Risk and benefit balanced reservoir seasonal flood season water level multi-objective decision control method

Abstract

The invention discloses a risk and benefit balanced reservoir seasonal flood season water level multi-objective decision control method, and particularly relates to the technical field of data management. A genetic algorithm and a support vector machine are introduced to carry out multi-generation optimization and nonlinear regression model training, a water level change trend and an optimal water discharge amount are respectively predicted to serve as input variables of a fuzzy control system, and a final continuous control amount is calculated by adopting a weighted fusion algorithm. The matching degree between the control command and the equipment is evaluated according to the equipment response delay and the discretization precision; and if the matching degree is low, the system dynamically adjusts the discretization rule and the drainage strategy, and performs accurate butt joint with the equipment through the optimized control command, so that the accuracy and efficiency of reservoir management are improved, the dam break risk is reduced, and the drainage operation of the reservoir is optimized.

Description

Technical Field

[0001] The invention relates to the technical field of data management, and in particular to a multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance. Background Art

[0002] Multi-objective decision-making control of reservoir flood season water level with risk-benefit balance refers to the process of multi-objective decision-making control in the reservoir management process by considering the water level control targets at different flood seasons, weighing the risks (such as the risk of floods) and benefits (such as reservoir water storage benefits, power generation benefits, etc.). Specifically, this control method aims to formulate appropriate water level control strategies at different flood season stages (such as the early, middle and late flood seasons) based on the reservoir's real-time water level, precipitation forecast, flow changes and other factors to ensure that the reservoir strikes a balance between safety and maximum benefit. Through the multi-objective decision-making model, the reservoir can make flexible adjustments when facing sudden risks and demands, which not only reduces the risk of dam breach, but also makes full use of the reservoir's resources.

[0003] In the existing technology, fuzzy logic is introduced to deal with the uncertainty and ambiguity in water level control. It does not require an accurate mathematical model, but sets fuzzy rules based on expert experience or historical data to adjust the reservoir water level in real time. For example, in the case of high water level, fuzzy control can set rules to gradually increase the amount of water released to reduce the risk of dam breach while taking into account the benefits of water storage and power generation. This technology can cope with the complexity and dynamic changes in reservoir management and has strong adaptability and flexibility.

[0004] The prior art has the following deficiencies: When the reservoir fuzzy control system is connected to the actual control equipment (such as gates, water pumps, etc.), there is an interface mismatch between the output continuous control quantity and the discrete commands received by the equipment. Specifically, the fuzzy control system usually generates refined control quantities (such as "50% water release"), but the equipment may only accept discrete control commands (such as opening and closing the gate), resulting in inconsistencies in operation. In addition, the fuzzy control system may not take into account the response delay or mechanical accuracy of the equipment, resulting in a deviation between the control decision and the actual execution, thereby affecting the accuracy and effectiveness of reservoir management. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-objective decision-making control method for reservoir flood season water level with a balance between risk and benefit, so as to solve the problems in the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance, comprising: Identify inputs that affect reservoir water level control, including current water level, precipitation, inflows, and reservoir storage; Input historical water level data, inflow flow and precipitation data, and optimize water level change prediction and optimal water release prediction by combining genetic algorithm and support vector machine model; The prediction results are passed as input to the fuzzy control system, and the weighted fusion algorithm is used to calculate the final continuous control quantity, including the water discharge volume and water discharge speed. The 50% water discharge volume output by the fuzzy control system is converted into the 50% open state of the equipment. According to the discretization accuracy of the control command and the device response delay, the interface matching degree between the continuous control quantity and the device receiving the discrete command is evaluated, specifically including: normalizing the discrete command control accuracy index and the execution response delay abnormality index so that they are both between [0,1], and calculating the interface matching degree between the continuous control quantity and the device receiving the discrete command according to the normalized discrete command control accuracy index and the execution response delay abnormality index; If the matching degree is low, the control decision will be continuously adjusted and optimized according to the evaluation results, including adjusting the discretization rules and water release strategy, and sending the adjusted discretization control command to the actual control equipment of the reservoir to complete the reservoir water release.

[0007] Preferably, a genetic algorithm is used for multi-generation optimization to predict future water level changes and output predicted water level change trend values, specifically: historical water level, inflow flow and precipitation data are collected, and a training data set is constructed, the independent variables of the data set include water level, inflow flow and precipitation data at several moments in the past, and the dependent variable is the predicted water level at a future moment; a genetic algorithm is used for multi-generation optimization to find the optimal solution by simulating the natural selection process; specifically including: initializing the population and randomly generating multiple candidate solutions, using mean square error MSE as the fitness function, selecting individuals with high fitness for crossover and mutation operations, generating a new generation of individuals, the optimization process is stopped by a termination condition, and finally a set of optimal parameters are obtained; using the optimal parameters, predicting the water level change trend at future moments as the first input variable of the fuzzy control system.

