A distributed intelligent control system for water supply and demand balance in arid regions

Abstract

This invention discloses a distributed intelligent control system for balancing water supply and demand in arid regions, specifically relating to the field of intelligent control. It includes a water supply and demand data acquisition module, a water supply and demand forecasting module, a distributed control decision-making module, an execution control and feedback module, and a system monitoring and emergency response module. This invention employs a distributed control architecture, with each regional edge controller operating independently and making collaborative decisions. This eliminates the single-point-of-failure risk of centralized architectures, significantly improving system reliability and emergency response speed. It quantitatively predicts future water demand using an ARIMA time series model and achieves multi-objective balance by combining it with an optimized allocation mathematical model. Compared to manual experience-based scheduling, this approach is more scientific and rational, effectively improving water resource utilization efficiency and reducing operating costs. An anomaly detection rule base and emergency plan base can automatically identify faults and quickly execute emergency strategies, greatly enhancing the system's ability to cope with emergencies and its control adaptability.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and more specifically, to a distributed intelligent control system for balancing water supply and demand in arid regions. Background Technology

[0002] Existing water resource allocation systems in arid regions typically employ a combination of centralized monitoring and manual dispatching. The system first collects real-time data using sensors such as flow meters and pressure gauges installed at various water sources and water usage areas. This data is then aggregated via wired or wireless communication networks to a SCADA system in the central control room. Dispatchers in the central control room manually assess the current supply and demand situation based on real-time data displayed on screens, historical experience, and water usage plans reported by each water usage area. When insufficient water supply is detected, dispatchers notify on-site operators via telephone or walkie-talkie to manually adjust valve openings or start / stop water pumps, or remotely issue control commands to on-site PLC controllers via the SCADA system. The PLC controllers then drive electric valves and frequency converters to adjust the water flow. Some systems incorporate simple expert rule bases; when monitored data triggers preset threshold conditions, the system automatically generates alarm messages to alert dispatchers or executes pre-programmed fixed control strategies, such as automatically activating a backup water pump when the pressure at a water source falls below a set value. The entire dispatching process relies heavily on centralized decision-making in the central control room, with equipment in each area passively receiving and executing commands from higher authorities, lacking autonomous coordination capabilities.

[0003] The aforementioned existing technologies have the following shortcomings: First, the centralized architecture makes the central control room a single point of failure in the system. Once the central server or communication network fails, the entire system will lose its scheduling capability and be unable to cope with emergencies. Second, manual scheduling relies on the experience and judgment of operators, resulting in slow response times and a high risk of decision-making errors. This is especially true in complex scheduling scenarios involving multiple water sources and water usage areas, where it is difficult for humans to calculate the optimal flow allocation scheme in a short time. Third, existing systems lack the ability to predict future water demand and can only respond passively based on the current situation. They cannot adjust water supply strategies in advance, leading to frequent water shortages or water waste. Fourth, the fixed expert rules lack flexibility and cannot adapt to the dynamic changes in water supply and demand in arid regions. When the actual operating conditions deviate from the applicable range of the preset rules, the system control effect drops significantly. Fifth, there is a lack of horizontal communication and coordination mechanisms between the execution equipment in different areas. When a water supply anomaly occurs in a certain area and support from adjacent areas is needed, instructions must be forwarded through the central control room, which increases response delay and reduces the system's emergency response capability. Sixth, the existing system is relatively crude in terms of priority management, usually only setting simple two or three levels of priority. It cannot achieve refined load allocation when water resources are severely scarce, resulting in some important water needs not being guaranteed. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a distributed intelligent control system for water resource supply and demand balance in arid areas, which solves the problems mentioned in the background art through the following scheme.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a distributed intelligent control system for water resource supply and demand balance in arid regions, comprising: a water resource supply and demand data acquisition module, a water resource supply and demand forecasting module, a distributed control decision module, an execution control and feedback module, and a system monitoring and emergency response module; The water resource supply and demand data acquisition module includes a water source flow monitoring unit, a water demand acquisition unit, and a data preprocessing unit. The water source flow monitoring unit is used to acquire instantaneous flow data of each water source in real time. The water demand acquisition unit is used to collect water demand information of each water use area. The data preprocessing unit is used to clean and standardize the collected instantaneous flow data of each water source and the water demand information of each water use area, and transmit the processed data to the water resource supply and demand prediction module for analysis. The water resource supply and demand forecasting module includes a short-term demand forecasting unit, a water supply capacity assessment unit, and a supply-demand gap calculation unit. The short-term demand forecasting unit forecasts water demand for the next two hours based on historical water consumption data and real-time water consumption data from water demand information for each water-using area, obtaining a water demand forecasting result. The water supply capacity assessment unit calculates the water supply capacity of each water source based on available water volume, instantaneous flow rate data, and pressure data. The supply-demand gap calculation unit compares the water demand forecasting result with the water supply capacity to obtain supply-demand gap data, and transmits the supply-demand gap data and water demand forecasting result to the distributed control decision module. The distributed control decision module includes a distributed coordination unit, an optimal allocation calculation unit, and a control command generation unit. The distributed coordination unit divides the system into multiple control zones based on the supply-demand gap data, water demand forecasts, and water supply capacity. The optimal allocation calculation unit performs optimal allocation calculations based on the water supply capacity of the water source within each control zone, the water demand forecasts for the water-using areas, and water use priority information to obtain the optimal water supply allocation result. The control command generation unit generates control commands based on the optimal water supply allocation result and transmits them to the execution control and feedback module. The execution control and feedback module includes an actuator drive unit, an execution status monitoring unit, and a deviation calculation and correction unit. The actuator drive unit receives control commands and drives valves and pumping stations to perform regulation. The execution status monitoring unit collects the operating status parameters of valves and pumping stations to obtain actual execution values. The deviation calculation and correction unit compares the actual execution values ​​with the target values ​​corresponding to the control commands to obtain execution deviations and corrects the control commands. The execution control and feedback module feeds back the actual execution values ​​and execution deviations to the water resource supply and demand data acquisition module to form a closed-loop control. The system monitoring and emergency response module includes an anomaly detection unit, an emergency strategy triggering unit, and an emergency control execution unit. The anomaly detection unit is used to identify system anomalies. The emergency strategy triggering unit is used to trigger emergency strategies according to the anomaly type. The emergency control execution unit is used to generate emergency control instructions and synchronize them to the distributed control decision module and the execution control and feedback module to adjust the control strategy and ensure the needs of critical water use areas.

