Intelligent flow distribution control method and system of self-supporting gate

Abstract

The invention relates to the technical field of automatic adjustment of water distribution, in particular to an intelligent flow distribution control method and system for a self-supporting type flashboard gate, and the method comprises the steps: obtaining liquid level data, carrying out the preprocessing of the liquid level data, calculating the theoretical value of the initial height of a water head on a weir according to the obtained liquid level data, and carrying out the calculation of the theoretical value. Adjusting the opening height of the downward-opening type gate according to the theoretical value, and adjusting the adjusting amount of the downward-opening type gate through the fuzzy rule base so as to adjust the height of the flow cross section; initial flow distribution is carried out based on the adjusted opening height, a state space and an action space are defined, the opening number of self-supporting flashboards is adjusted through a reinforcement learning algorithm, and the height of the flow cross section is adjusted; according to the flow distribution method, the target function is constructed and the optimization problem is solved by combining the multivariable predictive control algorithm with the state space model, so that the accuracy of flow distribution is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic water volume distribution, and in particular to an intelligent flow distribution control method and system for a self-supporting flap gate. Background Art

[0002] In the current water treatment field, efficient and accurate water volume distribution is crucial for ensuring the stable operation of the entire treatment process, improving the treatment effect, and reducing costs. When the traditional water treatment process performs multi-point water distribution of the total influent, it is severely restricted by various factors. From the perspective of channel layout, complex terrain and spatial limitations make it difficult to design and construct the water distribution channels according to the ideal hydraulic conditions. For example, in some areas with large terrain undulations, it is difficult to accurately control the channel slope, which easily leads to uneven water flow velocities, thereby affecting the flow distribution uniformity at each water distribution point. In a treatment site with limited space, the orientation and layout of the channels often need to accommodate other facilities, and it is impossible to achieve the optimal water distribution path planning.

[0003] In addition, the existing flow regulation methods have obvious deficiencies in terms of timeliness and accuracy. In real-time monitoring, the commonly used level gauges and flow meters have limited accuracy and slow response speeds, and are unable to capture water flow changes in a timely manner, resulting in a regulatory lag and making it difficult to meet the requirements for rapid flow adjustment during the water treatment process. As a result, the flow rates at each water distribution point cannot be adjusted in a timely manner, affecting the treatment effect. In terms of accuracy, the existing technical means are difficult to achieve precise flow distribution under complex working conditions. The traditional flow calculation models are too simplified and do not fully consider the actual characteristics of water flow, such as turbulent flow and boundary layer effects, resulting in a large deviation between the calculation results and the actual flow rates. In actual operation, affected by various factors such as the roughness of the channel wall and the change of water flow pressure, the traditional flow regulation methods cannot effectively compensate for these factors, making it difficult for the flow rates at each water distribution point to reach the designed proportion, affecting the operation efficiency and treatment quality of the entire water treatment system. At present, an intelligent flow distribution control method and system for a self-supporting flap gate are needed. Summary of the Invention

[0004] In order to solve the problems of low accuracy and uneven flow distribution in water volume distribution, the present invention provides an intelligent flow distribution control method and system for a self-supporting flap gate.

[0005] In the first aspect, an intelligent flow distribution control method for a self-supporting flap gate provided by the present invention adopts the following technical solutions:

[0006] An intelligent flow distribution control method for a self-supporting flap gate includes:

[0007] Obtain liquid level data and preprocess the liquid level data;

[0008] Calculate the theoretical value of the initial head height on the weir based on the acquired liquid level data, including calculating the theoretical value of the height by dynamically adjusting parameters using the turbulent flow correction coefficient and roughness coefficient;

[0009] Adjust the opening height of the bottom-opening gate according to the theoretical value, and adjust the adjustment amount of the bottom-opening gate through the fuzzy rule base;

[0010] Perform preliminary flow distribution based on the adjusted opening height, including defining the state space and action space, and adjusting the opening state of the gate plate using the reinforcement learning algorithm;

[0011] Use the preliminary distribution result for multi-parameter collaborative adjustment, including performing precise flow adjustment using the multivariable predictive control algorithm;

[0012] Feedback and optimize the control result of the flow adjustment.

[0013] Further, the calculating the theoretical value of the initial head height on the weir based on the acquired liquid level data includes calculating the Reynolds number according to the hydraulic diameter and the dynamic viscosity of the fluid, calculating the turbulent flow correction coefficient and the roughness coefficient dynamic adjustment parameter respectively based on the Reynolds number and the Manning coefficient of the current channel, and calculating the theoretical value of the initial head height on the weir using the theoretical formula, and the theoretical formula is expressed as:

[0014] h 理论 =H - h1 - h2 + Δh 紊流 +Δh 糙率 ,

[0015] Wherein, H represents the liquid level data, h1 represents the installation height of the water distribution hole, h2 represents the opening height of the bottom-opening gate, Δh 紊流 represents the turbulent flow correction coefficient, and Δh 糙率 represents the roughness coefficient dynamic adjustment parameter.

[0016] Further, the adjusting the opening height of the bottom-opening gate according to the theoretical value includes determining the target value of the adjustment height according to the theoretical value, calculating the real-time liquid level deviation using the liquid level data, and calculating the deviation change rate using the difference method, defining the fuzzy linguistic variables and fuzzy subsets of the liquid level deviation and the deviation change rate, determining the membership degree through the membership function, constructing the fuzzy rule base and determining the fuzzy output of the adjustment amount of the bottom-opening gate through the fuzzy synthesis operation.