[0008] Preferably, an SVM regression model is used to predict the optimal water release under different conditions and output the predicted water release value, specifically: model training is performed using the support vector regression method in a support vector machine (SVM); during the training process, current water level, inflow flow and precipitation forecast data are collected as independent variables, and the optimal water release of the reservoir is used as the dependent variable; by constructing a training data set, input features include the water level, inflow flow and precipitation at the current moment, and the target output is the optimal water release; a Gaussian radial basis kernel function is used to process nonlinear problems, optimize the SVM regression model, minimize the loss function, and find the optimal decision boundary and regression hyperplane; after the training is completed, the SVM regression model predicts the optimal water release based on the current water level, inflow flow and precipitation forecast data, and uses the predicted result as the second input variable of the fuzzy control system.

[0009] Preferably, in the reservoir water level control, the final continuous control quantity is obtained by combining the first input variable and the second input variable of the fuzzy control system for weighted fusion, specifically: the first input variable is the future water level change trend W(t+1) predicted by the genetic algorithm, that is, the predicted value of the future water level; the second input variable is the optimal water release predicted by the support vector machine SVM regression model , that is, the amount of water that needs to be released under current conditions; is the water level change trend W(t+1) and the predicted water release volume Set the weight coefficient, is the weight of water level change trend, is the weight of the optimal water release; the sum of the weight coefficients is 1; according to the set weight coefficients, the weighted average formula is used to merge the two input variables into a comprehensive control quantity: ;in: It is a comprehensive control quantity, which is a control signal calculated based on the future water level change trend and predicted water release volume.

[0010] Preferably, the fuzzy control system is based on the integrated control quantity after fusion Calculate the final water release volume and speed, including: The input Convert into fuzzy sets and define a set of fuzzy rule bases; According to the fuzzy rule base, a fuzzy output is inferred in combination with the input fuzzified data, which represents the fuzzy decision of the water release amount or speed; The defuzzification method is used to transform the fuzzy output into the actual continuous control quantity.

[0011] Preferably, the accuracy of the continuous control quantity of the fuzzy control system when it is converted into a discrete command executed by the device is analyzed to generate a discrete command control accuracy index, and the method for obtaining the discrete command control accuracy index is: The continuous control quantity output by the fuzzy control system is expressed as ; The actual discrete control commands executed by the device are represented as ; Calculate the error of each discrete command. It is calculated by the difference between the continuous control quantity and the device execution command, and the expression is: ; Calculate the weighted error value by weighted error method , the expression is: ; Where: z is the number of discretized control commands, is the weight coefficient of the ith control command, is the error of the ith discrete command, and the discrete command control accuracy index is calculated. The calculation expression is: ;in: Controls the precision exponent for discrete commands.

[0012] Preferably, the execution response delay abnormality index is generated after analyzing the time response delay required for the device to actually execute the command after the control command is issued. The method for obtaining the execution response delay abnormality index is: Collect the historical data of the device's response delay, select k nearest neighbors from the historical response delay data, and then use the Euclidean distance to calculate the difference between the current response delay and the historical delay data by calculating the difference between the response delay at the adjacent moment and the response delay at the current moment. , the expression is: ;in: It's a historical moment. is the response delay at the current moment, It's a historical moment Response delay; According to the calculated distance, the k historical response delay values ​​with the smallest distance are selected as the nearest neighbors, and the average response delay of the selected k nearest neighbors is calculated. The expression is: ;in: is the average value of the historical response delay at the current time t, and k is the number of selected neighbors. The execution response delay anomaly index is calculated as follows: ;in: is the standard deviation of historical response delay data, and ERD is the execution response delay anomaly index.

[0013] Preferably, the interface matching degree between the acquired continuous control quantity and the device accepting discrete commands is compared with a preset matching degree reference threshold. If the interface matching degree between the continuous control quantity and the device accepting discrete commands is greater than or equal to the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the device accepting discrete commands is high, and a normal matching signal is generated at this time; if the interface matching degree between the continuous control quantity and the device accepting discrete commands is less than the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the device accepting discrete commands is low, and a matching abnormality signal is generated, and the discretization rules and the water release strategy need to be adjusted.

