A resource scheduling method and system based on water data

Abstract

This invention discloses an intelligent resource scheduling method and system based on water data. It achieves real-time acquisition and preprocessing of multimodal water data through edge computing nodes, employs an adaptive sampling strategy and differential privacy protection to ensure data quality and security, and implements tamper-proof storage of water quality data. A demand prediction model based on a spatiotemporal graph neural network is used, employing spatiotemporal convolution and gated attention mechanisms to achieve accurate regional water use prediction. An improved multi-objective particle swarm optimization algorithm is used, aiming at water supply stability, energy consumption minimization, and leakage suppression, combined with dynamic inertial weights and fuzzy membership functions to select the optimal scheduling scheme. Virtual simulation is performed through a digital twin system to visualize pressure anomalies and recommend intervention strategies. Pump operation is optimized in conjunction with time-of-use electricity pricing strategies, and chemical dosing is dynamically adjusted based on water quality predictions. This method can improve water supply stability, reduce energy consumption and leakage rate, and significantly improve the reliability and economy of the water system.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a resource scheduling method and system based on water data. Background Technology

[0002] With the acceleration of urbanization, water systems face severe challenges such as supply-demand imbalance, aging pipe networks, excessive energy consumption, and delayed response to emergencies. Traditional dispatching methods suffer from the following technical bottlenecks:

[0003] Existing systems cannot effectively integrate cross-domain data from pressure sensors (fixed sampling rate), smart water meters (privacy sensitive), and water quality monitoring (anti-tampering requirements). For example, in 2024, a provincial capital city misjudged water demand during a rainstorm due to a lack of linkage with meteorological data, resulting in a 12-hour water outage in three urban areas.

[0004] Conventional SCADA systems use centralized data processing. In a water plant pipe burst incident, the delay from data acquisition to the central server response was as high as 8 minutes, missing the best time for handling the situation.

[0005] There have been cases where pH value records were altered to evade penalties when water quality monitoring data was stored in traditional databases.

[0006] The technical solution of this invention addresses the aforementioned pain points by leveraging edge computing and blockchain to ensure data quality, ST-GNN to achieve spatiotemporal joint prediction, an improved MOPSO algorithm to balance multi-objective optimization, and a digital twin system to achieve second-level emergency response, forming a complete technological closed loop. Application in a pilot city shows that this system can improve water supply stability by 40%, reduce energy consumption by 18%, and decrease leakage rate by 35%, demonstrating significant technological advancements. Summary of the Invention

[0007] To achieve the above objectives, this application provides the following technical solution:

[0008] According to a first aspect of the present invention, the present invention claims protection for a resource scheduling method based on water data, comprising:

[0009] S1, Collect multimodal water affairs data, and perform data preprocessing on the multimodal water affairs data through edge computing nodes;

[0010] S2, input the preprocessed multimodal water data into the water demand prediction model, and output the water demand prediction results for each zone;

[0011] S3, Set optimization objectives, and use multi-objective particle swarm optimization algorithm to correct the water demand prediction results of the partition to obtain a water resource scheduling scheme;

[0012] S4, The water resource scheduling scheme is virtually simulated, and the water resource scheduling scheme is dynamically adjusted according to the simulation results;

[0013] S5. Monitor the implementation effect of the adjusted water resource scheduling plan in real time and feed it back to the water demand prediction model for online learning.

[0014] Furthermore, S1 also includes:

[0015] Real-time acquisition of water supply network pressure data, user-end smart water meter water consumption data, water source quality monitoring data, meteorological environmental data, and industrial user water use plan data;

[0016] Data preprocessing is performed through edge computing nodes, including noise filtering, spatiotemporal alignment, and missing value imputation.

[0017] S2 further includes: inputting the preprocessed multimodal water data into a water demand prediction model based on the spatiotemporal graph neural network ST-GNN model;

[0018] By combining the attention mechanism to dynamically adjust the feature weights of different regions and time periods, the regional water demand forecast results for the next 24 hours are output.