[0017] Further, the adjusting the adjustment amount of the bottom-opening gate through the fuzzy rule base includes solving the fuzzy output using the centroid method to obtain the adjustment amount of the bottom-opening gate, and the calculation formula of the adjustment amount of the bottom-opening gate is:

[0018]

[0019] Wherein, μ iDenoted as the membership degree of the regulation amount fuzzy subset i, x i Denoted as the quantization value corresponding to the fuzzy subset.

[0020] Furthermore, the method for adjusting the opening state of the gate using the reinforcement learning algorithm includes: applying the deep Q-network reinforcement learning algorithm to find the optimal gate opening strategy, approximately estimating the Q-value function by constructing a neural network, at each time step, selecting an action according to the current state, and calculating the reward value according to the feedback value after executing the action. The reward value is determined according to the matching degree between the flow distribution result and the target flow and the water level situation. The reward value calculation formula is:

[0021]

[0022] where q i Denoted as the actual water inflow of the i-th water tank, Denoted as the target water inflow, h i Denoted as the real-time water level of the i-th water tank, Denoted as the optimal water level, and n represents the total number of water tanks.

[0023] Furthermore, the method for performing multi-parameter collaborative regulation using the preliminary distribution result further includes defining a system state vector and a control input vector based on the historical data and actual operation data of the water distribution system, and then establishing a linear discrete-time state space model. According to the state space model, the output equation is determined. The state space model formula is:

[0024] x(k + 1) = Ax(k) + Bu(k) + w(k),

[0025] where A is denoted as the m×m state transition matrix, B is denoted as the m×p input matrix, which reflects the influence of the control input on the system state, w(k) is denoted as the process noise vector, used to consider the uncertainties and interference factors in the system, x(k) is denoted as the system state vector, u(k) is denoted as the control input vector, m represents the number of state variables, and p represents the number of control inputs.

[0026] Furthermore, the method for performing precise flow regulation using the multivariable predictive control algorithm includes using the multivariable predictive control algorithm, setting the prediction horizon N p and the control horizon N c , combining the output equation of the state space model to construct an objective function, and solving the optimization problem of the objective function at each sampling moment to obtain the optimal control input sequence for the next N c time steps, and using the optimal control input sequence for precise flow regulation.

[0027] Further, the feedback and optimization of the control result of the flow regulation include calculating an error signal based on the output of the state space model and the actual output of the system, and making adjustments according to the error signal. The adjustment law is expressed as:

[0028]

[0029] where represents the change rate of the parameter vector at time k, Γ represents the adaptive gain matrix, represents the regression vector, including information on the system state and input, and e T (k) represents the transpose of the error signal vector.

[0030] In a second aspect, an intelligent flow distribution control system for a self-supporting flap gate includes:

[0031] Data acquisition module: including a liquid level gauge, which is installed on one side of the water distribution channel near each water distribution hole. The liquid level gauge is used to monitor the liquid level data in the water distribution channel in real time to calculate the head height over the weir, providing a basis for the automatic control system to judge the target water distribution point;

[0032] Water flow channel module: including a water distribution channel and a water tank. The water distribution channel serves as the main inlet channel and is located at the starting end of the system, carrying all the water to be distributed and providing a water source for subsequent water tanks. There are usually multiple water tanks or compartments, which are the target objects of water volume distribution. The required water inflow for each water tank is different or in a changing state. The water distribution channel is connected to each water tank through water distribution holes;

[0033] Gate control module: including a self-supporting flap gate and a bottom-opening gate. The self-supporting flap gate is installed on one side of the water distribution channel outside the water distribution hole and is composed of multiple gate plates, a gate frame, a gate rod, a hoist, a bushing, a sealing cover, a vent hole, and a water inlet hole. The gate plates can be opened independently, and the water inlet width is controlled by changing the number of opened gate plates;

[0034] Processing module: including an automatic control system. The automatic control system is used to receive the data transmitted by the liquid level gauge, and after analysis and processing, according to the preset algorithm and logic, send control signals to the self-supporting flap gate and the bottom-opening gate to achieve precise control of the gate opening, adjust the height and width of the water passing section, and thus regulate the flow rate.

[0035] Furthermore, the gate frame and the gate plate are made of lightweight materials such as engineering plastics. The inside of the gate plate is hollow, and the buoyancy is overcome by the incoming water to reduce the power and power consumption of the hoist. The sealing cover is used to prevent the odor in the pool from emitting, and a transparent window is provided on the shaft sleeve for observing the opening height of the gate rod. The bottom-opening gate is installed on one side of the pool inside the water distribution hole for controlling the incoming water height. The hoist of the self-supporting flap gate and the driving device of the bottom-opening gate are connected to the automatic control system through control lines to receive control signals.

[0036] In summary, the present invention has the following beneficial technical effects:

[0037] 1. The present invention can accurately collect liquid level data by obtaining liquid level data in real time through a liquid level gauge and preprocessing it, providing a reliable basis for subsequent flow regulation. When calculating the theoretical value of the initial head height over the weir, the turbulent flow correction coefficient and the roughness dynamic adjustment parameter are considered, making the calculation result more in line with the actual water flow situation. Compared with the traditional calculation method, the calculation accuracy is improved, and thus the accuracy of flow distribution is enhanced.

