Energy management method, apparatus, device, and medium for water energy coupling control

Abstract

This application discloses an energy management method, device, equipment, and medium for water-energy coupling control. The method improves the global allocation efficiency of water and energy resources, enhances the capacity for new energy absorption, ensures the ecological security of the watershed, and improves the robustness of energy management in complex scenarios by constructing a digital twin and quantifying the water-energy coupling relationship, deploying a hierarchical collaborative control architecture, adopting ecological and benefit multi-objective game optimization, performing water-energy bidirectional conversion buffer control, and introducing reinforcement learning to achieve self-learning iteration.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of water energy management technology, and in particular to an energy management method, apparatus, equipment, and medium for water energy coupling control. Background Technology

[0002] Currently, the industry generally adopts a model that separates water resource scheduling and energy regulation for the coupled management of hydropower and energy systems. Specifically, hydropower scheduling focuses on maximizing runoff utilization efficiency and power generation economic benefits, with its core objective being to optimize reservoir operation and improve hydropower conversion efficiency; while energy network scheduling focuses on the power supply stability of the power system and the capacity for renewable energy absorption. These two scheduling systems lack an effective coordination mechanism at the planning, operation, and control levels, leading to increasingly prominent problems such as low hydropower conversion efficiency and limited renewable energy absorption capacity.

[0003] Furthermore, hydrological runoff, renewable energy output, and electricity load all exhibit significant dynamic changes. Coupled with constraints such as watershed ecological flow, the conflict between hydropower development and ecological protection is difficult to reconcile. When the energy system faces drastic fluctuations, existing management methods are clearly insufficient in terms of the flexibility of control response and system robustness, making it difficult to achieve the optimal global allocation of water and energy resources. Summary of the Invention

[0004] In view of this, embodiments of this application provide at least one energy management method, apparatus, device, and medium for hydropower coupling control.

[0005] The technical solution of this application embodiment is implemented as follows: On one hand, embodiments of this application provide an energy management method for hydropower coupling control, the method comprising: Acquire watershed hydrological, energy, and ecological data, construct a water-energy coupled sensing dataset, and establish a digital twin of watershed hydrology and energy coupling based on the dataset. Quantify the spatiotemporal coupling correlation between runoff, water level changes, energy output, and load demand, and generate a water-energy coupled dynamic response model. Based on the aforementioned hydropower coupling dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed, and a hierarchical control strategy including medium- and long-term scheduling plans and real-time adjustment strategies is formulated. Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the game between ecological protection objectives and energy efficiency objectives, and a hydropower coupling scheduling strategy is generated. Implement water energy bidirectional conversion buffer regulation, and control pumped storage units and hydropower units to carry out bidirectional conversion of electricity and water according to the dynamic changes of new energy output and grid load, so as to smooth energy system fluctuations. Historical scheduling data and real-time operational feedback are collected, and the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy are iteratively optimized through a reinforcement learning agent to form a closed-loop energy management system.

[0006] On the other hand, embodiments of this application provide an energy management device for water-powered coupling control, the method comprising: The processing module is used to acquire watershed hydrological data, energy data and ecological data, construct a water-energy coupled sensing dataset, and establish a digital twin of watershed hydrology and energy coupling based on the dataset, quantify the spatiotemporal coupling correlation between runoff, water level changes and energy output and load demand, and generate a water-energy coupled dynamic response model. Based on the aforementioned hydropower coupling dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed, and a hierarchical control strategy including medium- and long-term scheduling plans and real-time adjustment strategies is formulated. Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the game between ecological protection objectives and energy efficiency objectives, and a hydropower coupling scheduling strategy is generated. The execution module is used to perform the two-way conversion buffer regulation of hydropower. Based on the dynamic changes in the output of new energy sources and the load of the power grid, it controls the pumped storage units and hydropower units to carry out the two-way conversion of electrical energy and water, so as to smooth the fluctuations of the energy system. Historical scheduling data and real-time operational feedback are collected, and the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy are iteratively optimized through a reinforcement learning agent to form a closed-loop energy management system.

[0007] In another aspect, embodiments of this application provide a computer device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements some or all of the steps in the energy management method for hydropower coupling control described above.

[0008] In another aspect, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements some or all of the steps in the energy management method for hydropower coupling control described above.

[0009] In another aspect, embodiments of this application provide a computer program including computer-readable code. When the computer-readable code is run in a computer device, the processor in the computer device executes some or all of the steps in the energy management method for hydropower coupling control described above.

[0010] In another aspect, embodiments of this application provide a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, it implements some or all of the steps in the energy management method for hydropower coupling control described above.

[0011] This application embodiment improves the global allocation efficiency of water and energy resources, enhances the capacity for new energy absorption, ensures the ecological security of the basin, and improves the robustness of energy management in complex scenarios by constructing a digital twin and quantifying the coupling relationship between water and energy, deploying a hierarchical collaborative control architecture, adopting ecological and benefit multi-objective game optimization, performing two-way conversion buffer control of water and energy, and introducing reinforcement learning to achieve self-learning iteration.

[0012] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0014] Figure 1 A schematic diagram illustrating the implementation process of an energy management method for hydroelectric coupling control provided in an embodiment of this application; Figure 2 A logic diagram of an energy management method for hydroelectric coupling control provided in an embodiment of this application; Figure 3 A schematic diagram of the composition structure of an energy management device for water-energy coupling control provided in an embodiment of this application; Figure 4 This is a schematic diagram of the hardware entity of a computer device provided in an embodiment of this application. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0017] The terms “first / second / third” are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that “first / second / third” may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used herein is for descriptive purposes only and is not intended to limit the scope of this application.

[0019] This application provides an energy management method for hydropower coupling control, which can be executed by a processor of a computer device. The computer device can refer to a server, laptop, tablet, desktop computer, smart TV, set-top box, mobile device (e.g., mobile phone, portable video player, personal digital assistant, dedicated messaging device, portable gaming device), or other similar computer equipment. Figure 1 This application provides a schematic diagram of the implementation process of an energy management method for hydropower coupling control, as illustrated in the embodiments of this application. Figure 1 As shown, the method includes: Step 101: Obtain watershed hydrological data, energy data, and ecological data; construct a water-energy coupled sensing dataset; and establish a digital twin of watershed hydrology and energy coupling based on the dataset. Quantify the spatiotemporal coupling correlation between runoff, water level changes, energy output, and load demand, and generate a water-energy coupled dynamic response model.

[0020] In this embodiment, watershed hydrological data reflects the state and processes of the water cycle within the watershed, such as river flow, water level, rainfall, and runoff forecast information; energy data reflects the operational status of the energy system, such as the output power of hydropower stations, photovoltaic power stations, wind farms, grid load, and the state of charge of energy storage systems; ecological data reflects the ecological environment status of the watershed, such as ecological flow and water quality indicators. The hydro-energy coupling sensing dataset is a unified data set formed by fusing various types of collected data to describe the relationship between hydropower and electrical energy. The digital twin is a virtual simulation model constructed by the system that can map the operational status of the physical world in real time. The spatiotemporal coupling correlation degree is a numerical indicator quantified by the system, describing the degree of mutual influence between hydrological parameters and energy parameters at different spatial locations and time points. The hydro-energy coupling dynamic response model is a mathematical model established by the system that can simulate and predict the dynamic behavior of the hydropower system under different input conditions.