[0019] S3 further includes:

[0020] With water supply stability, energy consumption minimization, and pipeline leakage suppression as optimization objectives, an improved multi-objective particle swarm optimization algorithm (MOPSO) is used to correct the water demand prediction results of the zoning and generate a water resource scheduling scheme. The MOPSO algorithm introduces physical constraints of water equipment and a dynamic reward and punishment mechanism.

[0021] S4 further includes:

[0022] The water resource scheduling plan is synchronized to the water affairs digital twin system for virtual simulation. Based on the simulation results, the water resource scheduling plan is dynamically adjusted and then sent to the execution terminal.

[0023] Furthermore, S1 also includes:

[0024] The edge computing node adopts an adaptive sampling strategy, and the sampling frequency of the pressure data is positively correlated with the pressure fluctuation rate of the pipeline network.

[0025] Differential privacy protection is implemented for user water consumption data by adding noise that conforms to a Laplace distribution;

[0026] Blockchain-based evidence storage ensures the immutability of water quality data from water sources. The evidence includes the collection time, location, and sensor ID.

[0027] Furthermore, S2 also includes:

[0028] The spatiotemporal graph neural network (ST-GNN) model includes:

[0029] The spatiotemporal convolution module is used to extract pipeline network topology features;

[0030] Gated attention module dynamically weights the differences in water usage patterns between industrial and residential users;

[0031] The online learning module updates model parameters in real time based on scheduling feedback data;

[0032] When the pipeline leakage rate is detected to exceed the threshold, a tiered emergency dispatch is triggered to prioritize water supply to critical nodes.

[0033] Automatically generate valve shut-off schemes to isolate leaking areas, and combine hydraulic models to calculate the optimal water replenishment path.

[0034] Furthermore, S3 also includes:

[0035] The improved multi-objective particle swarm optimization algorithm (MOPSO) includes:

[0036] Dynamic inertia weight: The weight value is adaptively adjusted according to the variance of the water supply pressure;

[0037] Constraint handling mechanism: The limit on the number of pump station start-stop cycles is transformed into a penalty function;

[0038] Pareto solution selection: Selecting the optimal scheduling scheme based on fuzzy membership functions.

[0039] Furthermore, S4 also includes:

[0040] The robustness of the water resource scheduling scheme under extreme weather conditions was simulated in the water affairs digital twin system.

[0041] Visualize areas of abnormal pipeline pressure through an AR interface and recommend manual intervention strategies;

[0042] The water resource scheduling scheme introduces a time-of-use electricity pricing strategy, prioritizing the activation of high-energy-consuming pump groups during off-peak electricity hours;

[0043] The dosage of water treatment chemicals is dynamically adjusted based on water quality prediction results.

[0044] According to a second aspect of the present invention, the present invention claims protection for a resource scheduling system based on water data, comprising:

[0045] One or more processors;

[0046] A memory that stores one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the resource scheduling method based on water data.

[0047] This invention discloses an intelligent resource scheduling method and system based on water data. It achieves real-time acquisition and preprocessing of multimodal water data through edge computing nodes, employs an adaptive sampling strategy and differential privacy protection to ensure data quality and security, and implements tamper-proof storage of water quality data. A demand prediction model based on a spatiotemporal graph neural network is used, employing spatiotemporal convolution and gated attention mechanisms to achieve accurate regional water use prediction. An improved multi-objective particle swarm optimization algorithm is used, aiming at water supply stability, energy consumption minimization, and leakage suppression, combined with dynamic inertial weights and fuzzy membership functions to select the optimal scheduling scheme. Virtual simulation is performed through a digital twin system to visualize pressure anomalies and recommend intervention strategies. Pump operation is optimized in conjunction with time-of-use electricity pricing strategies, and chemical dosing is dynamically adjusted based on water quality predictions. This method can improve water supply stability, reduce energy consumption and leakage rate, and significantly improve the reliability and economy of the water system. Attached Figure Description

[0048] Figure 1 A flowchart illustrating a resource scheduling method based on water data claimed in an embodiment of this application;

[0049] Figure 2 This is a structural diagram of a resource scheduling method based on water data, which is claimed in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0051] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative positional relationships and movements between components in a specific orientation (as shown in the figures). If the specific orientation changes, the directional indications also change accordingly. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0052] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0053] According to the first embodiment of the present invention, referring to Figure 1 This invention claims protection for a resource scheduling method based on water data, comprising:

