Reservoir flood season water level dynamic multi-objective optimization control system and method

Abstract

The invention relates to the technical field of reservoir optimization scheduling, in particular to a dynamic multi-objective optimization control system and method for the water level of a reservoir in a flood season. The method comprises the following steps: acquiring real-time monitoring data of a reservoir; extracting current water level state information according to the real-time monitoring data of the reservoir; establishing a reservoir water balance basic model based on the current water level state information; analyzing pipeline water supply flow change characteristics according to the reservoir water balance basic model; identifying response lag time of pipeline fluid inertia according to pipeline water supply flow change characteristics; predicting a water supply line pressure fluctuation risk based on the response lag time; and determining the water level regulation and control advance according to the water supply pipeline pressure fluctuation risk and the response lag time. According to the method, the multi-objective optimization model is adopted, the flood control safety margin and the water supply guarantee degree can be balanced, and it is ensured that the reservoir can cope with the flood risk in the flood season and can stably meet the water supply requirement.

Description

Technical Field

[0001] The present invention relates to the technical field of reservoir optimization and dispatching, and in particular to a dynamic multi-objective optimization control system and method for reservoir water level during flood season. Background Art

[0002] The flood season refers to the period of time when river water levels regularly rise due to seasonal precipitation, ice melt, and snowmelt within the basin. For reservoirs, the flood season presents a period of heightened flood risk and requires strengthened management and scheduling. Dynamic multi-objective optimization control of reservoir water levels during the flood season involves dynamically adjusting reservoir water levels through scientific scheduling strategies and optimization methods, taking into account multiple objectives such as flood control, power generation, water supply, and irrigation.

[0003] However, traditional multi-objective optimization methods for dynamic reservoir water levels during flood season often suffer from the following issues: feedback hysteresis caused by the inertia of water supply pipelines. For urban water supply reservoirs, water must be distributed to water plants through long pipelines, and pipeline filling takes a long time. The optimization model assumes instantaneous water supply response and does not model pipeline fluid inertia. This can easily cause water hammer-dispatching resonance when water demand suddenly changes, leading to oscillations in the water flow rate at the distribution plant. In cascade reservoir systems, when a downstream reservoir fails, flood waves propagate upstream. Existing optimization models only consider downstream flood propagation, but the negative wave velocity caused by a dam failure can reach 1.5 times that of the positive wave velocity, and existing systems lack the computational capacity to handle such reverse waves. Summary of the Invention

[0004] Based on this, it is necessary for the present invention to provide a dynamic multi-objective optimization control system and method for reservoir water level during flood season to solve at least one of the above technical problems.

[0005] To achieve the above objectives, a dynamic multi-objective optimization control method for reservoir water level during flood season is provided, comprising the following steps: Step S1: Acquire real-time monitoring data of the reservoir; extract current water level status information according to the real-time monitoring data of the reservoir; establish a basic model of reservoir water balance based on the current water level status information; Step S2: Analyze the pipeline water supply flow rate variation characteristics based on the reservoir water balance basic model; identify the response lag time of pipeline fluid inertia based on the pipeline water supply flow rate variation characteristics; and predict the water supply pipeline pressure fluctuation risk based on the response lag time; Step S3: determining a water level control lead time based on the water supply pipeline pressure fluctuation risk and the response lag time; formulating a water level control early intervention strategy based on the water level control lead time; and calculating a target water level control execution sequence based on the water level control early intervention strategy. Step S4: constructing a water level control gradient smooth curve based on the target water level control execution sequence; monitoring the probability of pipeline water hammer effect based on the water level control gradient smooth curve; adjusting parameters to ensure that the pipeline pressure wave and the control instruction do not resonate based on the probability of pipeline water hammer effect, and obtaining the reservoir outflow control parameters; Step S5: Evaluate the flood control safety margin of the reservoir during the flood season according to the reservoir outflow control parameters, and adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season to achieve dynamic multi-objective optimization control of the reservoir water level during the flood season.

[0006] The present invention also provides a dynamic multi-objective optimization control system for reservoir water level during flood season, which is characterized in that it is used to execute the dynamic multi-objective optimization control method for reservoir water level during flood season as described above, and the dynamic multi-objective optimization control system for reservoir water level during flood season includes: The water balance analysis module is used to obtain real-time monitoring data of the reservoir; extract the current water level status information based on the real-time monitoring data of the reservoir; and establish the basic model of the reservoir water balance based on the current water level status information; The pipeline pressure risk assessment module is used to analyze the pipeline water supply flow variation characteristics based on the reservoir water balance basic model; identify the response lag time of pipeline fluid inertia based on the pipeline water supply flow variation characteristics; and predict the water supply pipeline pressure fluctuation risk based on the response lag time; The water level control prediction module is used to determine the water level control lead time based on the water supply pipeline pressure fluctuation risk and response lag time; formulate a water level control early intervention strategy based on the water level control lead time; and calculate the target water level control execution sequence based on the water level control early intervention strategy; The water hammer effect control module is used to construct a water level control gradient smoothing curve based on the target water level control execution sequence; monitor the probability of pipeline water hammer effect based on the water level control gradient smoothing curve; and adjust parameters based on the pipeline water hammer effect probability to ensure that the pipeline pressure wave and the control instruction do not resonate, thereby obtaining the reservoir outflow control parameters; The flood control dynamic optimization control module is used to evaluate the flood control safety margin of the reservoir during the flood season based on the reservoir outflow control parameters, and to adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season, thereby realizing dynamic multi-objective optimization control of the reservoir water level during the flood season.

[0007] The present invention effectively improves the safety, regulation efficiency and water supply guarantee capacity of the reservoir during the flood season through precise water level regulation, real-time monitoring and multi-objective optimization control. First, the method ensures that the reservoir can accurately control water level changes in a complex flood season environment by gradually deducing the execution sequence based on the target water level regulation, avoiding the risk of excessive water storage or excessive water release. By refining the water level changes, gate opening adjustments and water level regulation rates at each time node, the reservoir's regulation system can respond more accurately to sudden rainfall and flow changes, ensuring a balance between flood control goals and water supply needs. Secondly, the method identifies the flood control safety margin in advance by predicting the reservoir water level change trend and comparing it with the flood level warning line during the flood season, allowing reservoir managers to take effective preventive measures before the flood occurs to avoid the reservoir facing the risk of over-storage due to excessively high water levels. Furthermore, by combining meteorological data with inflow information to predict future flood conditions, this method not only improves reservoir flood forecasting capabilities but also provides a more scientific basis for reservoir operation, enabling reservoirs to flexibly adjust control strategies under different flood season scenarios, ensuring rapid and effective water level reduction and mitigating flood risks. Regarding pipeline water hammer monitoring, this method effectively monitors and reduces its frequency by adjusting and optimizing reservoir outflow control parameters in real time, thus preventing damage to pipeline facilities caused by pressure waves and water shocks caused by rapid flow fluctuations. By calculating changes in water velocity, pressure wave propagation characteristics, and the risk of water hammer, reservoir outflow control parameters can be adjusted in real time to ensure the stability and safety of the pipeline system. This control measure not only improves reservoir outflow efficiency but also prevents sudden damage to the pipeline system, providing strong guarantees for the long-term operation of the reservoir. Furthermore, through the application of a multi-objective optimization model, this method dynamically balances the two key objectives of flood control safety margin and water supply security. By comprehensively assessing reservoir storage capacity and water supply security, it is possible to ensure both the safety of reservoir water level regulation during the flood season and the integrity of the reservoir's water supply capacity after floods. This balancing mechanism avoids the problems associated with previous single-objective control methods, such as excessive focus on flood control safety while neglecting water supply needs, or excessive emphasis on water supply while neglecting flood control. By dynamically adjusting reservoir outflow control parameters, this method automatically optimizes reservoir regulation strategies based on actual flood conditions at different times. This allows the reservoir to effectively respond to floods during the flood season while maintaining a stable water supply and meeting downstream water demand. This dynamic multi-objective optimization control method not only improves the accuracy of reservoir operation but also enhances its adaptability to extreme flood conditions. Combining advanced forecasting technology, real-time data monitoring, and intelligent optimization algorithms, it enables the reservoir to respond rapidly in complex environments. Furthermore, by mitigating safety hazards such as overflow and leakage caused by improper water level regulation, this method reduces the risk of dam failure and protects the lives and property of people downstream.In practical applications, with the accumulation of data from multiple flood seasons and the continuous optimization of models, the reservoir's dispatching capacity and safety will continue to improve, ensuring the reservoir's sustainability in long-term operation. In summary, the above methods, through multi-dimensional regulation and optimization, not only enhance the reservoir's flood control capacity but also ensure the continuity and stability of water supply, playing a vital role in the dynamic regulation of water levels during flood seasons. By precisely controlling water levels, scientifically predicting flood conditions, monitoring water hammer effects, and dynamically optimizing outflow control parameters, the reservoir can flexibly adjust its regulation strategy in the face of complex and changing flood season environments, maximizing the safety and water supply needs of the reservoir and its surrounding areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 This is a schematic flow chart of the steps of a dynamic multi-objective optimization control method for reservoir water level during flood season according to the present invention; Figure 2 for Figure 1 Detailed step flow diagram of step S1; Figure 3 for Figure 1 Detailed step flow chart of step S3 in FIG. DETAILED DESCRIPTION

[0009] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0010] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0011] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0012] To achieve this, please refer to Figures 1 to 3 The present invention provides a dynamic multi-objective optimization control method for reservoir water level during flood season, the method comprising the following steps: Step S1: Acquire real-time monitoring data of the reservoir; extract current water level status information according to the real-time monitoring data of the reservoir; establish a basic model of reservoir water balance based on the current water level status information; The embodiment of the present invention accesses monitoring equipment deployed in different areas of the reservoir in real time, including but not limited to water level meters, rain gauges, flow meters and evaporation sensors, to obtain real-time monitoring data of the reservoir, which include the current reservoir water level, water surface evaporation, inflow, rainfall, etc.; extracts the current water level status information from it, such as water level height, change rate and the corresponding relationship between water level and inflow; then, based on the current water level status information, the water balance relationship is used, that is, the inflow of water per unit time is subtracted from the outflow of water and evaporation loss, to obtain the reservoir water level change trend, and establish a basic reservoir water balance model. The model can use differential relationships to describe the process of water volume changing over time. For example, in a typical application scenario, real-time data is updated once a minute. When the inflow is 500 cubic meters per second, the evaporation is 5 mm per day, and the current water level is 120 meters, the water level change trend is simulated by accumulating changes, and the foundation is laid for subsequent prediction and analysis.