[0021] The system first acquires hydrological, energy, and ecological data from hydrological monitoring stations, energy plants, and power grid monitoring nodes deployed within the watershed. This multi-source data is then cleaned, synchronized in time, and fused to construct a unified hydro-energy coupling sensing dataset. Based on this dataset, the system utilizes physical models of watershed topography, turbine characteristics, and power grid flow, along with data-driven models, to establish a digital twin of the watershed's hydrological and energy coupling, capable of real-time mapping the physical system's state. On top of this digital twin, the system analyzes the interrelationships between runoff, water level changes, energy output, and load demand, quantifying their spatiotemporal coupling correlation, and ultimately generating a dynamic response model for hydro-energy coupling that can accurately predict the marginal benefits and constraints of hydropower conversion.

[0022] Step 102: Based on the hydropower coupled dynamic response model, deploy a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level, and formulate a hierarchical control strategy that includes medium- and long-term scheduling plans and real-time adjustment strategies.

[0023] In this embodiment, the basin-level is the highest level in the system architecture, responsible for formulating macro-level strategies covering the entire basin over a long timescale; the power plant-level is the intermediate level, responsible for coordinating the real-time operation of one or more power plants; and the equipment-level is the lowest level, responsible for directly controlling the instantaneous actions of specific equipment. The hierarchical control strategy is a complete scheme formulated by the system according to the responsibilities of different levels, containing control instructions at different timescales. The medium- and long-term scheduling plan is a strategy formulated at the basin level to plan reservoir operation and power generation quotas for the next few days or even months. The real-time adjustment strategy is a strategy at the power plant-level and equipment-level to adjust equipment operating parameters in real time based on current operating conditions.

[0024] Based on the spatiotemporal coupling relationship revealed by the generated hydropower coupled dynamic response model, the system deploys a three-tiered collaborative decision-making and control architecture encompassing the basin level, power plant level, and equipment level. Under this architecture, the system first establishes a medium- to long-term scheduling plan, including cascade reservoir scheduling and power generation quotas, based on hydrological forecasts and energy consumption targets, through a basin-level overall control layer. Then, the power plant level coordination layer receives basin-level instructions and, combined with real-time data, dynamically allocates the output of each unit and the charging and discharging power of energy storage, forming a real-time adjustment strategy to address intraday fluctuations. Finally, the equipment level execution layer distributes the power plant level instructions to specific equipment, executing millisecond-level fine-grained adjustments, thus forming a complete hierarchical control strategy covering different time scales.

[0025] Step 103: Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the ecological protection objective and the energy benefit objective to generate a hydropower coupling scheduling strategy.

[0026] In this embodiment, the ecological flow hard constraint is the minimum flow requirement that the system must strictly meet during the scheduling process to ensure the health of the watershed ecosystem. The multi-objective game theory algorithm is a mathematical method used by the system to handle multiple conflicting optimization objectives; it finds an equilibrium point by simulating the game process between multiple decision-makers. The ecological protection objective is the watershed ecological indicator that the system aims to improve, such as the ecological flow compliance rate. The energy efficiency objective is the economic or operational indicator that the system aims to improve, such as the comprehensive utilization efficiency of hydropower and the renewable energy absorption rate. The hydropower coupling scheduling strategy is the final set of specific operational instructions generated by the system to guide each execution node in hydropower conversion.

[0027] The system treats the hard constraint of ecological flow as an insurmountable bottom line, while simultaneously optimizing for both watershed ecological protection and energy efficiency goals. Employing a multi-objective game theory algorithm, such as a non-dominated sorting genetic algorithm, the system iteratively optimizes and dynamically adjusts the weights of different objectives within a solution space encompassing multiple dimensions, including ecological flow, power generation efficiency, and power supply reliability, thereby generating a Pareto-optimal solution set. Combining grid dispatch requirements with ecological assessment results, the system selects the optimal solution from this set that achieves a game-theoretic equilibrium between ecological protection and energy efficiency, thus generating the final hydropower coupled dispatch strategy.

[0028] Step 104: Perform bidirectional hydropower conversion buffer regulation. Based on the dynamic changes in new energy output and grid load, control the pumped storage units and hydropower units to perform bidirectional conversion of electrical energy and water in order to smooth out energy system fluctuations.

[0029] In this embodiment, the hydropower bidirectional conversion buffer regulation is a dynamic adjustment and control method that utilizes reservoirs and energy storage devices to realize the conversion of electrical energy into the potential energy of water for storage and the conversion of water's potential energy into electrical energy for release. New energy output refers to the power generated by renewable energy power generation equipment such as photovoltaics and wind power. Grid load refers to the total power consumed by all electrical equipment in the power system. A pumped storage unit is a device that can pump water from a low-lying area to a high-lying area to store electrical energy and release the water to generate electricity when needed. A hydroelectric generator unit is a device that uses the impact of water flow on a turbine to drive a generator to generate electricity.

[0030] The system implements bidirectional hydropower conversion and buffering regulation, monitoring the dynamic changes in renewable energy output and grid load in real time. When the system detects excess renewable energy output and a low grid load, it controls pumped storage units to start, converting excess electrical energy into the potential energy of water and storing it in upstream reservoirs. When the system detects insufficient renewable energy output or a peak grid load, it controls hydropower units to generate more electricity, rapidly converting the stored water energy into electrical energy to supplement the power supply gap. Through this bidirectional conversion mechanism, the system can effectively utilize the regulation capabilities of reservoirs and energy storage equipment to smooth out fluctuations in renewable energy output and ensure power balance in the grid.

[0031] Step 105: Collect historical scheduling data and real-time operation feedback, and iteratively optimize the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy through a reinforcement learning agent to form a closed-loop energy management system.

[0032] In this embodiment, historical scheduling data refers to all operation instructions, status parameters, and effect feedback data recorded by the system during its past operation. Real-time operation feedback refers to the current operating status information collected by the system from field equipment, such as equipment power, water level, and valve opening. The reinforcement learning agent is a decision-making model built based on reinforcement learning algorithms within the system, which can learn the optimal strategy through trial and error with the environment.

[0033] The system continuously collects historical scheduling data and real-time operational feedback, using this data as training samples. These samples are then used to train an internal reinforcement learning agent, enabling it to learn optimal control strategies under different hydrological and energy scenarios. Based on the agent's learning results, the system automatically iteratively optimizes the parameters of the hydro-energy coupling dynamic response model and the generation logic of the hydro-energy coupling scheduling strategy. Through this continuous learning and self-updating mechanism, the system achieves a complete closed loop from perception, modeling, decision-making, execution to optimization, enabling energy management strategies to continuously adapt to environmental changes and achieve self-evolution and self-adaptation.

[0034] This application embodiment improves the global allocation efficiency of water and energy resources, enhances the capacity for new energy absorption, ensures the ecological security of the basin, and improves the robustness of energy management in complex scenarios by constructing a digital twin and quantifying the coupling relationship between water and energy, deploying a hierarchical collaborative control architecture, adopting ecological and benefit multi-objective game optimization, performing two-way conversion buffer control of water and energy, and introducing reinforcement learning to achieve self-learning iteration.