[0054] S1, Collect multimodal water affairs data, and perform data preprocessing on the multimodal water affairs data through edge computing nodes;

[0055] S2, input the preprocessed multimodal water data into the water demand prediction model, and output the water demand prediction results for each zone;

[0056] S3, Set optimization objectives, and use multi-objective particle swarm optimization algorithm to correct the water demand prediction results of the partition to obtain a water resource scheduling scheme;

[0057] S4, The water resource scheduling scheme is virtually simulated, and the water resource scheduling scheme is dynamically adjusted according to the simulation results;

[0058] S5. Monitor the implementation effect of the adjusted water resource scheduling plan in real time and feed it back to the water demand prediction model for online learning.

[0059] Furthermore, S1 also includes:

[0060] Real-time acquisition of water supply network pressure data, user-end smart water meter water consumption data, water source quality monitoring data, meteorological environmental data, and industrial user water use plan data;

[0061] Data preprocessing is performed through edge computing nodes, including noise filtering, spatiotemporal alignment, and missing value imputation.

[0062] S2 further includes: inputting the preprocessed multimodal water data into a water demand prediction model based on the spatiotemporal graph neural network ST-GNN model;

[0063] By combining the attention mechanism to dynamically adjust the feature weights of different regions and time periods, the regional water demand forecast results for the next 24 hours are output.

[0064] S3 further includes:

[0065] With water supply stability, energy consumption minimization, and pipeline leakage suppression as optimization objectives, an improved multi-objective particle swarm optimization algorithm (MOPSO) is used to correct the water demand prediction results of the zoning and generate a water resource scheduling scheme. The MOPSO algorithm introduces physical constraints of water equipment and a dynamic reward and punishment mechanism.

[0066] S4 further includes:

[0067] The water resource scheduling plan is synchronized to the water affairs digital twin system for virtual simulation. Based on the simulation results, the water resource scheduling plan is dynamically adjusted and then sent to the execution terminal.

[0068] Furthermore, S1 also includes:

[0069] The edge computing node adopts an adaptive sampling strategy, and the sampling frequency of the pressure data is positively correlated with the pressure fluctuation rate of the pipeline network.

[0070] Differential privacy protection is implemented for user water consumption data by adding noise that conforms to a Laplace distribution;

[0071] Blockchain-based evidence storage ensures the immutability of water quality data from water sources. The evidence includes the collection time, location, and sensor ID.

[0072] In this embodiment, the hardware deployment for monitoring water supply network pressure involves deploying 20 wireless pressure sensors (range 0-1.6MPa, accuracy ±0.3%FS) at key nodes of the main pipeline network, using the NB-IoT transmission protocol.

[0073] The sensor has a built-in triaxial vibration detection module that can distinguish between real pressure fluctuations and mechanical vibration interference.

[0074] Normal pressure range: 0.25-0.45 MPa; sampling reference frequency: 5 minutes / time.

[0075] When the volatility of two consecutive data points exceeds 8%, the system automatically switches to high-frequency sampling at 1 second per sampling, continuing until the volatility drops to less than 5%, at which point it returns to the base frequency.

[0076] The user end uses smart water meter data, and the data sources include:

[0077] Residential users: 500,000 ultrasonic water meters (±1% accuracy), uploading data every 15 minutes.

[0078] Industrial users: 200 electromagnetic flowmeters (±0.5% accuracy), data uploaded every 5 minutes.

[0079] Privacy protection measures include:

[0080] Laplace noise (ε=0.1) is added to residential user data, while industrial user data is only aggregated and encrypted due to the need for accurate measurement.

[0081] Adding noise ensures that individual user data is irretrievable, but the total regional error is less than 3%.

[0082] The monitoring parameters for water quality monitoring at water sources include:

[0083] The five basic parameters (pH, turbidity, residual chlorine, COD, and ammonia nitrogen) were collected every 10 minutes.

[0084] Heavy metal indicators (lead, arsenic, etc.) are tested every 2 hours.

