An intelligent water affair cloud edge end collaborative control system and method based on an industrial microgrid

Abstract

The application discloses a kind of wisdom water affair cloud edge end collaborative control system and method based on industrial microgrid, including the following steps: cloud platform obtains historical operation data, meteorological data and electricity price data, generates day-ahead scheduling plan, photovoltaic power prediction result, load prediction result and electricity price reference information;Edge controller combines the scheduling plan, prediction result and real-time operation data to establish water-energy coupling scheduling model, generates variable frequency water pump speed regulation strategy and energy storage control strategy;Power prediction response module corrects control strategy according to photovoltaic power change and electricity price change, generates equipment control instruction;Equipment end executes control instruction and feedback operation data, and cloud platform updates prediction model according to feedback data, forms closed-loop collaborative control. Through the collaborative scheduling of water supply system and industrial microgrid, the application realizes efficient consumption of renewable energy under the constraint of water supply safety and economic operation of microgrid.

Description

Technical Field

[0001] This invention relates to the field of collaborative control technology for smart water management and industrial microgrids, and in particular to a cloud-edge-device collaborative control system and method for smart water management based on industrial microgrids. Background Technology

[0002] With the increasing number of electrical equipment such as variable frequency pumps, aeration fans, and chemical dosing pumps in water plants, pumping stations, and sewage treatment plants, the electricity load of water systems continues to grow. Simultaneously, distributed energy sources such as photovoltaic power generation and energy storage devices are gradually being integrated into water facilities, forming industrial microgrids with water supply loads, renewable energy, and energy storage. How to improve the capacity for renewable energy absorption and reduce electricity purchase costs while ensuring water supply security has become a crucial issue in smart water management operations.

[0003] Existing smart water management control systems typically focus on monitoring and controlling water process parameters such as water level, pressure, and water quality. Energy-side systems usually manage photovoltaic power generation, energy storage charging and discharging, or electricity load independently, lacking a unified collaborative scheduling mechanism. When photovoltaic output fluctuates, electricity prices change, or water pump loads change, existing systems struggle to integrate clear water tank levels, pipeline pressure, water quality indicators, and microgrid power status for timely optimization. This can easily lead to problems such as curtailment of solar power, increased peak-hour electricity purchase costs, or uneconomical water pump operation.

[0004] Existing control methods largely rely on centralized cloud computing or local single-point control. Cloud-based centralized control suffers from insufficient real-time performance, making it difficult to respond quickly to power fluctuations at the grid connection point; local control, on the other hand, lacks comprehensive utilization of meteorological, electricity price, and historical operating data, making it difficult to formulate optimized strategies for the next scheduling cycle. Furthermore, existing technologies typically do not incorporate the regulating capacity of variable frequency pumps and clear water tanks as adjustable loads in microgrid control, failing to fully utilize the water level margin of the clear water tanks to proactively respond to photovoltaic power fluctuations and peak-valley electricity price changes.

[0005] Therefore, how to provide a smart water management cloud-edge-device collaborative control system and method based on industrial microgrids is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a smart water management cloud-edge-device collaborative control system and method based on industrial microgrids. This invention fully utilizes cloud-edge-device collaborative control, water-energy coupled scheduling, power prediction response, and real-time closed-loop control technologies to construct a collaborative control architecture that integrates cloud-based predictive optimization, edge-based real-time decision-making, and device-side rapid execution. By integrating historical operating data, meteorological data, electricity price data, and real-time operating data, a water-energy coupled scheduling model is established. Combined with photovoltaic power prediction, load prediction, and clear water reservoir regulation capabilities, the collaborative optimization control of water supply load and microgrid energy supply is achieved. Under the premise of ensuring water supply safety and water quality compliance, this invention improves the renewable energy absorption capacity, reduces electricity purchase costs and microgrid operating costs, and simultaneously enhances the system's real-time response capability, operational stability, and intelligent control level.

[0007] A smart water management cloud-edge-device collaborative control system and method based on an industrial microgrid, according to an embodiment of the present invention, includes the following modules: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller communicates with the cloud platform to receive day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves and electricity price reference signals, and collects real-time power at the microgrid grid connection point, water level in the clear water tank, pipeline pressure, water quality indicators and operating status of water equipment. The water-energy coupled scheduling module, located in the edge controller, generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operating status. The power prediction response module, located in the edge controller, generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. On the device side, it communicates with the edge controller, controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The model training engine, located on a cloud platform, updates the prediction model based on equipment operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, and electricity price reference signal for the next scheduling cycle.

[0008] Optionally, the cloud platform includes a meteorological-electricity price data interface, a historical database, a microgrid energy management module, a water process optimization module, and the model training engine; The meteorological-electricity price data interface is used to acquire meteorological data and electricity price data, and to write the meteorological data and electricity price data into the historical database; The historical database is used to store historical operating data, meteorological data, electricity price data, and equipment operating data generated by the device and uploaded by the edge controller; The model training engine is used to update the prediction model using historical operating data, meteorological data, electricity price data, and equipment operating data as training samples, and to generate photovoltaic power prediction curves and load prediction curves based on the updated prediction model. The microgrid energy management module is used to generate microgrid-side energy dispatch results based on photovoltaic power prediction curves, load prediction curves, and electricity price reference signals. The water process optimization module is used to generate start-up and shutdown arrangements for water equipment based on the energy dispatch results on the microgrid side and the operational needs of the water system. The cloud platform is used to generate day-ahead scheduling plans based on the energy dispatch results of the microgrid and the start-up and shutdown arrangements of water equipment. After receiving new equipment operation data, it updates the prediction model through the model training engine to generate the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle.

[0009] Optionally, the equipment includes a variable frequency water pump, a water quality analyzer, a smart meter, a photovoltaic inverter, and an energy storage converter; The edge controller is used to send the variable frequency water pump speed command to the variable frequency water pump, and the variable frequency water pump performs speed adjustment according to the variable frequency water pump speed command. The edge controller is also used to send the power setting value of the energy storage converter to the energy storage converter, and the energy storage converter performs charging control or discharging control according to the power setting value of the energy storage converter. The water quality analyzer is used to collect water quality indicators, the smart meter is used to collect real-time power at the microgrid grid connection point, the photovoltaic inverter is used to feed back photovoltaic power generation operation data, and the variable frequency water pump and energy storage converter are used to feed back the corresponding equipment operating status. The device generates equipment operation data based on water quality indicators, real-time power of the microgrid grid connection point, photovoltaic power generation operation data, and equipment operation status, and uploads it to the cloud platform through the edge controller for the model training engine to update the prediction model.

[0010] Optionally, the cloud platform and the edge controller use the MQTT publish / subscribe protocol to transmit day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals. The transmitted data is in Protobuf format and carries timestamps, confidence levels, and validity periods. The edge controller and the device end use OPC UA or Modbus TCP to read and write device parameters. When the system includes multiple edge controllers, the multiple edge controllers share distributed data through DDS, and the edge controller uploads status information to the cloud platform every 500ms. The instructions issued by the cloud platform carry sequence numbers.

[0011] Optionally, the water-energy coupling scheduling module is used to divide the scheduling cycle into multiple scheduling periods according to a preset scheduling time interval, and to acquire the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operation status in each scheduling period; The water-energy coupled scheduling module determines the water supply safety boundary based on the water level of the clear water tank, the pressure of the pipeline network, the water quality indicators and the operating status of the water equipment. It determines the microgrid power balance boundary based on the photovoltaic power prediction curve, the load prediction curve, the electricity price reference signal and the real-time power of the microgrid grid connection point. Based on the water supply safety boundary and the microgrid power balance boundary, a water-energy coupled scheduling model is established. The water-energy coupled scheduling model uses the start-stop status, speed setpoint, energy storage converter power setpoint, and grid connection point power of each variable frequency water pump in each scheduling period as scheduling variables, and the clear water tank level, pipeline pressure, water quality indicators, and microgrid grid connection point power meeting the corresponding boundary conditions in each scheduling period as constraints to generate variable frequency water pump speed regulation strategies and energy storage converter power setpoints that meet water supply safety requirements.

[0012] Optionally, the water-energy coupled scheduling model is a mixed integer linear programming model, and the preset scheduling time interval is 15 minutes; The constraints include: the water level in the clear water tank is not lower than the preset lower limit and not higher than the preset upper limit; the pipeline pressure is not lower than the preset minimum pressure; the number of times each variable frequency water pump stops per unit time does not exceed the preset number of stops; and the water quality indicators are within the preset water quality compliance range. The optimization objective of the mixed-integer linear programming model is to minimize the sum of electricity purchase cost, equipment loss cost, and renewable energy abandonment penalty cost in each scheduling period, while satisfying the constraints. The water-energy coupled scheduling module obtains the start-stop status and speed setpoint of each variable frequency water pump in each scheduling period by solving a mixed integer linear programming model, as well as the power setpoint of the energy storage converter in each scheduling period. It then combines the start-stop status and speed setpoint to generate a variable frequency water pump speed regulation strategy, and generates the power setpoint of the energy storage converter in each scheduling period as the energy storage converter power setpoint.