[0006] The technical effects and advantages of this invention are as follows: 1. This invention adopts a distributed control architecture, decomposing the global scheduling task into multiple edge controllers for independent execution. Each control area has complete data acquisition, decision calculation, and execution control capabilities. When the edge controller or communication link of a certain area fails, other areas can still operate normally, avoiding the problem of a single point of failure causing the entire system to be paralyzed in a centralized architecture. This significantly improves the reliability and fault tolerance of the system. The inter-area consistency algorithm implemented through the distributed coordination unit enables each edge controller to make control decisions quickly locally without waiting for instructions from the central server, greatly shortening the response time from data acquisition to control execution. The edge controllers communicate directly through a ring network. When a certain area needs support from adjacent areas, they can directly negotiate the scheduling scheme, eliminating the intermediate link of forwarding through the center, further accelerating the emergency response speed, and enabling the system to respond to sudden water supply situations in a timely manner. 2. The water resource supply and demand forecasting module of this invention uses the ARIMA time series model combined with meteorological correction factors to quantitatively predict water demand for the next 2 hours. This changes the limitation of existing technologies that can only passively respond to the current state, giving the system a forward-looking scheduling capability. It can adjust the water supply strategy in advance to avoid supply and demand imbalance. The short-term demand forecasting unit and the water supply capacity assessment unit provide an accurate data foundation for subsequent optimized allocation, enabling the system to activate backup water sources or adjust pipeline pressure before the arrival of peak demand. The optimized allocation mathematical model established by the distributed control decision module comprehensively considers multiple factors such as water transmission cost, energy consumption, and user priority. The optimal flow allocation scheme obtained by solving the Lagrange multiplier method is more scientific and reasonable than manual experience scheduling. It not only ensures the water demand of high-priority users, but also minimizes system operating costs and energy consumption, and improves the utilization efficiency of water resources. The constraints in the optimization model ensure that the allocation scheme achieves multi-objective balance under the premise of meeting physical constraints, avoiding the problem of local optimization and global suboptimal that may be caused by simple rule scheduling. 3. This invention sets detailed priority classifications of 1-5 levels for each water use area in the water demand acquisition unit, and sets differentiated minimum satisfaction rates for different priorities in the optimization allocation calculation unit, achieving refined load allocation management. Compared with the crude two- or three-level priority division of existing technologies, this system can more rationally allocate limited water supply resources when water resources are scarce, ensuring that critical water needs are prioritized while also taking into account the basic needs of general users. The execution control and feedback module collects execution data such as valve opening, pump station speed, and pipeline flow in real time through the execution status monitoring unit, and dynamically corrects execution deviations by combining the incremental PID algorithm of the deviation calculation and correction unit, forming a complete closed-loop control loop, which significantly improves control accuracy and makes actual The water supply flow rate can accurately track the target value, and the closed-loop feedback mechanism enables the system to have adaptive adjustment capabilities. It can automatically compensate for the impact of external disturbances such as actuator wear and changes in pipeline resistance. Compared with the open-loop control or fixed rule control of existing technologies, this system shows stronger adaptability and robustness when facing changes in operating conditions. The abnormal detection rule library and emergency plan library established by the system monitoring and emergency handling module can automatically identify various abnormal situations such as water supply interruption, equipment failure, and pipeline leakage and trigger corresponding emergency strategies. The emergency control execution unit can quickly adjust the control strategy to implement emergency measures such as water source switching and load reduction. Compared with existing technologies that rely on manual judgment and manual operation, the emergency handling of this system is more timely and accurate, effectively improving the system's ability to cope with emergencies. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the overall modular structure of the present invention.

[0008] Figure 2 This is a schematic diagram of the overall unit structure of the present invention.

[0009] Figure 3 This is a schematic diagram of the water resource supply and demand forecasting module of the present invention.

[0010] Figure 4 This is a schematic diagram of the distributed control decision module structure of the present invention.

[0011] Figure 5 This is a schematic diagram of the execution control and feedback module structure of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] refer to Figures 1-5 The distributed intelligent control system for water supply and demand balance in arid regions, as shown, includes: a water supply and demand data acquisition module, a water supply and demand forecasting module, a distributed control decision module, an execution control and feedback module, and a system monitoring and emergency response module.

[0014] The water resource supply and demand data acquisition module includes a water source flow monitoring unit, a water demand acquisition unit, and a data preprocessing unit. The water source flow monitoring unit is used to acquire instantaneous flow data of each water source in real time. The water demand acquisition unit is used to collect water demand information of each water use area. The data preprocessing unit is used to clean and standardize the collected instantaneous flow data of each water source and the water demand information of each water use area, and transmit the processed data to the water resource supply and demand prediction module for analysis.

[0015] The water supply source flow monitoring unit is equipped with an electromagnetic flow meter and a pressure sensor at each water source point. The flow meter has a measurement range of 0-500m. 3 / h, accuracy ±0.5%, pressure sensor range 0-1.6MPa, sampling frequency once every 10 seconds, data is sent to the local edge controller via LoRa wireless module, the edge controller records water source number S_i (i=1, 2, ..., n), instantaneous flow rate Q_si(t), pressure P_si(t), available water volume V_si and timestamp T.

