Water resource adaptation management methods and systems that coordinate supply and demand in the context of climate change

Abstract

This invention relates to a method and system for water resource adaptation management with coordinated supply and demand in the context of climate change. Belonging to the fields of automatic control and electronic digital data processing, the method utilizes real-time sensing of climate prediction data and hydrological monitoring data, and dynamically adapts to changes in climate factors using inflow and demand correction models to achieve accurate prediction of future water resource conditions. The prediction results are then input into a control decision model to construct a multi-objective optimization problem aimed at balancing supply and demand and maximizing reservoir storage capacity. This problem generates coordinated supply and demand control commands for directly driving physical actuators. Finally, the commands are transmitted via a communication network to a control actuator network consisting of reservoir gates and pipeline valve groups, driving their execution and completing closed-loop control of the water resource system. This invention effectively improves the dynamic adaptability, robustness, and control accuracy of the water resource control system to climate change through climate correction, coordinated supply and demand optimization, and adaptive learning.

Description

Technical Field

[0001] This invention relates to the fields of automatic control and electronic digital data processing technology, specifically to a method and system for water resource adaptation management that coordinates supply and demand in the context of climate change. Background Technology

[0002] In existing technologies, water resource scheduling methods primarily focus on addressing changes in the physical properties of water supply networks. For example, Chinese invention patent CN 118759832 B discloses a data processing-based water resource scheduling method. This method mainly analyzes pipeline attributes and historical water supply data to calculate a network blockage index reflecting the degree of blockage caused by deposits such as scale and rust within the pipelines. Based on this index, it dynamically adjusts the proportional gain coefficient of the PID controller to achieve closed-loop stable control of the water pressure at the end of the network. While such methods can effectively address localized water pressure fluctuations caused by pipeline aging and blockage, their technological scope is limited to the internal physical state of the water supply network. Their core is a passive compensation mechanism for deterioration of the system's internal state.

[0003] However, against the backdrop of climate change, the core challenge of water resource management has shifted from static pipeline distribution efficiency to dynamic systemic supply-demand imbalance risks. The aforementioned existing technologies do not consider the dynamic impacts of key climate factors such as temperature, rainfall, and evapotranspiration on both the supply and demand sides of regional water resources. They cannot provide climate-corrected forecasts of future water inflows, nor can they assess the impact of climate change on water-demanding sectors such as agriculture and ecology. Furthermore, they lack the ability to coordinate and optimize the supply side (e.g., reservoir releases) and the demand side (e.g., user water allocation) at the macro level.

[0004] Therefore, there is an urgent need for a new smart water resource management solution that can proactively adapt to climate change and achieve coordination between supply and demand. Summary of the Invention

[0005] In order to solve the problems of the prior art, this invention provides a method and system for water resource adaptation management that coordinates supply and demand in the context of climate change.

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: a water resource adaptation management method that coordinates supply and demand under the background of climate change, comprising the following steps:

[0007] S1. Data Preparation Steps: Obtaining the Future Daily forecast temperature Predicted rainfall and predicted evapotranspiration Time-series data; obtain the current water level of the reservoir as monitored in real time by a physical sensor network. and current river flow ;

[0008] S2. Climate-corrected inflow calculation steps: Input climate data into the pre-configured inflow correction model to calculate the climate-corrected predicted inflow. ;

[0009] S3. Climate-corrected water demand calculation steps: Input climate data into the pre-configured water demand correction model to calculate the climate-corrected predicted water demand. ;

[0010] S4. Control Instruction Generation Steps: ... and The input is fed into the control decision model, which solves and generates supply-side and demand-side control commands for directly driving and controlling actuators.

[0011] S5. Closed-loop control steps: The control command is sent to the corresponding control actuator network through the communication network to drive it to perform the corresponding physical operation in order to realize closed-loop control of the water resource system.

[0012] In one specific implementation of the first aspect, the calculation rule for the inflow correction model in step S2 is as follows:

[0013]

[0014] The correction coefficient is determined according to the following rules:

[0015]

[0016]

[0017]

[0018] in, , , ;

[0019] in This is the temperature correction factor. This is a rainfall correction factor. Evapotranspiration correction factor This serves as a reference value for temperature. This serves as a reference value for rainfall. This serves as a reference value for evaporation. This is the temperature correction factor at time t. This is a rainfall correction factor. This is the evapotranspiration correction factor. This refers to the historical water volume.

[0020] In one specific implementation of the first aspect, the calculation rule for the water demand correction model in step S3 is as follows:

[0021]

[0022] For agricultural water use, , ;

[0023] in This is the temperature-based water demand correction factor. This is the water demand correction factor for evapotranspiration. This is the temperature baseline value used in water demand calculations. As the baseline value for evaporation, This represents historical water demand.