[0013] Step S2: Analyze the pipeline water supply flow rate variation characteristics based on the reservoir water balance basic model; identify the response lag time of pipeline fluid inertia based on the pipeline water supply flow rate variation characteristics; and predict the water supply pipeline pressure fluctuation risk based on the response lag time; Based on the reservoir water balance basic model established in step S1, the embodiment of the present invention analyzes the water flow variation characteristics of each major pipeline in the water supply system (such as the diversion tunnel and the flood discharge culvert). The embodiment mainly extracts the statistical characteristics of the change frequency, change amplitude, and change rate by performing time series analysis on the historical flow variation curve. For example, in an application scenario, if the fluctuation amplitude of the peak pipeline flow reaches 20% within 24 hours, the variation characteristic is considered to be significant. Furthermore, the response lag time of the pipeline fluid inertia is identified, that is, the time delay between the issuance of a command and the response to a significant flow change due to the large water mass and large inertia of flow velocity changes. The inertia lag time can be obtained by fitting the actual flow change curve after applying a small control command. For example, in a typical scenario, the response lag time is approximately 3 to 7 minutes. On this basis, based on the current lag time data and the flow variation amplitude, the risk range of abnormal pressure fluctuations in the water supply pipeline is predicted. The risk prediction can be comprehensively judged based on the flow mutation rate, lag time, and water hammer model. If a sudden increase in flow exceeding 15% within 5 minutes is detected, it is determined that there is a high risk of pressure fluctuation.

[0014] Step S3: determining a water level control lead time based on the water supply pipeline pressure fluctuation risk and the response lag time; formulating a water level control early intervention strategy based on the water level control lead time; and calculating a target water level control execution sequence based on the water level control early intervention strategy. In this embodiment of the present invention, based on the water supply pipeline pressure fluctuation risk and fluid response lag time analyzed in step S2, a water level control lead time is further determined. The so-called water level control lead time refers to the amount of time required to issue a pre-instruction in advance before the water level is officially controlled based on the response characteristics of the pipeline fluid. This amount of time is typically set based on the maximum lag time plus a system safety margin (generally taking a redundancy of 10%-20%). For example, if the lag time is measured to be 5 minutes in an application, the lead time may be set to 6 minutes. Subsequently, a water level control early intervention strategy is formulated. Specifically, based on factors such as predicted load changes and rainfall intensity in the weather forecast, a small adjustment instruction is issued 6 minutes in advance to avoid water hammer caused by sudden adjustment. Finally, based on the early intervention strategy, in combination with conditions such as the current water level, the target water level change rate, and the minimum adjustable step size, a target water level control execution sequence is calculated. The execution sequence includes the control start time, the control amplitude, and the control duration. For example, to meet the requirement of a target water level drop of 0.5 meters, the execution sequence may include a fine-tuning operation of lowering the target water level by 5 centimeters every 10 minutes for 3 consecutive hours.

[0015] Step S4: constructing a water level control gradient smooth curve based on the target water level control execution sequence; monitoring the probability of pipeline water hammer effect based on the water level control gradient smooth curve; adjusting parameters to ensure that the pipeline pressure wave and the control instruction do not resonate based on the probability of pipeline water hammer effect, and obtaining the reservoir outflow control parameters; In this embodiment of the present invention, based on the target water level control execution sequence calculated in step S3, a water level control gradient smoothing curve is constructed. This smoothing curve performs interpolation processing and first-order derivative continuity correction on water level change points in the control sequence to ensure a smooth rate of change of the water level during the control process and avoid sudden changes. For example, cubic spline interpolation is used to make the water level change curve continuous in time and without sharp turns. Based on the constructed gradient smoothing curve, combined with a flow change characteristic model and fluid inertia response parameters, the probability of water hammer effect in the water supply pipeline is monitored in real time. Water hammer effect refers to pressure shock caused by a sudden change in water flow velocity. The monitoring method can collect the pipeline pressure change rate and determine a high risk when the change rate exceeds a set threshold (e.g., 10% change per second). If a high water hammer risk is detected, a resonance risk analysis is performed based on the relationship between the current pressure fluctuation frequency and the control instruction frequency. By adjusting the amplitude, rhythm, or distribution of interpolation nodes of the water level control instruction, the pressure wave frequency is ensured to be away from the control signal frequency to avoid resonance. This ultimately determines reservoir outflow control parameters that are suitable for the current reservoir and pipeline operating conditions, such as the optimal opening ratio and timing of the flood discharge gate.

[0016] Step S5: Evaluate the flood control safety margin of the reservoir during the flood season according to the reservoir outflow control parameters, and adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season to achieve dynamic multi-objective optimization control of the reservoir water level during the flood season.

[0017] In the embodiment of the present invention, based on the reservoir outflow control parameters determined in step S4, the flood control safety margin of the reservoir is further evaluated during the flood season. The safety margin is defined as the dynamic margin between the current reservoir storage capacity, the upstream water forecast, and the downstream flood discharge capacity. The evaluation method adopts a dynamic simulation deduction method to simulate the water level change process under different rainfall scenarios in the next 24 hours, and update the safety margin index in real time. For example, the safety warning line is set to 135 meters. When the simulation predicts that the maximum water level does not exceed 134 meters and the flood discharge flow is controlled within the downstream safety flow limit, it is judged to be safe. If the safety margin is insufficient, the gate opening will be dynamically adjusted according to real-time monitoring and simulation results. Combined with the predicted flood peak arrival time, part of the reservoir capacity will be released first, while ensuring that no flood disasters occur downstream. In specific applications, for example, if a rainstorm with a peak flow of 800 cubic meters per second is expected within the next three hours, the gate opening can be increased to 30% two hours before the rainstorm arrives to free up sufficient reservoir capacity in advance. The entire water level regulation process realizes dynamic multi-objective optimization control of the reservoir water level during the flood season, taking into account goals such as flood control safety, downstream protection and rational use of water resources.

[0018] Preferably, step S1 includes the following steps: Step S11: Water level, flow, and rainfall sensors are deployed at the reservoir inlet, the middle of the reservoir area, and the spillway to periodically collect raw monitoring data and upload them to a data concentrator; The embodiment of the present invention provides the data foundation for implementing a dynamic multi-objective optimization control method for reservoir water levels during flood season. First, sensors are deployed in key areas of the reservoir. Ultrasonic water level meters and electromagnetic flow meters are installed at the main inlet of the reservoir to measure the inflow water level and flow rate. A group of radar water level gauges is deployed in the central reservoir area as a main control water level reference. Automatic rain gauges are deployed to monitor local rainfall in the reservoir area. Flow meters and water level gauges are installed at the spillway to monitor flood discharge in real time. All sensor equipment uses industrial-grade equipment with an IP68 protection rating and can adapt to high temperature and high humidity environments. To ensure data collection continuity, the sensors are set to a 5-minute collection cycle. Each collection includes a timestamp, measurement value, and status identifier. Each sensor regularly transmits raw monitoring data to a data concentrator located in the reservoir management center via a LoRa wireless communication module or a 4G gateway. The data concentrator has breakpoint resume and local cache functions to ensure data integrity and reserves an interface for subsequent data preprocessing.

[0019] Step S12: performing outlier detection on the original monitoring data to obtain real-time monitoring data of the reservoir; In the embodiment of the present invention, for the original monitoring data uploaded to the data concentrator in step S11, the data structure of each type of data is first standardized. For example, the water level, flow, and rainfall data are uniformly converted into data forms marked with UTC time standards and in units of meters, cubic meters per second, and millimeters. Then, outlier detection is carried out. The methods used include physical boundary verification (such as the normal range of water level is limited to between 100 meters and 160 meters above sea level), time continuity verification (such as the water level change between any two consecutive samples does not exceed 0.5 meters), and a sliding average test in the statistical method (the mean and standard deviation are calculated using a 30-minute sliding window. If a single value exceeds 3 times the standard deviation, it is judged to be abnormal). For the abnormal data found, they are eliminated, linearly interpolated, or retained as they are and marked according to different situations. Finally, a real-time monitoring data set of the reservoir with a complete structure and cleaned of outliers is formed, laying the foundation for subsequent water level status information extraction.

[0020] Step S13: extracting the reservoir real-time monitoring data, the inflow, outflow and rainfall from the reservoir real-time monitoring data, thereby obtaining the current water level status information; The embodiment of the present invention uses the real-time monitoring data of the reservoir obtained in step S12 to further extract key indicators in a targeted manner, namely the water level value, inflow, outflow and rainfall information of the reservoir at the current moment, wherein the water level value selects the reading of the radar water level gauge in the middle of the reservoir area as the main reference data, the inflow is directly measured by the electromagnetic flowmeter at the inlet, the outflow is provided by the spillway flowmeter, and the rainfall is the cumulative rainfall within 5 minutes of multiple automatic rain gauges arranged in the reservoir area and the average value is calculated. To ensure consistency, all data are aligned into a standardized format by time. For example, in the application scenario, if at 10:00 on July 10, 2025, the water level is 145.6 meters, the inflow is 320 cubic meters per second, the outflow is 180 cubic meters per second, and the rainfall is 1.2 mm, then these data are uniformly organized into the "current water level status information" record for the next step of analysis and call.

[0021] Step S14: Calculating the water level change rate based on the water level and flow changes over several time periods and determining the rising or falling trend based on the real-time monitoring data of the reservoir, thereby obtaining water level trend data; Based on the current water level status information extracted in step S13, the embodiment of the present invention selects 3 hours, 6 hours and 12 hours as characteristic time periods, and dynamically calculates the water level change rate. The specific operation is: take the water level difference between the starting point and the end point of each time period, divide it by the corresponding time period length, and thus obtain the water level change rate per unit time. At the same time, the water level change trend is determined based on the positive or negative sign of the rate. For example, if the water level rises from 145.6 meters to 146.2 meters within a 3-hour period, the change rate is 0.2 meters per hour, which is determined to be a rising trend of the water level; if the water level drops, it is determined to be a falling trend. If the change is less than a preset threshold (such as ±0.05 meters / hour), it is considered that the water level remains stable. Finally, water level trend data including water level change rates and trend states at different time scales are obtained, which provides a basis for evaluating the future dynamic changes of the reservoir water level and formulating optimized scheduling strategies.