[0013] Optionally, the power prediction response module is used to receive the photovoltaic power prediction curve, the electricity price reference signal, and the variable frequency water pump speed adjustment strategy, and determine the photovoltaic power change status within a preset time period according to the time sequence of the photovoltaic power prediction curve. The photovoltaic power change states include photovoltaic power rising state, photovoltaic power falling state, and photovoltaic power peak state. The power prediction response module is also used to determine the electricity price peak state based on the electricity price reference signal. The power prediction response module corrects the variable frequency pump speed setpoint for the corresponding scheduling period in the variable frequency pump speed regulation strategy based on the photovoltaic power rising state, photovoltaic power falling state, photovoltaic power peak state, and electricity price peak state, and generates a time-segmented variable frequency pump speed setpoint sequence.

[0014] Optionally, the preset future duration is 4 hours, and the time interval between adjacent prediction points in the photovoltaic power prediction curve is 5 minutes; During the scheduling period corresponding to the photovoltaic power rise state or photovoltaic power peak state, the power prediction response module determines whether the water level in the clear water tank is lower than the preset upper limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, the variable frequency pump speed setting value for the corresponding scheduling period is modified based on the variable frequency pump speed adjustment strategy to form the first speed setting value. During the scheduling period corresponding to the photovoltaic power decline state or the electricity price peak state, the power prediction response module determines whether the water level of the clear water tank is higher than the preset lower limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, the variable frequency pump speed setting value of the corresponding scheduling period is modified based on the variable frequency pump speed adjustment strategy to form a second speed setting value. The power prediction response module generates a time-segmented variable frequency pump speed setting sequence based on the first speed setting value and the second speed setting value.

[0015] Optionally, the power prediction response module determines the adjustable water level margin and the adjustable water level margin of the clear water tank based on the water level of the clear water tank, the preset upper limit of the water level, and the preset lower limit of the water level, and determines the regulating volume of the clear water tank based on the adjustable water level margin and the adjustable water level margin of the clear water tank. The power prediction response module characterizes the regulating volume of the clear water tank as a virtual energy storage capacity formed by the variable frequency water pump and the clear water tank, and limits the speed increase corresponding to the first speed setting value and the speed decrease corresponding to the second speed setting value according to the virtual energy storage capacity. The power prediction response module is used to perform speed change rate constraint processing on the time-sharing variable frequency water pump speed setting sequence. The speed change rate of the variable frequency water pump between adjacent control times does not exceed ±2% of the rated speed per second, and the time-sharing variable frequency water pump speed setting sequence after speed change rate constraint processing is used as a candidate variable frequency water pump speed command sequence. The power prediction response module is also used to collect the real-time power of the microgrid grid connection point with a collection period of 200ms, and determine the power setpoint of the grid connection point according to the day-ahead scheduling plan. The real-time power at the microgrid grid connection point is compared with the grid connection point power setpoint to generate the grid connection point power deviation. The grid connection point power deviation is then separated into high-frequency power components and low-frequency power components by a low-pass filter. The power setting value of the energy storage converter is generated or corrected based on the high-frequency power component, and the speed correction value of the variable frequency water pump is generated based on the low-frequency power component. The variable frequency pump speed correction is superimposed on the candidate variable frequency pump speed command sequence to generate the variable frequency pump speed command. The energy storage converter responds to the high-frequency power component corresponding to the photovoltaic power fluctuation, and the variable frequency pump responds to the low-frequency power component corresponding to the load change.

[0016] Optionally, a smart water management cloud-edge-device collaborative control method based on an industrial microgrid, used to implement the smart water management cloud-edge-device collaborative control system based on an industrial microgrid as described in any one of claims 1-9, includes: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller receives the day-ahead scheduling plan, photovoltaic power forecast curve, load forecast curve and electricity price reference signal, and collects the real-time power of the microgrid grid connection point, water level of the clear water tank, pipeline pressure, water quality indicators and water equipment operation status; The edge controller establishes a water-energy coupled scheduling model based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, water level of clear water tank, pipeline pressure, water quality indicators and water equipment operation status, and generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the water-energy coupled scheduling model. The edge controller generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. The equipment controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The edge controller uploads the device operation data to the cloud platform. The cloud platform updates the prediction model based on the device operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle.

[0017] The beneficial effects of this invention are: This invention establishes a cloud-edge-device collaborative control architecture suitable for industrial microgrids and smart water systems by combining cloud-based predictive optimization, edge real-time control, and rapid execution at the device end. This architecture enables the cloud to generate scheduling plans and forecasts based on historical operating data, meteorological data, and electricity price data. The edge controller can make real-time decisions based on the power of the microgrid grid connection point, the water level of the clear water tank, the pipeline pressure, water quality indicators, and equipment operating status. The device end can quickly execute control commands and provide feedback on operating data, thereby avoiding the problem of the separation between existing water control and energy management and improving the overall collaborative control capability of the system.

[0018] This invention establishes a water-energy coupled scheduling model to jointly optimize the speed of variable frequency pumps and the power of energy storage converters under constraints such as water supply safety, water quality compliance, stable pipeline pressure, and pump start-stop restrictions. This enables the water supply load to be coordinated with changes in photovoltaic power generation, energy storage equipment, and electricity prices, thereby reducing peak-hour electricity purchases, minimizing equipment losses, and improving the economic efficiency of microgrid operation while ensuring the continuity and safety of water supply.

[0019] This invention utilizes a power prediction response and clear water tank regulation capacity mechanism to use the adjustable load formed by the variable frequency water pump and the clear water tank to smooth out photovoltaic power fluctuations. When photovoltaic power increases, the water pump speed is increased to raise the water level in the clear water tank and absorb excess electricity. When photovoltaic power decreases or electricity prices rise, the water pump speed is reduced to utilize the water storage capacity to reduce power consumption from the grid, thereby improving the renewable energy absorption capacity, reducing the risk of curtailment, and enhancing the system's proactive response to peak and valley changes in electricity prices and power fluctuations.

[0020] This invention uses equipment operation data feedback and model training engine to update the prediction model, enabling the cloud platform to continuously correct the prediction results and scheduling plan for the next scheduling cycle based on the actual operation. This forms a closed-loop control process of "prediction-scheduling-execution-feedback-update", thereby improving prediction accuracy, control stability and operation intelligence, reducing reliance on manual scheduling, and enhancing the adaptive operation capability of the smart water system under complex operating conditions. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall architecture diagram of a smart water management cloud-edge-device collaborative control system based on an industrial microgrid proposed in this invention. Figure 2 This is a flowchart illustrating the modeling and solution process of the water-energy coupled scheduling model in this invention. Figure 3 This is a schematic diagram illustrating the control method in this invention, which utilizes a variable frequency water pump and a clear water tank to form a virtual energy storage system and regulate the water speed. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0023] refer to Figures 1-3 A smart water management cloud-edge-device collaborative control system and method based on industrial microgrids includes the following steps: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller communicates with the cloud platform to receive day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves and electricity price reference signals, and collects real-time power at the microgrid grid connection point, water level in the clear water tank, pipeline pressure, water quality indicators and operating status of water equipment. The water-energy coupled scheduling module, located in the edge controller, generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operating status. The power prediction response module, located in the edge controller, generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. On the device side, it communicates with the edge controller, controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The model training engine, located on a cloud platform, updates the prediction model based on equipment operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, and electricity price reference signal for the next scheduling cycle, so as to realize water-energy coordinated scheduling and cloud-edge-device closed-loop coordinated control between the water supply system and the industrial microgrid.