[0035] Optionally, step 101 includes: Step 1011: Deploy hydrological monitoring stations to collect flow data, water level data, and rainfall data; deploy hydropower unit status sensors to collect unit operating parameters; deploy wind and solar power station output acquisition terminals to collect photovoltaic output data and wind power output data; deploy energy storage battery management system to collect energy storage charging and discharging information; and deploy power grid load monitoring nodes to collect load data and electricity price data.

[0036] In this embodiment, a hydrological monitoring station is a fixed facility used to monitor and record hydrological elements within a watershed. Flow data is numerical information describing the volume of water passing through a cross-section of a river per unit time. Water level data is numerical information describing the height of the surface of a river, lake, or other water body relative to a reference surface. Rainfall data is numerical information describing the depth of rainwater falling to the ground within a certain time period. A hydropower unit status sensor is an electronic device installed on a hydropower generator unit to measure its operating status parameters. Unit operating parameters are specific values ​​describing the working state of the hydropower unit, such as speed, vibration, and temperature. A wind and solar power station output acquisition terminal is a device installed in a photovoltaic power station or wind farm to collect its power generation data in real time. Photovoltaic output data is numerical information describing the actual power generation of a photovoltaic power station. Wind power output data is numerical information describing the actual power generation of a wind farm. An energy storage battery management system is an electronic system used to monitor and manage the charging and discharging process, status, and safety of energy storage battery packs. Energy storage charging and discharging information is data describing parameters such as charging and discharging power, energy capacity, and status of the energy storage system. A power grid load monitoring node is a device installed in the power grid to measure and record electrical load in real time. Load data describes the total power consumption of the power grid at a given moment. Electricity price data describes the electricity trading prices over different time periods.

[0037] The system first deploys hydrological monitoring stations specifically for collecting flow, water level, and rainfall data within the watershed. Simultaneously, it deploys hydropower unit status sensors to collect operating parameters such as speed, vibration, and temperature. Next, it deploys power generation data acquisition terminals for wind and solar power plants to collect photovoltaic power output data and wind power output data. Furthermore, it deploys an energy storage battery management system to collect energy storage charging and discharging information. Finally, it deploys grid load monitoring nodes to collect grid load and electricity price data. Through these deployments, the system establishes a comprehensive perception capability for multi-source data, including hydrological, energy, and ecological data.

[0038] Step 1012: Collect the flow rate data, water level data, rainfall data, unit operating parameters, photovoltaic output data, wind power output data, energy storage charging and discharging information, load data, and electricity price data according to millisecond-level, minute-level, and hour-level time scales, respectively.

[0039] In this embodiment, the system collects data based on different time scales, taking into account varying data characteristics. Specifically, the system collects data with extremely high real-time requirements at the millisecond level, such as instantaneous fluctuations in energy storage charging and discharging information and grid load data. The system collects data requiring moderate real-time performance at the minute level, such as short-term changes in hydropower unit operating parameters and wind and solar power output data. The system collects relatively slow-changing data at the hourly level, such as flow data, water level data, rainfall data, and electricity price data. Through this multi-time-scale acquisition strategy, the system can accurately capture the characteristics of different dynamic processes in the hydropower coupling system.

[0040] Step 1013: The collected data is synchronized in time, imputed for missing values, and fused from multiple sources through an edge computing gateway to obtain a water energy coupled sensing dataset.

[0041] In this embodiment, the edge computing gateway is a computing device deployed near the data source for localized processing of the collected data. Time synchronization is the process of aligning data from different devices and sensors on a timeline. Missing value imputation is a technique for filling data gaps caused by faults or interference during data acquisition. Multi-source fusion is the process of integrating data from different sources, formats, and time scales to form a unified dataset. The hydro-energy coupled sensing dataset is a comprehensive data set containing multi-dimensional information on hydrology, energy, and ecology, after cleaning, synchronization, and fusion.

[0042] The system inputs all collected data into an edge computing gateway. This gateway first synchronizes the data from different devices to ensure all data is on the same time base. The system then imputes missing values ​​in the collected data, for example, by using linear interpolation or averaging historical data. Finally, the system performs multi-source fusion, integrating the synchronized and completed flow data, water level data, power output data, and other types of data into a structured, unified dataset. This final dataset is the hydropower coupled sensing dataset, providing a foundation for subsequent modeling and decision-making.

[0043] Step 1014: Based on watershed topographic data, hydrogeological data, turbine output characteristic model, wind and solar power prediction data, energy storage charging and discharging information, and power grid flow data, establish a spatiotemporal coupled correlation model for runoff flow, water level changes, power generation output, and load demand. Generate watershed-level, station-level, and equipment-level digital twins based on the spatiotemporal coupled correlation model to map the real-time operating status of the physical system.

[0044] In this embodiment, watershed topographic data refers to digital information describing the geographical features of the watershed, such as mountains, rivers, and canyons. Hydrogeological data describes the geological and hydrological conditions of the watershed, such as groundwater distribution and aquifer characteristics. The turbine output characteristic model is a mathematical model describing the output power of a turbine under different head and flow conditions. Wind and solar power prediction data are predicted values ​​of the power generation of photovoltaic power plants and wind farms over a future period. Power grid flow data is numerical information describing the voltage, current, power distribution, and flow conditions at each node in the power grid. The spatiotemporal coupling correlation model is a mathematical model describing the mutual influence and constraints between runoff, water level, power generation, and load in the temporal and spatial dimensions. A watershed-level digital twin is a high-fidelity mapping of the physical watershed system in digital space, capable of simulating the hydrological and energy dynamics of the entire watershed. A station-level digital twin is an accurate mapping of a single physical station, such as a hydropower station, wind farm, or photovoltaic power station, in digital space. Equipment-level digital twins are virtual copies of individual physical devices such as water turbines and energy storage converters in the digital space.

[0045] The system utilizes a hydropower coupled sensing dataset, combined with watershed topographic data, hydrogeological data, turbine output characteristic models, wind and solar power prediction data, energy storage charging and discharging information, and power grid flow data, to jointly establish a spatiotemporal coupled correlation model between runoff, water level changes, power generation output, and load demand. This model quantifies the complex dependencies between various elements. Based on this spatiotemporal coupled correlation model, the system further generates three levels of digital twins: a watershed-level digital twin to simulate the macroscopic state of hydropower coupling across the entire watershed; a station-level digital twin to simulate the operation of a single hydropower station or renewable energy station; and an equipment-level digital twin to simulate the real-time response of key equipment such as turbines and energy storage converters. These three levels of digital twins work together to map the operating status of the physical system in real time, providing a high-precision virtual simulation environment for subsequent decision-making and control.

[0046] This application embodiment deploys multi-source sensors and acquisition terminals, and performs multi-timescale data acquisition and edge computing fusion to construct a high-precision, high-real-time water energy coupling sensing dataset. Based on this dataset, a spatiotemporal coupling correlation model is established and a three-layer digital twin is generated, thereby enabling full-dimensional, full-time-space accurate sensing and digital mapping of the water energy coupling physical system.