[0024] In one specific implementation of the first aspect, the control decision model in step S4 is the following mathematical optimization problem:

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] in, ,and ;

[0031] in Let be the water level at time t. For the target water level, This is the minimum allowable water level for the reservoir. This is the maximum allowable water level of the reservoir. For actuator opening, For the minimum opening of the actuator, For the maximum opening of the actuator, For the discharge flow function, As ecological base flow, This is the supply and demand balance coefficient. To optimize the target weight coefficient, It is the acceleration due to gravity. This is the discharge coefficient.

[0032] In one specific implementation of the first aspect, ecological base flow It is a dynamic value, which is based on the probability of extreme drought. Adjustments will be made:

[0033] .

[0034] Secondly, a water resource adaptation management system that coordinates supply and demand in the context of climate change includes:

[0035] Physical sensor networks include water level sensors deployed in rivers, flow sensors deployed in water pipelines, and smart water meters at the user end.

[0036] The control actuator network includes servo motors installed on reservoir gates, electric regulating valves in the water supply network, and solenoid valves in irrigation valve assemblies;

[0037] The server is communicatively connected to the physical sensor network and the control actuator network. Its memory stores a program that, when executed by the processor, is used to implement a water resource adaptation management method that coordinates supply and demand in the context of climate change.

[0038] In one specific implementation of the second aspect, when the program in the server is executed by the processor, an adaptive learning step is also performed after implementing the method:

[0039] Record predicted water demand Compared with actual water demand deviation sequence ;

[0040] Using ARIMA The model fits and predicts the deviation sequence:

[0041]

[0042] And utilize the predicted bias The subsequent water demand forecasts will be revised as follows: ;

[0043] in These are the autoregressive parameters of the ARIMA model. For moving average parameters, For the backoff operator, It is a white noise sequence.

[0044] In one specific implementation of the second aspect, the system further includes a fault diagnosis and fault-tolerant control module, which is configured to:

[0045] When at the preset time No confirmation signal was received from the actuator, or its actual opening degree. With instruction opening The deviation consistently exceeds the tolerance. achieve At this time, the actuator is determined to be faulty, and the fault-tolerant control strategy is activated.

[0046] In one specific embodiment of the second aspect, the system further includes a visualization and human-computer interaction interface, configured to: provide a web graphical interface for displaying system status and warning information, and provide a manual intervention interface allowing operators to adjust the optimization target weights as described in claim 4. and And conduct a simulation run.

[0047] Thirdly, a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements a water resource adaptation management method that coordinates supply and demand in the context of climate change.

[0048] The beneficial effects of this invention are as follows:

[0049] 1. This invention constructs a climate-corrected prediction model for water inflow and demand, and establishes an optimized decision-making mechanism that coordinates supply and demand, thereby achieving more accurate dynamic simulation and forward-looking regulation of the water resource system. This method effectively overcomes the prediction distortion caused by climate fluctuations under the traditional static management model, significantly improving the scientific nature and accuracy of water resource planning and allocation, and ensuring a balance between water supply and demand from the system's source.

[0050] 2. By introducing multi-objective optimization functions and dynamic constraints, this system can automatically generate coordinated instructions that take into account both supply-side regulation and demand-side management under complex and ever-changing climatic conditions. This not only achieves balanced optimization of multiple objectives such as maximizing reservoir storage capacity and minimizing user water shortage, but also ensures the ecological health of rivers through dynamic ecological baseflow constraints, thereby maintaining the comprehensive benefits and sustainability of the water resource system even under extreme hydrological conditions.

[0051] 3. This invention integrates adaptive learning and fault diagnosis and tolerance mechanisms, forming a closed-loop control system with self-evolution and high reliability. By continuously correcting prediction deviations and intelligently responding to equipment failures, the system can continuously optimize its decision-making quality and ensure the reliable execution of control commands, ultimately forming a smart water resource management platform with strong robustness and long-term adaptability. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the system framework of the present invention.

[0053] Figure 2 This is a schematic diagram of the water resource adaptation management method of the present invention.

[0054] Figure 3 This is a schematic diagram of the water inflow correction model of the present invention.

[0055] Figure 4 This is a schematic diagram of the water demand correction model of the present invention.

[0056] Figure 5 This is a schematic diagram of the optimization process of the control decision model of the present invention.

[0057] Figure 6 This is a schematic diagram of the adaptive learning process of the present invention.

[0058] Figure 7 This is a schematic diagram of the fault diagnosis and fault-tolerant control logic of the present invention. Detailed Implementation

[0059] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0060] like Figures 1 to 7 The method and system for water resource adaptation management that coordinates supply and demand in the context of climate change are shown.