[0024] In this embodiment, the cloud platform includes a meteorological-electricity price data interface, a historical database, a microgrid energy management module, a water process optimization module, and the model training engine; The meteorological-electricity price data interface is used to acquire meteorological data and electricity price data, and to write the meteorological data and electricity price data into the historical database; The historical database is used to store historical operating data, meteorological data, electricity price data, and equipment operating data generated by the device and uploaded by the edge controller; The model training engine is used to update the prediction model using historical operating data, meteorological data, electricity price data, and equipment operating data as training samples, and to generate photovoltaic power prediction curves and load prediction curves based on the updated prediction model. The microgrid energy management module is used to generate microgrid-side energy dispatch results based on photovoltaic power prediction curves, load prediction curves, and electricity price reference signals. The water process optimization module is used to generate start-up and shutdown arrangements for water equipment based on the energy dispatch results on the microgrid side and the operational needs of the water system. The cloud platform is used to generate a day-ahead scheduling plan based on the energy dispatch results of the microgrid and the start-up and shutdown arrangements of water equipment. After receiving new equipment operation data, it updates the prediction model through the model training engine to generate the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle. In Example 1, the cloud platform is deployed on a central server, a public cloud, or a private cloud. The cloud platform includes a meteorological-electricity price data interface, a historical database, a microgrid energy management module, a water process optimization module, and a model training engine. The meteorological-electricity price data interface is used to acquire external meteorological data and electricity price data according to a preset cycle. The meteorological data includes one or more of the following: solar irradiance, ambient temperature, humidity, cloud cover, wind speed, and weather type; the electricity price data includes time-of-use electricity price, real-time electricity price, or day-ahead electricity price; after acquiring the data, the meteorological-electricity price data interface adds a collection timestamp and data source identifier to each data point and writes it into the historical database. The historical database is used to store historical operating data, meteorological data, electricity price data, and equipment operating data generated by the equipment and uploaded by the edge controller. The historical operating data includes historical photovoltaic output, historical electricity load, historical water pump speed, historical clear water tank water level, historical pipeline pressure, historical water quality indicators, and historical energy storage charging and discharging power. The equipment operating data includes the operating status of the variable frequency water pump, the grid connection point power collected by the smart meter, the turbidity, residual chlorine, and pH value collected by the water quality analyzer, the photovoltaic power generation power fed back by the photovoltaic inverter, and the charging and discharging power fed back by the energy storage converter. The model training engine reads historical operating data, meteorological data, electricity price data, and equipment operating data from historical databases. It aligns data from different sources according to timestamps and processes missing, outlier, and duplicate data. For short-term missing data, it can use interpolation between adjacent time points or preserve the previous valid value to fill in the gaps. For data that significantly exceeds the equipment's rated range, it is marked as outlier data and removed from the training samples. The model training engine uses historical meteorological data, historical photovoltaic power output, and time characteristics as input samples, and uses future photovoltaic power output as the prediction target to train or update the photovoltaic power prediction model; it uses historical electricity load, historical water pump operating status, clear water tank water level, pipeline pressure, and time characteristics as input samples, and uses future water system load as the prediction target to train or update the load prediction model. The time characteristics include one or more of the following: hour, date, weekday / non-weekday, and season. Specifically, the photovoltaic power prediction curve is used to characterize the change in photovoltaic output over the next 4 hours, with a prediction point interval of 5 minutes; the load prediction curve is used to characterize the change in water load within the future scheduling cycle, with a scheduling interval of 15 minutes. After receiving new equipment operation data uploaded by the edge controller, the model training engine writes the new equipment operation data into the historical database and retrains or incrementally updates the prediction model according to a preset update cycle, so that the photovoltaic power prediction curve and load prediction curve generated in the next scheduling cycle can reflect the latest operating status. The microgrid energy management module calculates the photovoltaic consumption demand, grid-connected power demand, and energy storage charging and discharging demand for each scheduling period based on the photovoltaic power forecast curve, load forecast curve, and electricity price reference signal, and generates the microgrid-side energy scheduling results. The water process optimization module generates the water equipment start-up and shutdown schedule based on the clear water tank water level demand, pipeline pressure demand, and water quality compliance demand. The cloud platform integrates the microgrid-side energy scheduling results and the water equipment start-up and shutdown schedule to generate the day-ahead scheduling plan, and sends the day-ahead scheduling plan, photovoltaic power forecast curve, load forecast curve, and electricity price reference signal to the edge controller.

[0025] In this embodiment, the equipment includes a variable frequency water pump, a water quality analyzer, a smart meter, a photovoltaic inverter, and an energy storage converter; The edge controller is used to send the variable frequency water pump speed command to the variable frequency water pump, and the variable frequency water pump performs speed adjustment according to the variable frequency water pump speed command. The edge controller is also used to send the power setting value of the energy storage converter to the energy storage converter, and the energy storage converter performs charging control or discharging control according to the power setting value of the energy storage converter. The water quality analyzer is used to collect water quality indicators, the smart meter is used to collect the real-time power of the microgrid grid connection point, the photovoltaic inverter is used to feed back photovoltaic power generation operation data, and the variable frequency water pump and energy storage converter are used to feed back the corresponding equipment operation status. The device generates equipment operation data based on water quality indicators, real-time power of the microgrid grid connection point, photovoltaic power generation operation data, and equipment operation status, and uploads it to the cloud platform through the edge controller for the model training engine to update the prediction model. In Example 1, the equipment includes a variable frequency water pump, a water quality analyzer, a smart meter, a photovoltaic inverter, and an energy storage converter. Depending on the actual configuration of the water plant or pumping station, the equipment may also include an aeration blower, a dosing pump, or other water treatment process equipment. After receiving the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal from the cloud platform, the edge controller collects real-time operating data from the equipment and generates variable frequency pump speed commands and energy storage converter power setpoints based on the calculation results of the water-energy coupling scheduling module and the power prediction response module. After receiving a speed command, the variable frequency pump (VFP) driver adjusts the pump speed according to the command to change the pump flow rate and power consumption. The energy storage converter (EVC) receives a power setpoint and performs charging or discharging control accordingly. A water quality analyzer collects water quality indicators such as turbidity, residual chlorine, and pH; a smart meter collects real-time power, voltage, current, and power factor at the microgrid connection point; a photovoltaic inverter provides feedback on photovoltaic power generation, inverter operating status, and fault status; and the VFP and EDC provide feedback on their respective start / stop status, operating power, fault status, and execution results. The device generates device operation data from the collected data and execution results, and uploads it to the cloud platform via the edge controller. After receiving the device operation data, the cloud platform writes it into the historical database as samples for the model training engine to update the prediction model. This forms a closed-loop process of "cloud generating prediction and scheduling results - edge controller generating control commands - device executing and providing feedback - cloud updating prediction model".

[0026] In this embodiment, the cloud platform and the edge controller use the MQTT publish / subscribe protocol to transmit day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals. The transmitted data is in Protobuf format and carries timestamps, confidence levels, and validity periods. The edge controller and the device end use OPC UA or ModbusTCP to read and write device parameters. When the system includes multiple edge controllers, the multiple edge controllers share distributed data through DDS, and the edge controller uploads status information to the cloud platform every 500ms. The instructions issued by the cloud platform carry sequence numbers. In Example 1, the cloud platform and the edge controller are connected via 4G, 5G, or wired networks, and the MQTT publish / subscribe protocol is used to transmit day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals. The cloud platform acts as the publisher, and the edge controller acts as the subscriber. The edge controller subscribes to the scheduling topics of the corresponding site to receive data from the cloud platform. The data distributed by the cloud platform is encapsulated in Protobuf format. The encapsulated fields include site number, message type, generation timestamp, validity period, confidence level, sequence number, and data payload. Among them, the timestamp is used to determine the data generation time, the confidence level is used to indicate the reliability of the prediction results, the validity period is used to limit the time range in which the data can be used, and the sequence number is used to distinguish the order of instructions and avoid repeating old instructions. The edge controller communicates with the device via OPC UA or Modbus TCP to read and write device parameters. For variable frequency water pumps, the edge controller can write the speed setting value to the water pump driver via OPC UA. For smart meters, photovoltaic inverters, and energy storage converters, the edge controller can read real-time parameters such as power, voltage, current, operating status, and fault status via OPC UA or Modbus TCP. When the system includes multiple water plants, pumping stations, or wastewater treatment plants, edge controllers at different sites share data in a distributed manner via DDS. The shared data includes site load, adjustable pump capacity, clear water tank level, photovoltaic output, energy storage status, and grid connection power, supporting multi-site coordinated scheduling. Specifically, the edge controller uploads status information to the cloud platform every 500ms. This status information includes the edge controller's online status, communication quality, device connection status, current control mode, and key operating parameters. Control commands issued by the cloud platform carry serial numbers. Upon receiving a control command, the edge controller verifies the serial number and validity period. If the control command is a valid new command, it is executed and a confirmation message is returned. If the control command has already been executed or has expired, it is discarded, thereby improving control reliability.