[0047] Optionally, step 1013 includes: Step 10311: Configure the data cache queue and time synchronization reference clock of the edge computing gateway.

[0048] In this embodiment, the edge computing gateway is a network device located at the data acquisition front end, possessing computing and storage capabilities. It connects sensors to the upper-layer system to achieve data preprocessing and forwarding. The data cache queue is a first-in, first-out (FIFO) memory area within the edge computing gateway used for temporary storage of acquired data. The time synchronization reference clock is a high-precision, unified time source commonly referenced by all devices within the system, such as a clock obtained through the Global Positioning System or Network Time Protocol (NAT).

[0049] The system first initializes and configures the edge computing gateway. It creates one or more data buffer queues within the gateway to temporarily store raw data collected in real-time from various sensors and monitoring terminals. The system sets a time synchronization reference clock for the gateway, which is calibrated by receiving an external standard time signal to ensure that its time deviation from other devices within the system is within acceptable limits. This step lays the foundation for subsequent processing of all data within a unified time frame.

[0050] Step 10312: The flow rate data, photovoltaic power output data, wind power output data and energy storage charging and discharging information collected on a millisecond time scale, the unit operating parameters, load data and electricity price data collected on a minute time scale, and the water level data and rainfall data collected on an hourly time scale are respectively stored in the data cache queue.

[0051] In this embodiment, flow rate data reflects the volume of water flowing in a river or pipeline per unit time, such as cubic meters per second. Photovoltaic output data is the real-time active power output of a photovoltaic power station. Wind power output data is the real-time active power output of a wind farm. Energy storage charging and discharging information includes data such as the charging power, discharging power, and state of charge of the energy storage system. Generator unit operating parameters include the speed, guide vane opening, vibration, and other status data of the hydroelectric generator unit. Load data is the total active power demand of the power grid or the area where electricity is consumed. Electricity price data is real-time or time-of-use electricity price information published by the electricity market or power grid. Water level data is the elevation of the surface of a water body such as a reservoir or river relative to a certain reference surface. Rainfall data is the amount of rainfall per unit time.

[0052] The system collects data at different time scales based on the rate of change of data types and business needs. Rapidly changing data such as flow rate, photovoltaic output, wind power output, and energy storage charging / discharging information are collected at millisecond-level frequencies and immediately stored in the data cache queue of the edge computing gateway. Rapidly changing unit operating parameters, load data, and electricity price data are collected at minute-level frequencies and stored in the same data cache queue. Slowly changing data such as water level and rainfall are collected at hourly-level frequencies and also stored in this queue. All data is stored in the same queue in chronological order of collection time.

[0053] Step 10313: Time-align all data in the data cache queue according to the unified timestamp of the time synchronization reference clock.

[0054] In this embodiment, a timestamp is precise time information appended to the data, identifying the moment the data was collected. Time alignment is the process of correcting the timestamps of data from different sources and at different collection times to make them comparable and correlated on a unified timeline.

[0055] After retrieving all data sequentially from the data cache queue, the system adds or corrects a timestamp for each data entry. The system's configured time synchronization reference clock serves as the sole reference, uniformly converting the original time information carried by the data, which may have originated from different devices' local clocks, into timestamps consistent with this reference clock. Through this operation, the system ensures that all data collected at the millisecond, minute, and hourly levels has timestamps based on the same precise time source, thus achieving accurate data alignment in the time dimension.

[0056] Step 10314: Perform outlier detection on the time-aligned data and remove data points that exceed the preset reasonable range.

[0057] In this embodiment, outlier detection is the process of identifying values ​​in a dataset that are significantly different from the majority of data and do not conform to expected patterns or physical laws. The preset reasonable range is an acceptable upper and lower limit interval for each data type, pre-defined by the system based on device range, historical data statistics, or physical constraints.

[0058] After time alignment is completed, the system performs outlier detection on each data point. For different types of monitoring data, such as flow rate, power output, and water level, the system calls upon their corresponding preset reasonable ranges. The system compares the value of each data point with this preset range. If a data point's value exceeds this range—for example, a negative water level value or photovoltaic power output exceeding the rated power at night—the system determines that data point as an outlier and removes it from the dataset.

[0059] Step 10315: The missing data generated after the removal is filled in using a linear interpolation method to generate complete time series data.

[0060] In this embodiment, missing data refers to numerical points in the time series that are absent due to outlier removal, sensor malfunction, or other reasons during data acquisition or processing. Linear interpolation is a simple completion method that uses two adjacent valid data points before and after a missing data point to estimate the value of the missing point by calculating the equation of a straight line between them. Time series data is a series of data points arranged in chronological order, with each data point corresponding to a specific timestamp.

[0061] After removing outlier data points, gaps appeared in the original time series. To maintain the continuity and integrity of the time series, the system initiated a data completion process. The system located each missing data point and found the two nearest valid data points before and after it. The system used a linear interpolation method, assuming that the change between these two valid data points was linear, and then estimated the value of the missing data point at the corresponding time point by calculating the slope of the line between them. The system filled in the missing value in the gap, thus generating a complete time series without any missing data.

[0062] Step 10316: The hydrological data, energy data and ecological data in the time series data obtained after time synchronization and missing value completion are fused from multiple sources according to the preset data fusion rules to generate a water energy coupled sensing dataset containing hydrological feature fields, energy feature fields and ecological feature fields.

[0063] In this embodiment, hydrological data refers to monitoring data related to the water cycle, such as flow rate, water level, and rainfall. Energy data refers to data related to energy production, transmission, and consumption, such as photovoltaic output, wind power output, load, and electricity price. Ecological data refers to data reflecting the ecological environment status of the watershed, such as ecological flow and water body indicators. Predefined data fusion rules are a set of predefined logic or algorithms used to combine, associate, and format data from different data types that are temporally and spatially related. Hydrological feature fields are specific columns in the fused dataset used to store hydrological data or its feature values. Energy feature fields are specific columns in the fused dataset used to store energy data or its feature values. Ecological feature fields are specific columns in the fused dataset used to store ecological data or its feature values. The hydro-energy coupling sensing dataset is the final generated structured data set used to describe the coupling state and characteristics of the hydro-energy and energy systems.

[0064] The system categorizes the complete time-series data, obtained after time synchronization and missing value completion, according to its attributes. The system identifies data points belonging to hydrological, energy, and ecological categories. Based on preset data fusion rules, such as using the same timestamp, the system associates and combines hydrological, energy, and ecological data from the same watershed or station. The system organizes these associated data into a new structured dataset with clearly defined fields, including hydrological feature fields for storing hydrological characteristics, energy feature fields for storing energy characteristics, and ecological feature fields for storing ecological characteristics. Ultimately, the system generates a multi-dimensional, multi-type information-coupled water and energy sensing dataset, providing a data foundation for subsequent model building and decision analysis.

[0065] This application embodiment achieves the effect of processing raw data from different sensors, different time scales, and possibly containing noise and missing data into a high-quality, time-consistent, attribute-complete, and structured hydropower coupling sensing dataset by configuring an edge computing gateway, multi-scale acquisition and time alignment, outlier removal and linear interpolation completion, and multi-source data fusion. This provides a reliable data foundation for subsequent accurate construction and optimization decision-making of hydropower coupling dynamic response models.