[0061] Part 1: System Overall Architecture

[0062] The system of this invention adopts a layered architecture design, including a physical perception layer, a data transmission layer, a platform processing layer, and an application interaction layer.

[0063] The physical sensing layer, composed of various sensors and devices, is responsible for collecting real-time status data of the water resource system. Specifically, it includes:

[0064] Water level monitoring sensor: A pressure-type water level gauge is used, with a measurement range of 0-50 meters and an accuracy of ±0.1%FS. It is installed in front of the reservoir dam and at important river sections. During installation, ensure that the sensor is installed vertically to avoid direct impact from water flow.

[0065] Flow monitoring equipment includes electromagnetic flow meters and ultrasonic flow meters. Electromagnetic flow meters are used for pipeline flow monitoring with an accuracy class of 0.5; ultrasonic flow meters are used for open channel flow monitoring, employing the time-of-flight measurement principle with an accuracy of ±1%.

[0066] Water quality monitor: Equipped with a multi-parameter water quality monitor, the monitoring indicators include pH value (measurement range 0-14, accuracy ±0.1), dissolved oxygen (measurement range 0-20mg / L, accuracy ±0.1mg / L), turbidity (measurement range 0-1000NTU, accuracy ±2%), etc.

[0067] Meteorological monitoring station: Construct automatic weather stations to monitor meteorological elements such as temperature (measurement range -50℃ to +60℃, accuracy ±0.1℃), precipitation (measurement resolution 0.1mm, accuracy ±2%), wind speed and direction, and relative humidity.

[0068] Smart water meters: IoT smart water meters are installed at the water user's end, using NB-IoT communication, with a metering accuracy level of 2 and remote valve control function.

[0069] The data transmission layer employs a hybrid communication network to ensure reliable data transmission.

[0070] Wired communication: Fiber optic communication is used between fixed sites, with a transmission rate of not less than 100Mbps and a latency of less than 10ms.

[0071] Wireless communication: A hybrid network of 4G / 5G and NB-IoT is adopted. 4G / 5G is used for high-volume data transmission, while NB-IoT is used for low-power wide-area coverage. The communication protocol uses MQTT version 3.1.1, with a heartbeat interval of 60 seconds and a QoS level of 1 to ensure that messages are delivered at least once.

[0072] The platform processing layer is based on a cloud computing architecture and adopts a microservice design pattern.

[0073] Computing resources: Docker containerized deployment, CPU configuration of no less than 16 cores, memory of no less than 64GB, and storage using SSD hard disks with a capacity of no less than 2TB.

[0074] Software environment: The operating system is CentOS 7.6, the database is PostgreSQL 12.0, the cache is Redis 6.0, and the message queue is RabbitMQ 3.8.

[0075] Part Two: Inflow Correction Model

[0076] Model Principles and Construction Logic

[0077] The core idea of ​​the water inflow correction model is to quantify the nonlinear impact of key climate factors (temperature, rainfall, evapotranspiration) on natural water inflow and to dynamically correct the predicted water inflow based on historical statistics.

[0078] Basic water inflow It is usually calculated from hydrological models (such as the Xin'anjiang model and the Sacramento model) or historical runoff sequences from the same period. It reflects the expected water inflow under the absence of extreme climate anomalies.

[0079] Climate Correction: The model uses three correction factors The combined effects of temperature, rainfall, and evapotranspiration on water inflow are captured separately. Their product... As a comprehensive influencing factor, the baseline water inflow is scaled.

[0080] Temperature has an impact, but in the long term it will reduce glacial water storage; abnormally low temperatures may slow melting or increase permafrost, reducing water inflow. ).

[0081] Rainfall impact: Rainfall is the most direct positive correlation factor with water inflow. The model uses a piecewise function to distinguish the contribution under different rainfall intensities; torrential rains may produce a larger runoff coefficient due to infiltration saturation.

[0082] Evaporation has an impact.

[0083] Parameter determination method

[0084] The thresholds and coefficients in the model need to be determined using historical data-driven methods to ensure that they are applicable to a specific watershed.

[0085] Data preparation:

[0086] Collect historical daily data for the past 10-20 years in the watershed, including: temperature. Rainfall Potential evaporation (Can be calculated using the Penman formula), and the corresponding measured water inflow. .

[0087] Parameter calibration method:

[0088] Calculate the baseline water inflow: Input historical climate data into the hydrological model to obtain the simulated water volume. sequence.

[0089] Calculate the actual correction factor: For each day in history, calculate the actual composite correction factor. .

[0090] Establish relationships and determine thresholds:

[0091] Will Separately with the day Perform scatter plot analysis and multivariate nonlinear regression.