[0014] In the above technical solution, the technical effects and advantages provided by the present invention are: The present invention provides a multi-objective decision-making control method for reservoir flood season water level with balanced risk and benefit. By combining genetic algorithm and support vector machine to optimize water level change prediction and optimal water release prediction, intelligent regulation of reservoir water level and water release is realized. The method provides accurate control quantity for fuzzy control system by weighted fusion predicted water level change trend and optimal water release, and optimizes the interface matching degree between continuous control quantity and device acceptance of discrete commands by evaluating discrete command control accuracy and execution response delay abnormality index, thereby reducing the deviation between control decision and device execution. By continuously adjusting the discretization rules and water release strategies, accurate water release and safe management of reservoirs are ensured, the efficiency of reservoir management and the accuracy of water level regulation are effectively improved, management risks caused by equipment response delay or control interface mismatch are avoided, and scientific decision support is provided for reservoir management during flood season. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] For examples, see Figure 1 As shown, the multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance described in this embodiment includes: Identify inputs that affect reservoir water level control, including current water level, precipitation, inflows, and reservoir storage; Input historical water level data, inflow flow and precipitation data, and optimize water level change prediction and optimal water release prediction by combining genetic algorithm and support vector machine model; The prediction results are passed as input to the fuzzy control system, and the weighted fusion algorithm is used to calculate the final continuous control quantity, including the water discharge volume and water discharge speed. The 50% water discharge volume output by the fuzzy control system is converted into the 50% open state of the equipment. According to the discretization accuracy of the control command and the device response delay, the interface matching degree between the continuous control quantity and the discrete command received by the device is evaluated; If the matching degree is low, the control decision will be continuously adjusted and optimized according to the evaluation results, including adjusting the discretization rules and water release strategy, and sending the adjusted discretization control command to the actual control equipment of the reservoir to complete the reservoir water release.

[0019] Determining the inputs that affect reservoir water level control is the basis for effective reservoir management. These inputs provide critical real-time information for reservoir water level control decisions. The following is a detailed description of each input: The current water level refers to the height of the water surface in the reservoir, usually measured in meters. It is the most direct indicator that affects the reservoir's water release decision and water storage strategy. The current water level determines whether the reservoir is close to the overflow level or whether there is enough water to release water. Too high a water level may lead to the risk of dam failure, while too low a water level may affect the functions of the reservoir such as water supply and power generation. Therefore, accurate current water level data is the core of formulating a reservoir water level control strategy. Real-time monitoring is carried out through water level sensors and telemetry systems, and data is updated regularly.

[0020] Precipitation refers to the total amount of precipitation within a reservoir basin within a certain period of time, usually in millimeters (mm). Precipitation directly affects the water input to the reservoir basin. Heavy rain or continuous precipitation will increase the amount of water in the rivers in the basin, thereby increasing the inflow of the reservoir. Conversely, precipitation decreases in the dry season, and the water supply in the basin also decreases accordingly. The reservoir needs to adjust the water level control strategy according to the change in precipitation. Precipitation information is obtained through meteorological stations, satellite remote sensing technology or meteorological forecast data. Reservoirs usually use short-term and long-term precipitation forecast data for prediction.

[0021] Inflow refers to the amount of water that flows into a reservoir per unit time, usually in cubic meters per second (m³ / s). This is a dynamic indicator that is affected by factors such as precipitation, upstream river flow, and melting ice and snow. Inflow directly determines the rate of rise in the reservoir water level. Too much inflow may cause the water level to rise rapidly and increase the risk of dam breach; too little inflow may cause the reservoir water level to be too low to meet downstream demand. Accurate inflow data is essential for reservoir management and helps to estimate future water level changes in the reservoir. Inflow water is calculated using flow meters, river monitoring stations, remote sensing technology, or basin hydrological models. Real-time monitoring systems obtain inflow data based on the hydrological characteristics of the reservoir basin.

[0022] Reservoir storage capacity refers to the amount of water stored in the reservoir, usually in cubic meters (m³). It is affected not only by the current water level, but also by factors such as the capacity of the reservoir, the design height, and the range of water level regulation. The storage capacity of a reservoir determines the reservoir's water storage capacity and scheduling flexibility. When the reservoir has sufficient water storage, the reservoir has more water available for downstream use (such as power generation, irrigation, water supply, etc.) and has the ability to regulate the water level. When the water storage capacity is low, the reservoir may need stricter water level control to prevent the water level from dropping too quickly due to excessive or too fast water release. The water storage capacity of a reservoir is calculated using real-time water level data and the reservoir capacity curve. The capacity curve is a relationship diagram between the water level and water storage given in the reservoir design, which is used to calculate the water storage capacity based on the current water level.

[0023] When using genetic algorithm (GA) for reservoir water level prediction, the goal is to predict the future water level change trend value based on historical water level data, inflow flow and precipitation through an optimization process so as to serve as the first input variable of the fuzzy control system.

[0024] In order to predict the water level change trend, the following historical data must be collected first: Historical water level data: includes water level records over a period of time (such as the past few days or months). Inflow data: includes reservoir inflow data over that period of time. Precipitation data: includes precipitation data over that period of time.

[0025] Construct a training dataset. Each record in the dataset will contain: Independent variables (input features): Historical water level: water level data at several times in the past. Inflow: inflow data at several times in the past. Precipitation: precipitation data at several times in the past.

[0026] Dependent variable (target output): Water level at a future time (W(t+1)): predicted water level data, which serves as the prediction target of the model.