[0066] Optionally, step 102 includes: Step 1021: At the basin-level overall control layer, using hydrological forecast data, wind and solar power output prediction data and annual grid absorption targets within a preset time period as inputs, and basin ecological base flow, reservoir flood control capacity and grid safety constraints as limitations, formulate a medium- and long-term scheduling plan for cascade reservoirs and a general control strategy for hydropower conversion.

[0067] In this embodiment, the basin-level overall control layer is the decision-making level responsible for overall coordination and medium-to-long-term planning; the preset time period is the time length set by the system for prediction and planning, such as 7 days or one month; the hydrological forecast data is the system's prediction results of hydrological elements such as future runoff and water level in the basin; the wind and solar power output prediction data is the system's prediction results of future photovoltaic and wind power generation; the annual grid absorption target is the system's annual renewable energy consumption target obtained from the grid; the basin ecological base flow is the minimum flow threshold set by the system to maintain the health of the river ecosystem; the reservoir flood control capacity is the reservoir volume reserved by the system for flood prevention; the grid security constraints are the power, frequency, and other restrictions set by the system to ensure the stable operation of the grid; the medium-to-long-term scheduling plan for cascade reservoirs is the system's arrangement for water storage, release, and power generation of multiple cascade reservoirs in the basin over a period of time in the future; and the overall control strategy for hydropower conversion is the system's overall principle and framework for converting hydropower into electrical energy or electrical energy into hydropower.

[0068] The system is initiated and operates at the basin-level overall control level. First, the system acquires hydrological forecast data, wind and solar power output prediction data, and the annual grid absorption target for a preset time period, using these data as inputs for decision-making. Simultaneously, the system incorporates basin ecological baseflow, reservoir flood control capacity, and grid security constraints as limitations for decision-making. Based on these inputs and constraints, the system performs calculations and planning to formulate medium- and long-term scheduling plans for cascade reservoirs and overall control strategies for hydropower conversion. This step is the top-level decision-making stage in the hierarchical collaborative decision-making and control architecture, providing macro-level guidance to lower levels.

[0069] Step 1022: At the station-level coordination layer, the medium- and long-term scheduling plan of the cascade reservoirs and the overall control strategy for hydropower conversion issued by the basin-level overall control layer are received. Combining real-time hydrological data, real-time wind and solar power output data and real-time load data, the output of each hydropower unit, the charging and discharging power of the energy storage system and the grid connection ratio of new energy sources are dynamically allocated. The medium- and long-term scheduling plan is corrected through a rolling optimization algorithm to adapt to intraday meteorological fluctuations and load fluctuations.

[0070] In this embodiment, the station-level coordination layer is the decision-making level responsible for regional coordination and intraday regulation; dynamic allocation is the process by which the system flexibly adjusts resource allocation based on real-time conditions; the output of each hydropower unit is the system's set value for the power generation of each hydropower generator unit; the charging and discharging power of the energy storage system is the system's set value for the charging or discharging power of the energy storage device; the new energy grid connection ratio is the proportion of new energy power generation such as photovoltaic and wind power that the system allows to be connected to the grid; the rolling optimization algorithm is a calculation method by which the system corrects the plan through continuous prediction and adjustment; intraday meteorological fluctuations are the weather changes monitored by the system within a day, such as the fluctuations in photovoltaic output caused by changes in cloud cover; and load fluctuations are the changes in electricity demand monitored by the system within a day.

[0071] At the station-level coordination layer, the system receives the medium- and long-term scheduling plans for cascade reservoirs and the overall control strategy for hydropower conversion from the basin-level overall control layer. Based on this, the system performs dynamic calculations by combining real-time hydrological data, real-time wind and solar power output data, and real-time load data. The system dynamically allocates instructions to set specific values ​​for the output of each hydropower unit, the charging and discharging power of the energy storage system, and the proportion of new energy grid connection. To adapt to intraday meteorological and load fluctuations, the system employs a rolling optimization algorithm to periodically revise the medium- and long-term scheduling plans, ensuring a high degree of consistency between the plans and actual operating conditions. This step serves as a bridge between macro-level planning and micro-level execution.

[0072] Step 1023: At the equipment-level execution layer, the dynamic allocation instructions generated by the station-level coordination layer are converted into control signals for the turbine governor, energy storage converter, and wind and solar inverters, and millisecond-level fine-tuning is performed.

[0073] In this embodiment, the device-level execution layer is the execution layer responsible for directly controlling the physical equipment; the dynamic allocation command is a control command generated by the station-level coordination layer that includes specific power setpoints; the turbine governor control signal is an electrical signal sent by the system to the turbine governor to adjust the turbine speed and output; the energy storage converter control signal is an electrical signal sent by the system to the energy storage converter to control the charging and discharging state and power of the energy storage system; the wind and solar inverter control signal is an electrical signal sent by the system to the wind and solar inverter to control the grid-connected power of photovoltaic and wind power; and the millisecond-level fine-tuning is the process by which the system precisely adjusts the equipment operating parameters in a very short time.

[0074] The system receives dynamic allocation instructions generated by the station-level coordination layer at the device-level execution layer. The system converts these instructions into specific physical control signals, including control signals for the hydro turbine governor, energy storage converter, and wind and solar inverters. The system then sends these control signals to the corresponding physical devices, performing millisecond-level fine-tuning, thereby achieving precise control of each hydropower unit, energy storage system, and wind and solar power generation device. This step is the lowest-level execution stage in the entire control architecture, directly impacting the physical world.

[0075] Step 1024: Collect the operating status parameters of each device and feed them back to the station-level coordination layer and the watershed-level overall control layer to form a hierarchical closed-loop control architecture.

[0076] In this embodiment, the operating status parameters are data collected by the system from the device end that reflect its current working status, such as speed, power, temperature, voltage, etc.; the hierarchical closed-loop control architecture is a structure in which the system feeds back the execution results to the upper decision layer, forming a complete control loop from decision-making to execution and then to feedback.

[0077] After completing adjustments at the device-level execution layer, the system immediately collects the operating status parameters of each device. These parameters are then used as feedback signals and uploaded to the station-level coordination layer and the basin-level overall control layer. The upper-level decision-making layer evaluates the control effectiveness based on this feedback information and adjusts subsequent scheduling plans and control strategies accordingly. In this way, the system constructs a hierarchical closed-loop control architecture, realizing a complete cycle from perception, decision-making, execution to feedback, ensuring the accuracy and adaptability of the control.

[0078] This application embodiment achieves millisecond-level precise control from macro-planning to micro-execution through a hierarchical collaboration and closed-loop feedback mechanism at the basin-level overall control layer, station-level coordination layer, and equipment-level execution layer. This enhances the flexibility and robustness of the hydropower coupled energy management system across multiple time scales.

[0079] Optionally, step 103 includes: Step 1031: Define a set of multi-objective functions, which includes the objectives of maximizing the compliance rate of watershed ecological flow, maximizing the comprehensive utilization efficiency of hydropower, maximizing the absorption rate of new energy sources, and maximizing the reliability of power supply.