[0022] Step S15: Calculating the water storage coefficient, evaporation loss rate, and leakage loss rate based on the water level trend data to obtain water balance model parameters; The embodiment of the present invention is based on the water level trend data obtained in step S14, and further combines the inflow, outflow and rainfall information to calculate the key parameters of the water balance model. The water storage coefficient refers to the amplitude of the water level change caused by the unit flow change, which can be obtained by statistically analyzing the ratio of the unit inflow to the water level change in several time periods. For example, if the total inflow within 6 hours is 6900 cubic meters and the water level rises by 0.3 meters, the water storage coefficient is 1 meter for every 23,000 cubic meters of water level rise; the evaporation loss rate is based on It is calculated based on the daily evaporation data from meteorological monitoring and the water surface area of ​​the reservoir. For example, if the daily evaporation monitored by the evaporation pan is 7 mm and the water surface area of ​​the reservoir is 5 square kilometers, the daily evaporated water volume is 35,000 cubic meters, which is converted into the reservoir capacity loss rate. The leakage loss rate is calculated by comparing the water level changes in the same period in history with the balance of inflow and outflow and rainfall. For example, the leakage loss calculated for a certain period of time is an average daily 0.01% of the reservoir capacity. Finally, a complete parameter set of water storage coefficient, evaporation loss rate, and leakage loss rate is formed to provide support for dynamic water volume simulation.

[0023] Step S16: Constructing a reservoir water balance basic model based on the current water level status information and water balance model parameters.

[0024] The embodiment of the present invention uses the water balance principle to construct a basic model of reservoir water balance based on the current water level status information extracted in step S13 and the water balance model parameters calculated in step S15. The specific modeling method is as follows: based on the initial water volume of the reservoir, the water volume increment of the reservoir in each time step (such as 5 minutes) is considered. The increment is obtained by subtracting the sum of the outflow flow, evaporation loss and leakage loss from the sum of the inflow water volume and rainfall. The reservoir capacity change is updated by continuously superimposing the water volume increment of each time step, and the real-time water level change is synchronously converted. In the scenario, for example, the current water level is 145.6 meters, the initial reservoir capacity is 20 million cubic meters, and within 5 minutes, the inflow is 1,600 cubic meters, the outflow is 900 cubic meters, the rainfall contributes 120 cubic meters, and the evaporation and leakage lose a total of 10 cubic meters. The net increase in water volume in this step is 810 cubic meters, and the updated reservoir capacity is 20,000,810 cubic meters. The corresponding water level rises slightly, and finally forms a basic water balance model that can dynamically reflect the changes in water volume and water level trends in the reservoir during flood season, laying a solid data and model foundation for the subsequent multi-objective optimization and regulation.

[0025] Preferably, analyzing the pipeline water supply flow variation characteristics according to the reservoir water balance basic model in step S2 includes: Generate theoretical water supply flow sequence based on the historical and real-time water level forecast sequence and outflow control parameters pre-acquired by the reservoir water balance basic model input; Compare the theoretical water supply flow sequence with the inflow flow in the reservoir real-time monitoring data, and calculate the flow error sequence of the two at the same time point; Calculate the mean, variance, and maximum deviation of the flow error series, and evaluate the prediction accuracy of the reservoir water balance basic model and the systematic deviations in the pipeline water delivery process; The flow rate change rate is calculated based on a fixed time window in the theoretical water supply flow sequence to obtain the pipeline water supply flow rate change rate sequence; Generate a flow change characteristic curve based on the pipeline water supply flow change rate sequence; Extract key inflection points and their corresponding moments from the flow change characteristic curve, and evaluate the effective change of each inflection point in combination with the systematic offset; The effective change before and after each inflection point and the flow change characteristic curve are integrated into the pipeline water supply flow change characteristics.

[0026] The embodiment of the present invention is based on the reservoir water balance basic model established above. First, the pre-acquired historical water level prediction sequence, real-time water level prediction sequence and outflow control parameters are input for simulation calculation. The historical water level prediction sequence can be obtained by modeling the daily water level change trend of the same flood season in the past three years. The real-time water level prediction sequence is comprehensively estimated by the water level change and the weather forecast rainfall in the past three days. The outflow control parameters include the gate opening scheduling plan and the maximum outflow capacity of the spillway. Through the water balance basic model, the theoretical water supply flow is calculated in each simulation time step according to the dynamic balance relationship between the inflow flow, evaporation, leakage and outflow flow, that is, the optimal discharge or water supply flow under the current predicted water level conditions, thereby generating a theoretical water supply flow sequence. The sequence is generally recorded in hours for subsequent comparison and feature extraction. Subsequently, the generated theoretical water supply flow sequence is compared moment by moment with the inflow flow sequence from the aforementioned real-time monitoring data. The difference between the theoretical water supply flow and the actual inflow flow at each time point is calculated, resulting in a flow error sequence. The units of this error sequence are typically cubic meters per second. A positive flow error indicates that water supply exceeds inflow, posing a potential risk of water level decline; a negative error indicates insufficient water supply, posing a risk of water accumulation. This process is automated using a Python script, and the error results are stored for subsequent statistical analysis. For the resulting flow error sequence, the mean, variance, and maximum deviation are further calculated over the entire sequence. The mean reflects the overall supply-demand discrepancy, the variance measures the magnitude of the error fluctuation, and the maximum deviation identifies extreme anomalies. In the mean statistics, if the mean deviates from zero by more than a ±5% threshold, it is determined that the water balance model has a systematic bias and requires further calibration. These statistics can be used to assess the prediction accuracy of the reservoir water balance model under the current flood season conditions and to preliminarily determine whether there are unmodeled losses in the pipeline water supply system, such as hidden leakage or abnormally enhanced evaporation, to ensure more targeted adjustments to subsequent control parameters. Next, a fixed time window (e.g., 6 hours) is slid across the theoretical water supply flow sequence to calculate the rate of change of the flow within each window. This is the final flow minus the initial flow divided by the window time width. The unit of the rate of change is cubic meters per second per hour. This rate of change is used to reflect the dynamic regulation characteristics of the water supply strategy. For example, if the outflow is increased or decreased before and after heavy rainfall, significant fluctuations can be seen in the rate of change. Based on the pipeline water supply flow rate change sequence obtained above, the local weighted regression (LOWESS) method is used for curve smoothing to generate a continuous flow change characteristic curve that removes high-frequency noise interference. The LOWESS method can eliminate random disturbances caused by measurement errors while retaining the main change trend, making subsequent inflection point extraction and trend analysis more accurate and reliable. The characteristic curve is generally presented graphically to intuitively display the changing process of water supply flow control during the flood season.From the smoothed flow change characteristic curve, an inflection point detection algorithm based on the change of the first-order derivative is further used to identify all local maxima, minima, and inflection point locations, and record the time point corresponding to each inflection point. Subsequently, combined with the aforementioned statistical systematic offset, the effective change amount correction is performed on the change amount at each inflection point. That is, if the systematic offset is positive, the inflection point change amount should be subtracted from the offset mean, and vice versa, to ensure that the key change characteristics finally extracted reflect the actual supply and demand regulation effect rather than the influence of systematic errors. The effective change amount information before and after each inflection point extracted is integrated with the overall flow change characteristic curve to form a complete set of pipeline water supply flow change characteristic parameter sets. This characteristic set includes the inflection point location, corresponding time, inflection point change amplitude, change trend direction before and after (increasing or decreasing), and change duration, etc., and is ultimately used as the input for the subsequent water level dynamic multi-objective optimization scheduling model to support the balanced regulation of the three objectives of water level safety, water supply demand, and flood discharge risk during the flood season.

[0027] Preferably, the step S2 of identifying the response lag time of the pipeline fluid inertia according to the pipeline water supply flow rate variation characteristics includes: For each flow change process curve in the pipeline water supply flow change characteristics, mark the time when the flow adjustment instruction is issued and the time when the actual flow begins to change, and calculate the time difference to form the initial response delay sequence; Fit the flow change process curve of each section to determine the time difference between the moment when the flow change rate reaches the peak and the moment when the command is issued, forming a peak response delay sequence; Calculate the time required for the flow change curve in the pipeline water supply flow change characteristics to change from the beginning to stability, and form a stable process time series; According to the initial response delay sequence, peak response delay sequence and stable process time series, group statistics based on different flow rate change amplitudes are performed to obtain hysteresis characteristic statistical data; The response lag time of pipeline fluid inertia is determined based on hysteresis characteristic statistics.

[0028] In an embodiment of the present invention, for each flow change process curve, the corresponding control scheduling log is first traced back to extract the issuance time of each flow adjustment instruction. These instruction times are usually derived from the operation records of the scheduling system, including manual operations or automatic control system instructions, such as flood gate opening adjustment, pump station start or stop, etc.; then, in the flow change curve, the first time point when the actual flow begins to change continuously is identified, which is usually determined by detecting that the flow change rate exceeds a preset small threshold (such as a change of more than 0.1 cubic meters per second per hour) in three consecutive time steps; the time difference between each pair of instruction issuance time and the actual flow change start time is calculated to form an initial response delay sequence, which is used to reflect the time lag characteristics of the system from receiving the instruction to generating an observable response. Next, for each of the above-mentioned flow change process curves, a piecewise fitting method is used to perform local polynomial fitting on each curve, usually using a second-order polynomial fitting, so as to smoothly and accurately capture the changing trend of the flow change rate; by taking the derivative of the fitting curve, the curve of the flow change rate changing with time is calculated, and further finds the moment when the flow change rate reaches the maximum value in each change process, that is, the time point when the change is most drastic; then, the time difference between the peak change moment and the time when the corresponding flow adjustment instruction is issued is calculated to form a peak response delay sequence; this sequence reflects the time required for the system to reach the maximum response speed after the scheduling instruction is issued, and is an important indicator of the inertial response of the fluid system. Continue to analyze the flow change process curve of each section to identify the time it takes from the beginning of the flow change to basic stability. The specific method is to find the initial point of change in the curve and track the subsequent flow changes. When the flow change amplitude is lower than the set threshold (such as less than 0.05 cubic meters per second per hour) in five consecutive time steps, it is determined that a stable state has been reached, and the flow change in the stable state is approximately constant; calculate the time difference between the moment the flow starts to change and the moment it is determined to be stable, and obtain the stable process time of each flow change section, and summarize it to form a stable process time series; this time series is used to evaluate the time characteristics required for the pipeline system to reach a new stable state when facing flow adjustments of different amplitudes. Subsequently, the obtained initial response delay series, peak response delay series and stable process time series are grouped and counted according to the flow change amplitude, that is, they are divided into three intervals according to the absolute value of the flow change: small change (less than 5 cubic meters per second), medium change (5 to 15 cubic meters per second) and large change (greater than 15 cubic meters per second); in each interval, the mean, standard deviation and maximum value of each lag indicator (initial response delay, peak response delay, stabilization time) are counted respectively to describe the statistical characteristics of the system response under flow control of different amplitudes; this process uses Python's Pandas library for data grouping and statistics, and visualizes the results in the form of histograms or box plots to facilitate the identification of abnormal lags or phenomena with significantly different lag characteristics under specific amplitude changes.Based on the above hysteresis characteristic statistical data, the comprehensive response lag time of pipeline fluid inertia is determined. The specific method is as follows: first, in the small, medium and large change groups, the weighted average of the initial response delay mean and the peak response delay mean is calculated respectively. The weight is set according to the actual number of occurrences to reflect the impact of the frequency of occurrence of various changes in actual operation; second, combined with the median of the stabilization process time, the total lag time required from the issuance of the instruction to the basic stabilization of the system under typical flow adjustment scenarios is comprehensively evaluated; finally, the comprehensive lag time under the small change scenario is used as the minimum response limit, the comprehensive lag time under the large change scenario is used as the maximum response limit, and the lag time under the medium change scenario is used as the reference setting of the optimization control model, which is used to consider the advance compensation of the timing of issuing control instructions in the subsequent dynamic multi-objective optimization scheduling of water levels.