[0016] The water demand acquisition unit installs soil moisture sensors in agricultural irrigation areas, measuring depths of 0-40cm with an accuracy of ±2%. In industrial water areas, smart water meters (pulse output type, 0.01m³ resolution) are installed. In domestic water areas, ultrasonic water meters (nominal diameter DN15-DN300) are deployed. Each water area is equipped with a demand acquisition terminal, recording the water area number D_j (j=1, 2, ..., m), real-time water consumption Q_dj(t), cumulative water consumption V_dj, demand priority L_j (levels 1-5, with level 1 being the highest), and water type C_j (including agricultural, industrial, and domestic water use). Data is uploaded via an NB-IoT network.

[0017] In this embodiment, it should be specifically noted that, among the priority requirements: L_j=Level 1: Class I water use, including drinking water supply, medical and health water, and fire-fighting water; L_j=Level 2: Class II water use, including residential water use (excluding drinking water), water use for schools and public facilities, and drinking water for livestock; L_j=Level 3: Three categories of water use, including agricultural irrigation water, general industrial production water, and commercial service water; L_j=Level 4: Class IV water use, including municipal greening water, road washing water, and industrial cooling circulation replenishment water; L_j=5 level: Class 5 water use, including water for landscape and environment, water for car washing, and water for construction.

[0018] The data preprocessing unit receives data from all acquisition terminals, first performs outlier detection, and marks any instantaneous flow data Q(t) as an anomaly when |Q(t)-Q(t-1)|>3σ, then corrects it using linear interpolation, where σ is the standard deviation of historical data, and the correction formula is... Then, data synchronization processing is performed to unify data from different sampling frequencies into a 1-minute time granularity; next, unit standardization is performed to convert all instantaneous flow data into m³ / min units; finally, the water supply source dataset {S_i, Q_si, P_si, V_si, T} and the water demand dataset {D_j, Q_dj, V_dj, L_j, C_j, T} are encapsulated into standardized data packets according to timestamps and transmitted to the water resource supply and demand forecasting module via TCP / IP protocol.

[0019] The water resource supply and demand forecasting module includes a short-term demand forecasting unit, a water supply capacity assessment unit, and a supply-demand gap calculation unit. The short-term demand forecasting unit forecasts water demand for the next two hours based on historical water consumption data and real-time water consumption data from water demand information for each water-using area, obtaining a water demand forecasting result. The water supply capacity assessment unit calculates the water supply capacity of each water source based on available water volume, instantaneous flow rate data, and pressure data. The supply-demand gap calculation unit compares the water demand forecasting result with the water supply capacity to obtain supply-demand gap data, and transmits the supply-demand gap data and water demand forecasting result to the distributed control decision module.

[0020] The short-term demand forecasting unit uses time series analysis to predict water demand for the next two hours. First, historical water usage data is arranged in time series as {Q_dj(t-τ)}, where τ = 1, 2, ..., 120 represents the data from the past 120 minutes. For each water usage area D_j, an ARIMA(p, d, q) prediction model is established, where p is the autoregressive order (3), d is the difference order (1), and q is the moving average order (2). The prediction formula is:

[0021] in Water usage area At any moment Predicted water consumption Indicates the next 120 minutes, These are the autoregressive coefficients. The moving average coefficient is... The white noise error term is used. The model parameters are estimated using the least squares method and are updated every 24 hours.

[0022] This model employs time series analysis to predict future water consumption by utilizing the autocorrelation and moving average characteristics of historical water use data. The model structure is ARIMA(3,1,2), which includes three autoregressive terms, a first-order difference term, and two moving average terms. The autoregressive component reflects the inertial characteristics of water consumption, indicating a linear dependence between current water consumption and water consumption at the previous three time points. This dependence stems from the continuity of water use behavior; for example, agricultural irrigation, once started, often continues for a period of time, and water use in industrial production processes also exhibits continuity. The moving average component is used to capture the impact of random shocks, meaning that fluctuations in water consumption caused by certain sudden factors will gradually decay over time. This design effectively handles measurement errors and short-term disturbances. The prediction time window is set at 120 minutes, based on the actual needs of water supply scheduling in arid regions. A two-hour prediction span provides sufficient preparation time for scheduling decisions while ensuring that prediction accuracy does not significantly decrease due to an excessively long time span.

[0023] For agricultural water use with added meteorological correction factors, the correction formula is:

[0024] in This is the revised forecast of water consumption. For reference evapotranspiration (mm / day), This is the historical average evapotranspiration. The real-time temperature (°C) The historical average temperature For correction factor, and Historical actual water consumption data, along with corresponding evapotranspiration and temperature data, were collected and determined using multiple linear regression; finally, the total predicted demand was calculated. Output the demand forecast sequence for the next 120 time points.

[0025] The reference evapotranspiration correction term reflects the direct impact of atmospheric moisture demand on water consumption. In arid regions, evapotranspiration is the core factor driving agricultural irrigation demand. When real-time evapotranspiration is higher than the historical average, the rate of water loss from crops and soil accelerates, requiring increased irrigation to maintain water balance. The correction factor of 0.15 indicates that for every 1 unit increase in evapotranspiration, water demand increases by 15%. The temperature correction term reflects the combined effect of temperature on various water use behaviors. High temperatures not only directly increase evapotranspiration but also increase the demand for industrial cooling water and residential water. The relatively small correction factor of 0.08 reflects the strength of temperature as an indirect factor.

[0026] The water supply capacity assessment unit calculates the water supply capacity for the next 2 hours for each water source S_i based on the current available water volume V_si, instantaneous flow rate Q_si(t), and pressure P_si(t). First, it calculates the theoretical maximum water supply flow rate. Based on the principles of pipeline hydraulics, the relationship between water supply flow rate and pressure is as follows:

[0027] in The theoretical maximum water supply flow rate of water source S_i The flow coefficient of water source S_i, in units of , The pressure of water source S_i at time t; This formula is based on Bernoulli's equation and continuity equation in fluid mechanics. Flow rate is proportional to the square root of pressure because flow velocity is proportional to the square root of pressure difference, and flow rate equals flow velocity multiplied by the flow area. The flow coefficient K encompasses the combined effects of pipe geometry, friction coefficient, and local resistance coefficient. Differences in K values ​​among different water sources reflect differences in their hydraulic conditions. Deep well water sources typically have smaller K values ​​due to their large lifting height, while surface reservoirs have relatively larger K values ​​due to gravity flow. This formula is applicable only when the pipe network is in steady-state flow and pressure fluctuations are not drastic. In actual water supply systems in arid regions, these conditions are generally met. Then calculate the sustainable water supply duration using the following formula:

[0028] in For sustainable water supply duration (minutes). The current available water volume for water source S_i; when At what time (minutes), the actual water supply capacity of the water source is:

[0029] when At the specified time, the actual water supply capacity is:

[0030] Due to operational constraints of the water supply equipment, the actual water supply capacity needs to be multiplied by the equipment efficiency. The corrected water supply capacity was obtained. Calculate the total water supply capacity of all water sources. .