[0080] In this embodiment, the multi-objective function set is a combination of multiple objective functions that need to be optimized simultaneously. These functions are used to quantify performance indicators in different aspects, such as ecological protection, energy efficiency, and power supply stability. The objective of maximizing the ecological flow compliance rate of the watershed is an objective function that aims to ensure that the ecological flow at key sections within the watershed reaches or exceeds preset ecological protection standards through scheduling strategies, thereby safeguarding the health of the river ecosystem. The objective of maximizing the comprehensive utilization efficiency of hydropower is an objective function that aims to maximize the comprehensive benefits of hydropower resources in power generation, water supply, and ecology, such as the electricity generated per unit volume of water or its optimized allocation among different uses. The objective of maximizing the renewable energy absorption rate is an objective function that aims to maximize the grid's capacity to accept renewable energy generation such as wind and solar power, thereby reducing wind and solar curtailment. The objective of maximizing power supply reliability is an objective function that aims to maximize the power system's ability to maintain a stable power supply in response to load fluctuations and emergencies through optimized scheduling strategies, such as minimizing load losses or voltage exceedance risks.

[0081] The system first defines a set of multi-objective functions, which consists of four specific objective functions: maximizing the watershed ecological flow compliance rate, maximizing the comprehensive utilization efficiency of hydropower, maximizing the renewable energy absorption rate, and maximizing power supply reliability. These objective functions define the optimal direction to be pursued in the subsequent optimization process from four dimensions: ecological protection, resource utilization, clean energy absorption, and system security.

[0082] Step 1032: Define a set of constraints, which includes reservoir flood control limit water level constraints, hydropower unit output limit constraints, power grid cross-sectional transmission capacity constraints, and ecological base current hard constraints.

[0083] In this embodiment, the constraint set is a combination of multiple hard constraints that must be satisfied during the optimization process. These constraints define the feasible domain of the scheduling strategy and ensure operational safety and compliance. The reservoir flood control limit water level constraint is one constraint, stipulating that the reservoir water level must be maintained below a certain safe upper limit during the flood season or specific periods to ensure the flood control safety of the dam and downstream areas. The hydropower unit output upper limit constraint is another constraint, stipulating that the power generation of each hydropower unit cannot exceed its design or operational maximum value to ensure equipment safety. The power grid cross-section transmission capacity constraint is yet another constraint, specifying the maximum power that a specific transmission section or line in the power system can safely transmit to prevent line overload. The ecological baseflow hard constraint is yet another constraint, stipulating that the river channel must guarantee a water flow of no less than a certain minimum flow rate at all times to meet the basic survival needs of the river ecosystem.

[0084] The system defines a set of constraints, which consists of four specific constraints: reservoir flood control limit water level constraint, hydropower unit output limit constraint, power grid cross-sectional transmission capacity constraint, and ecological base current hard constraint. These constraints together constitute a hard boundary for safety and ecology. Any generated scheduling strategy must strictly satisfy these conditions; otherwise, it will be considered an infeasible solution.

[0085] Step 1033: Embed an adaptive weighted multi-objective game algorithm to dynamically adjust the first weight coefficient corresponding to the ecological protection objective and the second weight coefficient corresponding to the energy efficiency objective.

[0086] In this embodiment, the adaptive weighted multi-objective game algorithm is an optimization algorithm that dynamically adjusts the importance of different objectives based on the progress of the optimization process or environmental changes, thereby simulating the game process to find an equilibrium solution. The first weight coefficient is a numerical parameter used in the optimization algorithm to quantify the relative importance of ecological protection objectives compared to other objectives; a larger value indicates that the ecological protection objective is given higher priority. The second weight coefficient is a numerical parameter used in the optimization algorithm to quantify the relative importance of energy efficiency objectives compared to other objectives; a larger value indicates that the energy efficiency objective is given higher priority.

[0087] The system incorporates an adaptive weighted multi-objective game theory algorithm. The core function of this algorithm is to dynamically adjust the first weight coefficient corresponding to the ecological protection objective and the second weight coefficient corresponding to the energy efficiency objective. Through this dynamic adjustment, the algorithm can simulate the game between these two conflicting objectives during the optimization process, assigning different priorities to different objectives at different stages or in different scenarios, thereby guiding the search towards a more balanced solution space.

[0088] Step 1034: The non-dominated sorting genetic algorithm is used to iteratively optimize the multi-objective function set and the constraint set to generate the Pareto optimal solution set.

[0089] In this embodiment, the non-dominated sorting genetic algorithm is an evolutionary algorithm that solves multi-objective optimization problems by simulating natural selection and genetic mechanisms and combining them with a non-dominated sorting strategy. It can simultaneously optimize multiple conflicting objectives. Iterative optimization computation refers to the process by which the algorithm repeatedly performs operations such as selection, crossover, and mutation to gradually improve the fitness of individuals in the population, thereby approximating the optimal solution or Pareto front of the problem. The Pareto optimal solution set is a set of solutions in which no single solution can further improve any other objective function without compromising at least one other objective function; it represents a series of non-dominated equilibrium schemes in a multi-objective optimization problem.

[0090] The system employs a non-dominated sorting genetic algorithm to iteratively optimize the set of multi-objective functions and constraints. In each iteration, the algorithm performs non-dominated sorting and crowding calculations on candidate solutions to maintain solution diversity and convergence, and generates new solutions through selection, crossover, and mutation operations. After multiple iterations, the algorithm ultimately generates a Pareto optimal solution set, which contains a series of non-dominated solutions that achieve the best trade-offs among the four objectives of ecology, energy efficiency, energy consumption, and reliability while satisfying all constraints.

[0091] Step 1035: Obtain power grid dispatch demand data and ecological assessment result data, and select the optimal hydropower coupling dispatch strategy from the Pareto optimal solution set based on the power grid dispatch demand data and the ecological assessment result data.

[0092] In this embodiment, grid dispatch demand data reflects the operating status and future plans of the power system, such as load forecasting, tie-line power plans, and spinning reserve capacity requirements, and is used to guide dispatch decisions. Ecological assessment results data are obtained after quantitatively evaluating the current or future ecological status of the watershed, such as the ecological flow compliance rate at key sections and the suitability index for aquatic habitats, and are used to assess the ecological impact of dispatch strategies. The optimal hydropower coupling dispatch strategy is a specific dispatch scheme selected from the Pareto optimal solution set. This scheme best simultaneously satisfies grid dispatch demands and ecological protection requirements, and guides the actual execution of hydropower resource dispatch.

[0093] The system acquires power grid dispatch demand data and ecological assessment results. Then, based on the acquired power grid dispatch demand data (e.g., requirements for power load and reserve capacity) and ecological assessment results (e.g., real-time evaluation of ecological flow compliance rates), the system filters from the Pareto optimal solution set. The system selects the solution from the Pareto optimal solution set that satisfies the requirements for safe and stable power grid operation while also achieving the optimal or most acceptable level of ecological assessment results, and identifies this as the optimal hydropower coupling dispatch strategy.

