sequence of each gate. The execution module is used to execute the first control action in the future control action sequence for each gate, wait for the control interval, update the initial state of the full flow prediction model, and repeat the above steps.

[0023] Compared with the prior art, the present invention has the following beneficial effects: (1) There is currently no predictive model that can be used for the control of open full flow process in canal-tunnel systems. The proposed implicit volume method discrete Saint-Venant model is of great significance for the fine control of canal-tunnel systems under emergency conditions.

[0024] (2) The present invention uses a predictive model to predict and control the open flow process of the canal-tunnel system. The objective function considers water level deviation and over-limit penalty, which helps to avoid the risk of overflow in the open channel section and pressure over-limit in the tunnel section caused by blind operation.

[0025] (3) The present invention fully considers the real-time hydraulic fluctuation state of the system, takes into account both structural safety and water-saving economy, and dynamically adjusts the control target, which can significantly reduce the scheduling safety risk or unnecessary water loss caused by empirical fixed control target. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the channel-tunnel series system and the open flow process control method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the alternating open and full flow water conveyance system according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the discretization method of the prediction model for the canal pool in an embodiment of the present invention; Figure 4 This is a diagram showing the gate operation process results according to an embodiment of the present invention; Figure 5 These are diagrams showing the hydrodynamic process results of various channels and pools in embodiments of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0028] Example 1 like Figure 1 As shown, this embodiment provides a model-predictive-based method for controlling the full-flow process in a channel-tunnel. This embodiment is based on... Each control gate is divided into There are several series of canals and pools, each of which can be an open channel or a tunnel. Each canal and pool is equipped with a water distribution outlet and a drainage gate. The control gate and drainage gate participate in emergency regulation. During the regulation process, the pressure process in the tunnel exhibits strong nonlinearity and strong transient characteristics. Furthermore, due to the strong hydraulic coupling between the series of canals and pools, they must be treated as a canal-tunnel system for unified emergency regulation.

[0029] A centralized model predictive control scheme is adopted for controller design, using a target optimization-trajectory tracking control framework. The target optimization layer establishes an optimization problem considering both safety and water conservation objectives to optimize the target water levels of each canal and pool within the scheduling section, obtaining a more suitable control target. Due to the open-full flow transition process in the canal-tunnel system, significant hydrodynamic fluctuations occur. To ensure safety during these fluctuations, the control target often needs to be set at a lower value to maintain a larger safety margin. However, blindly maintaining a low target water level after the water level gradually stabilizes would waste the system's safety storage capacity and cause unnecessary water loss during drainage. Based on the contradiction between operational safety and economy in the control target setting, the real-time hydrodynamic fluctuation state of the system is incorporated into the control target optimization problem. This allows for lowering the target water level to ensure safety when the system is unstable, and raising the target water level to store water and reduce losses when the system tends towards a steady state. In practice, the current hydraulic fluctuation index of the system is evaluated by measuring the hydrodynamic state data through sensors deployed in the water conveyance canal-tunnel system. Based on the fluctuation index, the safety weight and water-saving weight of the optimization problem are calculated and adjusted. Solving the optimization problem yields the optimal control objective that conforms to the current system state. Specifically, the following steps are included: Step 1: Determine the safe operation limits of each open channel and tunnel within the accident dispatch section as state constraints; Step 2: Based on the aforementioned state constraints, estimate the number of channels and pools required for dispatching according to the location and scale of the sudden accident and the available storage capacity of each upstream channel and pool, and delineate the scope of the accident dispatching segment. Step 3: Read the real-time monitoring data of each canal and pool in the accident dispatch section, quantify the system fluctuation state using the hydraulic fluctuation evaluation index, establish an optimization problem that considers safety and water conservation, determine the safety weight and water conservation weight in the optimization problem based on the fluctuation index, and solve to obtain the optimal control objective; Step 4: Establish a full-flow prediction model based on the channel parameters of the accident dispatch section; Step 5: Using the optimal control objective as input, the Mingman Flow Prediction Model is used to predict the hydrodynamic state variables under the given gate action sequence. The action sequence of each gate in the accident scheduling section is used as the decision variable. At the same time, gate action constraints are set to establish a predictive optimization control problem and solve it to obtain the future control action sequence of each gate. Step 6: Each gate executes the first control action in the future control action sequence, waits for the control interval, updates the initial state of the full flow prediction model, and repeats the above steps.