[0027] In this embodiment, the water-energy coupling scheduling module is used to divide the scheduling cycle into multiple scheduling periods according to a preset scheduling time interval, and to acquire the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operation status in each scheduling period. The water-energy coupled scheduling module determines the water supply safety boundary based on the water level of the clear water tank, the pressure of the pipeline network, the water quality indicators and the operating status of the water equipment. It determines the microgrid power balance boundary based on the photovoltaic power prediction curve, the load prediction curve, the electricity price reference signal and the real-time power of the microgrid grid connection point. Based on the water supply safety boundary and the microgrid power balance boundary, a water-energy coupled scheduling model is established. The water-energy coupled scheduling model uses the start-stop status, speed setpoint, energy storage converter power setpoint, and grid connection point power of each variable frequency water pump in each scheduling period as scheduling variables, and the clear water tank level, pipeline pressure, water quality indicators, and microgrid grid connection point power meeting the corresponding boundary conditions in each scheduling period as constraints to generate variable frequency water pump speed regulation strategies and energy storage converter power setpoints that meet water supply safety requirements. In Example 1, the water-energy coupled scheduling module is located in the edge controller and is used to establish a water-energy coupled scheduling model at a preset scheduling time interval. The preset scheduling time interval can be set to 15 minutes, and a scheduling cycle can include multiple consecutive scheduling periods, such as dividing the next 4 hours into 16 scheduling periods, or dividing the next 24 hours into 96 scheduling periods; During each scheduling period, the water-energy coupled scheduling module reads the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal issued by the cloud platform, and reads the real-time power of the microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operation status collected by the edge controller. The water-energy coupled scheduling module determines the water supply safety boundary based on the clear water tank level, pipeline pressure, water quality indicators, and the operating status of water equipment. The water supply safety boundary includes the allowable range of clear water tank level, the allowable range of pipeline pressure, the range of water quality compliance, and the allowable range of pump start-stop cycles. The water-energy coupled scheduling module also determines the microgrid power balance boundary based on the photovoltaic power prediction curve, load prediction curve, electricity price reference signal, and the real-time power at the microgrid grid connection point. This microgrid power balance boundary is used to limit the balance between photovoltaic output, load power, energy storage charging and discharging power, and grid connection point power drawdown during each scheduling period. In modeling, the water-energy coupled scheduling model uses the start / stop status, speed setpoint, energy storage converter power setpoint, and grid connection point power drawdown of each variable frequency pump in each scheduling period as scheduling variables. Specifically, the start / stop status of the variable frequency pump is a discrete variable, representing whether the pump is running or stopped during the corresponding scheduling period; the variable frequency pump speed setpoint is a continuous variable, representing the target speed of the pump during the corresponding scheduling period; the energy storage converter power setpoint is a continuous variable, representing the charging or discharging power of the energy storage system; and the grid connection point power drawdown is a continuous variable, representing the power that the microgrid draws from or feeds back to the external grid. The output of the water-energy coupled scheduling model includes a variable frequency pump speed regulation strategy and an energy storage converter power setpoint. The variable frequency pump speed regulation strategy includes the start / stop status and speed setpoint of each variable frequency pump during each scheduling period; the energy storage converter power setpoint includes the charging or discharging power of the energy storage converter during each scheduling period. The edge controller generates the variable frequency pump speed command and energy storage control parameters required for subsequent real-time control based on this output.

[0028] In this embodiment, the water-energy coupled scheduling model is a mixed integer linear programming model, and the preset scheduling time interval is 15 minutes; The constraints include: the water level in the clear water tank is not lower than the preset lower limit and not higher than the preset upper limit; the pipeline pressure is not lower than the preset minimum pressure; the number of times each variable frequency water pump stops per unit time does not exceed the preset number of stops; and the water quality indicators are within the preset water quality compliance range. The optimization objective of the mixed-integer linear programming model is to minimize the sum of electricity purchase cost, equipment loss cost, and renewable energy abandonment penalty cost in each scheduling period, while satisfying the constraints. The water-energy coupling scheduling module obtains the start-stop status and speed setpoint of each variable frequency water pump in each scheduling period by solving a mixed integer linear programming model, as well as the power setpoint of the energy storage converter in each scheduling period. It then combines the start-stop status and speed setpoint to generate a variable frequency water pump speed regulation strategy, and generates the power setpoint of the energy storage converter in each scheduling period as the energy storage converter power setpoint. In Example 1, the water-energy coupled scheduling model is represented by a mixed-integer linear programming model. This model has a time resolution of 15 minutes and simultaneously considers water supply security constraints and microgrid economic objectives within each scheduling cycle. Water supply safety constraints include clear water tank level constraints, pipeline pressure constraints, pump start / stop constraints, and water quality constraints. The clear water tank level constraint can be expressed as: ,in, Indicates the first The water level in the clear water reservoir during each scheduling period. This indicates the preset lower limit of the water level. This indicates the preset upper limit of the water level. Pipeline pressure constraints can be expressed as: ,in, Indicates the first Pipeline pressure during each scheduling period This indicates the preset minimum pressure; specifically, the pressure at the end user should not be lower than 0.14 MPa. Pump start / stop constraints limit the number of times each variable frequency pump can stop per unit time; specifically, each variable frequency pump should not stop more than 6 times per hour. Water quality constraints limit turbidity, residual chlorine, and pH value to within preset water quality standards. The microgrid power balance constraint is used to define the relationship between photovoltaic (PV) output, load power, energy storage power, and grid connection point power drawdown. Specifically, the microgrid grid connection point power drawdown is determined based on the water load power, energy storage charging and discharging power, and PV power. When PV power exceeds the local load absorption capacity, the model reduces curtailment of PV power by increasing the speed of the variable frequency water pump or controlling energy storage charging. When PV power decreases or electricity prices increase, the model reduces grid connection by decreasing the speed of the variable frequency water pump or controlling energy storage discharging. The economic optimization objective of the water-energy coupled dispatch model is to minimize the sum of electricity purchase cost, equipment loss cost, and renewable energy curtailment penalty cost, while satisfying water supply security constraints and microgrid power balance constraints. This optimization objective can be expressed as: summing the electricity purchase cost, grid connection point power taken under the weighted electricity price, equipment loss cost, and wind / solar curtailment penalty cost for each dispatch period throughout the entire dispatch cycle, and minimizing this total cost. During the solution process, if a mixed-integer linear programming full solution is used, the Gurobi solver can be called to obtain a globally feasible or optimal solution. When the edge controller's computing resources are limited or a fast response is required, continuous variables can also be relaxed and the OSQP solver can be called to obtain approximate control results. When the solver does not obtain a feasible solution within a preset computing time, the edge controller uses preset priority rules for degraded control. The preset priority rules include prioritizing the satisfaction of clear water tank water level constraints, prioritizing the satisfaction of pipeline pressure constraints, and prioritizing the satisfaction of water quality constraints. Economic optimization is then performed after the above constraints are satisfied. In the above manner, the water-energy coupling scheduling module can obtain the start-stop status and speed setting value of each variable frequency water pump in each scheduling period, as well as the power setting value of the energy storage converter in each scheduling period. It can also combine the start-stop status and speed setting value of each variable frequency water pump to generate a variable frequency water pump speed regulation strategy, and output the power setting value of the energy storage converter as the power setting value of the energy storage converter.

[0029] In this embodiment, the power prediction response module is used to receive the photovoltaic power prediction curve, the electricity price reference signal and the variable frequency water pump speed adjustment strategy, and determine the photovoltaic power change status within a preset time period according to the time sequence of the photovoltaic power prediction curve. The photovoltaic power change states include photovoltaic power rising state, photovoltaic power falling state, and photovoltaic power peak state. The power prediction response module is also used to determine the electricity price peak state based on the electricity price reference signal. The power prediction response module corrects the variable frequency pump speed setpoint for the corresponding scheduling period in the variable frequency pump speed regulation strategy based on the photovoltaic power rising state, photovoltaic power falling state, photovoltaic power peak state and electricity price peak state, and generates a time-segmented variable frequency pump speed setpoint sequence. In Example 1, the power prediction response module is located in the edge controller and is used to receive the photovoltaic power prediction curve and electricity price reference signal sent from the cloud platform, and to receive the variable frequency water pump speed regulation strategy output by the water-energy coupling scheduling module. The photovoltaic power prediction curve is arranged in chronological order with multiple prediction points, and each prediction point includes the prediction time and the corresponding photovoltaic prediction power. The power prediction response module calculates the change in photovoltaic power between adjacent prediction points according to the time sequence of the prediction points. When the photovoltaic predicted power of the later prediction point is greater than that of the earlier prediction point and the difference exceeds a preset rising threshold, the corresponding time period is determined to be a photovoltaic power rising state; when the photovoltaic predicted power of the later prediction point is less than that of the earlier prediction point and the difference exceeds a preset falling threshold, the corresponding time period is determined to be a photovoltaic power falling state; when the photovoltaic predicted power is higher than a preset peak power threshold or higher than the average photovoltaic predicted power within a preset future time period by a certain proportion, the corresponding time period is determined to be a photovoltaic power peak state.

[0030] The power prediction response module also determines the peak electricity price state based on the electricity price reference signal. When the electricity price during a certain scheduling period is higher than the preset electricity price threshold, or when the scheduling period is in the peak segment of the time-of-use electricity price, the scheduling period is determined to be a peak electricity price state. After identifying the photovoltaic power rising state, photovoltaic power falling state, photovoltaic power peak state, and electricity price peak state, the power prediction response module matches the above states with the corresponding scheduling time periods in the variable frequency pump speed regulation strategy, and corrects the variable frequency pump speed setpoint within the corresponding scheduling time period to generate a time-sharing variable frequency pump speed setpoint sequence. The photovoltaic prediction and electricity price signals generated by the cloud platform can be converted into a time-sharing speed control plan that can be executed on the edge side.