[0094] This application embodiment defines a set of multi-objective functions and a set of constraints, embeds an adaptive weighted multi-objective game algorithm combined with a non-dominated sorting genetic algorithm for iterative optimization, and then selects the optimal strategy based on the results of power grid scheduling and ecological assessment. This achieves the effect of dynamically balancing ecological protection and energy benefits while meeting hard constraints such as ecological base flow and power grid security, and automatically generating the optimal hydropower coupling scheduling scheme.

[0095] Optionally, step 104 includes: Step 1041: Monitor wind and solar power output data and grid load data in real time to determine whether there is a situation where the output of new energy is excessive and the grid load is lower than the preset load lower limit threshold.

[0096] In this embodiment of the application, the new energy output data is numerical information representing the real-time output power of new energy power generation facilities such as wind farms and photovoltaic power stations; the grid load data is numerical information representing the total power consumption carried by the grid at the current moment; the preset load lower limit threshold is a power value limit set in advance by the system to determine whether the grid is in a low load state.

[0097] The system acquires real-time data on renewable energy output and grid load through sensors and acquisition terminals deployed at wind and solar power plants and key grid nodes. The system compares the acquired real-time grid load data with a preset lower load threshold stored internally, while simultaneously analyzing the magnitude of renewable energy output data. It comprehensively determines whether two conditions are met simultaneously: renewable energy output is excessive, meaning the renewable energy output data exceeds a certain set standard; and the grid load is below the preset lower load threshold. This determination serves as the logical starting point for triggering subsequent hydropower bidirectional conversion buffer regulation actions.

[0098] Step 1042: When it is determined that the output of new energy is in excess, the pumped storage unit is started to convert the excess electrical energy into hydropower and store it in the upstream reservoir, and the reservoir capacity regulation capacity of the cascade reservoir is used to smooth the fluctuation of new energy output.

[0099] In this embodiment, a pumped storage unit is a reversible pump-turbine unit that can convert electrical energy into the potential energy of water for storage; an upstream reservoir is a reservoir area located at a higher elevation and with water storage capacity in a cascade hydropower station group; a cascade reservoir is a system composed of a series of reservoirs distributed along the upstream and downstream of a river and having regulation capabilities; reservoir capacity regulation capability refers to the ability of a cascade reservoir to change its reservoir capacity state by adjusting the water storage and release, thereby smoothing out water volume fluctuations; new energy output fluctuations refer to the phenomenon of unstable and rapid changes in wind power and photovoltaic power generation caused by factors such as weather.

[0100] When the system determines that there is an overcapacity of renewable energy output, it immediately issues a start-up command to the pumped-storage units. The pumped-storage units operate in electric motor mode, absorbing excess electrical energy from the grid to drive pumps that draw water from downstream or lower elevations to upstream reservoirs, thus converting electrical energy into the potential energy of the water for storage. Simultaneously, the system utilizes the overall capacity regulation of the cascade reservoir group to coordinate the water storage and release rhythms of each reservoir, absorbing grid power fluctuations caused by the overcapacity of renewable energy output and thus mitigating the instability of renewable energy output.

[0101] Step 1043: When it is determined that the output of new energy sources is insufficient or the grid load is higher than the preset load limit threshold, the cascade hydropower stations are dispatched to generate more power to convert water energy into electrical energy to supplement the power supply gap of the grid.

[0102] In this embodiment, insufficient new energy output refers to a state where the actual output power of new energy power generation facilities is lower than their rated capacity or lower than a certain low power threshold preset by the system, and cannot meet the grid demand; the preset load upper limit threshold is a power value limit set in advance by the system to determine whether the grid is in a high load state; a cascade hydropower station is a collection of a series of hydropower stations developed in cascades along the upstream and downstream of a river; the grid power supply gap refers to the difference between the grid load demand and the total power supply capacity of all current power generation resources.

[0103] When the system determines that the current renewable energy output is insufficient, or that the grid load data exceeds the preset load limit threshold, the system immediately issues an increase generation command to the cascade hydropower station group. The cascade hydropower stations respond to the command by adjusting the turbine guide vane opening and other methods to increase power generation, rapidly converting the potential energy of the water stored in the reservoir into electrical energy and injecting it into the grid. This supplements the power supply gap caused by insufficient renewable energy output or peak grid load, ensuring power balance and power supply reliability.

[0104] Step 1044: Based on the real-time perception of the physical system's operating status by the digital twin, update the water energy bidirectional conversion buffer control strategy according to a preset rolling time interval.

[0105] In this embodiment, the digital twin is a high-fidelity digital mapping model of the physical system in virtual space, capable of synchronizing the operating status of the physical system in real time. The operating status of the physical system refers to a set of parameters collected by sensors that characterize the current working status of physical entities such as hydropower stations, power grids, and reservoirs, such as water level, flow rate, power generation, and voltage. The preset rolling time interval is a fixed duration, such as 15 minutes, pre-set by the system for periodically performing update operations. The hydropower bidirectional conversion buffer control strategy is a set of rules and parameters formulated by the system to guide the direction and intensity of energy conversion between pumped storage and hydropower generation.

[0106] The system continuously acquires real-time operational status data of the physical system from the digital twin, including reservoir water level, turbine operating conditions, and power grid frequency. At preset rolling time intervals, such as every 15 minutes, the system automatically reads the latest physical system operational status data and, combined with the latest renewable energy output forecasts and load forecasts, recalculates and updates the hydropower bidirectional conversion buffer control strategy. This update process is a dynamic optimization process, designed to ensure that the control strategy remains synchronized with the actual situation of the physical system, thereby improving the accuracy of control.

[0107] Step 1045: Combine runoff forecast data to plan the water storage and release rhythm of upstream reservoirs in advance, so as to achieve spatiotemporal coordination between the hydropower conversion rhythm, the new energy output fluctuation curve and the power grid load change curve.

[0108] In this embodiment, runoff forecast data refers to the numerical information of river runoff predicted by the system using tools such as hydrological models over a future period; the upstream reservoir's water storage rhythm refers to the planned rate and timing of the reservoir's water level rise; the upstream reservoir's water release rhythm refers to the planned rate and timing of the reservoir's water level fall; the hydropower conversion rhythm refers to the timing and rate of the system's pumped storage and hydropower generation energy conversion operations; the new energy output fluctuation curve is a graphical representation of the change in new energy power generation over time; the grid load change curve is a graphical representation of the change in grid power consumption over time; and spatiotemporal coordination refers to the system coordinating hydropower conversion operations at different time points and spatial locations to match the changing trends of new energy output and grid load.

[0109] When formulating and updating the hydropower bidirectional conversion buffer and control strategy, the system calls upon and analyzes runoff forecast data. Based on the prediction of future runoff, the system plans the water storage and release rhythms of upstream reservoirs in advance. By coordinating these rhythms, the system ensures that the overall hydropower conversion rhythm of pumped storage and hydropower generation is synchronized and matched with the fluctuation curves of new energy output and grid load changes in the time dimension, and coordinated with the reservoir capacity status and geographical location of different reservoirs in the spatial dimension. This approach ensures that hydropower conversion operations can effectively offset fluctuations in new energy output and respond to changes in grid load, achieving globally optimal resource allocation.