[0092] Temperature threshold: can be analyzed and The relationship curve is used to select the inflection point as the threshold. For example, It can be set as the starting point for the effects of high temperature (e.g., 25℃). It can be set as the starting point for the effects of low temperatures (e.g., 5°C). Coefficient , This is obtained through regression analysis, for example... It could be 0.02 / ℃, meaning that for every ℃ above 0.02, the value is... At one point, the water inflow increased by 2%.

[0093] Rainfall threshold , : It can be set as the minimum rainfall required to generate effective runoff (e.g., 5 mm / day). This can be set to the rainfall amount that generates strong runoff (e.g., 25 mm / day). Coefficient , , It reflects the runoff coefficient in different rainfall intensity ranges.

[0094] Evapotranspiration coefficient :pass and The negative correlation is determined. For example, It can be set to 0.01 / mm, which means that for every 1mm increase in daily potential evapotranspiration, the incoming water volume decreases by 1%.

[0095] Part Three: Water Demand Correction Model

[0096] Model Principles and Construction Logic

[0097] The core idea of ​​the water demand correction model is to identify the sensitivity of different water-using sectors (such as agriculture, industry, domestic and ecological) to climate conditions and establish quantitative relationships for prediction correction.

[0098] Basic water demand It is determined by water resource planning, historical water use habits, or economic development forecasts, and is a relatively static benchmark value.

[0099] Climate correction: The model focuses on correcting agricultural water demand, which is most sensitive to climate, while other sectors can be simplified as appropriate.

[0100] Agricultural water demand correction: Agricultural irrigation water demand and crop evapotranspiration ( ) directly related, and Reference crop evapotranspiration determined by climatic factors ( ) and crop coefficient ( This is jointly determined. (Formula) It is a simplified empirical model, in which:

[0101] This represents the amount of additional irrigation water needed due to the hot and dry climate.

[0102] This represents the substitution effect of effective rainfall for irrigation, reducing net water demand.

[0103] Parameter determination method

[0104] parameter and The calibration needs to be determined based on the specific crop type and irrigation area.

[0105] Data preparation: Collect historical daily data of the crop growing season in the study area: Effective rainfall (usually a proportion of total rainfall) and measured irrigation water usage. .

[0106] Parameter calibration method:

[0107] Determine basic water demand: It can be taken as the average irrigation water volume over many years or the planned water requirement divided according to the crop growth period.

[0108] Regression analysis: Establishing the regression equation (in (For error). The parameters can be obtained by fitting using the least squares method. and .

[0109] Physical meaning of the parameters:

[0110] This can be understood as the comprehensive response coefficient, and its value is close to that of the crop coefficient. The product of factors such as irrigated area and irrigation efficiency. For rice, The value may be as high as 1.0-1.3; for water-saving irrigated wheat, it may be 0.6-0.8.

[0111] This can be understood as the effective utilization coefficient of rainfall, with a value between 0 and 1, indicating what proportion of rainfall can be directly utilized by crops. For sandy soils, it might be 0.6, and for clay soils, it might be 0.8.

[0112] Part Three: Regulation Decision Model

[0113] Model Principles and Construction Logic

[0114] This model is a multi-objective constrained optimization problem. Its goal is to coordinate the optimization of the supply side (reservoir water release) and the demand side (user water allocation) under various physical and safety constraints, so as to achieve efficient, fair and sustainable use of water resources.

[0115] Objective function:

[0116]

[0117] First item (supply and demand balance): Minimize the predicted water demand. Compared with actual water supply The sum of squares of the differences. This reflects the core objective of the system to meet user needs, and the squared term results in a more severe penalty for large deviations.

[0118] The second item (reservoir storage): Maximize the reservoir's final water storage capacity. This reflects the system's ability to store water during periods of abundance and replenish water during periods of scarcity, as well as its long-term water resource security goals. and It is a weighting coefficient used to weigh these two competing objectives.

[0119] Constraints:

[0120] Water balance constraint: This is the most critical physical constraint, ensuring a balance between water supply and demand within the system.

[0121] Reservoir capacity constraints: Ensure that the reservoir water level operates between the dead storage capacity and the flood control limit / normal storage level.

[0122] Ecological baseflow constraints: Ensuring the health of the river's ecological environment is a manifestation of sustainable development goals.

[0123] Water supply capacity constraints: The water supply volume cannot exceed the maximum water conveyance capacity of the channel or pipeline.

[0124] Parameter determination method

[0125] Weighting coefficient , :

[0126] Methodology: This is a multi-criteria decision problem.

[0127] Expert experience method: This method is jointly determined by water resource management experts based on regional water resource policies and the current drought situation. For example, during the high-water season, a set of... During drought periods, to ensure basic water supply, the ratio can be set to 2:1 or higher.

[0128] Analytic Hierarchy Process (AHP): Construct a judgment matrix to compare the two criteria of "ensuring water supply" and "reservoir energy storage" pairwise, and scientifically determine the weights by calculating eigenvectors.