[0031] The equipment efficiency is calculated by actually measuring the input and output power of the equipment. ,in For effective output power, This refers to the input power.

[0032] The sustainable water supply duration reflects the water source's capacity to continue operating under its current output capacity. For water sources with limited volume, such as water towers and reservoirs, this duration directly determines the maximum time that water supply can be maintained without replenishment. Water sources with a sustainable duration of less than 120 minutes are considered to have insufficient reserves and require priority consideration for replenishment or restriction of output. This judgment criterion is consistent with the prediction time window, ensuring that the time scale of prediction and control is matched. Water sources with a sustainable duration of more than 120 minutes can be supplied continuously at maximum flow rate without worrying about depletion of reserves. This segmented processing method simplifies subsequent optimization calculations and improves real-time decision-making efficiency.

[0033] The supply-demand gap calculation unit calculates the supply-demand gap for each time point t+h in the next 120 time points:

[0034] when This indicates a water shortage. This indicates that the water supply is sufficient. This indicates a balance between supply and demand; This formula reflects the difference between water supply capacity and demand at a certain moment. Dividing the total water supply capacity by 120 is to convert the total amount over two hours into the average water supply capacity per minute, which is then directly compared with the predicted demand per minute. A positive gap indicates that supply is insufficient to meet demand, requiring a reduction in demand or an increase in supply. A negative gap indicates that supply exceeds demand, with a surplus of water available, which can be used to replenish water storage facilities or reduce energy consumption. A zero gap is the ideal balance, but it is difficult to achieve precisely in actual operation. The trend of the instantaneous gap is more important than the absolute value. If the gap continues to widen at multiple consecutive time points, it indicates that the supply-demand imbalance is worsening, requiring timely intervention. Then calculate the cumulative gap for the next 2 hours:

[0035] Supply and demand gap data , Forecast demand data and water supply capacity data After encapsulation, it is transmitted to the distributed control and decision-making module via a message queue.

[0036] The cumulative gap is calculated by summing all instantaneous gaps over the next 120 minutes to obtain the cumulative supply-demand difference over the entire forecast period. This indicator reflects the overall severity and duration of the supply-demand imbalance. When the cumulative gap is positive and large, it indicates a long-term water shortage, requiring the activation of emergency water sources or large-scale demand-side management. When the cumulative gap is negative, it indicates an ample water supply, which can be supplemented by reserves. This indicator provides a basis for formulating medium- and long-term scheduling strategies and, together with the instantaneous gap, constitutes decision-making information at both the short-term and medium-term levels.

[0037] The distributed control decision module includes a distributed coordination unit, an optimal allocation calculation unit, and a control command generation unit. The distributed coordination unit divides the system into multiple control areas based on the supply and demand gap data, water demand forecast results, and water supply capacity. The optimal allocation calculation unit performs optimal allocation calculations based on the water supply capacity of the water source within the control area, the water demand forecast results of the water use area, and water use priority information to obtain the optimal water supply allocation result. The control command generation unit generates control commands based on the optimal water supply allocation result and transmits them to the execution control and feedback module.

[0038] The distributed coordination unit divides the entire water supply system into K control zones R_k (k=1, 2, ..., K), with one edge controller configured in each zone. The zone division principle is that the pipeline connection distance between the water source and the water-using area within the same zone does not exceed 5km, and each zone contains a certain number of water-using areas. m / K The edge controllers are connected via an Ethernet ring network to form a distributed control network. The coordination process employs a consensus algorithm, with each edge controller maintaining a state variable x_k(t) representing the supply-demand balance state of the control area R_k. The update formula is:

[0039] in For the region At any moment The state of supply and demand equilibrium. In order to cooperate with the region Adjacent region sets, For the connection weights between regions, The value is 1 if the elements are directly adjacent, otherwise it is 0. The coupling coefficient is... This is the local adjustment coefficient. For the region The desired state of supply and demand equilibrium (target is 0).

[0040] The According to the consensus theory of multi-agent systems, the following must be satisfied:

[0041] in The largest eigenvalue of the Laplacian matrix of the system is taken as the eigenvalue of the Laplacian matrix of the regional connectivity graph for the regional topology of this system. This ensures a balance between system convergence speed and stability.

[0042] The local adjustment coefficient is designed using a Lyapunov function:

[0043] Find its reciprocal The stability condition can be obtained as follows:

[0044] in In this embodiment, the maximum degree of the node is... ,Pick It meets the requirement of quickly tracking the desired state.

[0045] The distributed coordination unit implements a multi-agent consensus algorithm. Each regional controller maintains a state variable representing the supply and demand balance of its region. This state variable is continuously updated through two feedback mechanisms: first, information coupling between adjacent regions, which gradually makes the states of each region converge; second, local feedback adjustment, which drives its own state to move closer to the balance target. The state update formula adopts a discrete-time first-order dynamic model. The first term is the maintenance term, representing the system inertia; the second term is the coupling term, which reflects the influence of adjacent regions on the local region. Its strength is determined by the coupling coefficient and the connection weight between regions; the third term is the local adjustment term, which applies a correction force according to the current deviation. The connection weight matrix describes the topological relationship between regions. Directly adjacent regions have a weight of 1, indicating that they can directly exchange information and water resources. Non-adjacent regions have a weight of 0, indicating that there is no direct connection. This sparse topology structure significantly reduces the communication burden.