[0030] In step 2 of this embodiment, based on the safety requirements of the upstream canal and the closing rate limits of each gate, the current flow rate of the upstream and downstream gates in the emergency dispatch section, i.e., the time required for safe gate closure, is determined. The canal and pool area of ​​the emergency dispatch section is divided according to the scale of the accident. In principle, the current available safe storage capacity of the dispatch section should be close to the difference between the inlet and outlet water volume during the emergency dispatch period, i.e.:

[0031] in, , Representing canals and pools respectively i Maximum safe storage capacity and current storage capacity, in cubic meters (m³) 3 ; , These represent the upstream inflow and downstream outflow of the accident dispatch segment at the moment the accident occurred, respectively, in meters (m). 3 / s; , These represent the closing times of the upstream and downstream gates in the accident dispatching section, respectively, in seconds. Indicates the various water distribution points during the accident dispatch period. r Flow rate, unit m 3 / s; Indicates the duration of the accident dispatch period, in seconds (s). Indicates the number of canals and pools in the accident dispatch section; Indicates the number of water distribution points in the accident dispatch section; This indicates the water outlet number of the accident dispatch section.

[0032] In this embodiment, water level sensor data from each canal and pool upstream of the accident point are read, and the water level of each canal and pool is calculated based on the water level lines of each canal and pool. i Current stock According to the safety head limit of each canal and pool Calculation of canals i Maximum safe storage capacity in still water conditions The formula for calculating storage capacity is as follows:

[0033] In the formula: , These represent the canal / pool number and the sensor number, respectively. Indicates canal pool i water level sensor Observations, in meters; This represents the cross-sectional area of ​​the water passage, calculated based on the cross-sectional geometric parameters and water level, with the unit being m. 2 ; This indicates the average spacing between water level sensors, in meters (m). This indicates the number of water level sensors in a single canal.

[0034] Step 3 in this embodiment includes: (1) Read sensor data and calculate hydraulic fluctuation index:

[0035] in, Indicates the current time Canal Pond i Hydraulic fluctuation evaluation index, unit: m; Indicates the canal pool number; Indicates the current time, in seconds (s); This indicates the width of the sliding window, in seconds (s). Represents the time integral variable, in seconds (s). Indicates canal pool i Water level at the downstream control point, in meters (m). Indicates the current time The average water level at the downstream control point of the canal pool during the previous sliding window period, in meters.

[0036] (2) Establish an optimization problem considering safety and water conservation and solve it to obtain the control objective: The safety item considers the safety margin between the target water level and the upper limit of safety, while the water conservation item penalizes any unused safe storage capacity.

[0037] in, Indicates canal pool i The objective function used to address optimization problems that consider both safety and water conservation; Indicates canal pool i The safety weight varies with the time-varying hydrodynamic fluctuations and meets the following requirements. ; Indicates the canal pool i Water conservation weight; Indicates canal pool i The target water level at the downstream control point, in meters (m). Indicates the canal / pool in still water condition i The water level at the downstream control point is The storage capacity of the canal at that time, m 3 ; Indicates canal pool i The safety margin that the target water level setting must meet, in meters; Indicates canal pool i Safety limits on water head.

[0038] In step 4 of this embodiment, establishing the Mingman flow prediction model includes: Figure 2This diagram illustrates an alternating open and closed flow system, where both pressurized and unpressurized flow regimes coexist within the system. The boundary between open and closed flow is uncertain, and the flow regime at a cross-section may switch between these two states. The different governing equations for the two flow regimes make it difficult to predict the system's hydrodynamic state. In open and closed flow hydrodynamic simulation techniques, the Plesman slit method is often used to unify the governing equations.

[0039]

[0040] In the formula: , , These represent hydrodynamic state variables, fluxes, and source terms, respectively. The cross-sectional area of ​​the water passage is expressed in meters (m²). 2 ; The flow rate of a cross-section is expressed in cubic meters (m³). 3 / s; This represents the integral term of hydrostatic pressure at the cross section, in meters (m). 3 ; Expresses gravitational acceleration, unit m / s². 2 ; Indicates the bottom slope of the channel; dimensionless. Indicates hydraulic gradient, dimensionless; For unified form control equations, the volume method is often used to solve them in order to ensure conservation and stability. Here, based on the implicit volume method, the limitation of the Cronough condition on the time step can be overcome, and a discrete Saint-Venant model that can be used for predictive control is established.

[0041] Figure 3 This is a schematic diagram of the discretization method for the canal pool in the prediction model, where a single canal pool is spatially discretized as follows: For each spatially discrete point, the hydrodynamic state variables are updated using the implicit volume method:

[0042] In the formula: Indicates the discrete point number; Indicates the time corresponding to the state variable; Represents the distance between adjacent discrete points in space, in meters (m). This indicates the time interval for updating state variables, in seconds (s). Represents the numerical viscosity coefficient, according to Calculation, where , These are the cross-sectional flow velocity and the wave velocity, respectively. To avoid the iterative computation required for implicit updates of state variables, a Taylor expansion is performed on flux and source terms:

[0043] Finally, the implicit scheme volume method discrete Saint-Venant model is obtained:

[0044] In the formula: Represents the identity matrix; Further derivation of the discrete Saint-Venant model yields the prediction model:

[0045] In the formula: express The state update matrix at time step 1.