[0129] Implementation: The system default setting is... =0.7, =0.3. Operators can dynamically adjust this based on the warning information through the human-machine interface.

[0130] Dynamic ecological base flow :

[0131] Principle: In extreme drought conditions, in order to ensure basic water use for humans, the ecological baseflow standard can be appropriately and scientifically reduced.

[0132] benchmark value The minimum flow rate is usually determined using the Tennant method (e.g., taking 10% of the multi-year average flow rate) or based on hydraulic methods (e.g., the wetted perimeter method), and is the bottom line that must be guaranteed under normal circumstances.

[0133] Extreme drought probability The severity of drought can be calculated by analyzing the frequency of historical drought events or based on future climate forecasts (such as the Standardized Precipitation Index (SPI)). For example, when the predicted 7-day average SPI is less than -2.0, it is considered an extreme drought. 0.3 is acceptable.

[0134] Implementation: A certain river channel =5 m³ / s. When the system predicts... When the value is greater than 0.2, dynamic adjustment is initiated. =0.3, then =5×(1-0.3)=3.5 m3 / s.

[0135] Model solution:

[0136] This optimization model is a typical constrained quadratic programming (QP) problem.

[0137] Solution Algorithms: Mature optimization solvers are used, such as interior-point methods, effective set methods, or commercial / open-source optimization libraries (e.g., MATLAB's `quadprog` function, Python's `SciPy.optimize` module, or the `CVXPY` package). These algorithms can efficiently find the optimal solution that satisfies all constraints. and This generates specific control instructions.

[0138] Part Four: Control Instruction Generation Steps

[0139] This step is the brain of the invention, responsible for transforming the prediction results of the preceding steps into an executable control strategy.

[0140] Optimization solution and instruction conversion

[0141] After the control decision model is solved, the platform processing layer converts the numerical solution into specific equipment control commands.

[0142] Instruction generation logic:

[0143] Supply-side control directives:

[0144] Reservoir gate commands: Optimal discharge sequence output by the optimization model. Based on the reservoir's water level-capacity-discharge curve, the corresponding target gate opening is calculated. (Unit: meters or percentage).

[0145] Command format: Generates structured data containing [Device ID, Command Type, Target Value, Execution Time Window]. For example: [Gate_A01, OPEN_LEVEL, 65%, 2023-07-04T10:00:00Z] means that gate A01 will be opened to 65% at the specified time.

[0146] Demand-side control directives:

[0147] Channel / pipeline valve group instructions: Water distribution to each user based on the output of the optimization model. Based on the valve flow-opening characteristic curve, the target opening degree of each electric regulating valve or irrigation valve group is calculated. .

[0148] User water usage plan: For smart water meters with remote valve control capabilities, more refined instructions can be generated, such as [User_B25, FLOW_LIMIT, 0.5m]. 3 The string " / h, 2023-07-04T00:00:00Z / 2023-07-05T00:00:00Z" indicates that a 0.5m delay was implemented on the user on July 4th. 3 / hourly traffic limit.

[0149] Instruction security verification and simulation

[0150] Before the command is officially issued, the system will perform a final security check and simulation to prevent command conflicts or unsafe situations.

[0151] Safety constraint verification: Verify whether the expected reservoir water level after the command execution is within the specified range. and Between these points, does the river flow meet the dynamic ecological base flow requirements? Require.

[0152] Conflict detection: Check for contradictory operation commands in time and space (e.g., an instruction to open an upstream gate by a large margin when the downstream valve is not open).

[0153] Simulated operation ("digital twin"): Using the simulation function provided by the human-computer interaction interface, the operator can load the current system status and the sequence of instructions to be issued, conduct forward simulation, and intuitively observe the changing trend of the system status (water level, flow rate, reservoir capacity) over a period of time in the future. Only after confirming that there are no errors can the operation be approved.

[0154] Part 5: Instruction Issuance and Execution Steps

[0155] This step is the final step in the invention's transition from the digital world to the physical world, ensuring that the control intent is executed accurately and reliably.

[0156] Command issuance and confirmation

[0157] Communication delivery: The platform processing layer sends the verified control instructions to the corresponding control executors via the data transmission layer through a message queue (such as RabbitMQ).

[0158] For critical equipment such as reservoir gates and main canal valves, a high-priority message channel is used, and the QoS level is set to 2 to ensure that messages are delivered only once.

[0159] For a large number of distributed irrigation valve groups and smart water meters, the NB-IoT network is used for batch distribution.

[0160] Execution Confirmation: After receiving the command, the actuator drives the servo motor or solenoid valve to operate and feeds back the confirmation signal and actual opening degree / status to the platform processing layer in real time. The system records the command's "issued", "confirmed", and "executed" status.