[0029] Preferably, predicting the water supply line pressure fluctuation risk based on the response lag time in step S2 includes: The response lag time of pipeline fluid inertia is analyzed with the pre-acquired historical pressure fluctuation events to establish a preliminary corresponding relationship between lag time and pressure fluctuation. Based on the response lag time and the change characteristics of the pipeline water supply flow rate, a model of the fluid momentum change in the pipeline is established to calculate the theoretical pressure peak; Calculate the pipeline pressure bearing capacity and natural vibration frequency of the pipeline system based on the pre-acquired pipeline geometric parameters and material properties; Compare the theoretical pressure peak with the pipeline pressure bearing capacity to generate a pressure overlimit risk index; The pressure wave resonance risk index is evaluated using the preliminary correspondence combined with the matching degree of the natural vibration frequency and the response lag time; The pressure fluctuation risk of water supply pipelines is predicted based on the pressure overlimit risk index and pressure wave resonance risk index.

[0030] Based on the above-mentioned response lag time for determining the inertia of pipeline fluid, the embodiment of the present invention further performs a time correlation analysis between the response lag time and the pre-collected historical pressure fluctuation events. In specific implementation, the historical pressure anomaly events that occurred during the reservoir scheduling operation in the past three flood seasons are first sorted out, including pressure monitoring records triggered by pump station shutdown, rapid sluice closure, etc. These records contain the time when the pressure anomaly occurred and the pressure change amplitude; then, the flow adjustment instruction time before each historical pressure fluctuation event is compared, combined with the known response lag time, to retrospectively determine whether the event may be caused by a sudden change in flow rate due to the lag effect; by setting a maximum tolerance deviation (such as 5 seconds) window, it is determined whether there is a time correspondence, and finally the frequency and proportion of pressure fluctuations triggered in different lag time intervals are counted, and a preliminary correspondence between lag time and pressure fluctuation events is established, providing a basis for subsequent pressure fluctuation risk assessment. Based on the above time correlation analysis results, the pressure data series of each pipeline measuring point are further extracted from the reservoir real-time monitoring system. Reservoirs usually have multiple pressure monitoring points along the main water supply pipeline, such as water inlet, main trunk section, branch point and pump station outlet, and the sampling frequency is generally once per second. After extracting the data, the sliding window statistical method (such as 30 seconds as a window and a step size of 5 seconds) is used to calculate the characteristic quantities such as the pressure mean, standard deviation, and maximum change rate in each window. In addition, in order to characterize the intensity of pressure fluctuations, the pressure fluctuation amplitude index is introduced, that is, the ratio of the maximum pressure change per unit time to the initial value. At the same time, the time point when the mutation occurs is recorded to form a pressure fluctuation characteristic data set, which lays the data foundation for the subsequent coupling calculation with the fluid momentum change model. Based on the flow change characteristics and response lag time information extracted above, a momentum change model of the fluid in the pipeline is established, and the theoretical pressure peak is calculated. In the specific implementation, the fluid in the pipeline is first regarded as a one-dimensional non-steady flow system according to the control volume method, and the pipeline length, cross-sectional changes and friction losses are taken into account; in the momentum change equation, the momentum increment is equal to the pressure difference multiplied by the pipeline cross-sectional area minus the friction resistance term, plus the water mass multiplied by the flow velocity change rate multiplied by the time step; under the existing flow change curve and response lag time correction, the instantaneous flow velocity change rate is calculated and substituted into the above momentum balance relationship to reversely infer the instantaneous pressure change; the accumulated maximum pressure increment superimposed on the initial pressure is the theoretical pressure peak. Commonly used parameters in this step include pipeline diameter of 1 meter, initial flow velocity of 2 meters per second, fluid density of 1000 kilograms per cubic meter, pipeline roughness of 0.01 meters and other specific settings.After calculating the theoretical pressure peak based on the momentum change model, the geometric parameters (such as diameter, wall thickness, and length) and material properties (such as elastic modulus and yield strength) of the pipeline are extracted in combination with the pipeline design drawings and data provided by the reservoir management department. The thin-walled cylinder theory is used to calculate the maximum pressure-bearing capacity of the pipeline, that is, the maximum allowable internal pressure is equal to the material yield strength multiplied by the wall thickness and then divided by the pipeline radius, and corrections are made to consider the temperature influence coefficient and aging reduction coefficient. In addition, based on the wave theory, the natural vibration frequency of the pipeline is calculated as follows: the natural frequency of the pipeline is equal to the pressure wave propagation velocity divided by the pipeline length, where the pressure wave velocity can be calculated jointly based on the material elastic modulus and the water body bulk modulus. The actual parameters of this step are such as a pipeline wall thickness of 20 mm, a yield strength of 350 MPa, and a pipe length of 500 meters. The natural frequency is preliminarily estimated to be around 20 Hz. Subsequently, the theoretical pressure peak obtained in the previous step is compared with the pressure bearing capacity of the pipeline to calculate the pressure overlimit risk index. The specific method is: subtract the pressure limit pressure from the theoretical pressure peak. If the result is positive, there is an overpressure risk. The risk index is defined as the ratio of the pressure overlimit to the pressure bearing capacity, and is normalized between 0 and 1; if the theoretical pressure peak is less than the pressure bearing capacity, the overlimit risk index is set to zero, indicating that there is no direct overpressure risk; for example, if the pressure limit of a section of pipeline is 2.5 MPa and the theoretical pressure peak is 2.8 MPa, the overlimit risk index is (2.8-2.5) divided by 2.5, that is, 0.12, indicating that there is a moderate overlimit risk and it needs to attract the attention of the dispatcher. Furthermore, the preliminary correspondence between the lag time and historical pressure fluctuations established previously is used, combined with the known natural vibration frequency of the pipeline and the response lag time for matching, to evaluate the possible pressure wave resonance phenomenon, and calculate the pressure wave resonance risk index accordingly; the specific method is: if the lag time is close to an integer multiple of the natural vibration frequency, such as the period corresponding to the natural frequency of 20 Hz is 0.05 seconds, then the 5-fold period is 0.25 seconds, if the lag time is within the range of ±10% (that is, 0.225 seconds to 0.275 seconds), it is determined that there is a potential resonance risk; the resonance risk index is set according to the degree of matching, with a complete match being recorded as 1, a partial match being recorded as 0.5, and a large deviation being recorded as 0; for example, if the response lag time of a certain adjustment operation is 0.26 seconds and the natural frequency is 20 Hz, the resonance risk index is set to 0.5, indicating that there is a certain resonance-inducing risk.Finally, based on the pressure overlimit risk index and the pressure wave resonance risk index, the pressure fluctuation risk of the water supply pipeline under specific scheduling operations is comprehensively evaluated. The specific method is: weights are assigned to the pressure overlimit risk index and the resonance risk index respectively, with typical values ​​of 0.7 for pressure overlimit and 0.3 for resonance risk, and the weighted summation is used to obtain a comprehensive risk score; the risk level is distinguished according to the score, such as a score less than 0.3 is low risk, 0.3 to 0.6 is medium risk, and more than 0.6 is high risk; in actual application, when a high risk is predicted, the execution time of the corresponding scheduling instruction will be automatically postponed in the dynamic multi-objective optimization control strategy, or the flow change amplitude will be adjusted to avoid the risk of structural damage to the pipeline system due to pressure fluctuations, and ensure the safety of water level scheduling and system stability during flood season.

[0031] Preferably, step S3 includes the following steps: Step S31: setting the initial value range of the water level control advance amount corresponding to different risk levels according to the water supply pipeline pressure fluctuation risk and response lag time; After completing the analysis of the pressure fluctuation risk and response lag time of the water supply pipeline, the embodiment of the present invention formulates a preliminary response strategy for water level control according to different risk levels. First, based on the predicted pressure fluctuation risk index and response lag time, the initial value range of the water level control advance corresponding to different risk levels is set. Among them, the water level control advance refers to the time advance value for taking control actions in advance before a significant water level change may occur. For example, when the pressure fluctuation risk index is less than 0.3 and the response lag time is less than 5 minutes, the water level control advance is set between 10 minutes and 20 minutes; when the risk index is between 0.3 and 0.6 and the lag time is between 5 and 15 minutes, the advance range is set to 20 minutes to 40 minutes; when the risk index is higher than 0.6 and the lag time exceeds 15 minutes, the advance range is set to 40 minutes to 1 hour. Through this hierarchical setting, it is ensured that the response timing of water level control is reasonably arranged in different risk scenarios to avoid pipeline overpressure or uncontrolled vibration due to delayed response.