[0031] In this embodiment, the preset future duration is 4 hours, and the time interval between adjacent prediction points in the photovoltaic power prediction curve is 5 minutes. During the scheduling period corresponding to the photovoltaic power rise state or photovoltaic power peak state, the power prediction response module determines whether the water level of the clear water tank is lower than the preset upper limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, it modifies the variable frequency pump speed setting value for the corresponding scheduling period based on the variable frequency pump speed adjustment strategy to form a first speed setting value for absorbing excess photovoltaic power and raising the water level of the clear water tank. During the scheduling period corresponding to the photovoltaic power decline state or the electricity price peak state, the power prediction response module determines whether the water level of the clear water tank is higher than the preset lower limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, it modifies the variable frequency pump speed setting value for the corresponding scheduling period based on the variable frequency pump speed adjustment strategy to form a second speed setting value for reducing the grid-connected power consumption. The power prediction response module generates a time-segmented variable frequency pump speed setting sequence based on the first speed setting value and the second speed setting value.

[0032] In Example 1, the preset duration is 4 hours, and the time interval between adjacent prediction points in the photovoltaic power prediction curve is 5 minutes. The power prediction response module corrects the time period of the variable frequency pump speed adjustment strategy output by the water-energy coupling scheduling module based on the photovoltaic power prediction curve and the electricity price reference signal. When a certain scheduling period coincides with a rise or peak in photovoltaic power, the power prediction response module first determines whether the water level in the clear water tank is below the preset upper limit and whether the pipeline pressure is not below the preset minimum pressure. If the water level in the clear water tank is below the preset upper limit and the pipeline pressure is not below the preset minimum pressure, it indicates that the clear water tank still has storage space and the water supply pressure is safe. The power prediction response module then increases the variable frequency pump speed setting value for this scheduling period based on the original variable frequency pump speed setting value, forming the first speed setting value. The first speed setting value is used to increase the pump load in advance, consume the surplus photovoltaic power, and raise the water level in the clear water tank to a higher level. When a certain scheduling period is in a state of declining photovoltaic power or peak electricity price, the power prediction response module first determines whether the water level in the clear water tank is higher than the preset lower limit and whether the pipeline pressure is not lower than the preset minimum pressure. If the water level in the clear water tank is higher than the preset lower limit and the pipeline pressure is not lower than the preset minimum pressure, it means that the clear water tank has the conditions to release water storage capacity. The power prediction response module reduces the variable frequency pump speed setting value for this scheduling period based on the original variable frequency pump speed setting value to form a second speed setting value. The second speed setting value is used to reduce the power load of the pump and reduce the power taken from the external grid during peak electricity price or declining photovoltaic power periods. If the water level in the clear water tank has reached the preset upper limit, the power prediction response module will no longer increase the speed of the variable frequency pump; if the water level in the clear water tank is close to the preset lower limit, the power prediction response module will no longer decrease the speed of the variable frequency pump; if the pipeline pressure is lower than the preset minimum pressure, the variable frequency pump speed that meets the water supply pressure requirements will be restored or maintained first. Through this logic, the system can achieve photovoltaic power consumption and electricity price response while ensuring water supply security.

[0033] In this embodiment, the power prediction response module determines the adjustable water level margin and the adjustable water level margin of the clear water tank based on the water level of the clear water tank, the preset upper limit of the water level and the preset lower limit of the water level, and determines the regulating volume of the clear water tank based on the adjustable water level margin and the adjustable water level margin of the clear water tank. The power prediction response module characterizes the regulating volume of the clear water tank as a virtual energy storage capacity formed by the variable frequency water pump and the clear water tank, and limits the speed increase corresponding to the first speed setting value and the speed decrease corresponding to the second speed setting value according to the virtual energy storage capacity. The power prediction response module is used to perform speed change rate constraint processing on the time-sharing variable frequency water pump speed setting sequence. The speed change rate of the variable frequency water pump between adjacent control times does not exceed ±2% of the rated speed per second, and the time-sharing variable frequency water pump speed setting sequence after speed change rate constraint processing is used as a candidate variable frequency water pump speed command sequence. The power prediction response module is also used to collect the real-time power of the microgrid grid connection point with a collection period of 200ms, and determine the power setpoint of the grid connection point according to the day-ahead scheduling plan. The real-time power at the microgrid grid connection point is compared with the grid connection point power setpoint to generate the grid connection point power deviation. The grid connection point power deviation is then separated into high-frequency power components and low-frequency power components by a low-pass filter. The power setting value of the energy storage converter is generated or corrected based on the high-frequency power component, and the speed correction value of the variable frequency water pump is generated based on the low-frequency power component. The variable frequency pump speed correction is superimposed on the candidate variable frequency pump speed command sequence to generate the variable frequency pump speed command. The energy storage converter responds to the high frequency power component corresponding to the photovoltaic power fluctuation, and the variable frequency pump responds to the low frequency power component corresponding to the load change. In Example 1, the power prediction response module determines the adjustable water level margin of the clear water tank based on the water level, the preset upper limit of the water level, and the preset lower limit of the water level, and determines the regulating volume of the clear water tank based on the geometric parameters of the clear water tank. For a cross-sectional area of... The current water level in the clear water pool is The preset upper limit of the water level is The preset lower limit of the water level is At that time, the water level in the clear water tank can be adjusted upwards by a margin of [missing information]. The water level in the clear water tank can be lowered by a margin of [missing information]. Correspondingly, the volume of the clear water tank can be adjusted upwards to... The volume of the clear water tank can be adjusted downwards. When the system includes multiple clear water tanks, the regulating volume of each clear water tank can be summed to obtain the total regulating volume; The power prediction and response module characterizes the regulating volume of the clear water tank as a virtual energy storage capacity formed by the variable frequency pump and the clear water tank. Specifically, when the photovoltaic power is high, the system increases the speed of the variable frequency pump to convert electrical energy into a rise in the water level of the clear water tank; when the photovoltaic power decreases or the electricity price increases, the system decreases the speed of the variable frequency pump and uses the already raised water level in the clear water tank to maintain water supply, reducing the need for immediate power extraction. In one example, if the cross-sectional area of ​​the clear water tank is 1000m² 2 If the current water level is 0.8m below the preset upper limit, then the adjustable volume of the clear water tank can be increased to 800m³. 3 If the water plant's unit water supply power consumption is 0.30 kWh / m³ 3 Therefore, the transferable electricity corresponding to the adjustable volume of the clear water tank is approximately 240 kWh. This data is only an example to illustrate the calculation method of virtual energy storage capacity. The actual virtual energy storage capacity can be determined based on the size of the clear water tank, the current water level, the efficiency of the water pump, and the unit water supply power consumption. The power prediction response module limits the increase in pump speed corresponding to the first setpoint and the decrease in pump speed corresponding to the second setpoint based on the virtual energy storage capacity. In other words, if the adjustable volume of the clear water tank is small, the increase in pump speed is limited to prevent the water level from exceeding the preset upper limit; if the adjustable volume of the clear water tank is small, the decrease in pump speed is limited to prevent the water level from falling below the preset lower limit. The power prediction response module also performs speed change rate constraint processing on the time-sharing variable frequency pump speed setting sequence. Let the rated speed of the variable frequency pump be... The control cycle is The change in the variable frequency pump speed between adjacent control times shall not exceed This constraint ensures that the variable frequency pump speed regulation rate does not exceed ±2% of the rated speed per second, thereby suppressing water hammer and sudden changes in water pressure. In Example 1, the power prediction response module collects the real-time power of the microgrid's grid connection point with a collection period of 200ms. The grid connection point power setpoint can be determined by the day-ahead dispatch plan or the power output of the grid connection point from the water-energy coupling dispatch module. The power prediction response module compares the collected real-time power of the microgrid's grid connection point with the grid connection point power setpoint to generate the grid connection point power deviation. Specifically, the power deviation at the grid connection point can be expressed as: ,in, Indicates the first Real-time power of the microgrid grid-connected point at each data acquisition time. This indicates the grid connection point power setting value corresponding to the time of data acquisition. Indicates the power deviation at the grid connection point; The power prediction response module separates the grid connection point power deviation using a low-pass filter to obtain the low-frequency power component. The difference between the grid connection point power deviation and the low-frequency power component is then used as the high-frequency power component. Specifically, a first-order low-pass filter can be used, and its expression is as follows: ,in, For low-frequency power components, The filter coefficients range from 0 to 1; the high-frequency power components are... ; The power prediction response module generates or corrects the power setpoint of the energy storage converter based on high-frequency power components, enabling the energy storage converter to preferentially absorb or compensate for rapid power changes caused by photovoltaic fluctuations. The power prediction response module also generates a variable frequency pump speed correction based on low-frequency power components, enabling the variable frequency pump to respond to adjustment needs caused by changes in water load or slow power deviations. The power prediction response module superimposes the variable frequency pump speed correction amount onto the candidate variable frequency pump speed command sequence to generate the final variable frequency pump speed command. The edge controller writes the final variable frequency pump speed command to the variable frequency pump driver through OPC UA, and writes the energy storage converter power setpoint to the energy storage converter through OPC UA or Modbus TCP. The energy storage converter is used to respond to high frequency power components, and the variable frequency pump is used to respond to low frequency power components, realizing closed-loop control of the power at the microgrid grid connection point.