[0161] Fault diagnosis and fault-tolerant control (real-time intervention)

[0162] This process occurs synchronously with instruction execution and is handled by a dedicated fault diagnosis and fault tolerance control module.

[0163] Real-time monitoring: The module continuously monitors the command execution status.

[0164] Communication timeout: If the communication timeout occurs within the preset time... If no confirmation signal is received from the critical actuator within 30 seconds, it is determined to be a communication failure.

[0165] Execution deviation: If the actual opening degree fed back by the actuator... With instruction opening The deviation consistently exceeds the tolerance. (e.g., ±5%) to reach If it occurs 3 times (e.g., 3 times), it is determined to be a failure of the actuator.

[0166] Fault tolerance strategy:

[0167] Primary fault tolerance: In the event of a momentary communication failure, the command is automatically retransmitted (up to 3 times).

[0168] Secondary fault tolerance: If retransmission is ineffective or the actuator is determined to be faulty, a pre-set safety contingency plan will be immediately activated. For example:

[0169] Close the upstream main valve of the faulty valve.

[0170] The water distribution task handled by this valve can be dynamically switched to the backup pipeline or its water supply can be temporarily reduced according to priority.

[0171] Send alarm information (such as SMS or App push) to maintenance personnel, including device ID, fault type and fault tolerance measures already taken.

[0172] At the same time, the system will record this event in the database for subsequent equipment maintenance and analysis.

[0173] Closed-loop feedback and adaptive learning triggering

[0174] After the instruction is executed, the system enters the next loop.

[0175] Status Update: The physical sensor network continuously monitors reservoir water levels, river flow, and user water consumption. This new real-time data is collected and transmitted back to update the system status, serving as the initial conditions for the next round of decision-making.

[0176] Triggered learning: The deviation between the recorded actual water demand and the predicted water demand is input into the adaptive learning module to update the ARIMA model, continuously optimize the future prediction accuracy, and thus achieve the self-evolution of the system.

[0177] Example: Taking the Qinghe River Basin, a typical river basin in northern China, as an example, the method and system of this invention are applied to water resource management. This river basin is mainly used for agricultural irrigation, but also has water needs for urban and rural domestic and industrial use. In recent years, it has been affected by climate change, resulting in frequent seasonal droughts.

[0178] 1. Scenario Setting and Data Preparation

[0179] Basin Overview: The Qinghe River basin covers an area of ​​5000 km². 2 Upstream is the Qingyuan Reservoir (total capacity 120 million cubic meters). 3 Xingli reservoir capacity 80 million m³ 3 The middle and lower reaches are the main irrigation areas and towns.

[0180] System Configuration:

[0181] Physical sensor network: Water level and flow sensors are deployed at three key river sections in front of and downstream of the reservoir dam; electromagnetic flow meters are installed at the head of the main canal and at the entrances of three major irrigation areas; and IoT smart water meters are installed for 500 typical agricultural users.

[0182] Control actuator network: The gates of Qingyuan Reservoir are equipped with servo motors; electric regulating valves are installed at the head of the main canal and the three branch outlets.

[0183] Server: Deployed in the cloud, configured with a 16-core CPU, 64GB of memory, and running CentOS 7.6 operating system and PostgreSQL database.

[0184] Data input (taking July 1st of a certain year as an example):

[0185] Climate prediction data: Gridded forecast data for the next 7 days (July 1-7) are obtained, and after interpolation and aggregation, the daily average series of the watershed surface is obtained.

[0186] Predicted temperature : [28, 30, 32, 35, 36, 34, 33] °C;

[0187] Predicted rainfall : [2, 0, 0, 15, 5, 0, 0] mm;

[0188] Predicted evaporation :[5.0, 5.2, 5.5, 5.8, 6.0, 5.7, 5.5] mm;

[0189] Real-time monitoring data (July 1, 8:00 AM):

[0190] Current water level of Qingyuan Reservoir = 745.0 m (corresponding to a reservoir capacity of 55 million m³) 3 );

[0191] Current flow rate of the river channel at the downstream control section = 12.5 m 3 / s.

[0192] 2. Model Application and Calculation Process

[0193] 2.1 Climate-corrected water inflow calculation

[0194] Basic water inflow Based on hydrological model simulations, the estimated natural inflow sequence for the next 7 days is [80, 82, 85, 90, 88, 85, 83] million m³. 3 / day.

[0195] Model parameters: Based on historical data of this watershed, the water inflow correction model parameters are as follows:

[0196] =28, =0, =0.015, =0.01;

[0197] =10, =30, =0.005, =0.015, =0.025;

[0198] =0.008;

[0199] Calculation process (taking the 4th day, i.e., July 4th, as an example):

[0200] That day =35°C, =15mm, =5.8mm, =90.