[0046] The distributed coordination unit finally converges all x_k(t) to a globally consistent value through iterative calculation. Each edge controller obtains the water source set S_k={S_i|S_i∈R_k} and the water use area set D_k={D_j|D_j∈R_k} in its local area, and passes the local data {S_k, D_k, x_k} to the optimization allocation calculation unit.

[0047] The optimization allocation calculation unit establishes a water resource optimization allocation model within each control area R_k, defining the decision variable f_ij as the water supply flow rate (m³ / min) from water source S_i to water use area D_j. The optimization objective is to minimize water transmission energy consumption and maximize the satisfaction of high-priority users while meeting water demand; the objective function is:

[0048] in To supply water source to water use area Water supply flow rate decision variable Cost of water transport per unit flow rate (yuan / m³) 3 ), Pipeline distance (km) Energy consumption coefficient (yuan·min / m 6 ), This is the priority weight coefficient. Water usage area priority, Water usage area Demand satisfaction is defined as .

[0049] The constraints of the objective function include: 1. Water supply capacity constraints:

[0050] 2. Demand satisfaction constraints:

[0051] in The minimum satisfaction rate; 3. Pipeline flow constraints:

[0052] in For pipelines Rated flow rate; 4. Non-negativity constraint:

[0053] The Lagrange multiplier method is used to solve this optimization problem, and the Lagrange function is constructed as follows:

[0054] in For water supply capacity-constrained Lagrange multipliers, For the Lagrange multiplier that satisfies the demand constraint.

[0055] Take the partial derivative of f_ij with respect to f_ij and set it to zero:

[0056] The optimal solution is obtained:

[0057] Iteratively update the multipliers and Until convergence, the convergence condition is that the rate of change of the objective function in two consecutive iterations is less than 1%. Optimal traffic allocation scheme It is then passed to the control command generation unit.

[0058] The optimization allocation calculation unit solves a multi-objective constrained optimization problem. The inputs to this problem are the future demand provided by the prediction module, the water supply capacity provided by the evaluation module, and the balance objective given by the coordination unit. The output is the optimal water flow allocation scheme from each water source to each water-using area. The objective function contains two competing sub-objectives. The first sub-objective is to minimize the total operating cost, including water transmission cost and energy consumption cost. The water transmission cost is proportional to the pipeline length and flow rate, reflecting the economic cost of long-distance water transmission. The energy consumption cost is proportional to the square of the flow rate, conforming to the fluid dynamics law that pipeline resistance is proportional to the square of the flow velocity. The second sub-objective is to maximize the satisfaction of the needs of high-priority users through optimization. The priority weight coefficient is converted into an equivalent cost reduction; the constraints ensure the physical feasibility of the optimization solution. The water supply capacity constraint prevents any water source from operating under overload conditions, the demand satisfaction constraint ensures that each user receives at least a certain proportion of their demand, avoiding the extreme situation of complete water outage, the pipeline flow constraint reflects the physical capacity limit of the pipe network, and the non-negativity constraint is a natural attribute of the flow variable; the solution method uses the Lagrange multiplier method to transform the constrained optimization problem into an unconstrained problem. After introducing the Lagrange multipliers, the optimal solution of the original problem satisfies the Kuhn-Tucker condition. Taking the partial derivative of the objective function and setting it to zero yields a set of simultaneous equations. Since the objective function contains quadratic terms, this set of equations has a unique analytical solution.

[0059] The control command generation unit generates the control command according to the optimal flow allocation scheme. Generate opening commands for regulating valves on each pipeline, and generate start / stop and speed commands for each pump station; for the pipeline connecting water source S_i and water usage area D_j, the valve opening... With traffic The relationship is determined by the flow equation:

[0060] in The valve flow coefficient, The pressure difference across the valve; the current pressure difference. The valve opening command can be calculated from the pressure data in the water resource supply and demand data acquisition module.

[0061] in The valve flow coefficient, The pressure difference across the valve. This is the valve opening command (0-100%). The valve opening is limited to the range of 0-100%. If the calculated value exceeds the range, the boundary value will be used.

[0062] For pumping stations, based on the total flow rate required to be provided. The pump station speed is determined, and the relationship between the pump station flow rate and the speed is as follows:

[0063] in Let m be the pump station flow constant (m³ / (min·rpm)). The rotational speed is rpm.

[0064] The speed command is:

[0065] in The flow constant of the pump station is m² / (min·rpm). The speed command (rpm) is limited to the range of 600-1800 rpm. When, a pump station shutdown command is generated; when At that time, a pump station start command is generated.

[0066] The control command generation unit will generate a control command set. Encapsulated according to the ModbusTCP protocol, with the instruction number ID_cmd, target device address Addr, execution timestamp T_exec, and checksum CRC appended, it is transmitted to the execution control and feedback module via industrial Ethernet.

[0067] The physical basis of the valve flow equation in the control command generation unit is Bernoulli's equation and the continuity equation. When fluid flows through the valve, local resistance loss occurs due to the sudden contraction and expansion of the channel cross-section. This loss is proportional to the square of the flow velocity, which in turn is proportional to the flow rate. Therefore, the flow rate is proportional to the square root of the pressure difference. The flow coefficient comprehensively reflects the influence of various factors such as valve channel shape, roughness, and Reynolds number. The calculation of the valve opening command is essentially an inverse problem. Given the desired system response, i.e., the flow rate, the input, i.e., the valve opening, is deduced. Due to the monotonicity of the flow equation, this inverse problem has a unique solution. The existence of the pressure difference term means that the flow rate at the same opening will change with the water supply pressure. This requires the controller to be able to... It can acquire pressure sensor measurements in real time; the linear relationship between pump station flow and speed simplifies control design, and the calculation of speed command becomes a simple division operation without the need for complex iterative solutions. This linear characteristic also makes variable frequency speed regulation an effective means of energy saving for water pumps. Reducing flow by lowering speed is more energy-efficient than throttling by closing valves, because throttling converts energy into heat loss; the upper and lower limits of speed command protect equipment safety. When running at low speed, pump efficiency decreases and cavitation may occur. When running at high speed, mechanical wear increases and may exceed the rated power of the motor. Limiting the speed within a reasonable range extends equipment life. When the optimized flow demand exceeds the capacity of a single pump, multiple pumps need to be started and run in parallel.