[0027] The basic idea of ​​genetic algorithm is to simulate the process of natural selection to find the optimal solution. GA usually includes the following steps: Initialize the population: randomly generate multiple candidate solutions (individuals), and the genes of each individual represent a set of parameters to be optimized. For example, the genes of an individual may include the amount of water released from the reservoir, the coefficient of influence of precipitation on water level, the coefficient of influence of inflow on water level, etc.

[0028] Calculate the fitness function: Evaluate the fitness of each individual based on the prediction error, and set the mean square error MSE as the fitness function: ;in, is the water level obtained by the current individual prediction, is the actual historical water level data, and N is the number of data points.

[0029] Selection operation: Select individuals with higher fitness to enter the next generation based on their fitness values, usually using roulette wheel selection or tournament selection.

[0030] Crossover operation: A crossover operation is used to combine two parent individuals to generate a new offspring individual. For example, some gene fragments of the parent individuals are exchanged to generate a new combination.

[0031] Mutation operation: Perform mutation operations on some individuals and randomly change some gene values ​​to increase the diversity of the population and avoid falling into the local optimal solution.

[0032] Termination condition: When the preset generation number is reached or the fitness meets a certain threshold, the optimization process is stopped.

[0033] Through multiple generations of optimization by genetic algorithm, an optimal solution is finally obtained, and the model has the best prediction ability. Based on this solution, the optimal parameters obtained by genetic algorithm are used to predict the future water level change trend. The optimal parameters optimized by genetic algorithm are set as , these parameters will be used to predict water level changes in the future: ; Where: W(t) is the water level at the current moment, is the inflow flow at the current moment; P(t) is the precipitation at the current moment; Other factors that may affect the water level, such as temperature, wind speed, etc.; the predicted future water level change trend W(t+1) is output and used as the first input variable of the fuzzy control system.

[0034] The fuzzy control system will use the predicted water level trend W(t+1) to estimate the future water level and adjust the water release strategy based on this information. The system output (water release volume, water release speed, etc.) will be interfaced with the actual control equipment of the reservoir.

[0035] The support vector machine (SVM) is used to predict the optimal water release of the reservoir. The model is trained based on the current water level, inflow flow and precipitation forecast data, and then the prediction results are provided to the fuzzy control system so as to serve as the second input variable of the fuzzy control system.

[0036] In order to train the SVM regression model, the following data needs to be collected: Current water level: includes the water level information of the reservoir at a certain moment. Inflow flow: the amount of water flowing into the reservoir per unit time. Precipitation forecast data: the precipitation forecast data for a period of time in the future. These data will be used as independent variables in the training data set, and the optimal water release of the reservoir will be used as the dependent variable.

[0037] The training dataset will contain the following: Independent variables (input features): water level W(t) at the current moment, inflow flow at the current moment ;Precipitation at the current moment P(t); Dependent variable: optimal water release : At this moment, the optimal water release volume is calculated based on the current water level, inflow flow and precipitation forecast data.

[0038] The structure of the training dataset can be expressed as: ; Each data record corresponds to the input characteristics at a moment and the corresponding optimal water release amount.

[0039] Use the SVR (Support Vector Regression) method of SVM for regression training.

[0040] SVM regression models usually use different kernel functions (such as linear kernel, polynomial kernel, Gaussian radial basis kernel, etc.) to deal with nonlinear problems. The commonly used Gaussian radial basis kernel (RBF kernel) function is: ;in, is the width parameter of the kernel function, which controls the measure of similarity between data points. They represent two different features in the training set.

[0041] Using the constructed training data set, the SVM optimization process is used to learn the data pattern and find the support vector and regression hyperplane. The goal is to minimize the error of the regression model and find the optimal decision boundary.

[0042] The goal of the training process is to minimize the loss function: ;Where: w is the weight of the model; b is the bias term; C is the regularization parameter, which is used to control the balance between training error and model complexity; is the error of each sample, and m is the total number of samples.

[0043] Use support vectors (i.e. data points that have a greater impact on the model) to train the regression model to obtain the optimal prediction parameters.

[0044] Once the SVM regression model is trained, it is used to predict the optimal water release given the current water level, inflow flow and precipitation forecast data.

[0045] Through the SVM regression model, each input feature Make a prediction and output the corresponding optimal water release volume: ;in, is the optimal water release output by the model, and f is the output function of the model.

[0046] Optimal water release output from SVM regression model It will be used as the second input variable of the fuzzy control system. This variable will be passed to the fuzzy control system as input together with the water level change trend value predicted by the genetic algorithm.

[0047] In reservoir water level control, the final continuous control variables, such as water release volume and water release speed, can be obtained by combining the first input variable (future water level change trend) and the second input variable (predicted water release volume) of the fuzzy control system for weighted fusion.

[0048] The input of the fuzzy control system includes the following two main variables: The first input variable: the future water level change trend W(t+1) predicted by the genetic algorithm, that is, the predicted value of the future water level.

[0049] Second input variable: optimal water release predicted by support vector machine (SVM) regression model , that is, the amount of water that needs to be released under current conditions.