[0034] In this embodiment, a smart water management cloud-edge-device collaborative control method based on an industrial microgrid is used to implement the smart water management cloud-edge-device collaborative control system based on an industrial microgrid as described in any one of claims 1-9, comprising: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller receives the day-ahead scheduling plan, photovoltaic power forecast curve, load forecast curve and electricity price reference signal, and collects the real-time power of the microgrid grid connection point, water level of the clear water tank, pipeline pressure, water quality indicators and water equipment operation status; The edge controller establishes a water-energy coupled scheduling model based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, water level of clear water tank, pipeline pressure, water quality indicators and water equipment operation status, and generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the water-energy coupled scheduling model. The edge controller generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. The equipment controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The edge controller uploads the device operation data to the cloud platform. The cloud platform updates the prediction model based on the device operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle, so as to realize water-energy coordinated scheduling and cloud-edge-device closed-loop coordinated control between the water supply system and the industrial microgrid.

[0035] Example 2:

[0036] To verify the feasibility of this invention in practice, it was applied to a smart water management station equipped with an industrial microgrid. The station includes a water intake and lifting unit, a clear water tank, a pressurized water supply unit, a power distribution room, a photovoltaic power generation unit, and an energy storage unit. The station is equipped with multiple variable frequency water pumps, including a 160kW rated water supply pump, a 90kW rated circulation pump, and dosing pumps and aeration equipment for auxiliary process control. The effective volume of the clear water tank within the station is approximately 5000m³. 3 The normal operating water level range is set from 3.2m to 5.8m, with a target pressure control at the end of the pipeline network of no less than 0.14MPa, and the normal water supply pressure operating range maintained between 0.22MPa and 0.36MPa. A photovoltaic power generation device with a rated installed capacity of approximately 1.2MW is configured on the microgrid side, and the energy storage system has a rated power of 500kW and a rated capacity of 1000kWh. The microgrid is connected to the external power grid through a grid connection point. In this scenario, there is a significant mismatch between water load and photovoltaic output. When photovoltaic output is high, the traditional water pump control method still operates at a fixed water level or pressure, unable to actively increase the adjustable load, resulting in some photovoltaic power not being fully absorbed. Conversely, when electricity prices are high or photovoltaic output drops rapidly, the water pumps may still maintain a high operating power, leading to increased grid-connected power consumption and higher operating costs. Because the water control system and the microgrid energy management system operate separately, the original system can usually only focus on water level, pressure or power data separately. It is difficult to integrate the water level margin of the clear water tank, the safety of the pipeline pressure, the prediction of photovoltaic power and the change of electricity price into the control decision. This can easily lead to problems such as frequent start-up and shutdown of water pumps, large pressure fluctuations, delayed energy storage response and frequent manual intervention.

[0037] In this embodiment, the cloud platform is deployed on the station management center server or cloud server, and the edge controller is deployed in the station's power distribution room. Variable frequency water pumps, water quality analyzers, smart meters, photovoltaic inverters, and energy storage converters are connected to the edge controller as equipment. The cloud platform obtains data such as solar irradiance, ambient temperature, weather type, and time-of-use electricity price through meteorological and electricity price data interfaces, and retrieves historical data such as historical photovoltaic output, historical water supply flow, historical water pump speed, historical clear water tank water level, historical pipeline pressure, and historical grid-connected point power from the historical database. The model training engine aligns the above data according to timestamps and removes outliers that significantly exceed the equipment's rated range. For example, photovoltaic power values ​​exceeding the upper limit of photovoltaic installed capacity, water pump speed values ​​exceeding the upper limit of rated speed, and pressure values ​​with abnormal instantaneous fluctuations are marked as abnormal data. After data processing, the cloud platform generates a photovoltaic power prediction curve for the next 4 hours, with prediction points at 5-minute intervals, and generates a water load prediction curve and electricity price reference signal for the future scheduling cycle. The cloud platform further combines historical operating patterns and equipment operating constraints to generate day-ahead scheduling plans, and then sends the day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals to the edge controllers.

[0038] After receiving data from the cloud platform, the edge controller collects real-time data on the microgrid grid-connected power, clear water tank level, pipeline pressure, water quality indicators, and the operating status of water equipment. Water quality indicators include turbidity, residual chlorine, and pH value. The operating status of water equipment includes the start / stop status, current speed, fault status, and cumulative start / stop count of each variable frequency pump. The water-energy coupling scheduling module in the edge controller divides the scheduling cycle into multiple scheduling periods with a 15-minute interval. Within each scheduling period, the clear water tank level, pipeline pressure, water quality indicators, and pump start / stop count are used as the water supply safety boundary, while the photovoltaic predicted power, load predicted power, electricity price reference signal, energy storage converter power, and grid-connected power intake are used as the microgrid power balance boundary. The water-energy coupled scheduling module uses the start / stop status of the variable frequency water pump, the setpoint of the variable frequency water pump speed, the setpoint of the energy storage converter power, and the power taken from the grid connection point as scheduling variables. Under the conditions that the water level in the clear water tank is not lower than 3.2m and not higher than 5.8m, the pipeline pressure is not lower than 0.14MPa, the water quality indicators meet the operating requirements, and the number of pump shutdowns per hour for a single water pump does not exceed 6, the module generates a variable frequency water pump speed adjustment strategy and an energy storage converter power setpoint with the goal of minimizing the sum of electricity purchase cost, equipment loss cost, and renewable energy abandonment penalty cost. Through this process, the edge controller is no longer just passively executing fixed pressure or fixed liquid level control, but can jointly schedule water pumps and energy storage based on photovoltaic output, electricity price changes, and water supply security status.

[0039] The power prediction response module further utilizes the photovoltaic power prediction curve and electricity price reference signal to correct the variable frequency pump speed adjustment strategy. The photovoltaic power prediction curve shows that when the photovoltaic output gradually increases from approximately 320kW to approximately 980kW within a future operating range, and remains at a high output level for a considerable period, the power prediction response module determines that there is a photovoltaic power increase and peak state within this range. At this time, the edge controller reads the current water level in the clear water tank as 4.35m, with a preset upper limit of 5.8m, indicating that the clear water tank still has an upward adjustment margin of approximately 1.45m; based on the effective cross-sectional area of ​​the clear water tank being approximately 1000m²... 2 It is estimated that the adjustable volume of the clear water tank is approximately 1450 m³. 3 If the unit water supply power consumption is approximately 0.30 kWh / m³ 3 The adjustable capacity of this section corresponds to approximately 435 kWh of transferable electricity. Based on this, the power prediction response module increases the speed setpoint of some variable frequency pumps during the corresponding time period, raising the pump load from approximately 420 kW to between 560 kW and 620 kW. This increased water supply converts excess photovoltaic power into a rise in the water level of the clear water tank. Within this operating range, the water level in the clear water tank rose from 4.35 m to approximately 5.22 m, remaining below the preset upper limit. The pipeline pressure remained between 0.24 MPa and 0.31 MPa, and no risk of low-pressure water supply was observed.

[0040] When the photovoltaic power prediction curve shows that the photovoltaic output will decrease from approximately 900kW to approximately 260kW, or when the electricity price reference signal indicates that it is about to enter a high electricity price range, the power prediction response module determines that there is a photovoltaic power decline or an electricity price peak in this range. At this time, the edge controller reads that the current water level in the clear water tank is approximately 5.18m, which is 3.2m higher than the preset lower limit, indicating that the storage capacity can be released. Without lowering the pipeline pressure below 0.14MPa, the power prediction response module reduces the speed setpoint of some variable frequency pumps, causing the total operating power of the pumps to decrease from approximately 540kW to between approximately 360kW and 410kW, while utilizing the water level margin accumulated in the clear water tank during the previous operating range to maintain water supply demand. After control, the water level in the clear water tank gradually decreases to approximately 4.62m, the pipeline pressure is maintained between 0.23MPa and 0.29MPa, the terminal water supply pressure still meets the requirements, and the power drawn from the external grid during the high electricity price range decreases significantly. This process demonstrates that the variable frequency water pump and the clear water tank together form an adjustable load, making the clear water tank not only a water volume buffer facility, but also a regulating resource that can be used to transfer electrical loads during microgrid operation.