[0068] The execution control and feedback module includes an actuator drive unit, an execution status monitoring unit, and a deviation calculation and correction unit. The actuator drive unit receives control commands and drives valves and pumping stations to perform regulation. The execution status monitoring unit collects the operating status parameters of valves and pumping stations to obtain actual execution values. The deviation calculation and correction unit compares the actual execution values ​​with the target values ​​corresponding to the control commands to obtain execution deviations and corrects the control commands. The execution control and feedback module feeds back the actual execution values ​​and execution deviations to the water resource supply and demand data acquisition module to form a closed-loop control.

[0069] The actuator drive unit is configured with a field control cabinet for each control area. The control cabinet is equipped with a PLC controller and a multi-channel I / O module. The PLC controller receives control commands via Modbus TCP protocol. The PLC performs instruction parsing and validity verification. For valve control, the PLC outputs a 4-20mA current signal to the electric actuator. The relationship between the current and the valve opening degree is as follows:

[0070] in The output current (mA) is used to adjust the valve opening based on the current signal, and the adjustment time does not exceed 30 seconds.

[0071] The baseline value of 4mA represents the fully closed valve state. This is the minimum current value specified in the industry standard to ensure signal integrity. The coefficient 0.16 comes from the ratio of the signal range to the opening range, that is, 16mA divided by 100% equals 0.16mA per percentage point. When the valve opening is 50%, the output current is 12mA, which is exactly at the midpoint of the signal range. This linear relationship ensures the uniformity of control accuracy.

[0072] For pump station control, the PLC outputs a 0-10V voltage signal to the frequency converter. The relationship between voltage and speed is as follows:

[0073] Among them U i The output voltage (V) ranges from 0 to 10V, n min =600rpm, n max =1800rpm; The frequency converter adjusts the motor speed according to the voltage signal, and the response time is no more than 5 seconds. When a stop command is received, the PLC outputs a digital quantity 0 (low level) to the contactor coil to disconnect it; when a start command is received, it outputs a digital quantity 1 (high level) to close the contactor. All executed actions are recorded in the local log, including the instruction number ID_cmd, execution time T_actual, target value and device response code.

[0074] This formula uses normalization to map the speed range to the voltage range. The numerator represents the increment of the actual speed relative to the minimum speed, and the denominator represents the total range of speed variation. The ratio of the two is the normalized speed. Multiplying it by 10V gives the corresponding control voltage. This design allows different pump stations to be adapted simply by modifying the speed range parameters, and it has good versatility.

[0075] The core task of the actuator drive unit is to convert the control commands output by the decision module into standard signals that the actuator can recognize. Modern industrial control systems generally use 4-20mA current signals to control valves and 0-10V voltage signals to control frequency converters. The drive unit needs to establish a linear mapping relationship between the control quantity and the signal value to ensure accurate transmission of commands. For electric regulating valves, the opening range is 0-100%, and the corresponding control current is 4-20mA. When the opening is 0%, the output is 4mA, and when the opening is 100%, the output is 20mA, showing a linear relationship in between. The advantage of this design is that 4mA can be used as a zero-point signal to detect whether the circuit is broken, while the true zero signal will be lower than 4mA. For water pump frequency converters, the speed range is usually 600-1800rpm, corresponding to a control voltage of 0-10V. The speed and voltage have a linear relationship, and this mapping method conforms to industrial standards and facilitates equipment integration.

[0076] The execution status monitoring unit is equipped with an opening feedback sensor, specifically an angle encoder, installed on each valve, with an accuracy of ±0.5°, to measure the actual opening degree in real time. A speed sensor, specifically a Hall effect sensor, is installed at each pumping station with an accuracy of ±1 rpm to measure the actual speed in real time. Install flow meters at pipe joints to measure the actual flow rate. The monitoring data sampling period is 1 second, and it is uploaded to the PLC controller via the fieldbus. The PLC controller calculates the execution error.

[0077]

[0078]

[0079] when or or When an abnormal alarm is triggered, the abnormal information will be sent out simultaneously. Send to the upper-level control system.

[0080] For equipment exhibiting execution deviations, the deviation calculation and correction unit employs an incremental PID algorithm for real-time correction. Taking valve opening control as an example, the incremental calculation formula for the PID control quantity is as follows:

[0081] Where K p =1.2 is the proportionality coefficient, K i =0.3 is the integral coefficient, K d =0.05 is the differential coefficient, and t is the current sampling time.

[0082] Proportional Term Calculate the change in error relative to the previous error, multiply it by a proportional coefficient to obtain the corresponding control increment. The physical meaning of this term is to adjust the control intensity according to the rate of error change: increase the control intensity when the error increases, and decrease the control intensity when the error decreases. The proportional coefficient... This value, obtained through system identification and simulation optimization, ensures both rapid response and avoids excessive oscillation; integral term The current error is directly accumulated and multiplied by the integral coefficient to obtain the control increment. The physical meaning of this term is to eliminate the steady-state error of the system. As long as the error exists, the control quantity is continuously adjusted until the error is zero. The integral coefficient... Relatively small, to avoid excessive integral action leading to overshoot and oscillation; differential term The second-order difference of the error is calculated, approximately representing the rate of change of the error. The physical meaning of this term is to predict the trend of error change: when the error increases rapidly, the control input is increased in advance; when the error decreases rapidly, the control input is decreased in advance to avoid overshoot. (The differential coefficient is mentioned here.) The coefficient is relatively small because the differential action is sensitive to noise, and an excessively large coefficient would amplify the measurement noise.