[0038] 2. The present invention adjusts the opening height of the bottom-opening gate by using a fuzzy rule base and a PID control algorithm, and can flexibly and accurately adjust the gate adjustment amount according to the real-time liquid level deviation and the deviation change rate, avoiding over-adjustment or under-adjustment.

[0039] 3. The present invention uses the deep Q network in the reinforcement learning algorithm, estimates the Q value function by constructing a neural network, determines the reward value according to the matching degree between the flow distribution result and the target flow and the water level situation, and continuously optimizes the gate opening strategy, realizing the intelligence and high efficiency of the preliminary flow distribution.

[0040] 4. The present invention combines a multivariable predictive control algorithm with a state space model, sets the prediction time domain and the control time domain, constructs an objective function and solves the optimization problem, realizing the precise flow regulation with multi-parameter coordination and further improving the accuracy of flow distribution.

[0041] 5. The present invention prevents the odor in the pool from emitting through the sealing cover of the self-supporting flap gate. The transparent window on the shaft sleeve is convenient for observing the opening height of the gate rod. The bottom-opening gate and the self-supporting flap gate work together, and through the precise control of the automatic control system, the stable operation of the entire water distribution system is ensured. Description of the Drawings

[0042] Figure 1 is the overall flow schematic diagram of an intelligent flow distribution control method for a self-supporting flap gate according to Embodiment 1 of the present invention.

[0043] Figure 2 is the structural schematic diagram of an intelligent flow distribution control system for a self-supporting flap gate according to Embodiment 2 of the present invention.

[0044] Figure 3 It is a schematic structural diagram of the self-supporting plug gate in Embodiment 2 of the present invention.

[0045] Among them, 1. Down-opening gate; 2. Self-supporting plug gate; 201. Water inlet hole; 202. Gate plate; 203. Vent hole; 204. Gate rod; 205. Sealing cover; 206. Gate frame; 207. Electric actuator; 208. Bush; 3. Liquid level gauge; 4. Water distribution hole. Specific implementation manners

[0046] The present invention will be further described in detail below with reference to the accompanying drawings.

[0047] Embodiment 1

[0048] Referring to Figure 1 , an intelligent flow distribution control method for a self-supporting plug gate in this embodiment includes:

[0049] Obtain liquid level data and preprocess the liquid level data;

[0050] Calculate the theoretical value of the initial head height above the weir according to the obtained liquid level data, including calculating the theoretical value of the height by using the turbulent flow correction coefficient and the roughness dynamic adjustment parameter;

[0051] Adjust the opening height of the down-opening gate according to the theoretical value, and adjust the adjustment amount of the down-opening gate through the fuzzy rule base;

[0052] Perform preliminary flow distribution based on the adjusted opening height, including defining the state space and the action space, and using the reinforcement learning algorithm to adjust the opening state of the gate plate;

[0053] Use the preliminary distribution result for multi-parameter coordinated adjustment, including using the multivariable predictive control algorithm for precise flow adjustment;

[0054] Feedback and optimize the control result of the flow adjustment.

[0055] Specifically, an intelligent flow distribution control method for a self-supporting plug gate includes the following steps:

[0056] As Figure 1 shown, S1. Obtain liquid level data and preprocess the liquid level data;

[0057] In the water distribution system where the self-supporting slide gate is located, high-precision level meters and other sensor equipment are installed. The level meter is used to obtain real-time liquid level data in the water distribution channel. This data can reflect the depth of the water in the current channel and is an important basic information for flow distribution, which is crucial for accurate flow distribution. Perform preliminary processing and cleaning on the obtained liquid level data. Check the integrity of the data and eliminate data points with obvious errors or abnormalities, such as extreme values ​​caused by sensor failure. Smooth the data to eliminate noise interference in the data and make the data more stable and reliable. At the same time, standardize the data and unify the dimensions and range of the data so that subsequent calculations and analysis can be more accurate and efficient.

[0058] S2. Calculating the theoretical value of the initial water head height above the weir according to the acquired liquid level data, including calculating the theoretical value of the height using the turbulence correction coefficient and the roughness dynamic adjustment parameter;

[0059] In step S1, liquid level data including liquid level data H are collected by high-precision liquid level gauges. At the same time, the system also integrates historical data accumulated from the long-term operation of the water distribution system. Liquid level data H reflects the height of the current water level in the water distribution channel and is one of the key basic data for calculating the water head height on the weir. The installation height h1 of the water distribution hole from the bottom of the water distribution channel is a fixed geometric parameter. It is combined with liquid level data H to preliminarily determine the head height relationship when other factors are not considered.

[0060] When water flows in an actual channel, the flow state is complex and changeable, and turbulence is common. Turbulence will increase the energy loss of the water flow, thereby affecting the head height. The Reynolds number (Re) is an important indicator for judging the flow state of water flow, and its calculation formula is: Where ρ is the fluid density, v is the water velocity, and d is the hydraulic diameter. For circular pipes, d is the pipe diameter, and for non-circular channels, μ is the fluid dynamic viscosity. When the Reynolds number is small, the water flow is in a laminar state, the water flow is relatively stable, and the energy loss is relatively small; when the Reynolds number exceeds a certain critical value, the water flow turns into a turbulent state, strong vortices and mixing are generated inside the water flow, and the energy loss increases significantly. In order to more accurately consider the impact of turbulence on the head height, the turbulence correction coefficient Δh is introduced. 紊流 , which is dynamically calculated based on the water velocity and Reynolds number, specifically:

[0061] Δh 紊流 =k1×(Re-Re 临界 )α×v β ,

[0062] Among them, k1, α and β are coefficients determined according to different channel conditions and water flow characteristics, Re 临界is the critical Reynolds number. When the water flow velocity changes or the water flow state in the channel changes, resulting in a change in the Reynolds number, Δh 紊流 can be adjusted in real time to accurately reflect the influence of turbulence on the water head height.