[0032] Step S32: screening the applicability of the water level control advance amount in the initial value range of the water level control advance amount according to the current water level state information, thereby obtaining the initial value of the water level control advance amount; After setting the initial ranges for water level control lead times for different risk levels, the embodiment of the present invention performs a suitability screening based on the current reservoir's real-time water level status to determine a specific initial value for the water level control lead time. This suitability screening process first collects real-time data on the current water level, water level change trend (rise or fall rate), current water storage capacity, and flood discharge capacity. If the current water level is less than 1 meter from the flood control limit and the water level rise rate is greater than 10 centimeters per hour, a smaller lead time value within the initial range is prioritized to accelerate the control response. If the current water level is more than 2 meters from the flood control limit and the water level rise rate is less than 5 centimeters per hour, a larger lead time value within the initial range can be selected to reduce the energy consumption and operating costs associated with frequent control. For example, at a certain moment, if the current water level is 95% of the upper reservoir capacity limit and the rise rate is 15 centimeters per hour, the initial lead time value is determined to be 45 minutes, within the range of 40 minutes to 1 hour corresponding to the high risk. Through suitability screening, the water level control strategy is more consistent with the actual operating conditions of the reservoir.

[0033] Step S33: establishing a candidate set of water level control advance intervention strategies based on the initial value of the water level control advance amount, wherein the candidate set of water level control advance intervention strategies includes control start timing, control target water level, and control duration; After determining the initial value of the water level control lead time, the present embodiment further establishes a set of candidate water level control advance intervention strategies based on this initial value. The candidate intervention strategy set includes the control start timing (i.e., the number of minutes after the current moment to delay control), the control target water level (i.e., the desired water level to be lowered or controlled by this intervention), and the control duration (i.e., the duration required for the intervention). In practice, multiple strategies are set centered around the initial value. For example, using the initial value of 45 minutes as a benchmark, strategies are set to start 40 minutes, 45 minutes, and 50 minutes in advance. The control target water level is set to a drop of 0.5 meters, 0.8 meters, or 1.0 meters from the current water level. The control duration is calculated within a reasonable range based on the pump's drainage capacity or flood discharge flow, such as 1 hour, 1.5 hours, or 2 hours. This yields nine different combinations of candidate strategies, forming the candidate intervention strategy set. This step ensures a sufficient selection of strategies to meet diverse operational requirements, facilitating subsequent optimization and screening.

[0034] Step S34: Screening the candidate set of water level control early intervention strategies based on the matching between the water level change rate and the flow rate change rate, eliminating strategies with change rates exceeding the system response capability, and selecting strategies with high water level control smoothness and small pressure fluctuations as the water level control early intervention strategies; This embodiment of the present invention screens the candidate set of intervention strategies based on their compatibility with the water level change rate and flow rate change rate, eliminating strategies whose change rates exceed the system's responsiveness. The water level change rate refers to the height of the water level change per unit time, while the flow rate change rate refers to the change in the volume of inflow and outflow per unit time. During this screening process, the system responsiveness threshold is first determined based on the design limits of the reservoir discharge system, such as a maximum allowable flood discharge rate of 200 cubic meters per second and a maximum allowable water level drawdown rate of 30 centimeters per hour. The change rate corresponding to each strategy in the candidate set is compared with these thresholds to eliminate strategies that cause excessively rapid water level drawdown or discharge volume overload. For example, if a strategy requires a one-meter drop within one hour, and the calculated water level change rate is one meter per hour, this far exceeds the allowable value and should be eliminated. Furthermore, among the remaining strategies, strategies with smoother water level control and less pressure fluctuation are prioritized. These strategies are ranked based on the smoothness index of the water level curve and the predicted fluctuation amplitude of the water supply pressure. Ultimately, strategies with high smoothness and less pressure fluctuation are selected as the early intervention strategies for water level control. This screening not only ensures the safety of regulation, but also takes into account the stability of system operation.

[0035] Step S35: Calculate the target water level control execution sequence based on the water level control early intervention strategy.

[0036] After determining the final early intervention strategy for water level control, the embodiment of the present invention calculates the target water level control execution sequence based on the strategy. The specific operation is to refine the control start timing, control target water level and control duration into minute-level or higher resolution execution instructions in combination with the real-time water supply flow and the operating capacity of the flood discharge equipment. For example, if the strategy is to start after 45 minutes, the target water level drops by 0.8 meters, and the duration is 1.5 hours, then with a time step of 5 minutes, the corresponding flood discharge flow change instructions are planned according to the flood discharge valve opening adjustment curve and the drainage capacity of the pump station to form an execution sequence. The execution sequence clearly marks the control actions such as the gate opening ratio and the start and stop status of the pump station required for each time node, and dynamically calculates the water level change forecast. The execution sequence is adjusted in real time to cope with sudden inflow changes, ensuring that the water level change process is smoothly controlled and the pressure fluctuation is maintained within a safe range. The target water level control execution sequence finally formed can be directly used as an input instruction of the reservoir scheduling system to guide the actual flood season water level control operation.

[0037] Preferably, step S35 includes the following steps: Step S351: Divide the target control period according to the control start timing and control duration in the water level control early intervention strategy, and generate a control time node set; After determining the early intervention strategy for water level control, the embodiment of the present invention first divides the target control period according to the control start time and control duration, and generates a control time node set. Specifically, assuming the current time is 8:00, and the early intervention strategy for water level control requires starting at 8:45 and lasting for 1.5 hours, the control period is from 8:45 to 10:15. The entire control period is then divided into fixed time steps, for example, selecting a 5-minute step size, that is, a time node is generated every 5 minutes at 8:45, 8:50, 8:55, and until 10:15, forming a control time node set. The control time node set refers to a set of time points arranged at predetermined intervals and serves as the basis for subsequent water level changes and the formulation of control instructions. For example, the set generated in this example is {8:45, 8:50, 8:55, 9:00, …, 10:15}, totaling 22 nodes. The division granularity (step size) can be flexibly set according to the response speed of reservoir regulation and the system adjustment capability. If higher control accuracy is required, a smaller step size such as 1 minute can be used.

[0038] Step S352: Calculate the corresponding target water level value at each control time node in the control time node set in combination with the target water level change to form a time-water level comparison table; After obtaining a set of control time nodes, the embodiment of the present invention calculates the target water level value corresponding to each control time node based on the target water level change, thus forming a time-water level comparison table. Specifically, the total water level change is calculated based on the initial water level, target water level, and total control duration. This change is then distributed to each time node, either evenly or according to a weighted trend. For example, assuming the water level before control begins is 135.0 meters, the target water level is 134.2 meters, and the total drop is 0.8 meters, lasting 1.5 hours (90 minutes). With a total of 22 time nodes, the water level needs to drop by approximately 0.036 meters every 5 minutes. Based on this result, the changes are accumulated in chronological order to determine the target water level value for each node, forming a time-water level comparison table such as {8:45-134.964 meters, 8:50-134.928 meters, 8:55-134.892 meters, ..., 10:15-134.2 meters}. If the regulation requires nonlinear changes, such as slowness in the early stage, acceleration in the middle stage, and slowdown in the late stage, a weighted allocation method can also be used. For example, the water level drops by 20% in the first 30% of the time, drops by 60% in the middle 40% of the time, and drops by 20% in the last 30% of the time, to further improve the rationality and safety of water level control.

[0039] Step S353: Calculate the water level change and change rate between adjacent time nodes according to the time-water level comparison table to form a water level change sequence; Based on the above-mentioned time-water-level comparison table, the embodiment of the present invention further calculates the water level change and rate of change between adjacent time nodes, forming a water level change sequence. Specifically, for each pair of adjacent time nodes, the water level change is calculated by subtracting the water level value of the previous node from the water level value of the next node. This change is then divided by the time interval (e.g., 5 minutes) to obtain the water level change rate. For example, if the water level drops from 134.964 meters to 134.928 meters from 8:45 to 8:50, the change is -0.036 meters, and the rate of change is -0.036 meters / 5 minutes, or -0.0072 meters / minute. All adjacent nodes are processed sequentially to form a complete water level change sequence, such as [-0.036 meters, -0.036 meters, -0.036 meters, ...]. This water level change sequence reflects the expected change at each time step and serves as the basis for calculating gate opening adjustments during subsequent control execution, ensuring that control actions in each time period accurately match the target water level change.

[0040] Step S354: Calculate the corresponding gate opening adjustment amount based on the water level variation sequence and the control target water level to form a gate opening execution instruction sequence; This embodiment of the present invention calculates the corresponding gate opening adjustment based on the water level change sequence and the overall target water level, generating a gate opening execution instruction sequence. Specifically, the required water discharge per unit water level change is determined based on a curve relating the reservoir gate to outflow (typically derived from measured data or system calibration, such as the functional relationship between opening percentage and flow rate). Based on the current effective reservoir area, the corresponding volume of flood discharge per unit water level drop is calculated. This is then converted into the required gate opening change based on the discharge capacity of the flood discharge device. For example, if the reservoir has an effective area of ​​1 square kilometer (i.e., 1 million square meters), a 0.036-meter water level drop requires the discharge of 36,000 cubic meters of water. If the discharge capacity of a single gate at 50% opening is 10,000 cubic meters per hour, 3,000 cubic meters of water will need to be discharged within 5 minutes, requiring the opening to be adjusted to approximately 15% (determined by looking up or interpolating a curve). This method calculates the gate opening adjustment for each time point, forming a sequence of opening execution instructions, such as {8:45-15%, 8:50-15%, 8:55-15%, ...}. This sequence ensures that each step of water level change is accompanied by a corresponding opening adjustment, achieving detailed and controllable dynamic regulation.

[0041] Step S355: Generate a target water level control execution sequence according to the gate opening execution instruction sequence, the water level change sequence and the control time node set.

[0042] After generating a gate opening execution instruction sequence, the embodiment of the present invention combines the water level change sequence and the control time node set to ultimately generate a target water level control execution sequence. Specifically, at each control time node, the gate opening instruction, water level change, and time information are packaged into a control instruction unit, such as "At 8:45, adjust the gate to 15% opening, with an expected water level change of -0.036 meters." All instruction units are then arranged in sequence to form a complete target water level control execution sequence. The execution sequence not only contains operational instructions but can also include monitoring and feedback requirements. For example, real-time monitoring of water level changes within 5 minutes after each operation. If the change deviates from the expected value by more than 10%, the adjustment instruction is automatically triggered to improve the robustness and safety of actual operation. Ultimately, this sequence serves as direct input to the reservoir scheduling system, guiding the automatic control equipment to gradually complete the optimized water level control during the flood season, achieving dynamic balance control of the reservoir's multiple objectives: flood control, water supply, and ecology.