[0085] Each piece of data stored on the blockchain includes:

[0086] Timestamp (UTC standard time);

[0087] GPS coordinates (accuracy ±3m);

[0088] Sensor device ID (e.g., WQ-SENSOR-007);

[0089] Digital signature (ECDSA algorithm);

[0090] Evidence storage delay < 500ms, block confirmation time < 2 seconds.

[0091] Meteorological and environmental data are fused together and accessed via the meteorological bureau's API to obtain 72-hour forecasts (temperature, precipitation, humidity) and real-time monitoring of wind speed and evaporation from local micro-meteorological stations (used to correct water use prediction models).

[0092] The data interface for industrial user water use planning includes:

[0093] By connecting with the enterprise's ERP system, production plans and water reservations can be obtained (e.g., XX factory will increase production from 9:00 to 12:00 tomorrow and needs to increase water supply by 2000m³).

[0094] The data format adopts the ISO 8601 time standard, and the water volume unit is accurate to 0.1m³.

[0095] Edge computing node preprocessing includes noise filtering. For pressure data, a sliding window midpoint filter (window width = 5 sampling points) is used to eliminate transient pulse interference.

[0096] A lightning strike caused the data of sensor P-205 to suddenly drop to 0MPa, and the system automatically identified and removed it.

[0097] For water quality data, based on historical data distribution (e.g., pH is typically 6.5-8.5), data exceeding the 3σ range triggers a verification mechanism.

[0098] For spatiotemporal alignment, when aligning time, all data are aligned to the whole minute timestamp (e.g., 08:00:00), and lagging data is compensated by linear interpolation; when aligning space, scattered monitoring points are mapped to the pipeline network GIS topology map, and regional interpolation is achieved through Delaunay triangulation.

[0099] For handling missing values, the process is based on the fan processing rule table, as shown in Table 1.

[0100] Table 1 Missing Item Handling Rules

[0101] Missing type Handling method Single missing point (<5 minutes) linear interpolation Long-term absence (>30 minutes) Hydraulic model simulation completion Water quality data missing Take the average of data from other sensors at the same water source.

[0102] The adaptive sampling strategy is a dynamically adjusted logic:

[0103] The baseline sampling interval is 5 minutes. When the pressure change rate ΔP / P0 > 8%, the sampling frequency is increased to 1 Hz. Monitoring continues until ΔP / P0 < 5% for 10 minutes.

[0104] In a pipeline leak incident, the system switched to high-frequency sampling within 2 seconds after the pressure dropped by 12%, accurately capturing the leak characteristics.

[0105] For differential privacy protection, the noise addition rule is that when the actual value of the water volume Q of a residential user is 100m³, the output value after adding noise is in the range of 90-110m³ (ε=0.1).

[0106] Industrial user data is only annoyed after regional aggregation (ensuring total error <1%).

[0107] For blockchain evidence storage

[0108] Evidence storage data structure:

[0109] {

[0110] "timestamp": "2023-11-20T08:00:00Z",

[0111] "location": {"lat": 34.0522, "lng": -118.2437},

[0112] "sensor_id": "WQ-7890",

[0113] "parameters": {"pH": 7.2, "turbidity": 0.8},

[0114] "signature": "0x3a7b...e2c4"

[0115] }

[0116] During tamper-proof verification, if the pH data is maliciously modified from 7.2 to 8.5, the blockchain hash verification will fail (hash collision probability < 10^-18).

[0117] The following is an illustration using specific test cases:

[0118] Data processing during heavy rain:

[0119] Input data:

[0120] Pressure data: Data for 2 hours was missing due to damage at 3 monitoring points caused by flooding;

[0121] Water quality data: Turbidity increased sharply from 1.2 NTU to 12.8 NTU;

[0122] System response:

[0123] Start hydraulic model simulation at the edge node to complete the missing pressure data (error <5%).

[0124] Turbidity data triggers emergency blockchain evidence storage (the frequency of evidence storage is increased to once every minute).

[0125] Dynamically adjust the coagulant dosage at downstream water plants (increase from 20 mg / L to 35 mg / L).

[0126] Industrial user data privacy protection:

[0127] Test method:

[0128] The attacker obtained a residential user's annual water usage data sequence (with noise added).