[0041] To prevent water hammer and sudden pressure changes in the pipeline caused by rapid changes in pump speed, the power prediction response module performs speed change rate constraint processing on the time-sharing variable frequency pump speed setting sequence. In this embodiment, the rated speed of the variable frequency pump is 1480 r / min, and the speed regulation rate is limited to no more than ±2% of the rated speed per second. That is, the speed change between adjacent control moments is limited according to the control cycle. In actual operation, when the pump speed increases from 980 r / min to 1180 r / min, it does not switch directly in a step manner, but gradually increases according to the limited ramp curve. The maximum instantaneous fluctuation of the pipeline pressure is reduced from about 0.065 MPa under the original control method to about 0.028 MPa, and the pressure fluctuation amplitude is reduced by about 56.9%. At the same time, since the pump start-stop control is partially replaced by continuous speed regulation, the number of times a single pump starts and stops per hour is reduced from 4 to 6 times under the original control method to 1 to 3 times, reducing the risk of pump start-stop shock and equipment wear.

[0042] In terms of real-time closed-loop control, the power prediction response module collects the real-time power of the microgrid's grid connection point with a 200ms acquisition period and compares it with the grid connection point power setpoint to obtain the grid connection point power deviation. For rapid fluctuations in photovoltaic power caused by cloud shadows, the power prediction response module separates the grid connection point power deviation into high-frequency and low-frequency power components using a low-pass filter. The high-frequency power component is mainly responded to by the energy storage converter. For example, when the grid connection point power experiences a rapid deviation of approximately 180kW in a short period, the energy storage converter quickly adjusts its charging and discharging power within its allowable power range, compressing the grid connection point power deviation to within approximately 60kW. The low-frequency power component is used to generate the variable frequency pump speed correction. For example, when there is a continuous deviation between the load forecast and the actual water supply demand, the edge controller makes a small correction to the pump speed on the original candidate command sequence, so that the pump operating power gradually approaches the target load level. By employing a division of labor control approach—where the energy storage converter responds to rapid fluctuations and the variable frequency pump responds to slow deviations—the maximum fluctuation range of the grid-connected power is reduced from approximately ±210kW under the original control method to approximately ±75kW, a reduction of about 64.3%, significantly improving the operational stability of the microgrid.

[0043] During continuous commissioning and verification, the traditional control method was compared with the cloud-edge-device collaborative control method of this embodiment. Under the traditional control method, the water pump mainly operates according to fixed water level limits and pressure targets, photovoltaic power generation is mainly consumed locally, and the energy storage system operates only according to a preset charge and discharge plan. Under this method, the photovoltaic power generation is approximately 10500 kWh, of which approximately 8200 kWh is consumed by local load, approximately 1200 kWh enters the energy storage system, and approximately 1100 kWh is not fully utilized due to insufficient local load or limited energy storage capacity, resulting in a photovoltaic consumption rate of approximately 89.5%. With the system of this embodiment, the cloud platform predicts changes in photovoltaic output in advance, and the edge controller actively adjusts the variable frequency water pump speed within the safe water supply range and coordinates the power of the energy storage converter. The photovoltaic power generation is approximately 10680 kWh, of which approximately 9820 kWh is consumed by local load and the energy storage system, reducing the underutilized photovoltaic power to approximately 380 kWh, and increasing the photovoltaic consumption rate to approximately 96.4%. Under similar water supply conditions, the photovoltaic absorption rate increases by about 6.9 percentage points, and the amount of curtailed photovoltaic power decreases by about 65.5%.

[0044] Regarding electricity purchase costs, under the traditional control method, the electricity purchased from the external grid is approximately 6840 kWh, of which approximately 2460 kWh is purchased during the high-price period. Based on peak-hour prices of RMB 1.18 / kWh, average-hour prices of RMB 0.72 / kWh, and off-peak prices of RMB 0.38 / kWh, the total electricity purchase cost is approximately RMB 4930. After adopting the system of this embodiment, because the photovoltaic system actively increases the pump speed and raises the water level in the clear water tank during high-output periods, and reduces the pump speed and utilizes water storage capacity to maintain water supply during high-price periods, the electricity purchased from the external grid is reduced to approximately 5870 kWh, and the electricity purchased during high-price periods is reduced to approximately 1580 kWh. The total electricity purchase cost is reduced to approximately RMB 3960, a decrease of approximately 19.7%. With a roughly equivalent water supply, the unit water supply electricity cost is reduced from approximately RMB 0.148 / m³. 3 Reduced to approximately 0.119 yuan / m 3 This demonstrates that this embodiment can reduce operating costs without sacrificing water supply security.

[0045] Regarding water supply safety and water quality stability, after adopting the system of this embodiment, the water level in the clear water tank was consistently maintained between 3.62m and 5.34m, without reaching the preset lower or upper water level limits; the pressure at the end of the pipeline network was maintained between 0.21MPa and 0.34MPa, exceeding the minimum pressure requirement of 0.14MPa; turbidity was maintained between 0.20NTU and 0.38NTU, residual chlorine between 0.34mg / L and 0.58mg / L, and pH value between 7.12 and 7.48, with all water quality indicators within the preset compliance range. Compared with traditional control methods, the standard deviation of pipeline pressure decreased from approximately 0.041MPa to approximately 0.019MPa, and pressure stability improved by approximately 53.7%; the average operating efficiency of the water pump increased from approximately 72.6% to approximately 78.4%, and the unit water supply power consumption decreased from approximately 0.315kWh / m³. 3 Reduced to approximately 0.286 kWh / m 3 It decreased by approximately 9.2%.

[0046] Regarding model update effectiveness, the cloud platform continuously receives equipment operation data uploaded by the edge controller from the device end, including pump speed, grid-connected power, actual photovoltaic output, clear water tank water level, and pipeline pressure. The model training engine writes the new equipment operation data into the historical database and updates the prediction model. After continuous operation, the average absolute error of photovoltaic power prediction decreased from approximately 9.8% initially to approximately 6.1%, and the average absolute error of load prediction decreased from approximately 8.6% to approximately 5.4%. With improved prediction accuracy, the variable frequency pump speed adjustment strategy generated by the edge controller is closer to actual operating requirements, reducing the deviation between the scheduling plan and on-site execution, and reducing the number of manual interventions from an average of 4 to 6 times per shift under traditional control methods to 1 to 2 times per shift. Therefore, this embodiment forms a closed loop through device-side feedback and cloud-based model updates, enabling the system to gradually adapt to operational changes under different weather conditions, water supply loads, and electricity prices.

[0047] Through the above implementation process, this embodiment solves the problems of existing water control systems and microgrid energy management being independent, photovoltaic fluctuations being difficult to be actively absorbed by water loads, high pump operating costs during periods of high electricity prices, and insufficient real-time performance of centralized cloud control. This embodiment utilizes a cloud platform for prediction and scheduling, edge controllers for real-time modeling and closed-loop control, and equipment-side rapid execution and data feedback. While ensuring the safety of clear water tank levels, pipeline pressure, and water quality indicators, it improves the capacity for renewable energy absorption, reduces grid-connected power fluctuations and electricity purchase costs, and enhances the stability and intelligent operation level of the water system.

[0048] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart water management cloud-edge-device collaborative control system based on an industrial microgrid, characterized in that, include: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller communicates with the cloud platform to receive day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves and electricity price reference signals, and collects real-time power at the microgrid grid connection point, water level in the clear water tank, pipeline pressure, water quality indicators and operating status of water equipment. The water-energy coupled scheduling module, located in the edge controller, generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operating status. The power prediction response module, located in the edge controller, generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. On the device side, it communicates with the edge controller, controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The model training engine, located on a cloud platform, updates the prediction model based on equipment operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, and electricity price reference signal for the next scheduling cycle.

2. The smart water management cloud-edge-device collaborative control system based on industrial microgrids according to claim 1, characterized in that, The cloud platform includes a meteorological-electricity price data interface, a historical database, a microgrid energy management module, a water process optimization module, and the model training engine. The meteorological-electricity price data interface is used to acquire meteorological data and electricity price data, and to write the meteorological data and electricity price data into the historical database; The historical database is used to store historical operating data, meteorological data, electricity price data, and equipment operating data generated by the device and uploaded by the edge controller; The model training engine is used to update the prediction model using historical operating data, meteorological data, electricity price data, and equipment operating data as training samples, and to generate photovoltaic power prediction curves and load prediction curves based on the updated prediction model. The microgrid energy management module is used to generate microgrid-side energy dispatch results based on photovoltaic power prediction curves, load prediction curves, and electricity price reference signals. The water process optimization module is used to generate start-up and shutdown arrangements for water equipment based on the energy dispatch results on the microgrid side and the operational needs of the water system. The cloud platform is used to generate day-ahead scheduling plans based on the energy dispatch results of the microgrid and the start-up and shutdown arrangements of water equipment. After receiving new equipment operation data, it updates the prediction model through the model training engine to generate the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle.