[0043] Preferably, step S5 includes the following steps: Step S41: extracting the target water level value, gate opening, and execution time point at each time point based on the target water level control execution sequence to obtain preliminary water level control data; After completing the formulation of the target water level control execution sequence, the embodiment of the present invention first extracts the corresponding water level target value, gate opening, and execution time point at each time node to obtain preliminary water level control data. In specific implementation, the previously generated control execution sequence is parsed item by item. Each control instruction unit contains three data items: "time point, water level target value, and gate opening." These data items are extracted in chronological order and stored in a unified format, for example, in a table with the first column representing the execution time (e.g., 8:45, 8:50, etc.), the second column representing the corresponding water level target value (e.g., 134.964 meters, 134.928 meters, etc.), and the third column representing the gate opening percentage (e.g., 15%, 15%, etc.). The preliminary water level control data serves as the foundational data set for subsequent trend calculations, smoothing control actions, and evaluating water hammer effects. The "water level target value" represents the reservoir water level the system expects to reach at that time point, and the "gate opening" represents the opening percentage set by the system to operate the gate to achieve water level control.

[0044] Step S42: Calculating the rate of change sequence of water level and gate opening according to the preliminary water level control data; Based on the extracted preliminary water level control data, the present embodiment further calculates a sequence of water level and gate opening change rates. Specifically, for the target water level, a differential is performed between two adjacent time points: the next water level value minus the previous water level value, and then divided by the interval between the two time points (e.g., 5 minutes). This yields the water level change rate, typically expressed in meters per minute. For the gate opening, a similar differential is performed between adjacent nodes and divided by the interval to yield the gate opening change rate, expressed in percentage per minute. For an example, if the water level is 134.964 meters at 8:45 and 134.928 meters at 8:50, with a 5-minute interval, the water level change rate is (-0.036 meters) / 5 minutes, or -0.0072 meters / minute. If the gate opening is 15% at 8:45 and 15% at 8:50, the gate opening change rate is 0, indicating no gate opening change. Through this processing, two change rate sequences can be formed, which respectively describe the water level change trend and the gate adjustment trend, providing a basis for the next step of gradient curve construction.

[0045] Step S43: constructing an initial water level control gradient curve using the change rate sequence; The embodiment of the present invention uses the water level change rate sequence and gate opening change rate sequence obtained in the previous step to construct an initial water level control gradient curve. In specific implementation, time is used as the horizontal coordinate, and the water level change rate or gate opening change rate is used as the vertical coordinate. The change rate points of each time step are plotted in sequence, and then all points are connected in a broken line manner to form a preliminary gradient curve. The gradient curve reflects the law of the speed of change in the water level control process. The place with a high rate of change means that the water level adjustment amplitude is large and the speed is fast, and the place with a low rate of change means that the adjustment is gentle. For example, in certain periods of time, in order to quickly lower the water level to cope with sudden flood peaks, the gradient curve will drop sharply; when the water level approaches the target, in order to avoid overshoot, the rate of change approaches zero, making the curve flat. This initial gradient curve is an important basis for subsequent smoothing and optimization of control instructions.

[0046] Step S44: smoothing the initial water level control gradient curve to eliminate mutation points and discontinuity points, thereby forming a water level control gradient smooth curve; The embodiment of the present invention further smoothes the initial water level control gradient curve to eliminate mutation points and discontinuity points, thereby forming a water level control gradient smooth curve. In specific implementation, the gradient curve is processed using a sliding average method or a cubic spline interpolation method. Taking the sliding average as an example, a sliding window can be set, such as 3 time steps, and the rate of change of each time point is taken as the average of the rate of change of the previous and next time points and its own rate of change as the new rate of change, smoothing out sharp mutations and thus avoiding drastic fluctuations in water level changes during the control process. In the application, it was found that, for example, the water level change rate suddenly increased from -0.0072 m / min to -0.02 m / min at 8:55. After smoothing, it can be adjusted to -0.012 m / min, making the adjustment transition more natural. The formed water level control gradient smooth curve can ensure that the control action is continuous and smooth, reduce the system instability caused by frequent and rapid adjustments, and especially reduce the impact on the downstream river channel.

[0047] Step S45: monitoring the probability of water hammer effect in the pipeline based on the water level control gradient smooth curve; After obtaining a smoothed water level control gradient curve, embodiments of the present invention monitor the probability of water hammer in pipelines based on this curve. Water hammer refers to pressure fluctuations in water pipelines caused by rapid gate opening and closing or sudden flow changes, potentially causing pipeline damage or system failure. In specific implementations, the corresponding flow rate change rate is first calculated based on the water level change rate in each time period of the smoothed curve. Then, considering parameters such as the length, diameter, material, and water density of the pipeline system, an empirical formula or simulation model (such as a simplified Joukowsky formula) is used to calculate the potential instantaneous pressure rise. If the flow rate change rate within a certain time period is excessive, causing the pressure fluctuation amplitude to exceed a safety threshold (e.g., 80% of the pipeline's upper pressure limit), a high water hammer risk is determined for that time period. For example, if the predicted water hammer pressure increase within a 5-minute time step is 0.8 MPa, while the pipeline's maximum pressure limit is 1 MPa, this period warrants special attention. By scanning the entire time period, a water hammer probability distribution corresponding to each time step is generated.

[0048] Step S46: adjusting parameters to ensure that the pipeline pressure wave and the control instruction do not resonate based on the probability of the pipeline water hammer effect, and obtaining the reservoir outflow control parameters.

[0049] After obtaining a smoothed water level control gradient curve, embodiments of the present invention monitor the probability of water hammer in pipelines based on this curve. Water hammer refers to pressure fluctuations in water pipelines caused by rapid gate opening and closing or sudden flow changes, potentially causing pipeline damage or system failure. In specific implementations, the corresponding flow rate change rate is first calculated based on the water level change rate in each time period of the smoothed curve. Then, considering parameters such as the length, diameter, material, and water density of the pipeline system, an empirical formula or simulation model (such as a simplified Joukowsky formula) is used to calculate the potential instantaneous pressure rise. If the flow rate change rate within a certain time period is excessive, causing the pressure fluctuation amplitude to exceed a safety threshold (e.g., 80% of the pipeline's upper pressure limit), a high water hammer risk is determined for that time period. For example, if the predicted water hammer pressure increase within a 5-minute time step is 0.8 MPa, while the pipeline's maximum pressure limit is 1 MPa, this period warrants special attention. By scanning the entire time period, a water hammer probability distribution corresponding to each time step is generated.

[0050] It is particularly important that step S45 includes the following steps: Step S451: Calculating the water flow velocity change and pressure wave propagation characteristics in the pipeline based on the water level control gradient smooth curve and pre-acquired pipeline parameters; After completing the generation of the water level control gradient smoothing curve, the embodiment of the present invention then proceeds to calculate the change in water flow velocity and pressure wave propagation characteristics within the pipeline based on this curve and pre-acquired pipeline parameters. In specific implementation, it is first necessary to obtain the geometric parameters of the pipeline (such as pipeline diameter, length, material properties, etc.) and the physical parameters of the fluid (such as water density and viscosity, etc.). Then, based on the water level changes during the water level control process, the change in water flow velocity within the pipeline is calculated. During this process, the water flow velocity can be derived using a pipeline-based flow calculation formula. For example, the water flow velocity at each time point can be calculated based on the relationship between the water flow rate and the pipeline cross-sectional area. At the same time, combined with the principles of fluid mechanics, the effect of changes in water flow velocity on pressure wave propagation is further analyzed to obtain the pressure wave propagation characteristics within different time periods, including parameters such as wave velocity and amplitude. These are the basic data for evaluating the water hammer effect.

[0051] Step S452: establishing a mathematical model of pipeline water hammer effect based on the water flow velocity change and pressure wave propagation characteristics in the pipeline; After obtaining the water flow velocity change and pressure wave propagation characteristics in the pipeline, the embodiment of the present invention needs to establish a mathematical model of the pipeline water hammer effect. In specific implementation, a pipeline-based water hammer effect model is used, combined with the change in water flow velocity and the parameters of the pipeline, and the Joukowsky formula in fluid dynamics or a specific water hammer effect calculation model is used to describe how pressure waves are generated in the pipeline when the water flow changes drastically during the water level control process. This mathematical model takes into account the characteristics of the pipeline, such as pipeline length, diameter, wall roughness and other factors, and calculates the instantaneous pressure and flow rate changes of the fluid to infer the propagation speed of the pressure wave and its fluctuation range. The mathematical model can quantitatively evaluate the intensity of the water hammer effect and determine the peak value of the pressure wave that may be caused at different time nodes, which is crucial for subsequent control optimization.

[0052] Step S453: Calculate the water hammer pressure peak and occurrence frequency under the current water level control gradient smooth curve using a pipeline water hammer effect mathematical model to obtain water hammer effect risk data; The embodiment of the present invention is based on a mathematical model of the water hammer effect in pipelines. This model is further used to calculate the water hammer pressure peak and frequency of occurrence under the current water level control gradient smooth curve, thereby obtaining water hammer effect risk data. During specific implementation, based on the established water hammer effect model and combined with the current water level control information, the sudden change in water flow at each time point is simulated, and the water hammer pressure peak and the corresponding frequency of occurrence are calculated. For example, if the water level changes significantly and the gate opening changes rapidly during a certain period of time, it may cause the water flow in the pipeline to suddenly accelerate, thereby generating a large pressure wave. Through simulation or numerical calculation, the pressure peak (usually in MPa) and the frequency of water hammer occurrence (in times / hour) are determined, and based on this data, the water hammer effect risk that the pipeline may encounter is evaluated. Based on the risk assessment results, a parameter basis can be provided for subsequent regulation to ensure the stability of system operation.

[0053] Step S454: Calculate the probability of pipeline water hammer effect based on the water hammer effect risk data and the pipeline pressure bearing capacity.

[0054] The embodiment of the present invention further calculates the probability of a water hammer effect occurring in a pipeline based on water hammer risk data and the pipeline's pressure bearing capacity. During specific implementation, the pipeline's maximum pressure bearing capacity (such as the pressure limit and the strength of the pipeline material) is compared with the water hammer pressure peak to determine the probability of the pipeline experiencing a water hammer effect at different time points. If the calculated water hammer pressure peak exceeds the pipeline's pressure bearing capacity, the probability of a water hammer effect occurring during that period will be high, and vice versa. At this point, it is also necessary to comprehensively consider factors such as pipeline material, age, and flow rate changes. Through statistical methods, Monte Carlo simulation, and other means, the probability of a water hammer effect can be further optimized and calculated to obtain a numerical result reflecting the water hammer risk, which can be used to guide subsequent adjustments to control parameters.