[0083] The corrected opening command is:

[0084] The formula adds the correction opening degree of the previous moment to the increment calculated in this moment to obtain the new control command. The physical meaning of this incremental update method is to make fine adjustments on the basis of the original control, so as to avoid the impact of sudden changes in control quantity on the actuator. At the same time, if the controller fails or communication is interrupted, the actuator will remain in the current position and will not suddenly open or close completely, thus ensuring the safety of the system.

[0085] Will After limiting the output to the 0-100% range, the output is sent to the actuator drive unit for re-execution. For pump station speed control, a similar PID correction method is used, with the PID parameters being... The PID parameters are adaptively adjusted every 10 seconds, and the parameter values ​​are dynamically corrected based on the system response characteristics, taking all execution status data into account. and deviation data Encapsulated using a unified timestamp T_fb, the data is fed back to the data preprocessing unit of the water resource supply and demand data acquisition module via the TCP / IP protocol, serving as input data for the next cycle of control and forming a complete closed-loop control circuit.

[0086] The system monitoring and emergency response module includes an anomaly detection unit, an emergency strategy triggering unit, and an emergency control execution unit. The anomaly detection unit is used to identify system anomalies. The emergency strategy triggering unit is used to trigger emergency strategies according to the anomaly type. The emergency control execution unit is used to generate emergency control instructions and synchronize them to the distributed control decision module and the execution control and feedback module to adjust the control strategy and ensure the needs of critical water use areas.

[0087] The anomaly detection unit deploys anomaly detection algorithms on the central monitoring server, receiving all monitoring data in real time from the water resource supply and demand data acquisition module and the execution control and feedback module; it establishes an anomaly detection rule base, including the following anomaly types: 1) Abnormal water supply interruption: When the flow rate Q_si(t) of water source S_i remains below 1 for 5 consecutive minutes... The water supply was deemed interrupted at that time. 2) Equipment malfunction: When the actuator fails to respond to control commands three times consecutively or If the failure continues for more than 2 minutes, it is considered a device malfunction. 3) Abnormal pipeline leakage: When the difference between the inlet flow rate and the outlet flow rate of a pipeline exceeds 20% for 10 consecutive minutes, it is determined to be a leak.

[0088] For each type of anomaly, define an anomaly severity level. The formula for assessing the severity of anomalies is:

[0089] in These are the weighting coefficients. Indicates rounding up. Level 3 is the most severe.

[0090] When an anomaly is detected, an anomaly event record is generated {anomaly type (Anom_type), anomaly location (Loc), severity}. The detection time is T_anom}, which triggers the emergency strategy triggering unit.

[0091] The emergency strategy triggering unit maintains an emergency plan library, pre-defining emergency strategies for different anomaly types and severity levels. The strategy library adopts a decision tree structure, with the root node representing the anomaly type, the second level representing the severity, and the leaf nodes representing specific emergency measures. The emergency strategy selection logic is as follows: first, match the strategy branch according to the anomaly type Anom_type, and then determine the response level according to SEV_level. Emergency measures include: Measure A - Water Source Switching: Transfer the water supply task of the affected area to the backup water source; Measure B - Load Reduction: Suspend water supply to low-priority users in descending order of priority L_j, specifically reducing water consumption for users with priorities L_j=4 and L_j=5; Measure C - Emergency Water Replenishment: Start the emergency water tank or water transfer vehicle to supply water; Measure D - Isolate the Faulty Section: Close the valves at both ends of the faulty pipeline and activate the bypass pipeline. The strategy triggering condition adopts the IF-THEN rule, for example: IF Anom_type="Water Supply Interruption" AND SEV_level=3 THEN Execute Measure A + Measure C; The selected emergency measure set {Measure ID, Target Equipment, Action Parameters} is passed to the emergency control execution unit.

[0092] After receiving emergency measures, the emergency control execution unit generates emergency control instructions and overrides the normal control process. For measure A, which involves switching the water supply source, it recalculates the water source set R_k in the affected control area, removes the faulty water source from S_k, adds the backup water source S_backup to S_k, and calls the optimization allocation calculation unit of the distributed control decision module to resolve the optimization problem and obtain a new flow allocation scheme. New control commands are generated and sent to the execution control and feedback module. For measure B, load reduction, the water use areas with priorities L_j=4 and L_j=5 within the affected area are identified, and the total flow to be reduced is calculated.

[0093] in This represents the remaining available water supply capacity.

[0094] The reduction amount is allocated according to the proportion of demand from each low-priority user. The reduction traffic for D_j is:

[0095] in The low-priority user set includes: and Generate valve closing command. Send water to the corresponding water-using area. For emergency water replenishment under measure C, activate the outlet valve and booster pump of the emergency water tank and set the replenishment flow rate. Equal to the current supply and demand gap Duration of hydration ,in For emergency water tank capacity; for measure D, isolate the faulty section and send a signal to the valves at both ends of the faulty pipeline. The command to close is sent simultaneously to the bypass pipeline valve. The system initiates the activation command; all emergency control commands are given a priority flag (Priority=HIGH) and sent to the execution control and feedback module via a dedicated emergency channel (independent VLAN) to ensure priority execution; the emergency handling results {emergency measure ID, execution status, scope of impact, estimated recovery time} are synchronized to the distributed control decision module and the water resource supply and demand data acquisition module via the message bus. The distributed control decision module adjusts and optimizes the objectives and constraints according to the emergency status, and the water resource supply and demand data acquisition module marks the data during the emergency period during data preprocessing to avoid affecting the normal prediction model.