3. The smart water management cloud-edge-device collaborative control system based on industrial microgrids according to claim 1, characterized in that, The equipment includes a variable frequency water pump, a water quality analyzer, a smart meter, a photovoltaic inverter, and an energy storage converter; The edge controller is used to send the variable frequency water pump speed command to the variable frequency water pump, and the variable frequency water pump performs speed adjustment according to the variable frequency water pump speed command. The edge controller is also used to send the power setting value of the energy storage converter to the energy storage converter, and the energy storage converter performs charging control or discharging control according to the power setting value of the energy storage converter. The water quality analyzer is used to collect water quality indicators, the smart meter is used to collect the real-time power of the microgrid grid connection point, the photovoltaic inverter is used to feed back photovoltaic power generation operation data, and the variable frequency water pump and energy storage converter are used to feed back the corresponding equipment operation status. The device generates equipment operation data based on water quality indicators, real-time power of the microgrid grid connection point, photovoltaic power generation operation data, and equipment operation status, and uploads it to the cloud platform through the edge controller for the model training engine to update the prediction model.

4. The smart water management cloud-edge-device collaborative control system based on industrial microgrids according to claim 1, characterized in that, The cloud platform and the edge controller use the MQTT publish / subscribe protocol to transmit day-ahead scheduling plans, photovoltaic power forecast curves, load forecast curves and electricity price reference signals. The transmitted data is in Protobuf format and carries timestamps, confidence levels and validity periods. The edge controller communicates with the device via OPC UA or Modbus TCP to read and write device parameters. When the system includes multiple edge controllers, the multiple edge controllers share distributed data through DDS, and the edge controller uploads status information to the cloud platform every 500ms. The instructions issued by the cloud platform carry a sequence number.

5. The smart water management cloud-edge-device collaborative control system based on an industrial microgrid according to claim 1, characterized in that, The water-energy coupling scheduling module is used to divide the scheduling cycle into multiple scheduling periods according to a preset scheduling time interval, and to acquire the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, clear water tank water level, pipeline pressure, water quality indicators and water equipment operation status in each scheduling period. The water-energy coupled scheduling module determines the water supply safety boundary based on the water level of the clear water tank, the pressure of the pipeline network, the water quality indicators and the operating status of the water equipment. It determines the microgrid power balance boundary based on the photovoltaic power prediction curve, the load prediction curve, the electricity price reference signal and the real-time power of the microgrid grid connection point. Based on the water supply safety boundary and the microgrid power balance boundary, a water-energy coupled scheduling model is established. The water-energy coupled scheduling model uses the start-stop status, speed setpoint, energy storage converter power setpoint, and grid connection point power of each variable frequency water pump in each scheduling period as scheduling variables, and the clear water tank level, pipeline pressure, water quality indicators, and microgrid grid connection point power meeting the corresponding boundary conditions in each scheduling period as constraints to generate variable frequency water pump speed regulation strategies and energy storage converter power setpoints that meet water supply safety requirements.

6. The smart water management cloud-edge-device collaborative control system based on an industrial microgrid according to claim 5, characterized in that, The water-energy coupled scheduling model is a mixed integer linear programming model, and the preset scheduling time interval is 15 minutes. The constraints include: the water level in the clear water tank is not lower than the preset lower limit and not higher than the preset upper limit; the pipeline pressure is not lower than the preset minimum pressure; the number of times each variable frequency water pump stops per unit time does not exceed the preset number of stops; and the water quality indicators are within the preset water quality compliance range. The optimization objective of the mixed-integer linear programming model is to minimize the sum of electricity purchase cost, equipment loss cost, and renewable energy abandonment penalty cost in each scheduling period, while satisfying the constraints. The water-energy coupled scheduling module obtains the start-stop status and speed setpoint of each variable frequency water pump in each scheduling period by solving a mixed integer linear programming model, as well as the power setpoint of the energy storage converter in each scheduling period. It then combines the start-stop status and speed setpoint to generate a variable frequency water pump speed regulation strategy, and generates the power setpoint of the energy storage converter in each scheduling period as the energy storage converter power setpoint.

7. The smart water management cloud-edge-device collaborative control system based on industrial microgrids according to claim 1, characterized in that, The power prediction response module is used to receive the photovoltaic power prediction curve, the electricity price reference signal and the variable frequency water pump speed adjustment strategy, and determine the photovoltaic power change status within a preset time period according to the time sequence of the photovoltaic power prediction curve. The photovoltaic power change states include photovoltaic power rising state, photovoltaic power falling state, and photovoltaic power peak state. The power prediction response module is also used to determine the electricity price peak state based on the electricity price reference signal. The power prediction response module corrects the variable frequency pump speed setpoint for the corresponding scheduling period in the variable frequency pump speed regulation strategy based on the photovoltaic power rising state, photovoltaic power falling state, photovoltaic power peak state, and electricity price peak state, and generates a time-segmented variable frequency pump speed setpoint sequence.

8. A smart water management cloud-edge-device collaborative control system based on an industrial microgrid according to claim 7, characterized in that, The preset duration for the future is 4 hours, and the time interval between adjacent prediction points in the photovoltaic power prediction curve is 5 minutes. During the scheduling period corresponding to the photovoltaic power rise state or photovoltaic power peak state, the power prediction response module determines whether the water level in the clear water tank is lower than the preset upper limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, the variable frequency pump speed setting value for the corresponding scheduling period is modified based on the variable frequency pump speed adjustment strategy to form the first speed setting value. During the scheduling period corresponding to the photovoltaic power decline state or the electricity price peak state, the power prediction response module determines whether the water level of the clear water tank is higher than the preset lower limit and whether the pipeline pressure is not lower than the preset minimum pressure. If so, the variable frequency pump speed setting value of the corresponding scheduling period is modified based on the variable frequency pump speed adjustment strategy to form a second speed setting value. The power prediction response module generates a time-segmented variable frequency pump speed setting sequence based on the first speed setting value and the second speed setting value.

9. A smart water management cloud-edge-device collaborative control system based on an industrial microgrid according to claim 8, characterized in that, The power prediction response module determines the adjustable water level margin and the adjustable water level margin of the clear water tank based on the water level of the clear water tank, the preset upper limit of the water level, and the preset lower limit of the water level, and determines the regulating volume of the clear water tank based on the adjustable water level margin and the adjustable water level margin of the clear water tank. The power prediction response module characterizes the regulating volume of the clear water tank as a virtual energy storage capacity formed by the variable frequency water pump and the clear water tank, and limits the speed increase corresponding to the first speed setting value and the speed decrease corresponding to the second speed setting value according to the virtual energy storage capacity. The power prediction response module is used to perform speed change rate constraint processing on the time-sharing variable frequency water pump speed setting sequence. The speed change rate of the variable frequency water pump between adjacent control times does not exceed ±2% of the rated speed per second, and the time-sharing variable frequency water pump speed setting sequence after speed change rate constraint processing is used as a candidate variable frequency water pump speed command sequence. The power prediction response module is also used to collect the real-time power of the microgrid grid connection point with a collection period of 200ms, and determine the power setpoint of the grid connection point according to the day-ahead scheduling plan. The real-time power at the microgrid grid connection point is compared with the grid connection point power setpoint to generate the grid connection point power deviation. The grid connection point power deviation is then separated into high-frequency power components and low-frequency power components by a low-pass filter. The power setting value of the energy storage converter is generated or corrected based on the high-frequency power component, and the speed correction value of the variable frequency water pump is generated based on the low-frequency power component. The variable frequency pump speed correction is superimposed on the candidate variable frequency pump speed command sequence to generate the variable frequency pump speed command. The energy storage converter responds to the high-frequency power component corresponding to the photovoltaic power fluctuation, and the variable frequency pump responds to the low-frequency power component corresponding to the load change.

10. A smart water management cloud-edge-device collaborative control method based on an industrial microgrid, used to implement the smart water management cloud-edge-device collaborative control system based on an industrial microgrid as described in any one of claims 1-9, characterized in that, include: The cloud platform acquires historical operating data, meteorological data, and electricity price data, and generates day-ahead dispatch plans, photovoltaic power forecast curves, load forecast curves, and electricity price reference signals based on these data. The edge controller receives the day-ahead scheduling plan, photovoltaic power forecast curve, load forecast curve and electricity price reference signal, and collects the real-time power of the microgrid grid connection point, water level of the clear water tank, pipeline pressure, water quality indicators and water equipment operation status; The edge controller establishes a water-energy coupled scheduling model based on the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve, electricity price reference signal, real-time power of microgrid grid connection point, water level of clear water tank, pipeline pressure, water quality indicators and water equipment operation status, and generates variable frequency pump speed regulation strategy and energy storage converter power setpoint based on the water-energy coupled scheduling model. The edge controller generates variable frequency pump speed commands based on the photovoltaic power prediction curve and the variable frequency pump speed adjustment strategy. The equipment controls the operation of the variable frequency water pump according to the variable frequency water pump speed command, and controls the operation of the energy storage converter according to the power setting value of the energy storage converter, generating equipment operation data; The edge controller uploads the device operation data to the cloud platform. The cloud platform updates the prediction model based on the device operation data and generates the day-ahead scheduling plan, photovoltaic power prediction curve, load prediction curve and electricity price reference signal for the next scheduling cycle.

Images