[0046] In step 5 of this embodiment, the control target from step 3 is used as a reference trajectory input to the model predictive control layer to optimize the control actions. The optimized control objective is input into the model predictive control layer as the controller's reference trajectory. A model predictive control optimization problem is established to optimize the control actions and track the reference trajectory. The model predictive control layer uses the control gate 1 within the prediction domain... The action sequence of +1 and the flow process of each branch outlet serve as feedforwards, with the control gate 2~ and drainage gate 1~ Given the sequence of actions as decision variables, establish a control optimization problem:

[0047] st

[0048] in, This represents the objective function used for gate action optimization; Indicates the length of the prediction domain, in seconds; This represents the water level sequence (matrix format) of the downstream control points of each canal and pool within the prediction domain, in meters (m). Indicates each canal and pool i Target water level at the downstream control point (vector format), in meters; This represents the penalty term (vector form) for exceeding the peak value of each canal / pool within the prediction domain. , , These represent the penalty weights for the water level deviation path term, terminal term, and peak term, respectively. Indicates control gate The opening degree, in meters; , They represent the control gates respectively. Maximum and minimum opening, in meters; Indicates the sequence number of the control gate action; Indicates control gate Maximum opening and closing speed, in m / s; Indicates the interval between control actions of the gate, in seconds; Indicates the drainage gate The discharge flow rate, in meters 3 / s; , These represent the drainage gates. Maximum and minimum discharge flow rates, in m³ 3 / s.

[0049] The penalty for exceeding the peak value of each canal and pool within the predicted domain is as follows:

[0050] in, This indicates the safety head limit for each canal and pool.

[0051] Furthermore, a nonlinear optimization solver is used to solve the predictive optimization control problem.

[0052] This embodiment uses a canal-tunnel system from a project in Yunnan Province as an example to conduct a control performance simulation test of the control framework. The test system includes three series-connected canals and pools with a total length of 40 km, and both upstream and downstream boundaries are constant water level boundaries. The design geometry and hydraulic parameters of the canal-tunnel system are shown in Table 1.

[0053] Table 1. Geometric and hydraulic parameters for the design of the example canal-tunnel system.

[0054] Taking a sudden accident downstream of the system as an example, the downstream gate needs to be closed within 10 minutes to ensure rapid cutoff of water flow; at the same time, to ensure that the upstream normal operation of the canals and pools is not affected by the accident scheduling, the gate closing speed of the upstream canal and pool in the scheduling section cannot be too fast, and is set to 3 hours. The operating conditions are shown in Table 2.

[0055]

[0056] Based on the above information and data, a simulation of the open flow process control of the canal-tunnel system was conducted.

[0057] Figure 4 and Figure 5The diagrams show the operational processes of each gate in the emergency control system of the canal-tunnel system and the water level control effects of each canal pool. Based on the operating conditions, the closing times for control gates 1 and 4 are 3 hours and 10 minutes, respectively. To ensure the upstream water flow in the affected section is cut off, control gates 2 and 3 are also gradually closed during the scheduling period. To ensure the utilization of the canal pool's storage capacity, the inflow rate to all three canals is higher than the outflow rate; specifically, the gate operation is characterized by the upstream control gate having a longer closing time than the downstream control gate. Furthermore, the control gates can only distribute water among the canals and pools; any water exceeding the safe storage capacity of the canals and pools must be discharged through the drainage gates to ensure the safety of the canals and pools. The flow process lines of each sluice gate are all "n"-shaped curves: In the early stage of emergency dispatch, the downstream gates are closed urgently, and the upstream water flow cannot be cut off in time, resulting in a large amount of water remaining in the dispatch section canal pool. At this time, it is necessary to open the sluice gates to discharge some water to prevent the water head of the dispatch section canal pool from exceeding the limit. In the later stage of emergency dispatch, the upstream flow gradually decreases and the water level gradually stabilizes. The target water level obtained by the target optimization layer rises, and the canal pool needs to store water. At this time, the sluice gates are gradually closed.