[0129] Attempting to reconstruct the true value through reverse engineering;

[0130] result:

[0131] The restored data deviated from the actual value by an average of 22%, meeting the privacy protection requirement of ε=0.1.

[0132] This embodiment achieves reliable acquisition and secure preprocessing of multi-source water data through the above-mentioned technology combination, providing a high-quality data foundation for subsequent intelligent scheduling.

[0133] Furthermore, S2 also includes:

[0134] The spatiotemporal graph neural network (ST-GNN) model includes:

[0135] The spatiotemporal convolution module is used to extract pipeline network topology features;

[0136] Gated attention module dynamically weights the differences in water usage patterns between industrial and residential users;

[0137] The online learning module updates model parameters in real time based on scheduling feedback data;

[0138] When the pipeline leakage rate is detected to exceed the threshold, a tiered emergency dispatch is triggered to prioritize water supply to critical nodes.

[0139] Automatically generate valve shut-off schemes to isolate leaking areas, and combine hydraulic models to calculate the optimal water replenishment path.

[0140] In this embodiment, a spatiotemporal graph is constructed during the preparation of input data. The nodes are 185 pipeline nodes (including water source plants, booster stations, and user access points), and the edges are 230 pipelines. The weight is 1 / pipe diameter (DN200 pipeline weight = 0.005, DN400 = 0.0025).

[0141] Based on static characteristics (pipe diameter, material, elevation) and dynamic characteristics (pressure and flow rate over the past 24 hours);

[0142] Forecasting process (taking City A as an example):

[0143] Spatiotemporal feature extraction was performed, and spatial convolution was used to identify the water supply correlation between commercial areas (nodes 23-45) and industrial areas (nodes 78-92);

[0144] For example, the spatial influence coefficient of industrial zone node 81 on commercial zone node 35 is 0.17;

[0145] Temporal convolution was used to detect a 15-minute delay in the propagation of water consumption during the evening peak (18:00-20:00) in residential areas;

[0146] For emergency dispatch and response mechanisms, such as the bursting of a DN300 main pipeline (leakage rate exceeding the threshold of 15%), a tiered response process is adopted;

[0147] Leakage was detected by using a pressure sensor P-208 when a sudden drop of 0.2 MPa (lasting for 5 minutes) was detected.

[0148] ST-GNN identified N-67 as an anomalous node (99% confidence level);

[0149] Prioritize ensuring the safety of critical points, including hospitals (point 12, pressure maintained ≥0.25MPa) and fire stations (point 55, flow rate increased by 10%).

[0150] Automatic shut-off valves V15 / V22 / V37 isolate leaking areas; the number of affected users is reduced from 12,000 to 1,500.

[0151] An online learning mechanism is used, triggered based on feedback data, for example, when actual water consumption exceeds the predicted value by 10% three times consecutively:

[0152] For example, if the occupancy of a newly built residential community causes the evening peak water usage time in the residential area to be delayed by 30 minutes, the system will automatically mark the data during this period as high-priority training samples;

[0153] The dynamic adjustment items include: the time decay factor of the LSTM unit (original value 0.9 → adjusted to 0.85), and the initial value of the industrial user attention weight (0.7 → 0.65).

[0154] After the model update, the prediction error for the new cell decreased from 12% to 4%.

[0155] Specific test cases include sudden changes in water usage patterns during holidays:

[0156] Input conditions:

[0157] On the eve of the Spring Festival, water usage in industrial areas decreased by 60%, while it increased by 45% in residential areas.

[0158] Model response:

[0159] The gating attention system automatically reduces the weight of industrial zones to 0.1;

[0160] Spatiotemporal convolution detected enhanced spatial correlation of water use in residential areas (inter-node influence coefficient +20%).

[0161] The revised forecast outputs the peak time for the business district, which is now earlier at 16:00 (error <3%).

[0162] Multiple pipelines leaking simultaneously:

[0163] During the simulation test, nodes N-12 / N-45 were manually triggered to simultaneously reduce the pressure to zero.

[0164] The system generates a two-level emergency response plan within 45 seconds:

[0165] Level 1 isolation: Shut down V8 / V19;

[0166] Secondary water replenishment: Activate backup water source S3;

[0167] Pressure fluctuations at critical points are controlled within ±0.03 MPa.