[0063] The roughness of the channel wall has an important influence on the water flow resistance and head loss. During the long-term use of the channel, due to factors such as water flow scouring, sediment deposition, and biological attachment, the roughness of the channel wall will gradually change. In order to consider the influence of this change on the water head height in real time, the channel roughness coefficient adjustment parameter Δh 糙率 is introduced. The channel roughness coefficient is usually represented by the Manning coefficient n. Channels of different materials have different initial Manning coefficients. For example, the Manning coefficient of a concrete channel is generally between 0.011 and 0.017, while that of a stone-lined channel is between 0.017 and 0.025, and then Δh 糙率 is calculated. The calculation is realized according to the deformation of the Manning formula, that is, Δh 糙率 = k2×(n 当前 - n 初始 ) γ ×L δ , where k2, γ, and δ are coefficients determined according to the channel characteristics, n 初始 is the initial Manning coefficient of the channel, L is the channel length. To simplify the calculation, in this application, γ = 1 and δ = 1, and the formula becomes Δh 糙率 = k2×(n 当前 - n 初始 )×L. At this time, k2 is the only coefficient to be determined. Calculate k2, substitute the collected data into the formula, and according to the least squares principle, make the sum of the squares of the errors between the actually measured Δh 糙率 and the calculated value of the formula the smallest, so as to solve the value of k2.

[0064] After that, use the improved formula:

[0065] h 理论 = H - h1 - h2 + Δh 紊流 + Δh 糙率 ,

[0066] where H represents the liquid level data, h1 represents the installation height of the water distribution hole, h2 represents the opening height of the bottom-opening gate, Δh 紊流 represents the turbulence correction coefficient, and Δh 糙率 represents the roughness dynamic adjustment parameter, and calculate the theoretical value of the initial weir head height. Compared with the traditional calculation method, the traditional method often ignores the influence of turbulence and channel roughness change on the water head height, or uses a fixed empirical coefficient for estimation, and cannot accurately reflect the actual situation. While this method calculates Δh 紊流 and Δh 糙率 in real time., fully considering the energy loss and resistance change of water flow in complex channel environments.

[0067] S3. Adjust the opening height of the bottom-opening gate according to the theoretical value, and adjust the adjustment amount of the bottom-opening gate through the fuzzy rule base;

[0068] After calculating the theoretical value h 理论 of the initial head water level height on the weir in step S2, use this theoretical value as the key reference index to clarify the adjustment direction and target of the opening height h2 of the bottom-opening gate. The proximity of the actual value of the head water level height h to the theoretical value directly affects the accuracy and stability of subsequent flow distribution. The actual head water level height h can be expressed as h = H - h1 - h2, where H is the liquid level data, h1 is the installation height of the water distribution hole from the bottom of the water distribution channel. By adjusting h2, make h infinitely approach h 理论 , laying a foundation for realizing accurate flow distribution.

[0069] The liquid level gauge collects the liquid level data in the water distribution channel in real time at a high frequency to form a sequence H(t) of the actual liquid level changing with time. By comparing with the theoretical liquid level, calculate the real-time liquid level deviation e(t), and the formula is:

[0070] e(t) = H(t) - H 理论 (t)

[0071] where, H 理论 (t) represents the theoretical liquid level at the corresponding moment. At the same time, calculate the deviation change rate ec(t), and use the difference method for approximate calculation, that is where, Δt is the time interval for data collection by the liquid level gauge. By this way, obtain the change trend of the liquid level deviation with time. In order to convert the accurate liquid level deviation and deviation change rate into fuzzy quantities that can be processed by the fuzzy controller, define the corresponding fuzzy language variables, construct the fuzzy rule base, and each rule in the rule base adopts the form of "IF-THEN". When performing fuzzy inference, according to the fuzzy results of e(t) and ec(t) at the current moment, activate the corresponding rules in the rule base. According to the fuzzy output of the adjustment amount Δh2 of the bottom-opening gate obtained by fuzzy inference, use the defuzzification algorithm to convert it into an accurate adjustment amount. In this embodiment, the centroid method is used for defuzzification, and the calculation formula for the adjustment amount of the bottom-opening gate is:

[0072]

[0073] where, μ i represents the membership degree of the adjustment amount fuzzy subset i, x iIt is expressed as the quantization value corresponding to the fuzzy subset. When the liquid level deviation e(t) is large and the deviation change rate ec(t) is also large, the calculated value of Δh2 is large, which means that the adjustment amplitude of the downward-opening gate needs to be increased to quickly change the water head height above the weir and narrow the gap between the actual liquid level and the theoretical liquid level. When the liquid level deviation e(t) is small, the value of Δh2 is small, and a fine-tuning method is adopted to avoid unnecessary fluctuations in the water distribution system caused by excessive adjustment.