[0201] Calculate the correction factor:

[0202] ;

[0203] Therefore ;

[0204] ;

[0205] Overall correction factor: K = K1 × K2 × K3 = 1.105 × 1.075 × 0.9536 ≈ 1.132

[0206] Climate-corrected projected water inflow: Wc = Wb × K = 90 × 1.132 = 101.9 Wc = Wb × K = 90 × 1.132 = 101.9 million m³ 3 .

[0207] Similarly, the 7-day series is calculated to obtain... = [81.2, 84.5, 92.1, 101.9, 93.6, 87.8, 85.5] ten thousand m 3 It is evident that the combined effects of high temperatures and rainfall significantly increased the predicted water volume for the fourth day.

[0208] 2.2 Calculation of Climate-Modified Water Demand

[0209] Basic water demand Db: Based on the water use plan, the total basic water demand for the next 7 days is [65, 65, 65, 65, 65, 65, 65] ten thousand m³. 3 / day (of which agriculture accounts for 70%).

[0210] Model parameters: The agricultural water use correction parameters calibrated for the main crop (maize) in this watershed are as follows: =0.9, =0.7.

[0211] Calculation process (taking day 4 as an example):

[0212] That day =5.8mm, =15mm.

[0213] Corrected water demand for agriculture:

[0214] .

[0215] Convert mm to water volume (assuming an irrigated area of ​​500,000 mu, the conversion factor is 0.667 m). 3 / acre / mm), to get ≈-176,000 m 3 .

[0216] Climate-corrected projected water demand: 10,000 m 3 (Due to rainfall that day, the need for irrigation water was greatly reduced.)

[0217] Similarly, the 7-day series is calculated to obtain... = [63.5, 62.8, 62.0, 47.4, 58.5, 63.0, 63.5] ten thousand m 3 .

[0218] 2.3 Generation of Control Commands

[0219] Optimize model parameter settings:

[0220] Weighting coefficients (determined by AHP method, currently in the normal water period): =0.6 (focusing on supply and demand balance). =0.4 (focusing on water storage).

[0221] Ecological baseflow: baseline value =8.0m 3 / s. The meteorological drought index for the next 7 days shows normal, therefore... =0, dynamic ecological base flow = =8.0m 3 / s.

[0222] Reservoir constraints: Dead capacity corresponding =50 million m 3 Flood control reservoir capacity corresponding =72 million m 3 .

[0223] Model solution:

[0224] The next 7 days sequence, Sequence and initial storage capacity =55 million m 3 Decision-making model for various constraint inputs.

[0225] The quadratic programming problem is solved by calling the Python CVXPY library on the server.

[0226] Optimization results and instruction generation:

[0227] Reservoir discharge command S: The model outputs the optimal discharge sequence for the next 7 days as [8.5, 8.3, 10.1, 15.2, 9.0, 8.5, 8.2] m. 3 / s. Notably, on the 4th day, due to abundant water inflow and low water demand, instructions were given to increase water release for flood storage and power generation, while simultaneously ensuring ecological flow.

[0228] User water allocation instructions: Generate water allocation instructions to each irrigation district. For example, on the 4th day, allocate water to the "East Main Canal Irrigation District" from the planned 250,000 m³. 3 Reduced to 150,000 m³ 3 .

[0229] Water storage plan The model calculates that the reservoir's water storage is expected to increase to 56.5 million cubic meters by the end of the 7th. 3 This achieved the goal of storing more water while meeting demand.

[0230] 3. Instruction execution and system adaptation

[0231] Command Issuance and Execution (Step S5): The server transmits the leakage command (15.2 m) via the 4G network. 3 The actuator sends a water distribution command to the servo motor of the Qingyuan Reservoir gate ( / s) and then to each electric regulating valve. The actuator confirms the return of the control signal.

[0232] Adaptive learning:

[0233] The system recorded the actual water consumption of the "East Main Canal Irrigation Area" for the next three days as 625,000, 619,000, and 460,000 m³, respectively. 3 , and the predicted value There is a biased sequence = [-1.0, -0.9, -1.4] ten thousand m 3 .

[0234] The bias sequence is fitted using the ARIMA(1,1,1) model to predict future biases. Approximately -12,000 m 3 .

[0235] The system then makes rolling corrections to subsequent water demand forecasts based on this information. .

[0236] 4. Effects of the Example

[0237] This embodiment demonstrates that the system of the present invention is capable of: when facing complex climatic conditions of high temperatures and localized rainfall.

[0238] Accurate forecasting: Dynamically corrects water inflow and demand forecasts. The water inflow forecast for the 4th day is 13% higher than the baseline value, and the water demand forecast is 27% lower than the baseline value, making it more in line with reality.