[0168] This embodiment utilizes the multi-level feature extraction and dynamic adjustment capabilities of ST-GNN to achieve high-precision water demand prediction and rapid emergency response. The system has been running stably for 12 months at the B City Water Group, cumulatively reducing water leakage by 187,000 tons / year.

[0169] Furthermore, S3 also includes:

[0170] The improved multi-objective particle swarm optimization algorithm (MOPSO) includes:

[0171] Dynamic inertia weight: The weight value is adaptively adjusted according to the variance of the water supply pressure;

[0172] Constraint handling mechanism: The limit on the number of pump station start-stop cycles is transformed into a penalty function;

[0173] Pareto solution selection: Selecting the optimal scheduling scheme based on fuzzy membership functions.

[0174] In this embodiment, the objectives defined by the multi-objective function include: water supply stability, energy consumption minimization, and leakage suppression.

[0175] When the objective is water supply stability, ensure that the pressure deviation between each node and the demand pressure is minimized; when the objective is to minimize energy consumption, it is necessary to reduce the total power consumption of the pumping station (α is the electricity price coefficient); when the objective is to suppress leakage, it is necessary to control the leakage of the pipeline with the most severe leakage.

[0176] The decision variables include: pump station frequency (30~50Hz), valve opening (0%~100%), and water tank scheduling level (3~8m).

[0177] Physical constraints include: pump station start-up and shutdown ≤ 5 times per day, pipeline node pressure ≥ 0.2 MPa, and water flow velocity ≤ 2.5 m / s;

[0178] The improved MOPSO algorithm implementation sets dynamic inertia weight adjustment. The adjustment rule is as follows: the baseline weight w=0.7, when the pressure variance σ²>0.05 MPa²: w=0.9 (enhancing global search), and when σ²<0.01 MPa²: w=0.4 (strengthening local optimization).

[0179] During a certain peak period, the pressure variance reached 0.08 MPa², and the algorithm automatically increased the weight to 0.88, quickly finding a new solution;

[0180] Regarding the constraint handling mechanism adopted, the pump station start / stop penalty function is as follows:

[0181] Penalty = 10^6×max(0, N_{start-stop} - 5)^2;

[0182] A certain scheme starts and stops 7 times → penalty value = 4 × 10 6 (The particle is eliminated directly);

[0183] For Pareto solution selection, different membership function types are used depending on the objective. When the objective is water supply stability, the membership function type used is the trapezoidal function; when the objective is energy consumption minimization, the membership function type used is the exponential function; and when the objective is leakage suppression, the membership function type used is the trigonometric function.

[0184] The scheduling scheme generation process includes an initialization phase, which encodes particles. Each particle contains: 3 pump station frequencies (Hz), 8 valve opening degrees (%), and 2 target water tank levels (m).

[0185] The population is set to have 100 particles and a maximum of 50 iterations.

[0186] The optimization process includes revising the prediction results, receiving the predicted residential area value (+25% demand) from the ST-GNN output, and adjusting the industrial area water supply through MOPSO (reducing it by 8% to balance the total amount).

[0187] Dynamic rewards and penalties include a penalty value of += 5 × 10 when a certain plan reduces the pressure on critical hospitals to <0.22 MPa. 5 At the same time, the particle is eliminated;

[0188] In the 20th generation of non-dominated sorting, 32 Pareto solutions were obtained, and 15 solutions that violated the flow velocity constraints were removed.

[0189] The final output parameters of the scheme are shown in Table 2.

[0190] Table 2 Final Scheme Parameter Table

[0191] equipment Setting value Constraint Verification Pump station P3 42Hz (formerly 38Hz) Daily start / stop = 4 times Valve V7 65% opening Flow velocity = 2.3 m / s Water tank W2 6.8m water level Pressure = 0.28 MPa

[0192] Typical test cases include peak summer water usage combined with aging and leaking pipes.

[0193] Input conditions: Peak demand in residential areas +30%, DN400 pipe leakage of 3.5 m³ / h detected.

[0194] Dynamic weights are applied during the MOPSO optimization process, with the stress variance changing from 0.12 to w=0.92.