[0058] Throughout the scheduling period, the target head of each canal and pool varied with the hydraulic fluctuations of the system. In the initial stage of scheduling, due to the rapid changes in boundary conditions at the gates, the hydraulic state within the canals and pools was unstable, and the target water level was lowered to ensure structural safety during the forecast period. Subsequently, the hydraulic fluctuations gradually decreased, and the target water level rose, prompting the canals and pools to store water and reducing the waste of water caused by the outflow. Throughout the entire scheduling process, the head of each canal and pool did not exceed the safe upper limit, meeting the safety requirements of the open channel section and the tunnel section.

[0059] This invention proposes a model-based control framework for open-flow processes in canals and tunnels. For canal-tunnel systems requiring emergency dispatch, it establishes an open-flow prediction model that meets real-time control requirements. Adaptive target control is employed, adjusting the target head in real-time based on system hydraulic fluctuations. Through model prediction and rolling optimization, the real-time target trajectory is tracked, ensuring that the system's hydrodynamic state consistently meets structural safety and water-saving requirements. This enhances the dispatch safety and operational resilience of water conveyance systems, significantly contributing to the resilience of the national water network and ensuring water-saving economic benefits.

[0060] Example 2 This embodiment provides a model-predictive channel-tunnel full-flow process control system, including: The constraint determination module is used to determine the safe operation limits of each open channel and tunnel within the accident dispatch section, as state constraint conditions. The scheduling segment division module is used to estimate the number of channels and pools required to participate in the scheduling based on the state constraints, the location and scale of the sudden accident, and the available storage capacity of each upstream channel and pool, and to divide the scope of the accident scheduling segment. The target optimization module is used to read real-time monitoring data of each canal and pool within the accident dispatch section, quantify the system fluctuation state using hydraulic fluctuation evaluation index, establish an optimization problem considering safety and water conservation, determine the safety weight and water conservation weight in the optimization problem based on the fluctuation index, and solve for the optimal control target. The prediction model building module is used to establish a full-flow prediction model based on the channel parameters of the accident scheduling section. The predictive optimization control module is used to predict the hydrodynamic state variables under a given gate action sequence using the optimal control objective as input and the Mingman flow prediction model. It also uses the action sequence of each gate in the accident scheduling section as decision variables, sets gate action constraints, establishes a predictive optimization control problem, and solves the future control action sequence of each gate. The execution module is used to execute the first control action in the future control action sequence for each gate, wait for the control interval, update the initial state of the full flow prediction model, and repeat the above steps.

[0061] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0062] Obviously, those skilled in the art can make various modifications and variations to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations to the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention also intends to include these modifications and variations.

[0063] All other parts not described in detail are existing technologies.

Claims

1. A method for controlling a channel-tunnel open-air full-flow process based on model prediction, characterized in that, Includes the following steps: Determine the safe operation limits for each open channel and tunnel within the accident dispatch section, as state constraints. Based on the aforementioned state constraints, the number of channels and pools required for dispatching is estimated according to the location and scale of the sudden accident, as well as the available storage capacity of each upstream channel and pool, and the scope of the accident dispatching segment is divided. Real-time monitoring data from each canal and pool within the accident dispatch section are read. The system fluctuation state is quantified using hydraulic fluctuation evaluation indices. An optimization problem considering safety and water conservation is established. Based on the fluctuation indices, safety and water conservation weights are determined in the optimization problem, and the optimal control objective is obtained by solving the problem. The hydraulic fluctuation evaluation indices are: in, Indicates the current time Canal Pond i Hydraulic fluctuation evaluation index, unit: m; Indicates the canal pool number; Indicates the current time, in seconds (s); This indicates the width of the sliding window, in seconds (s). Represents the time integral variable, in seconds (s). Indicates canal pool i Water level at the downstream control point, in meters (m). Indicates the current time The average water level at the downstream control point of the canal during the previous sliding window period, in meters; The optimization problem considering safety and water conservation is as follows: in, Indicates canal pool i The objective function used to address optimization problems that consider both safety and water conservation; Indicates canal pool i The safety weight varies with the time-varying hydrodynamic fluctuations and meets the following requirements. ; Indicates the canal pool i Water conservation weight; Indicates canal pool i The target water level at the downstream control point, in meters (m). Indicates the canal / pool in still water condition i The water level at the downstream control point is The storage capacity of the canal at that time, m 3 ; Indicates canal pool i The safety margin that the target water level setting must meet, in meters; Indicates canal pool i Safety limit head; A full-flow prediction model is established based on the channel parameters of the accident dispatch section; Using the optimal control objective as input, the Mingman Flow Prediction Model is used to predict the hydrodynamic state variables under a given gate action sequence. The action sequence of each gate in the accident scheduling section is used as the decision variable. At the same time, gate action constraints are set to establish a predictive optimization control problem and solve it to obtain the future control action sequence of each gate. Each gate executes the first control action in the future control action sequence, waits for the control interval, updates the initial state of the full flow prediction model, and repeats the above steps.