[0050] These two input variables represent two key decision factors in water level control: the water level change trend reflects the dynamic changes of the reservoir water level and provides the necessary reference for future water release. The optimal water release forecast gives a specific water release value based on the current water level, inflow flow and precipitation forecast data.

[0051] The weighted fusion algorithm combines the two information sources by weighting the two input variables to obtain a comprehensive control quantity. The core idea of ​​weighted fusion is to calculate the final control decision by combining the influence of the two variables by setting appropriate weights. The specific steps are as follows: First, we need to calculate the water level change trend W(t+1) and the predicted water release volume Set the weight coefficient. The weight coefficient can be optimized based on actual conditions through experiments or historical data, or set according to the rules of the fuzzy control system. Set the weights of the two inputs: is the weight of water level change trend, is the weight of the optimal water release; the sum of the weight coefficients is required to be 1.

[0052] According to the set weight coefficient, the weighted average formula is used to merge the two input variables into a comprehensive control quantity: ;in: It is a comprehensive control quantity, which is a control signal calculated based on the future water level change trend and the predicted water release volume. It will be used as the input of the fuzzy control system to calculate the final continuous control quantity, such as water release volume and water release speed.

[0053] The fuzzy control system is based on the integrated control quantity after fusion Calculate the final water release volume and speed. The basic steps of fuzzy control include: Fuzzification: The input Convert it into a fuzzy set and define a set of fuzzy rule bases. For example: If the water level is higher, the amount of water released should be increased. If it is lower, the amount of water released should be reduced.

[0054] According to the fuzzy rule base, a fuzzy output is inferred from the input fuzzified data, which represents the fuzzy decision of the water release amount or speed.

[0055] Use defuzzification methods (such as the centroid method) to convert the fuzzy output into actual continuous control quantities (water release amount and water release speed).

[0056] After the fuzzy control system calculates the final continuous control quantity, if the control quantity (such as 50% water release) cannot be directly matched with the control equipment of the reservoir (such as gates, pumps) (for example, the equipment can only accept discrete switch commands), further dynamic adjustment is required: evaluate the degree of match between the continuous control quantity and the discrete commands that the equipment can accept. If the match is low, the control decision needs to be optimized. According to the adjustment rules of the control strategy, change the discretization accuracy of the water release and adjust the compensation for the device response delay. For example, if the device response delay is large, adjust the water release command in advance to ensure that the device can respond accurately at the appropriate time. Send the adjusted discrete control command to the actual control equipment of the reservoir, such as gates or pumps.

[0057] It should be noted here that the actual water level of the reservoir and the control effect are monitored through a real-time feedback mechanism. If the actual water level change is found to be inconsistent with expectations, the system will automatically adjust the fuzzy control rules and weight coefficients to optimize the water level management strategy of the reservoir.

[0058] In the reservoir control system, the fuzzy control system usually outputs a continuous control quantity (such as 50% water release), while the actual control equipment (such as gates, water pumps, etc.) can only accept discrete control commands (such as "open 50%"). Therefore, the continuous control quantity must be converted into discrete control commands that the device can understand, and matched according to the response delay of the device.

[0059] First, convert the continuous control quantity (such as 50% water release) output by the fuzzy control system into a discrete control command that can be executed by the device. Since the device can only accept discrete states (such as "open 50%" or "closed"), it is necessary to discretize the continuous output of the fuzzy control. The specific steps are as follows: Define the range of discrete control commands: Assume that the control command that the device can accept has multiple discrete states, such as: the gate is completely closed (0%); the gate is partially open (such as 20%, 40%, 60%, 80%); the gate is fully open (100%).

[0060] Discretization control quantity: Match the continuous control quantity (such as 50%) output by the fuzzy control system with the discrete command of the device. Commonly used discretization methods are rounding method or interval division method: If the continuous control amount is 50%, it will be rounded to the nearest discrete command, which is 50%. If the control amount is another value (such as 45%), it may be rounded to 40% or 50%.

[0061] For example: control amount 50% → discrete command: 50%; control amount 65% → discrete command: 60%; control amount 37% → discrete command: 40%.

[0062] Interface matching refers to the matching degree between the continuous output of the fuzzy control system and the discrete commands that the device can accept. In order to ensure that the device accurately executes the control command, the matching degree between the continuous control quantity and the discrete command must be evaluated.

[0063] The discretization accuracy of the control command refers to the accuracy of converting the continuous control quantity of the fuzzy control system into discrete commands that the device can execute. Higher discretization accuracy means that the device control state (such as water release volume, gate opening degree) can be finely adjusted, for example, the control quantity is discretized in smaller steps (such as adjustment every 5%). However, the higher the accuracy, the more frequent changes in control commands may be caused by subtle adjustments made by the device, increasing the response burden of the device and even causing unnecessary operational fluctuations. Therefore, the discretization accuracy needs to be balanced according to the actual device's responsiveness and operational stability to ensure that the control command is both accurate enough and avoids the adverse effects of frequent switching.