[0096] In this embodiment, the water resource supply and demand data acquisition module continuously collects flow and water demand data from various water sources, and transmits the preprocessed data to the water resource supply and demand forecasting module. The water resource supply and demand forecasting module performs short-term demand forecasting and water supply capacity assessment based on the collected data, calculates the supply and demand gap, and transmits the data to the distributed control decision module. The distributed control decision module decomposes tasks into control areas through distributed coordination, performs optimization allocation calculations within each area, generates control commands, and transmits them to the execution control and feedback module. The execution control and feedback module drives valves and pump stations to execute control commands, monitors the execution status, corrects deviations using a PID algorithm, and feeds back the execution data to the water resource supply and demand data acquisition module to form a closed loop. The system monitoring and emergency handling module continuously monitors the operating status, triggers emergency strategies when an anomaly is detected, executes emergency control, and synchronizes information to other modules to ensure stable system operation under abnormal conditions. All modules exchange data through standardized data interfaces and industrial communication protocols, forming a complete distributed intelligent control system to achieve dynamic balance control of water resource supply and demand in arid regions.

[0097] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A distributed intelligent control system for balancing water supply and demand in arid regions, characterized in that, include: The system includes a water resources supply and demand data acquisition module, a water resources supply and demand forecasting module, a distributed control decision-making module, an execution control and feedback module, and a system monitoring and emergency response module. The water resources supply and demand data acquisition module includes a water source flow monitoring unit, a water demand acquisition unit, and a data preprocessing unit. The water supply source flow monitoring unit is used to acquire the available water volume, instantaneous flow data and pressure data of each water supply source in real time; the water demand acquisition unit is used to collect water demand information of each water use area. The data preprocessing unit is used to clean and standardize the instantaneous flow data of each water supply source and the water demand information of each water use area, and transmit the processed data to the water resource supply and demand prediction module for analysis. The water resource supply and demand forecasting module includes a short-term demand forecasting unit, a water supply capacity assessment unit, and a supply-demand gap calculation unit. The short-term demand forecasting unit is used to forecast the water demand for the next two hours based on historical water consumption data and real-time water consumption data from water demand information for each water-using area, and obtain the water demand forecasting result. The water supply capacity assessment unit is used to calculate the water supply capacity of each water source based on the available water volume, instantaneous flow rate data, and pressure data. The supply-demand gap calculation unit compares the water demand forecasting result with the water supply capacity to obtain the supply-demand gap data, and transmits the supply-demand gap data and the water demand forecasting result to the distributed control decision module. The distributed control decision module includes a distributed coordination unit, an optimal allocation calculation unit, and a control command generation unit. The distributed coordination unit is used to divide the system into multiple control areas based on the supply and demand gap data, water demand forecast results, and water supply capacity. The optimal allocation calculation unit is used to perform optimal allocation calculations on a per-control-area basis, based on the water supply capacity of the water source within the control area, the water demand forecast results of the water-using area, and water use priority information, to obtain the optimal water supply allocation result. The control command generation unit generates control commands based on the optimal water supply allocation result and transmits them to the execution control and feedback module. The execution control and feedback module includes an actuator drive unit, an execution status monitoring unit, and a deviation calculation and correction unit. The actuator drive unit receives control commands and drives valves and pumping stations to perform regulation. The execution status monitoring unit collects the operating status parameters of valves and pumping stations to obtain actual execution values. The deviation calculation and correction unit compares the actual execution values ​​with the target values ​​corresponding to the control commands to obtain execution deviations and corrects the control commands. The execution control and feedback module feeds back the actual execution values ​​and execution deviations to the water resource supply and demand data acquisition module to form a closed-loop control. The system monitoring and emergency response module includes an anomaly detection unit, an emergency strategy triggering unit, and an emergency control execution unit. The anomaly detection unit is used to identify system anomalies. The emergency strategy triggering unit is used to trigger an emergency strategy based on the anomaly type of the system anomaly. The emergency control execution unit is used to generate emergency control instructions based on the emergency strategy and synchronize them to the distributed control decision module and the execution control and feedback module to adjust the control strategy and ensure the needs of critical water use areas.

2. The distributed intelligent control system for water resource supply and demand balance in arid regions according to claim 1, characterized in that: The water supply source flow monitoring unit includes flow sensors and pressure sensors installed at each water supply source to collect instantaneous flow data and operating pressure data of the corresponding water supply source. The instantaneous flow data and operating pressure data are then associated with the water supply source number and the collection time and sent to the data preprocessing unit.

3. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The water demand collection units are respectively deployed in agricultural water use areas, industrial water use areas and domestic water use areas, and are used to collect real-time water consumption, cumulative water consumption, water use type and water use priority information of the corresponding water use areas, and upload the water demand data to the data preprocessing unit through wireless communication.

4. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The data preprocessing unit sequentially performs abnormal data identification processing, time scale unification processing, and data unit standardization processing on the received instantaneous flow data, operating pressure data, and water demand data. After encapsulating the processed data according to a unified data format, it sends the data to the water resource supply and demand forecasting module.

5. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The short-term demand forecasting unit constructs a time series forecasting model based on historical water use data of each water use area, forecasts the water use demand of each water use area hourly within a preset time window, and outputs the forecast demand data sequence for the corresponding time period.

6. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The water supply capacity assessment unit assesses the water supply capacity of each water source within a preset time window based on the available water volume data, instantaneous flow data, and operating pressure data of each water source obtained by the water source flow monitoring unit in the water resource supply and demand data acquisition module.

7. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The optimization allocation calculation unit takes the control area as a unit, models the water supply relationship between the water supply source and the water use area within the control area, and calculates the water supply flow allocation scheme from each water supply source to each water use area under the conditions of satisfying the water supply capacity constraint and the water demand constraint.

8. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: The actuator drive unit includes a current output interface for valve control and a voltage output interface for pump station control. The actuator drive unit outputs a matching control signal to the corresponding valve actuator or pump station frequency converter according to the received control command.

9. A distributed intelligent control system for balancing water supply and demand in arid regions according to claim 1, characterized in that: When the system monitoring and emergency response module detects abnormalities in the water supply source, execution equipment, or pipeline operation, it generates corresponding emergency control commands based on preset rules for determining the type and severity of the abnormality, and sends the emergency control commands to the execution control and feedback module for execution.

Images