2. The method for controlling channel-tunnel full-flow process based on model prediction according to claim 1, characterized in that, The state constraint condition is: in, , Representing canals and pools respectively i Maximum safe storage capacity and current storage capacity, in cubic meters (m³) 3 ; , These represent the upstream inflow and downstream outflow of the accident dispatch segment at the moment the accident occurred, respectively, in meters (m). 3 / s; , These represent the closing times of the upstream and downstream gates of the accident dispatch section, respectively, in seconds. Indicates the various water distribution points during the accident dispatch period. r Flow rate, unit m 3 / s; Indicates the duration of the accident dispatch period, in seconds (s). Indicates the number of canals and pools in the accident dispatch section; Indicates the number of water distribution points in the accident dispatch section; This indicates the water outlet number of the accident dispatch section.

3. The method for controlling the full-flow process of a channel-tunnel based on model prediction according to claim 1, characterized in that, The open channel flow prediction model is an implicit volumetric method discretized Saint-Venant prediction model. The implicit volumetric method discretized Saint-Venant prediction model is based on the Pressman slit method to unify the control equations of open channel flow and pressure flow. The control equations are numerically discretized using the finite volume method, and the flux and source terms of the equations are Taylor expanded.

4. The method for controlling the full-flow process of a channel-tunnel based on model prediction according to claim 3, characterized in that, The Mingman flow prediction model predicts the hydrodynamic processes of each canal and pool within the prediction domain, and predicts the canal and pool within a single prediction step. Medium grid The state variables are updated as follows: in, , These represent the channel pool number and the discrete grid number of the open flow prediction model, respectively. This represents the change in hydrodynamic state parameters within a single-step prediction time interval. These represent the coefficient matrices for updating the state variables of the Mingman flow prediction model. The initial conditions of the Mingman flow prediction model are the water level and flow sensor measurements of each canal and pool in the scheduling section at the start of the prediction; the inner and outer boundary conditions are the flow rates through each gate and the flow rates at the water distribution outlets in the scheduling section.

5. The method for controlling channel-tunnel full-flow process based on model prediction according to claim 1, characterized in that, The predictive optimization control problem is: in, This represents the objective function used for gate action optimization; Indicates the length of the prediction domain, in seconds; This represents the water level sequence at the downstream control point of each canal and pool within the prediction domain, in meters (m). Indicates each canal and pool i The target water level at the downstream control point, in meters (m). This indicates the penalty for exceeding the peak value of each canal / pool within the prediction domain; , , These represent the penalty weights for the water level deviation path term, terminal term, and peak term, respectively.

6. The method for controlling the full-flow process of a channel-tunnel based on model prediction according to claim 5, characterized in that, The penalty for exceeding the peak value of each canal and pool within the predicted domain is as follows: in, This indicates the safety head limit for each canal and pool.

7. The method for controlling the full-flow process of a channel-tunnel based on model prediction according to claim 5, characterized in that, The predictive optimization control problem is solved using a nonlinear optimization solver.

8. A model-predictive channel-tunnel open flow process control system, characterized in that, include: The constraint determination module is used to determine the safe operation limits of each open channel and tunnel within the accident dispatch section, as state constraint conditions. The scheduling segment division module is used to estimate the number of channels and pools required to participate in the scheduling based on the state constraints, the location and scale of the sudden accident, and the available storage capacity of each upstream channel and pool, and to divide the scope of the accident scheduling segment. The target optimization module is used to read real-time monitoring data of each canal and pool within the accident dispatch section, quantify the system fluctuation state using hydraulic fluctuation evaluation index, establish an optimization problem considering safety and water conservation, determine the safety weight and water conservation weight in the optimization problem based on the fluctuation index, and solve for the optimal control target. The prediction model building module is used to establish a full-flow prediction model based on the channel parameters of the accident scheduling section. The predictive optimization control module is used to predict the hydrodynamic state variables under a given gate action sequence using the optimal control objective as input and the Mingman flow prediction model. It also uses the action sequence of each gate in the accident scheduling section as decision variables, sets gate action constraints, establishes a predictive optimization control problem, and solves the future control action sequence of each gate. The execution module is used to execute the first control action in the future control action sequence for each gate, wait for the control interval, update the initial state of the Mingman flow prediction model, and repeat the above steps. The model-predictive open channel-tunnel full flow process control system is used to execute the steps in the model-predictive open channel-tunnel full flow process control method according to any one of claims 1-7.

Images