[0064] The accuracy of the fuzzy control system's continuous control quantity converted into discrete commands executed by the device is analyzed to generate a discrete command control accuracy index. The method for obtaining the discrete command control accuracy index is as follows: The continuous control quantity (such as 50%) output by the fuzzy control system is expressed as .

[0065] The actual discrete control command executed by the device (such as "turn on 50%") is represented by , which is usually obtained by discretizing the continuous control quantity.

[0066] Calculate the error for each discrete command: Error of discrete command It is calculated by the difference between the continuous control quantity and the device execution command, and the expression is: If the error is 0, indicating that the continuous control quantity completely matches the discrete command. If it is greater than 0, it means there is an error and the system fails to match accurately.

[0067] Since different discretization commands may have different impacts on the system, it is necessary to assign a weight coefficient to each discretization error to reflect the importance of the command. Usually, the weight coefficient can be set according to the response sensitivity of the device, the range of control quantity, or the priority of the operation.

[0068] Weighted error calculation: Calculate the weighted error value using the weighted error method , the expression is: ; Where: z is the number of discretized control commands, is the weight coefficient of the ith control command, is the error of the ith discrete command. Calculate the discrete command control accuracy index, the calculation expression is: ;in: It is the discrete command control accuracy index, which reflects the system control accuracy. The larger the value, the higher the accuracy.

[0069] Device response delay refers to the time required for the device to actually execute the command after the control command is issued. When the response delay is large, the immediate control amount of the fuzzy control system may deviate from the actual effect when the device is executed. For example, if the water discharge output of the system is 50%, but the response delay of the device is 1 minute, the water discharge during actual operation may not be consistent with the system's expectations, thereby affecting the accuracy of water level control. In order to compensate for the impact of response delay on the control effect, the system can adjust the control decision in advance according to the delay time of the device to ensure that the device adjusts the state as expected. In addition, long response delays may require the system to design a more flexible and robust control strategy to reduce the negative impact of delays on the overall control effect.

[0070] After the control command is issued, the time response delay required for the device to actually execute the command is analyzed to generate the execution response delay abnormality index. The method for obtaining the execution response delay abnormality index is as follows: Collect historical data on the device's response delay and evaluate abnormal situations by calculating the similarity between the current moment's response delay and the historical moment's response delay. First, select k nearest neighbors (k similar moments) from the historical response delay data, then calculate the difference between the current response delay and the historical delay data using the Euclidean distance by calculating the difference between the response delay at the neighboring moment and the response delay at the current moment. , the expression is: ;in: is a historical moment (i=1,2,…,k). is the response delay at the current moment, It's a historical moment response delay.

[0071] According to the calculated distance, select the k historical response delay values ​​with the smallest distance as the nearest neighbors. k can be determined by the following method: If the historical data is relatively stable, you can choose a smaller k value (such as 3 or 5). If the historical data varies greatly, you can choose a larger k value (such as 10 or 15) to smooth out abnormal fluctuations.

[0072] Calculate the average response delay of the selected k nearest neighbors, expressed as: ;in: is the average value of the historical response delay at the current time t, and k is the number of selected neighbors. The execution response delay anomaly index is calculated as follows: ;in: is the standard deviation of historical response delay data (measures the volatility of device response delay), and ERD is the execution response delay anomaly index.

[0073] According to the calculated abnormality index ERD, determine whether there is an abnormal execution response delay. Common judgment methods include: If ERD>2, it means that the response delay at that moment is abnormal, which may cause a deviation between the control strategy and the device execution. If ERD≤1, it is considered that the response delay of the current device is within the normal range.

[0074] The discrete command control precision index and the execution response delay anomaly index are normalized so that they are both between [0,1]. The interface matching degree between the continuous control quantity and the device accepting discrete commands is calculated based on the normalized discrete command control precision index and the execution response delay anomaly index.

[0075] For example, the present invention can use the following formula to calculate the interface matching degree between the continuous control quantity and the discrete command received by the device, and the calculation expression is: ; In the formula, The interface matching degree between the continuous control quantity and the discrete commands accepted by the device, is the discrete command control accuracy index, ERD is the execution response delay abnormality index, is the weight coefficient of the discrete command control accuracy index and the execution response delay abnormality index (which can be optimized based on experimental experience or machine learning), and Both are greater than 0.

[0076] The interface matching degree between the obtained continuous control quantity and the discrete command accepted by the device is compared with the preset matching degree reference threshold. If the interface matching degree between the continuous control quantity and the discrete command accepted by the device is greater than or equal to the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the discrete command accepted by the device is high, and a normal matching signal is generated at this time; if the interface matching degree between the continuous control quantity and the discrete command accepted by the device is less than the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the discrete command accepted by the device is low, and a normal matching signal is generated at this time.