[0195] During constraint handling, all schemes that start and stop more than 5 times (penalty value > 10) are eliminated. 6 (), retain solutions with pressure ≥ 0.23 MPa;

[0196] Final choice: Energy consumption increased by 8%, but leakage was reduced by 12%;

[0197] After implementation in the water system of City C, this embodiment achieved an average annual energy saving of 150,000 kWh and reduced the leakage rate from 18.7% to 14.2%, verifying the engineering practicality of the improved MOPSO algorithm.

[0198] Furthermore, S4 also includes:

[0199] The robustness of the water resource scheduling scheme under extreme weather conditions was simulated in the water affairs digital twin system.

[0200] Visualize areas of abnormal pipeline pressure through an AR interface and recommend manual intervention strategies;

[0201] The water resource scheduling scheme introduces a time-of-use electricity pricing strategy, prioritizing the activation of high-energy-consuming pump groups during off-peak electricity hours;

[0202] The dosage of water treatment chemicals is dynamically adjusted based on water quality prediction results.

[0203] In this embodiment, the virtual simulation process uses typhoon weather as an example to load extreme scenarios;

[0204] Input parameters:

[0205] Wind speed 25 m / s (Force 10 wind);

[0206] Rainfall of 60 mm / h lasting for 6 hours;

[0207] Three areas at risk of tree collapse (coordinates X:125, Y:348, etc.);

[0208] Robustness testing includes stress anomaly detection:

[0209] A simulated tree falling caused pipe P-09 to break, which was detected by the digital twin system within 12 seconds;

[0210] The pressure at node N-45 dropped sharply from 0.32 MPa to 0.08 MPa;

[0211] Downstream flow increased abnormally by 220%;

[0212] Dynamic adjustment strategies include:

[0213] Automatic response to shut off valves V12 / V15 / V21;

[0214] Start the backup water source S2 (replenishment rate 500 m³ / h);

[0215] We recommend dispatching a repair team to the GPS coordinates (31.2304°N, 121.4737°E).

[0216] The simulation results are shown in Table 3.

[0217] Table 3 Output of Deduction Results

[0218] index Original plan Adjusted plan Improvement effect Number of affected users 8,200 households 1,500 households 81.7%↓ Recovery time 4.5 hours 2.2 hours 51.1%↓ Risk of water quality exceeding standards 2 communities 0 Completely avoid

[0219] AR-assisted decision-making applications are used to visualize stress anomalies;

[0220] Display elements include the overlay of the 3D model of the pipeline network with the actual scene, pressure gradient coloring (blue 0.3MPa → red <0.15MPa), and abnormal point pulse warning (frequency 2Hz).

[0221] According to the manual intervention guidelines, the inspector stares at the abnormal area for 3 seconds to trigger the menu, which displays the available operations, including immediately closing the nearest valve (distance prompt: V15 requires walking 23m) and calling for backup (automatically sending location + on-site video). After selection, an electronic work order (including operation QR code) is generated.

[0222] The time-of-use pricing strategy, including pump scheduling optimization, is shown in Table 4.

[0223] Table 4 Pump Set Scheduling Optimization Table

[0224] Time period Electricity price Original pump set operation Optimized operation 08:00-11:00 Peak electricity 1.2 yuan All 3 55kW pumps are running at full capacity Two units are in operation, and one is on standby. 13:00-15:00 Electricity price: 0.7 yuan Two pumps are running intermittently. Maintain 2 units, reduce frequency by 5%. 22:00-06:00 Off-peak electricity: 0.3 yuan One pump in operation Three pumps are replenishing water at full load.

[0225] In terms of actual results, the pass rate of residual chlorine at the end of the pipeline network increased from 89% to 97%, and the annual consumption of sodium hypochlorite decreased by 12.5 tons.

[0226] After being deployed at the D City Water Resources Bureau, this embodiment successfully reduced the impact area of ​​the burst pipe by 67% during a typhoon, verifying the reliability of the digital twin system under extreme conditions.

[0227] According to a second embodiment of the present invention, referring to Figure 2 This invention claims protection for a resource scheduling system based on water data, comprising:

[0228] One or more processors;

[0229] A memory that stores one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the resource scheduling method based on water data.