[0074] S4. Perform preliminary flow distribution based on the adjusted opening height, including defining the state space and the action space, and using the reinforcement learning algorithm to adjust the opening state of the gate plate;

[0075] After completing the fine adjustment of the opening height of the downward-opening gate in step S3, the water head height above the weir is close to the theoretical set value, and the water flow condition reaches a relatively stable state that meets the requirements of flow distribution. At this time, the stability of the water head height h above the weir ensures that when the water flow passes through the water distribution openings, the flow distribution can be carried out based on relatively accurate hydraulic conditions. First, define the state space and the action space. Among them, the water demand, real-time water level, water distribution priority, and flow distribution target of each water tank are used as the state space. The water demand of each water tank is obtained by deeply analyzing and predicting the historical data accumulated in the long term in step S1. Using the machine learning algorithm, in this embodiment, the ARIMA model in time series analysis is adopted, combined with factors such as water quality component changes and seasons and time periods, to accurately predict the water demand of different water tanks in a future period of time. The number of openings and the opening degree of the self-supporting flap gate are used as the action space. The number of openings of the gate plate determines the number of water passing sections, and the opening degree directly affects the area size of the water passing section.

[0076] Then, use the deep Q-network reinforcement learning algorithm to find the optimal gate plate opening strategy. By constructing a neural network to approximately estimate the Q-value function. In the water distribution system, the Q-value function represents the expected cumulative reward that can be obtained in the future after taking a certain action in the current state. Specifically, the DQN algorithm first randomly initializes the parameters of the neural network and then explores in the state space. At each time step, the algorithm selects an action (that is, adjusts the number of openings and the opening degree of the gate plate) according to the current state, executes the action, observes the feedback of the water distribution system, including the actual water inflow and water level changes of each water tank, and calculates the reward value based on these feedbacks. The setting of the reward value is the key of the algorithm, and it is based on the matching degree between the flow distribution result and the target flow. The reward value calculation formula is:

[0077]

[0078] Among them, q i represents the actual water inflow of the i-th water tank, represents the target water inflow, h iDenoted as the real-time water level of the $i$-th pool, Denoted as the optimal water level, $n$ represents the total number of pools. The closer the actual water inflow is to the target water inflow and the closer the water level is to the optimal water level, the higher the reward value.

[0079] The algorithm stores the current state, the executed action, the obtained reward, and the next state in the experience replay buffer. When the buffer accumulates a certain number of samples, the algorithm randomly extracts a batch of samples from the buffer for training, and updates the parameters of the neural network through the backpropagation algorithm, so that the Q-value function can more accurately estimate the future cumulative reward. As the training progresses, the algorithm gradually learns the optimal gate opening strategy, that is, in different system states, it can select the optimal gate opening action to maximize the long-term cumulative reward and achieve preliminary flow distribution.

[0080] S5. Use the preliminary distribution result for multi-parameter collaborative adjustment, including using the multi-variable predictive control algorithm for precise flow adjustment;

[0081] After completing the preliminary flow distribution, in order to make the flow distribution of the water distribution system more accurate, it is necessary to comprehensively consider multiple interrelated parameters. The liquid level is a key real-time feedback parameter in the water distribution system, which directly reflects the water storage situation of each pool. For example, when the liquid level of a certain pool is too low, it may mean that the water inflow of this pool is insufficient and the water distribution strategy needs to be adjusted in time; while too high a liquid level may pose a risk of overflow and the water inflow needs to be reduced. In order to achieve precise flow adjustment, it is necessary to establish a dynamic model required by the multi-variable predictive control (MPC) algorithm based on the historical data and actual operation conditions of the system. Assume that there are $n$ pools in the water distribution system, and the vector $x(k)$ represents the state of the system at time $k$, which includes parameters such as the liquid level, flow rate, and pressure of each pool, that is, $x(k)=[x_1(k),x_2(k),\cdots,x$ m (k)] T , where $m$ represents the number of state variables, and the vector $u(k)$ represents the control input of the system at time $k$, mainly including the opening degrees of the downward-opening gate and the self-supporting plug gate, that is, $u(k)=[u_1(k),u_2(k),\cdots,u$ p (k)] T , where $p$ is the number of control inputs. Then establish the following linear discrete-time state-space model, and the formula of the state-space model is:

[0082] $x(k + 1)=Ax(k)+Bu(k)+w(k)$,

[0083] Among them, A represents an m×m state transition matrix, B represents an m×p input matrix, which reflects the influence of control inputs on the system state, w(k) represents a process noise vector used to consider the uncertainties and disturbance factors existing in the system, x(k) represents a system state vector, u(k) represents a control input vector, m represents the number of state variables, p represents the number of control inputs used to consider the uncertainties and disturbance factors existing in the system, and the output equation can be expressed as:

[0084] y(k) = Cx(k) + v(k),

[0085] where y(k) is the output vector of the system at time k, containing measurable parameters such as the water levels and flow rates of each water tank, C is a q×m output matrix, q is the number of output variables, v(k) is a measurement noise vector. By identifying and estimating parameters from the historical data of the system, the specific values of matrices A, B, and C can be determined, thereby establishing a model that can accurately describe the dynamic characteristics of the system.