[0239] Collaborative optimization: The system generates collaborative control instructions to "release more water during the flood season and use less water in irrigation areas," which not only ensures downstream ecological and flood control safety but also increases reservoir storage capacity, thus reserving resources for subsequent water use.

[0240] Self-evolution: Through the adaptive learning module, prediction bias is continuously reduced, improving the long-term robustness of the system.

[0241] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A water resource adaptation management method that coordinates supply and demand under the background of climate change, characterized in that, Includes the following steps: S1. Data Preparation Steps: Obtaining the Future Daily forecast temperature Predicted rainfall and predicted evapotranspiration Time-series data; obtain the current water level of the reservoir as monitored in real time by a physical sensor network. and current river flow ; S2. Climate-corrected inflow calculation steps: Input climate data into the pre-configured inflow correction model to calculate the climate-corrected predicted inflow. ; S3. Climate-corrected water demand calculation steps: Input climate data into the pre-configured water demand correction model to calculate the climate-corrected predicted water demand. ; S4. Control Instruction Generation Steps: ... and The input is fed into the control decision model, which solves to generate supply-side and demand-side control commands for directly driving and controlling actuators. The control decision model is a mathematical optimization problem as follows: in, ,and ; in Let be the water level at time t. For the target water level, This is the minimum allowable water level for the reservoir. This is the maximum allowable water level of the reservoir. For actuator opening, For the minimum opening of the actuator, For the maximum opening of the actuator, For the discharge flow function, As ecological base flow, This is the supply and demand balance coefficient. To optimize the target weight coefficient, It is the acceleration due to gravity. This is the discharge coefficient; And ecological base flow It is a dynamic value, which is based on the probability of extreme drought. Adjustments will be made: in This represents the probability of extreme drought. This serves as the baseline value for ecological baseflow. S5. Closed-loop control steps: The control command is sent to the corresponding control actuator network through the communication network to drive it to perform the corresponding physical operation in order to realize the closed-loop control of the water resource system. S6. Adaptive learning step: Record predicted water demand. Compared with actual water demand deviation sequence ; Using ARIMA The model fits and predicts the deviation sequence: And utilize the predicted bias The subsequent water demand forecasts will be revised as follows: ;in These are the autoregressive parameters of the ARIMA model. For moving average parameters, For the backoff operator, It is a white noise sequence.

2. The water resource adaptation management method for supply and demand coordination under the background of climate change as described in claim 1, characterized in that: The calculation rules for the water inflow correction model in step S2 are as follows: The correction factor is determined according to the following rules: in, , , ; in This is the temperature correction factor. This is a rainfall correction factor. Evapotranspiration correction factor This serves as a reference value for temperature. This serves as a reference value for rainfall. This serves as a reference value for evaporation. This is the temperature correction factor at time t. Let be the rainfall correction factor at time t. Let be the evapotranspiration correction factor at time t. This refers to the historical water volume.

3. The water resource adaptation management method for supply and demand coordination under the background of climate change as described in claim 1, characterized in that: The calculation rules for the water demand correction model in step S3 are as follows: For agricultural water use, , ; in This is the temperature-based water demand correction factor. This is the water demand correction factor for evapotranspiration. This is the temperature baseline value used in water demand calculations. As the baseline value for evaporation, This represents historical water demand.

4. A water resource adaptation management system that coordinates supply and demand under the background of climate change, characterized by: include: Physical sensor networks include water level sensors deployed in rivers, flow sensors deployed in water pipelines, and smart water meters at the user end. The control actuator network includes servo motors installed on reservoir gates, electric regulating valves in the water supply network, and solenoid valves in irrigation valve assemblies; The server is communicatively connected to the physical sensor network and the control actuator network, and its memory stores a program that, when executed by a processor, is used to implement the water resource adaptation management method for supply and demand coordination in the context of climate change as described in any one of claims 1 to 3.

5. The water resource adaptation management system for supply and demand coordination under the background of climate change as described in claim 4, characterized in that: The system also includes a fault diagnosis and fault tolerance control module, which is configured as follows: When at the preset time No confirmation signal was received from the actuator, or its actual opening degree. With instruction opening The deviation consistently exceeds the tolerance. achieve At this time, the actuator is determined to be faulty, and the fault-tolerant control strategy is activated.

6. The water resource adaptation management system for supply and demand coordination under the background of climate change as described in claim 4, characterized in that: The system also includes a visualization and human-computer interaction interface, configured to: provide a web graphical interface for displaying system status and early warning information, and provide a manual intervention interface allowing operators to adjust and optimize target weights. and And conduct a simulation run.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the water resource adaptation management method for supply and demand coordination in the context of climate change as described in any one of claims 1 to 3.

Images