[0230] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0231] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0232] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A resource scheduling method based on water resources data, characterized in that, include: S1, Collect multimodal water affairs data, and perform data preprocessing on the multimodal water affairs data through edge computing nodes; S2, input the preprocessed multimodal water data into the water demand prediction model, and output the water demand prediction results for each zone; S3, Set optimization objectives, and use multi-objective particle swarm optimization algorithm to correct the water demand prediction results of the partition to obtain a water resource scheduling scheme; S4, The water resource scheduling scheme is virtually simulated, and the water resource scheduling scheme is dynamically adjusted according to the simulation results; S5. Monitor the implementation effect of the adjusted water resource scheduling plan in real time and feed it back to the water demand prediction model for online learning; S1 further includes: Real-time acquisition of water supply network pressure data, user-end smart water meter water consumption data, water source quality monitoring data, meteorological environmental data, and industrial user water use plan data; Data preprocessing is performed through edge computing nodes, including noise filtering, spatiotemporal alignment, and missing value imputation. S2 further includes: inputting the preprocessed multimodal water data into a water demand prediction model based on the spatiotemporal graph neural network ST-GNN model; By combining the attention mechanism to dynamically adjust the feature weights of different regions and time periods, the regional water demand forecast results for the next 24 hours are output. S3 further includes: With water supply stability, energy consumption minimization, and pipeline leakage suppression as optimization objectives, an improved multi-objective particle swarm optimization algorithm (MOPSO) is used to correct the water demand prediction results of the zoning and generate a water resource scheduling scheme. The MOPSO algorithm introduces physical constraints of water equipment and a dynamic reward and punishment mechanism. S4 further includes: The water resource scheduling plan is synchronized to the water affairs digital twin system for virtual simulation. Based on the simulation results, the water resource scheduling plan is dynamically adjusted and then sent to the execution terminal. S1 further includes: The edge computing node adopts an adaptive sampling strategy, and the sampling frequency of the pressure data is positively correlated with the pressure fluctuation rate of the pipeline network. Differential privacy protection is implemented for user water consumption data by adding noise that conforms to a Laplace distribution; Blockchain-based evidence storage ensures the immutability of water quality data from water sources. The evidence includes the collection time, location, and sensor ID.

2. The resource scheduling method based on water data according to claim 1, characterized in that, The S2 further includes: The spatiotemporal graph neural network (ST-GNN) model includes: The spatiotemporal convolution module is used to extract pipeline network topology features; Gated attention module dynamically weights the differences in water usage patterns between industrial and residential users; The online learning module updates model parameters in real time based on scheduling feedback data; When the pipeline leakage rate is detected to exceed the threshold, a tiered emergency dispatch is triggered to prioritize water supply to critical nodes. Automatically generate valve shut-off schemes to isolate leaking areas, and combine hydraulic models to calculate the optimal water replenishment path.

3. The resource scheduling method based on water data according to claim 2, characterized in that, The S3 further includes: The improved multi-objective particle swarm optimization algorithm (MOPSO) includes: Dynamic inertia weight: The weight value is adaptively adjusted according to the variance of water supply pressure; Constraint handling mechanism: The limit on the number of pump station start-stop cycles is transformed into a penalty function; Pareto solution selection: Selecting the optimal scheduling scheme based on fuzzy membership functions.

4. A resource scheduling method based on water data according to claim 2, characterized in that, The S4 further includes: The robustness of the water resource scheduling scheme under extreme weather conditions was simulated in the water affairs digital twin system. Visualize areas of abnormal pipeline pressure through an AR interface and recommend manual intervention strategies; The water resource scheduling scheme introduces a time-of-use electricity pricing strategy, prioritizing the activation of high-energy-consuming pump groups during off-peak electricity hours; The dosage of water treatment chemicals is dynamically adjusted based on water quality prediction results.

5. A resource scheduling system based on water resources data, characterized in that, include: One or more processors; A memory having stored one or more programs that, when executed by one or more processors, cause the one or more processors to implement a resource scheduling method based on water data according to any one of claims 1 to 4.

Images