[0086] Using the MPC algorithm, the optimal adjustment amounts of the bottom-opening gate and the self-supporting flap gate are determined by optimizing the objective function. The core idea of the MPC algorithm is that at each sampling time, based on the current system state, the output of the system within the next N p time steps is predicted, and the control input sequence within the next N c time steps is determined by optimizing the objective function, where N p is the prediction horizon, N c is the control horizon, and N c ≤ N p . The objective function comprehensively considers factors such as flow deviation, gate adjustment amplitude, and water level balance of each water tank, and its form can be expressed as:

[0087]

[0088] where y ref (k + i|k) represents the reference output vector of the future i time steps predicted at time k, containing the target flow rates, target water levels, etc. of each water tank, y(k + i|k) is the system output vector of the future i time steps predicted at time k, Q is a q×q output weight matrix used to adjust the importance of each output variable in the objective function, Δu(k + i|k) = u(k + i|k) - u(k + i - 1|k) is the change in the control input, and R is a q×q control input change weight matrix used to limit the drastic change of the control input.

[0089] At each sampling time k, the MPC algorithm solves the optimization problem of the above objective function to obtain the optimal control input sequence u c for the next N *(k|k), u * (k + 1|k), …, u * (k + N c - 1|k). Then, only apply the first control input u * (k|k) to the system. At the next sampling instant k + 1, repeat the above process to achieve rolling optimization control.

[0090] S6. Feedback and optimize the control result of the flow regulation.

[0091] Calculate the error signal based on the output of the state - space model and the actual output of the system, and make adjustments according to the error signal. The adjustment law is expressed as:

[0092]

[0093] Where, represents the rate of change of the parameter vector at time k, Γ represents the adaptive gain matrix, represents the regression vector, including information on the system state and input, and e T (k) represents the transpose of the error signal vector. After updating and optimizing the controller parameter vector through the adjustment law, use the updated parameters to generate the final control instruction. For example, for an electric hoist, convert the number of gate openings or the height of the gate opening into instructions such as the number of turns and rotation direction of the motor; for a hydraulically driven gate, convert the control parameters into control signals such as the pressure and flow rate of the hydraulic system.

[0094] Embodiment 2

[0095] The difference between this embodiment and Embodiment 1 is that this embodiment provides an intelligent flow distribution control system for a self - supporting flap gate, including:

[0096] As Figure 2 shown, a liquid level gauge 3, which is installed on one side of the water distribution channel near each water distribution hole. The liquid level gauge is used to monitor the liquid level data in the water distribution channel in real time to calculate the head height over the weir, providing a basis for the automatic control system to judge the target water distribution point;

[0097] Water flow channel module: It includes a water distribution channel and a water tank. The water distribution channel, as the main intake channel, is at the starting end of the system, carrying all the water to be distributed and providing a water source for subsequent water tanks. There are usually multiple water tanks or compartments, which are the target objects of water volume distribution. The required water intake of each water tank is different or in a changing state. The water distribution channel is connected to each water tank through water distribution holes;

[0098] Gate control module: It includes a self-supporting flap gate 2 and a bottom-opening gate 1. The self-supporting flap gate is installed on one side of the water distribution channel outside the water distribution hole. It consists of multiple gate plates, a gate frame, a gate rod, a hoist, a shaft sleeve, a sealing cover, a vent hole, and a water inlet hole. The gate plates can be opened independently, and the water inlet width is controlled by changing the number of opened gate plates.

[0099] Processing module: It includes an automatic control system. The automatic control system is used to receive the data transmitted by the liquid level gauge. After analysis and processing, according to the preset algorithm and logic, it sends control signals to the self-supporting flap gate and the bottom-opening gate to achieve precise control of the opening degree of the bottom-opening gate plate. The bottom-opening gate is installed on one side of the water tank inside the water distribution hole 4 and is used to control the water inlet height and control the number of opened gate plates of the self-supporting gate. The self-supporting gate is installed outside the water distribution hole 4. The hoist of the self-supporting flap gate and the driving device of the bottom-opening gate are connected to the automatic control system through control lines to receive control signals.

[0100] As Figure 3 shown, for the self-supporting flap gate, the gate plate 202 is installed in the gate frame 206, and the two cooperate with each other. The gate frame 206 plays a role in supporting and fixing the gate plate 202, provides a track and guidance for the movement of the gate plate 202. The gate plate 202 can move up and down in the gate frame 206 to achieve opening and closing actions. The gate plate 202 is connected to the gate rod 204 through a connecting piece, and the gate rod 204 is arranged perpendicular to the gate plate 202. Generally, each gate plate 202 corresponds to one gate rod 204. One end of the gate rod 204 is fixedly connected to the gate plate 202, and the other end passes through the shaft sleeve 208 and is connected to the hoist 207. The shaft sleeve 208 is installed on the top structure of the gate. The inner hole of the shaft sleeve 208 is adapted to the gate rod 204, and the gate rod 204 can slide freely in the shaft sleeve 208. The shaft sleeve 208 plays a role in supporting and guiding the gate rod 204, ensuring the vertical stability of the gate rod 204 during the up and down movement, and reducing the friction between the gate rod 204 and other components. A transparent window is provided on the shaft sleeve 208 to facilitate the operator to observe the opening height of the gate rod 204, so as to master the position state of the gate plate 202.