[0055] It is particularly important that step S46 includes the following steps: Step S461: Analyze the frequency relationship between the pipeline pressure wave and the control instruction according to the probability of pipeline water hammer effect, and identify the resonance time node; The embodiment of the present invention analyzes the frequency relationship between the pipeline pressure wave and the control instruction based on the probability of the water hammer effect in the pipeline, and identifies possible resonance time nodes. In specific implementation, first, based on the water hammer effect risk data calculated previously, combined with the natural frequency of the pipeline (such as the vibration frequency of the pipeline system), the frequency of the gate opening change during the water level control process is compared. If in certain periods of time, the frequency of rapid changes in water flow is close to or coincides with the natural frequency of the pipeline, it is possible to induce resonance. At this time, the pressure wave will be further amplified due to the resonance effect, which may cause pipeline damage or system failure. By analyzing the frequency components of the control instruction through Fourier transform and combining it with the natural frequency characteristics of the pipeline, these potential resonance time nodes are identified, providing a reference for subsequent adjustment of the gate opening or water level control rate.

[0056] Step S462: adjusting the gate opening change rate or water level control rate at the corresponding time point according to the resonance time node, thereby obtaining an optimized water level control parameter set; After identifying the resonance time node, the embodiment of the present invention needs to adjust the gate opening change rate or water level control rate at the corresponding time point to obtain an optimized water level control parameter set. During specific implementation, for the identified resonance time node, the gate opening change rate is adjusted during these time periods to avoid excessive flow rate changes or sharp fluctuations in water level during these time periods. For example, the gate opening change can be slowed down, or the water level adjustment time can be extended to avoid resonance between rapid water flow changes and the natural frequency of the pipeline. These adjustments will help to smooth the control process and reduce the risk of water hammer effect. Through the adjusted parameter set, the outflow of the reservoir can be effectively controlled and the pressure wave of the system can be prevented from exceeding the safety range.

[0057] Step S463: recalculating the probability of water hammer effect based on the optimized water level control parameter set, and verifying the resonance frequency of the pipeline pressure wave and the control instruction, thereby obtaining adjustment parameter data; The embodiment of the present invention further recalculates the probability of the water hammer effect based on the optimized water level control parameter set, and verifies the resonant frequency relationship between the pipeline pressure wave and the control instruction. In specific implementation, the water hammer pressure peak and frequency will be recalculated according to the adjusted water level control parameters to evaluate whether there is still a resonance risk, and verify whether the adjusted gate opening and water level change rate have effectively eliminated the resonance phenomenon. If the probability of the water hammer effect is greatly reduced after the adjustment, and the pressure wave of the pipeline no longer resonates with the frequency of the control instruction, the effectiveness of the adjustment can be confirmed. Through this process, it is ensured that the reservoir control operation can avoid pressure wave damage to the pipeline system while meeting the water level target.

[0058] Step S464: Generate reservoir outflow control parameters based on the adjustment parameter data.

[0059] Finally, this embodiment of the present invention generates reservoir outflow control parameters based on the adjusted parameter data. During implementation, the optimized water level control parameters are combined with relevant data on the pipeline's pressure bearing capacity to generate the final outflow control parameters. These parameters include the rate of change of gate opening during the control process, the water level control rate within a time period, and the pressure wave adjustment limit of the pipeline system. These ultimately generated control parameters ensure efficient and stable water level control during the reservoir's flood season, while preventing pipeline damage caused by the water hammer effect and ensuring the safety and reliability of the entire control process.

[0060] Preferably, step S5 includes the following steps: Step S51: Calculate the reservoir outflow at different time periods based on the reservoir outflow control parameters, and predict the future water level change trend in combination with the reservoir water balance basic model; The embodiment of the present invention needs to calculate the reservoir outflow in different time periods based on the reservoir outflow control parameters. The calculation of the outflow is based on the actual outflow control system and historical scheduling records of the reservoir. In reservoir regulation, outflow is a key parameter for the reservoir's water storage capacity and downstream water level control. Therefore, the outflow calculation usually takes into account factors such as the reservoir's water level changes, gate opening, and water release volume. In order to obtain the outflow in each time period, it can be calculated using the reservoir's water balance basic model. This model combines the reservoir's water storage capacity, precipitation in the reservoir area, inflow, and existing outflow control data to calculate the actual outflow at each time point. At the same time, combined with historical water level data and scheduling strategies, numerical simulation or dynamic simulation methods are used to predict the changing trend of the reservoir water level in future time periods. This process can provide a scientific basis for subsequent reservoir water level regulation during the flood season, ensuring the smooth completion of flood control and water supply tasks.

[0061] Step S52: Calculate the flood control safety margin of the reservoir during the flood season based on the comparison between the future water level change trend and the pre-acquired flood level warning line during the flood season; The embodiment of the present invention needs to compare the future water level change trend with the pre-acquired flood level warning line during the flood season to calculate the flood control safety margin of the reservoir. The flood control safety margin refers to whether the reservoir has sufficient flood control space at a specific water level to withstand possible flood rises. During implementation, the flood control safety margin is calculated by comparing the predicted results of the above-mentioned reservoir water level change trend with the flood level warning line during the flood season. This calculation usually involves two aspects: first, the reservoir water level must be lower than the warning line to avoid overflow or excessive pressure on the dam; second, based on the predicted water level change trend, the remaining controllable water level range of the reservoir is determined to ensure that under possible flood conditions, the reservoir still has a certain control capability. Based on this, the obtained flood control safety margin can be used in the subsequent multi-objective optimization model to adjust the reservoir control parameters.

[0062] Step S53: predicting the future flood situation based on the rainfall information and inflow in the reservoir real-time monitoring data and combined with meteorological information; The embodiments of the present invention are based on the rainfall information and inflow in the real-time monitoring data of the reservoir, combined with meteorological information, and need to predict the changing trend of future flood conditions. During implementation, rainfall data is first obtained through the real-time monitoring system of the reservoir. This data comes from the meteorological station in the basin where the reservoir is located and is usually collected on an hourly basis. At the same time, it is necessary to obtain inflow data within the reservoir basin. This data can be obtained using flow meters and hydrological models. Combined with meteorological forecast information, the precipitation and inflow in the next few days can be further predicted, thereby inferring the evolution of future flood conditions. Specifically, a rainfall-runoff model (such as the SCS-CN model or the unit hydrological response method) is used to infer the reservoir inflow in the next few days based on the historical rainfall and flow relationship. Combined with the rainfall data in the meteorological forecast, the development trend of the flood season is evaluated. This information will help reservoir managers to keep abreast of the impending flood situation in a timely manner and carry out effective prevention, control and scheduling.

[0063] Step S54: Based on the predicted flood situation, the water storage capacity and water supply security of the reservoir under the current outflow control parameters are evaluated, and a multi-objective optimization model of the reservoir water storage capacity and water supply security is constructed; The embodiment of the present invention evaluates the water storage capacity and water supply guarantee capacity of the reservoir under the current outflow control parameters based on the aforementioned flood forecast results. Specifically, it is first necessary to evaluate the current water storage status of the reservoir based on the predicted future inflow and rainfall data, combined with the water storage capacity of the reservoir, and estimate the water supply guarantee level in the future period in combination with historical water level records and water supply demand. If the water storage capacity of the reservoir is close to the warning line, or the predicted future water level changes may cause the reservoir to be over-stored, the system will automatically adjust the reservoir outflow control strategy to avoid excessive water storage in the reservoir and ensure the stability of the water supply capacity. At the same time, the system will also calculate the water supply guarantee level of the reservoir, that is, whether the reservoir can continue to provide sufficient water supply in the next few days, especially in extreme weather conditions. Based on the multi-objective optimization model of water storage capacity and water supply guarantee level, the operation capacity of the reservoir in the future flood season can be scientifically evaluated, and the balance between the dual goals of flood control and water supply can be ensured.

[0064] Step S55: Dynamically adjust the reservoir outflow control parameters through a multi-objective optimization model to balance the flood control safety margin and water supply security, and convert the adjusted reservoir outflow control parameters into gate opening adjustment instructions to perform dynamic multi-objective optimization control of the reservoir water level during flood season.

[0065] This embodiment of the present invention dynamically adjusts the reservoir's outflow control parameters based on the multi-objective optimization model of reservoir water storage and water supply security in step S54. The optimization model calculates the optimal set of reservoir outflow control parameters by considering both flood control safety margin and water supply security. During the optimization process, multi-objective optimization algorithms (such as genetic algorithms and particle swarm optimization) are typically used to weigh different outflow parameters, ensuring that flood control requirements are met while optimizing the reservoir's water supply security. The optimized outflow control parameters include the rate of change of gate opening and the temporal distribution of water release. Ultimately, these adjusted parameters are converted into gate opening adjustment instructions, ensuring that the reservoir can dynamically adjust the water level based on actual conditions, avoiding the risk of over-storage and meeting the dual objectives of flood control and water supply. Through this dynamic adjustment process, the reservoir can flexibly adjust its strategy based on different scenarios during the flood season to ensure reservoir safety and water supply capacity.

[0066] The present invention also provides a dynamic multi-objective optimization control system for reservoir water level during flood season, which is characterized in that it is used to execute the dynamic multi-objective optimization control method for reservoir water level during flood season as described above, and the dynamic multi-objective optimization control system for reservoir water level during flood season includes: The water balance analysis module is used to obtain real-time monitoring data of the reservoir; extract the current water level status information based on the real-time monitoring data of the reservoir; and establish the basic model of the reservoir water balance based on the current water level status information; The pipeline pressure risk assessment module is used to analyze the pipeline water supply flow variation characteristics based on the reservoir water balance basic model; identify the response lag time of pipeline fluid inertia based on the pipeline water supply flow variation characteristics; and predict the water supply pipeline pressure fluctuation risk based on the response lag time; The water level control prediction module is used to determine the water level control lead time based on the water supply pipeline pressure fluctuation risk and response lag time; formulate a water level control early intervention strategy based on the water level control lead time; and calculate the target water level control execution sequence based on the water level control early intervention strategy; The water hammer effect control module is used to construct a water level control gradient smoothing curve based on the target water level control execution sequence; monitor the probability of pipeline water hammer effect based on the water level control gradient smoothing curve; and adjust parameters based on the pipeline water hammer effect probability to ensure that the pipeline pressure wave and the control instruction do not resonate, thereby obtaining the reservoir outflow control parameters; The flood control dynamic optimization control module is used to evaluate the flood control safety margin of the reservoir during the flood season based on the reservoir outflow control parameters, and to adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season, thereby realizing dynamic multi-objective optimization control of the reservoir water level during the flood season.