[0077] When the evaluation results show that the interface matching degree between the continuous control quantity and the discrete commands received by the device is low, that is, the matching degree does not reach the preset reference threshold, the control decision needs to be continuously adjusted and optimized. This mainly includes adjusting the discretization rules and water release strategies to ensure that the device can better respond to the commands of the control system.

[0078] When the matching degree is low, you need to analyze the cause first. Possible reasons include but are not limited to the following: Discrete command control accuracy is not high, resulting in a large error between refined control and actual execution of the device. The device response delay is long and fails to keep up with the command changes of the fuzzy control system in time. The discretization rules do not match the device control capabilities. For example, the device cannot accurately respond to overly detailed control commands (for example, "50% water release" cannot be directly converted into device actions).

[0079] The discretization rule is the standard for converting the continuous control quantity of the fuzzy control system into a discrete command that the actual device can accept. The discretization rule can be adjusted in the following ways: If the device has limited responsiveness, you may need to adjust the discretization rules so that the refinement of the control command is adapted to the device's execution capability. For example, adjust "50% water release" to "50% or 40% water release" to reduce the complexity of the device control command. Adopt a more relaxed water release strategy, such as increasing the discretization accuracy from 10% to 20% or 30%, to reduce the frequency of each device adjustment.

[0080] Based on the actual execution capabilities of the device, a dynamic adjustment strategy can be adopted. For example: When the water level changes rapidly, a lower discretization accuracy is used to make the device respond more quickly; when the water level is relatively stable, the discretization accuracy can be increased to achieve more detailed control. Using a real-time feedback mechanism, the discretization rules are dynamically adjusted so that the system can better match the device in different situations.

[0081] Adjusting the water release strategy can make the reservoir release more efficient and reduce the difference with the equipment control capability. Optimizing the water release strategy can be done in the following ways: Based on historical data and real-time forecasts, the water release range of the reservoir is redefined to ensure that the equipment can be accurately controlled within this range. For example, if the current reservoir water level is high and strong precipitation is predicted in the future, the upper limit of the water release can be increased so that the equipment can effectively lower the water level. When the reservoir water level is low, the water release volume can be appropriately reduced to avoid excessive water release and waste of resources.

[0082] Based on real-time data (such as current water level, precipitation, inflow, etc.), dynamic water release strategies are used to cope with different flood and drought periods. Dynamic strategies include: automatic adjustment of water release frequency and magnitude, so that the equipment can adapt according to actual conditions. Introduce multi-level water release strategies (such as small, medium, and large water release) to cope with changes in different situations.

[0083] After implementing the adjusted control decision, the system needs to further optimize and adjust the control strategy based on real-time feedback information: obtain the current state of the reservoir in real time through sensors and monitoring systems, including key data such as water level and flow. The system can compare these data with the prediction model to adjust the water release in real time. By comparing with the actual execution results (equipment response), the discretization rules and water release strategies are continuously optimized: if it is found that the equipment cannot respond accurately at a certain stage, the discretization accuracy can be further optimized or the interval of the water release strategy can be adjusted. According to the actual performance of the equipment response delay, the time window of the control decision is adjusted to ensure that the matching degree between the continuous control quantity and the actual operation is improved.

[0084] After the discretization rules and water release strategies are adjusted and optimized, the final generated control commands need to be sent to the actual control devices of the reservoir (such as gates, water pumps, etc.) for execution. At this point, the control commands have been optimized based on the device response capability and interface matching, ensuring that the devices can be accurately executed and achieve the expected results.

[0085] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance, characterized by: include: Identify inputs that affect reservoir water level control, including current water level, precipitation, inflows, and reservoir storage; Input historical water level data, inflow flow and precipitation data, and optimize water level change prediction and optimal water release prediction by combining genetic algorithm and support vector machine model; The prediction results are passed as input to the fuzzy control system, and the weighted fusion algorithm is used to calculate the final continuous control quantity, including the water discharge volume and water discharge speed. The 50% water discharge volume output by the fuzzy control system is converted into the 50% open state of the equipment. According to the discretization accuracy of the control command and the device response delay, the interface matching degree between the continuous control quantity and the device receiving the discrete command is evaluated, specifically including: normalizing the discrete command control accuracy index and the execution response delay abnormality index so that they are both between [0,1], and calculating the interface matching degree between the continuous control quantity and the device receiving the discrete command according to the normalized discrete command control accuracy index and the execution response delay abnormality index; If the matching degree is low, the control decision is continuously adjusted and optimized according to the evaluation results, including adjusting the discretization rules and water release strategy, and sending the adjusted discretization control command to the actual control equipment of the reservoir to complete the reservoir water release.

2. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 1 is characterized by: Genetic algorithms are used for multi-generation optimization to predict future water level changes and output the predicted water level change trend value. Specifically, historical water level, inflow and precipitation data are collected and a training data set is constructed. The independent variables of the data set include water level, inflow and precipitation data at several moments in the past, and the dependent variable is the predicted water level at a future moment. A genetic algorithm is used for multi-generation optimization to find the optimal solution by simulating the natural selection process; specifically, the following steps are used: initializing the population and randomly generating multiple candidate solutions, taking the mean square error (MSE) as the fitness function, selecting individuals with high fitness for crossover and mutation operations, generating a new generation of individuals, and stopping the optimization process through termination conditions to ultimately obtain a set of optimal parameters; using the optimal parameters to predict the water level change trend at future moments as the first input variable of the fuzzy control system.

3. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 2 is characterized by: The SVM regression model is used to predict the optimal water release under different conditions and output the predicted water release value. Specifically, the model is trained through the support vector regression method in the support vector machine SVM; during the training process, the current water level, inflow flow and precipitation forecast data are collected as independent variables, and the optimal water release of the reservoir is used as the dependent variable; by constructing a training data set, the input features include the water level, inflow flow and precipitation at the current moment, and the target output is the optimal water release; the Gaussian radial basis kernel function is used to deal with nonlinear problems, optimize the SVM regression model, minimize the loss function, and find the optimal decision boundary and regression hyperplane; after the training is completed, the SVM regression model predicts the optimal water release based on the current water level, inflow flow and precipitation forecast data, and uses the prediction result as the second input variable of the fuzzy control system.

4. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 3 is characterized by: In the reservoir water level control, the final continuous control quantity is obtained by combining the first input variable and the second input variable of the fuzzy control system for weighted fusion. Specifically, the first input variable is the future water level change trend W(t+1) predicted by the genetic algorithm, that is, the predicted value of the future water level; the second input variable is the optimal water release predicted by the support vector machine SVM regression model. , that is, the amount of water that needs to be released under current conditions; is the water level change trend W(t+1) and the predicted water release volume Set the weight coefficient, is the weight of water level change trend, is the weight of the optimal water release; The sum of the weight coefficients is 1; according to the set weight coefficients, the weighted average formula is used to merge the two input variables into a comprehensive control quantity: ;in: It is a comprehensive control quantity, which is a control signal calculated based on the future water level change trend and predicted water release volume.

5. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 4 is characterized by: The fuzzy control system is based on the integrated control quantity after fusion Calculate the final water release volume and speed, including: The input Convert into fuzzy sets and define a set of fuzzy rule bases; According to the fuzzy rule base, a fuzzy output is inferred in combination with the input fuzzified data, which represents the fuzzy decision of the water release amount or speed; The defuzzification method is used to transform the fuzzy output into the actual continuous control quantity.

6. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 5 is characterized by: The accuracy of the fuzzy control system's continuous control quantity converted into discrete commands executed by the device is analyzed to generate a discrete command control accuracy index. The method for obtaining the discrete command control accuracy index is as follows: The continuous control quantity output by the fuzzy control system is expressed as ; The actual discrete control commands executed by the device are represented as ; Calculate the error of each discrete command. It is calculated by the difference between the continuous control quantity and the device execution command, and the expression is: ; Calculate the weighted error value by weighted error method , the expression is: ; Where: z is the number of discretized control commands, is the weight coefficient of the ith control command, is the error of the ith discrete command, and the discrete command control accuracy index is calculated. The calculation expression is: ;in: Controls the precision exponent for discrete commands.

7. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 6 is characterized by: After the control command is issued, the time response delay required for the device to actually execute the command is analyzed to generate the execution response delay abnormality index. The method for obtaining the execution response delay abnormality index is as follows: Collect the historical data of the device's response delay, select k nearest neighbors from the historical response delay data, and then use the Euclidean distance to calculate the difference between the current response delay and the historical delay data by calculating the difference between the response delay at the adjacent moment and the response delay at the current moment. , the expression is: ;in: It's a historical moment. is the response delay at the current moment, It's a historical moment Response delay; According to the calculated distance, the k historical response delay values ​​with the smallest distance are selected as the nearest neighbors, and the average response delay of the selected k nearest neighbors is calculated. The expression is: ;in: is the average value of the historical response delay at the current time t, and k is the number of selected neighbors. The execution response delay anomaly index is calculated as follows: ;in: is the standard deviation of historical response delay data, and ERD is the execution response delay anomaly index.

8. The multi-objective decision-making control method for reservoir flood season water level with risk and benefit balance according to claim 1 is characterized by: The interface matching degree between the obtained continuous control quantity and the discrete command received by the device is compared with a preset matching degree reference threshold. If the interface matching degree between the continuous control quantity and the discrete command received by the device is greater than or equal to the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the discrete command received by the device is high, and a matching normal signal is generated at this time; If the interface matching degree between the continuous control quantity and the discrete commands accepted by the device is less than the preset matching degree reference threshold, it means that the interface matching degree between the continuous control quantity and the discrete commands accepted by the device is low. At this time, a matching abnormality signal is generated, and the discretization rules and release strategies need to be adjusted.

Images