[0101] The gate opening and closing machine 207 is connected to the upper end of the gate rod 204. The present invention is an electric actuator 207, and the motor inside it converts the rotary motion into linear motion through the transmission mechanism, thereby driving the gate rod 204 to move up and down. When the motor rotates forward, the screw rotates, and the nut drives the gate rod 204 to rise, thereby opening the gate plate 202; when the motor rotates reversely, the gate rod 204 descends, closing the gate plate 202, and the sealing cover 205 covers the top of the gate, covering the connection between the gate rod 204 and the gate opening and closing machine 207 and other related parts. The sealing cover 205 is sealed and connected to the top edge of the gate frame 206 to prevent the odor from being emitted in the pool, and at the same time plays a role in protecting the internal mechanical parts, preventing dust, debris, etc. from entering and affecting the normal operation of the equipment. The vent 203 is set on the surface of the gate plate 202, and its function is to balance the air pressure inside and outside the gate, ensuring that the gate is not affected by the air pressure during the opening and closing process.

[0102] The self-supporting gate is mainly used to control the amount of water flow. Its working principle is based on the control of the opening quantity and position of the gate 202. When the automatic control system receives data from sensors such as liquid level meters, it will analyze and process it according to the preset algorithm and logic. For example, when it is necessary to increase the water supply to the pool, the automatic control system will determine the current flow rate and the status of each gate 202, and then send a control signal to the corresponding hoist 207.

[0103] After receiving the signal, the gate hoist 207 starts, and drives the gate rod 204 to rise through the transmission mechanism. Since the gate rod 204 is connected to the gate plate 202, the gate plate 202 will gradually open as the gate rod 204 rises. By controlling the number of different gate plates 202 opened, the water inlet width can be changed. The more gate plates 202 are opened, the larger the water inlet width is, and the greater the water flow rate is. On the contrary, when the flow rate needs to be reduced, the automatic control system will send a signal to close the gate plate 202 to the gate hoist 207. The gate hoist 207 drives the gate rod 204 to descend, and the gate plate 202 is closed accordingly, thereby reducing the water inlet width and reducing the water flow rate.

[0104] In the whole process, the water inlet 201 and the vent 203 play an auxiliary role. The water inlet 201 allows water to enter the gate 202, using the buoyancy principle to reduce the burden on the gate hoist 207; the vent 203 ensures the balance of air pressure inside and outside the gate, so that the gate 202 can move up and down smoothly. At the same time, the transparent window on the sleeve 208 makes it easy for the operator to observe the position of the gate rod 204 in real time, so as to understand the opening height of the gate 202 and ensure the accuracy of flow control.

[0105] The specific process of opening and closing the self-supporting sliding plate gate:

[0106] Opening process:

[0107] Signal reception: Based on the analysis of flow requirements, the automatic control system determines the number of gate plates 202 to be opened and sends an opening signal to the corresponding hoist 207. The signal is transmitted through the control line to the control unit of the hoist 207.

[0108] Motor start: For the electric actuator 207, after the control unit receives the signal, it starts the motor. The motor transmits power to the screw through transmission components such as a speed reducer.

[0109] Screw rotation: The screw starts to rotate. The nut cooperating with the screw can only move linearly along the axis of the screw under the action of the guiding device. Since the nut is connected to the gate rod 204, the rising of the nut drives the gate rod 204 to rise.

[0110] Rise of the gate plate 202: During the rising process of the gate rod 204, it drives the connected gate plate 202 to move upward, gradually leaving the initial position, thus opening the gate. The operator can observe the rising height of the gate rod 204 through the transparent window on the bushing 208 to determine the opening degree of the gate plate 202.

[0111] Repeated operation: If multiple gate plates 202 need to be opened, repeat the above steps until the opening quantity set by the automatic control system is reached, achieving the required water inlet width and flow rate.

[0112] Closing process:

[0113] Signal reception: Based on the change in flow requirements, the automatic control system determines the number of gate plates 202 to be closed and sends a closing signal to the corresponding hoist 207. The signal is also transmitted through the control line to the control unit of the hoist 207.

[0114] Motor reverse rotation: After the control unit receives the closing signal, it controls the motor to reverse rotate. When the motor reversely rotates, the screw rotates in the reverse direction, and the nut moves downward along the axis under the drive of the screw.

[0115] Descent of the gate plate 202: The descent of the nut drives the gate rod 204 to descend, and then makes the gate plate 202 move downward, gradually returning to the initial position to close the gate. During the descent process of the gate plate 202, the operator can observe the descent situation of the gate rod 204 through the transparent window to ensure that the gate plate 202 is completely closed.

[0116] Repeated operation: If multiple gate plates 202 need to be closed, repeat the above steps until the closing state set by the automatic control system is reached, completing the adjustment of the flow rate.

[0117] The above are all the preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. An intelligent flow distribution control method for a self-supporting flap gate, characterized in that Including: Obtain liquid level data and preprocess the liquid level data; Calculate the theoretical value of the initial head height over the weir according to the obtained liquid level data, including calculating the theoretical value of the height by dynamically adjusting parameters using the turbulent flow correction coefficient and roughness coefficient; Adjust the opening height of the bottom-opening gate according to the theoretical value, and adjust the adjustment amount of the bottom-opening gate through the fuzzy rule base; Conduct preliminary flow distribution based on the adjusted opening height, including defining the state space and action space, and adjusting the opening state of the gate plate using the reinforcement learning algorithm; Conduct multi-parameter collaborative adjustment using the preliminary distribution result, including performing precise flow adjustment using the multivariable predictive control algorithm; Feedback and optimize the control result of the flow adjustment.

2. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that The calculating the theoretical value of the initial head height over the weir according to the obtained liquid level data includes calculating the Reynolds number according to the hydraulic diameter and kinematic viscosity, respectively calculating the turbulent flow correction coefficient and the roughness coefficient dynamic adjustment parameter based on the Reynolds number and the Manning coefficient of the current channel, and calculating the theoretical value of the initial head height over the weir using the theoretical formula, and the theoretical formula is expressed as: h 理论 =H - h1 - h2 + Δh 紊流 +Δh 糙率 , Among them, H represents the liquid level data, h1 represents the installation height of the water distribution hole, h2 represents the opening height of the downward-opening gate, and Δh 紊流 represents the turbulent flow correction coefficient, and Δh 糙率 represents the dynamic roughness adjustment parameter.

3. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that The adjusting the opening height of the bottom-opening gate according to the theoretical value includes determining the target value of the adjustment height according to the theoretical value, calculating the real-time liquid level deviation using the liquid level data, and calculating the deviation change rate using the difference method, defining the fuzzy linguistic variables and fuzzy subsets of the liquid level deviation and the deviation change rate, determining the membership degree through the membership function, constructing the fuzzy rule base and determining the fuzzy output of the adjustment amount of the bottom-opening gate through the fuzzy composition operation.

4. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that, The adjusting the adjustment amount of the bottom-opening gate through the fuzzy rule base includes solving the fuzzy output using the centroid method to obtain the adjustment amount of the bottom-opening gate, and the calculation formula of the adjustment amount of the bottom-opening gate is: Among them, μ i represents the membership degree of the fuzzy subset i of the adjustment amount, and x i represents the quantization value corresponding to the fuzzy subset.

5. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that, The adjusting the opening state of the gate plate using the reinforcement learning algorithm includes: applying the deep Q-network reinforcement learning algorithm to find the optimal gate opening strategy, approximately estimating the Q-value function by constructing a neural network, at each time step, selecting an action according to the current state, and calculating the reward value according to the feedback value after executing the action, and the reward value is determined according to the matching degree between the flow distribution result and the target flow and the water level condition, and the calculation formula of the reward value is: Among them, q i represents the actual water inflow of the i-th pool, represents the target water inflow, h i represents the real-time water level of the i-th pool, represents the optimal water level, and n represents the total number of pools.

6. The intelligent flow distribution control method of a self-supporting flap gate according to claim 1, characterized in that, The conducting multi-parameter collaborative adjustment using the preliminary distribution result further includes defining the system state vector and the control input vector based on the historical data and actual operation data of the water distribution system, then establishing a linear discrete-time state space model, and determining the output equation according to the state space model, and the formula of the state space model is: x(k + 1) = Ax(k) + Bu(k) + w(k), where, A represents the state transition matrix of m×m, B represents the input matrix of m×p, reflecting the influence of the control input on the system state, w(k) represents the process noise vector, used to consider the uncertainties and interference factors existing in the system, x(k) represents the system state vector, u(k) represents the control input vector, m represents the number of state variables, and p represents the number of control inputs.

7. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that The precise flow regulation using the multivariable predictive control algorithm includes using the multivariable predictive control algorithm to set the prediction horizon N p and the control horizon N c , constructing an objective function by combining the output equation of the state space model, solving the optimization problem of the objective function at each sampling moment, and obtaining the optimal control input sequence for the next N c time steps, and using the optimal control input sequence for precise flow regulation.

8. The intelligent flow distribution control method of a self-supporting plug gate according to claim 1, characterized in that, The feedback and optimizing the control result of the flow adjustment includes calculating the error signal according to the output of the state space model and the actual output of the system, and making adjustments according to the error signal, and the adjustment law is expressed as: Among them, represents the rate of change of the parameter vector at time k, and Γ represents the adaptive gain matrix. represents the regression vector, including information on the system state and input, and e T (k) represents the transpose of the error signal vector.

9. An intelligent flow distribution control system for a self-supporting flap gate, which executes the method described in claim 1. It is characterized in that, Including: Data acquisition module: including a liquid level gauge, which is used to monitor the liquid level data in the water distribution channel in real time to calculate the head height over the weir; Water flow channel module: including a water distribution channel and a water tank. The water distribution channel serves as the main intake channel and is at the starting end of the system. The water tank is the target object for water volume distribution. The water distribution channel is connected to each water tank through a water distribution opening; Gate control module: including a self-supporting flap gate and a bottom-opening gate. The self-supporting flap gate is installed on one side of the water distribution channel outside the water distribution opening and is composed of multiple gate plates, a gate frame, a gate rod, a hoist, a bushing, a sealing cover, a vent hole, and a water inlet hole. The self-supporting flap gate controls the intake width by changing the number of opened gate plates; Processing module: including an automatic control system, which is used to receive the data transmitted by the liquid level gauge. After analysis and processing, according to the preset algorithm and logic, it sends control signals to the self-supporting flap gate and the bottom-opening gate to achieve precise control of the gate opening, adjust the height and width of the water passing section, and regulate the flow rate.

10. The intelligent flow distribution control system of a self-supporting plug gate according to claim 9, characterized in that, The bottom-opening gate is installed on one side of the water tank inside the water distribution opening and is used to control the intake height. The hoist of the self-supporting flap gate and the driving device of the bottom-opening gate are connected to the automatic control system through control lines and are used to receive control signals.

Images