[0067] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0068] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A dynamic multi-objective optimization control method for reservoir water level during flood season, characterized in that: The following steps are involved: Step S1: Acquire real-time monitoring data of the reservoir; extract current water level status information according to the real-time monitoring data of the reservoir; establish a basic model of reservoir water balance based on the current water level status information; Step S2: analyzing the pipeline water supply flow variation characteristics according to the reservoir water balance basic model; identifying the response lag time of the pipeline fluid inertia according to the pipeline water supply flow variation characteristics; Predicting water supply line pressure fluctuation risk based on response lag time; Step S3: determining the water level control advance amount according to the water supply pipeline pressure fluctuation risk and the response lag time; Formulate water level control early intervention strategy based on the advance amount of water level control; Calculate the target water level control execution sequence based on the water level control early intervention strategy; Step S4: constructing a water level control gradient smooth curve based on the target water level control execution sequence; monitoring the probability of pipeline water hammer effect based on the water level control gradient smooth curve; adjusting parameters to ensure that the pipeline pressure wave and the control instruction do not resonate based on the probability of pipeline water hammer effect, and obtaining the reservoir outflow control parameters; Step S5: Evaluate the flood control safety margin of the reservoir during the flood season according to the reservoir outflow control parameters, and adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season to achieve dynamic multi-objective optimization control of the reservoir water level during the flood season.

2. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Water level, flow, and rainfall sensors are deployed at the reservoir inlet, the middle of the reservoir area, and the spillway to periodically collect raw monitoring data and upload them to a data concentrator; Step S12: performing outlier detection on the original monitoring data to obtain real-time monitoring data of the reservoir; Step S13: extracting the reservoir real-time monitoring data, the inflow, outflow and rainfall from the reservoir real-time monitoring data, thereby obtaining the current water level status information; Step S14: Calculating the water level change rate based on the water level and flow changes over several time periods and determining the rising or falling trend based on the real-time monitoring data of the reservoir, thereby obtaining water level trend data; Step S15: Calculating the water storage coefficient, evaporation loss rate, and leakage loss rate based on the water level trend data to obtain water balance model parameters; Step S16: Constructing a reservoir water balance basic model based on the current water level status information and water balance model parameters.

3. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 2 is characterized in that: Analyzing the pipeline water supply flow variation characteristics according to the reservoir water balance basic model in step S2 includes: Generate theoretical water supply flow sequence based on the historical and real-time water level forecast sequence and outflow control parameters pre-acquired by the reservoir water balance basic model input; Compare the theoretical water supply flow sequence with the inflow flow in the reservoir real-time monitoring data, and calculate the flow error sequence of the two at the same time point; Calculate the mean, variance, and maximum deviation of the flow error series, and evaluate the prediction accuracy of the reservoir water balance basic model and the systematic deviations in the pipeline water delivery process; The flow rate change rate is calculated based on a fixed time window in the theoretical water supply flow sequence to obtain the pipeline water supply flow rate change rate sequence; Generate a flow change characteristic curve based on the pipeline water supply flow change rate sequence; Extract key inflection points and their corresponding moments from the flow change characteristic curve, and evaluate the effective change of each inflection point in combination with the systematic offset; The effective change before and after each inflection point and the flow change characteristic curve are integrated into the pipeline water supply flow change characteristics.

4. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 3 is characterized in that: The response lag time of identifying the pipeline fluid inertia according to the pipeline water supply flow rate variation characteristics in step S2 includes: For each flow change process curve in the pipeline water supply flow change characteristics, mark the time when the flow adjustment instruction is issued and the time when the actual flow begins to change, and calculate the time difference to form the initial response delay sequence; Fit the flow change process curve of each section to determine the time difference between the moment when the flow change rate reaches the peak and the moment when the command is issued, forming a peak response delay sequence; Calculate the time required for the flow change characteristic curve in the pipeline water supply flow change characteristics to change from the beginning to stability, and form a stable process time series; According to the initial response delay sequence, peak response delay sequence and stable process time series, group statistics based on different flow rate change amplitudes are performed to obtain hysteresis characteristic statistical data; The response lag time of pipeline fluid inertia is determined based on hysteresis characteristic statistics.

5. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 4 is characterized in that: Predicting the water supply line pressure fluctuation risk based on the response lag time in step S2 includes: The response lag time of pipeline fluid inertia is analyzed with the pre-acquired historical pressure fluctuation events to establish a preliminary corresponding relationship between lag time and pressure fluctuation. Based on the response lag time and the change characteristics of the pipeline water supply flow rate, a model of the fluid momentum change in the pipeline is established to calculate the theoretical pressure peak; Calculate the pipeline pressure bearing capacity and natural vibration frequency of the pipeline system based on the pre-acquired pipeline geometric parameters and material properties; Compare the theoretical pressure peak with the pipeline pressure bearing capacity to generate a pressure overlimit risk index; The pressure wave resonance risk index is evaluated using the preliminary correspondence combined with the matching degree of the natural vibration frequency and the response lag time; The pressure fluctuation risk of water supply pipelines is predicted based on the pressure overlimit risk index and pressure wave resonance risk index.

6. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 5 is characterized in that: Step S3 includes the following steps: Step S31: setting the initial value range of the water level control advance amount corresponding to different risk levels according to the water supply pipeline pressure fluctuation risk and response lag time; Step S32: screening the applicability of the water level control advance amount in the initial value range of the water level control advance amount according to the current water level state information, thereby obtaining the initial value of the water level control advance amount; Step S33: establishing a candidate set of water level control advance intervention strategies based on the initial value of the water level control advance amount, wherein the candidate set of water level control advance intervention strategies includes control start timing, control target water level, and control duration; Step S34: Screening the candidate set of water level control early intervention strategies based on the matching between the water level change rate and the flow rate change rate, eliminating strategies with change rates exceeding the system response capability, and selecting strategies with high water level control smoothness and small pressure fluctuations as the water level control early intervention strategies; Step S35: Calculate the target water level control execution sequence based on the water level control early intervention strategy.

7. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 6 is characterized in that: Step S35 includes the following steps: Step S351: Divide the target control period according to the control start timing and control duration in the water level control early intervention strategy, and generate a control time node set; Step S352: Calculate the corresponding target water level value at each control time node in the control time node set in combination with the target water level change to form a time-water level comparison table; Step S353: Calculate the water level change and change rate between adjacent time nodes according to the time-water level comparison table to form a water level change sequence; Step S354: Calculate the corresponding gate opening adjustment amount based on the water level variation sequence and the control target water level to form a gate opening execution instruction sequence; Step S355: Generate a target water level control execution sequence according to the gate opening execution instruction sequence, the water level change sequence and the control time node set.

8. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 7 is characterized in that: Step S5 includes the following steps: Step S41: extracting the target water level value, gate opening, and execution time point at each time point based on the target water level control execution sequence to obtain preliminary water level control data; Step S42: Calculating the rate of change sequence of water level and gate opening according to the preliminary water level control data; Step S43: constructing an initial water level control gradient curve using the change rate sequence; Step S44: smoothing the initial water level control gradient curve to eliminate mutation points and discontinuity points, thereby forming a water level control gradient smooth curve; Step S45: monitoring the probability of water hammer effect in the pipeline based on the water level control gradient smooth curve; Step S46: adjusting parameters to ensure that the pipeline pressure wave and the control instruction do not resonate based on the probability of the pipeline water hammer effect, and obtaining the reservoir outflow control parameters.

9. The dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 8, characterized in that: Step S5 includes the following steps: Step S51: Calculate the reservoir outflow at different time periods based on the reservoir outflow control parameters, and predict the future water level change trend in combination with the reservoir water balance basic model; Step S52: Calculate the flood control safety margin of the reservoir during the flood season based on the comparison between the future water level change trend and the pre-acquired flood level warning line during the flood season; Step S53: predicting the future flood situation based on the rainfall information and inflow in the reservoir real-time monitoring data and combined with meteorological information; Step S54: Based on the predicted flood situation, the water storage capacity and water supply security of the reservoir under the current outflow control parameters are evaluated, and a multi-objective optimization model of the reservoir water storage capacity and water supply security is constructed; Step S55: Dynamically adjust the reservoir outflow control parameters through a multi-objective optimization model to balance the flood control safety margin and water supply security, and convert the adjusted reservoir outflow control parameters into gate opening adjustment instructions to perform dynamic multi-objective optimization control of the reservoir water level during flood season.

10. A dynamic multi-objective optimization control system for reservoir water level during flood season, characterized in that: Used to execute the dynamic multi-objective optimization control method for reservoir water level during flood season according to claim 1, the dynamic multi-objective optimization control system for reservoir water level during flood season comprises: The water balance analysis module is used to obtain real-time monitoring data of the reservoir; extract the current water level status information based on the real-time monitoring data of the reservoir; and establish the basic model of the reservoir water balance based on the current water level status information; The pipeline pressure risk assessment module is used to analyze the pipeline water supply flow variation characteristics based on the reservoir water balance basic model; identify the response lag time of pipeline fluid inertia based on the pipeline water supply flow variation characteristics; and predict the water supply pipeline pressure fluctuation risk based on the response lag time; The water level control prediction module is used to determine the water level control lead time based on the water supply pipeline pressure fluctuation risk and response lag time; formulate a water level control early intervention strategy based on the water level control lead time; and calculate the target water level control execution sequence based on the water level control early intervention strategy; The water hammer effect control module is used to construct a water level control gradient smoothing curve based on the target water level control execution sequence; monitor the probability of pipeline water hammer effect based on the water level control gradient smoothing curve; and adjust parameters based on the pipeline water hammer effect probability to ensure that the pipeline pressure wave and the control instruction do not resonate, thereby obtaining the reservoir outflow control parameters; The flood control dynamic optimization control module is used to evaluate the flood control safety margin of the reservoir during the flood season based on the reservoir outflow control parameters, and to adjust the reservoir gate opening based on the dynamic balance decision of the flood control safety margin of the reservoir during the flood season, thereby realizing dynamic multi-objective optimization control of the reservoir water level during the flood season.

Images