[0110] This application embodiment achieves dynamic buffering and control of bidirectional hydropower conversion through real-time monitoring, status judgment, precise scheduling, dynamic updates, and advance planning. This results in smoothing out fluctuations in new energy output, ensuring grid power balance, improving hydropower utilization efficiency, and enhancing system control robustness.

[0111] Optionally, step 105 includes: Step 1051: Construct a deep reinforcement learning agent, using historical hydropower coupling scheduling data, real-time operation status data, and regulation effect evaluation data as the training sample set.

[0112] In this embodiment, the deep reinforcement learning agent is a computational model combining deep neural networks and reinforcement learning algorithms, used to learn decision-making strategies based on environmental states; historical hydropower coupled scheduling data is recorded data of coordinated scheduling of water resources and energy in the past time period; real-time operation status data is the operating parameters of hydropower stations, new energy power plants and energy storage systems at the current moment; regulation effect evaluation data is the evaluation results of the executed scheduling strategies in terms of ecological and energy benefits; the training sample set is a dataset composed of the above data used for model training.

[0113] The system first constructs a deep reinforcement learning agent, which serves as the core decision-making module responsible for learning the regulation strategies of the hydropower coupling system. The system then collects historical hydropower coupling scheduling data, real-time operational status data, and regulation effect evaluation data, integrating these data into a unified training sample set for subsequent model training. In this way, the system provides the agent with the fundamental data source for learning.

[0114] Step 1052: The training sample set is classified and labeled according to the hydrological scenarios of the high water season, the normal water season, the low water season, the peak-valley load energy scenario, and the new energy output fluctuation scenario.

[0115] In this application embodiment, the hydrological scenario of the high-water season is a hydrological condition with a large runoff; the hydrological scenario of the normal water season is a hydrological condition with a moderate runoff; the hydrological scenario of the low-water season is a hydrological condition with a small runoff; the peak-valley load energy scenario is the change in grid load during peak and valley periods; the new energy output fluctuation scenario is the fluctuation in wind or solar power generation caused by weather changes; and the classification labeling is to label each sample in the training sample set as the corresponding scenario category.

[0116] The system classifies and labels the constructed training sample set according to its data characteristics, categorizing them into hydrological scenarios during high-water seasons, normal-water seasons, low-water seasons, peak-valley load energy scenarios, and new energy output fluctuation scenarios. Each sample is assigned one or more scenario labels; for example, a sample may belong to both the low-water season hydrological scenario and the peak-valley load energy scenario. This classification and labeling lays the foundation for subsequent targeted training.

[0117] Step 1053: Use the training sample set after classification and labeling to train the deep reinforcement learning agent, so that the deep reinforcement learning agent learns the optimal control strategy under different hydrological scenarios and different energy scenarios.

[0118] In this embodiment, model training is the process of optimizing the parameters of a deep reinforcement learning agent using a set of training samples after classification and labeling; the optimal control strategy is a scheduling decision scheme that maximizes ecological and energy benefits under specific hydrological and energy scenarios.

[0119] The system uses a categorized and labeled training sample set to train a deep reinforcement learning agent. During training, the system iterates repeatedly to enable the deep reinforcement learning agent to learn optimal control strategies under different hydrological scenarios, such as high-water season, normal-water season, and low-water season, as well as different energy scenarios, such as peak-valley load and fluctuations in renewable energy output. These strategies include hydropower unit output allocation, energy storage charging and discharging scheduling, and renewable energy consumption ratios, thereby improving the system's decision-making capabilities under various complex scenarios.

[0120] Step 1054: Based on the daily scheduling execution feedback data, the coupling matching parameters of the hydropower coupling dynamic response model, the dynamic weight parameters of the multi-objective game algorithm, and the logical parameters of the hydropower bidirectional conversion buffer regulation are iteratively updated through the deep reinforcement learning agent.

[0121] In this embodiment, the daily scheduling execution feedback data refers to the data collected from actual operation after the system executes the daily scheduling strategy, including indicators such as the ecological flow compliance rate, new energy absorption rate, and power supply reliability; the hydropower coupling dynamic response model is a mathematical model describing the spatiotemporal coupling relationship between water resources and energy systems; the coupling matching parameter is a variable in the model used to quantify the correlation between runoff, power generation, and load; the multi-objective game algorithm is an optimization algorithm used to balance ecological protection and energy benefits; the dynamic weight parameter is an adjustable coefficient representing the relative importance of ecological and energy objectives in the algorithm; the hydropower bidirectional conversion buffer regulation is a regulation mechanism by which the system achieves mutual conversion between hydropower and electrical energy through pumped storage and increased unit power generation; and the logic parameter is a variable controlling the execution rhythm and amplitude of this mechanism.

[0122] The system iteratively updates itself using a trained deep reinforcement learning agent based on daily scheduling execution feedback data. By analyzing the feedback data, the system adjusts the coupling matching parameters in the hydropower coupling dynamic response model to more accurately reflect the correlation between runoff, water level, power generation, and load. Simultaneously, the system updates the dynamic weight parameters in the multi-objective game algorithm, making the balance between ecological protection and energy efficiency more adaptable to real-time scenarios. The system also optimizes the logical parameters for the bidirectional hydropower conversion buffer regulation, such as the rhythm of water storage or release, to smooth out energy fluctuations. Through this daily iteration, the system continuously improves the adaptability of its scheduling strategy.

[0123] Step 1055: Full update of the model parameters of the deep reinforcement learning agent according to the preset quarterly cycle to achieve self-adaptation and self-optimization of the scheduling strategy.

[0124] In this embodiment, the preset quarterly cycle is a time interval set by the system to be executed once every three months; the model parameters are the weights and biases of the deep neural network in the deep reinforcement learning agent; full update is to retrain the entire model using newly collected data and replace the original parameters; self-adaptation is the system's ability to automatically adapt to different hydrological and energy scenarios; and self-optimization is the process by which the system automatically improves the performance of the scheduling strategy.

[0125] The system performs a full update of the deep reinforcement learning agent's model parameters according to a preset quarterly cycle. It collects all scheduling data, operational status, and feedback from the past quarter as new training samples to retrain the deep reinforcement learning agent, replacing the original model parameters. Through this periodic full update, the system achieves self-adaptation and self-optimization of scheduling strategies, ensuring consistently high energy management performance in the face of long-term hydrological changes and energy market fluctuations.

[0126] This application embodiment constructs a deep reinforcement learning agent and uses training samples with classification labels for model training. Combined with daily iterative updates and quarterly full updates, the system realizes the self-learning, self-adaptation and self-optimization of scheduling strategies, and achieves the technical effect of improving the robustness of the hydropower coupling system in complex hydrological and energy scenarios, the balance between ecological protection and energy benefits, and the capacity for new energy consumption.

[0127] Optionally, refer to Figure 2 The model training iteration process of this application includes: building a watershed hydropower coupled energy management platform to realize full-process visualization of data acquisition, model calculation, strategy generation, command issuance, and status feedback; real-time display of watershed hydrological status, energy output, ecological flow compliance and new energy consumption rate; support for manual intervention and strategy adjustment by dispatchers; and forming a complete closed loop of perception-modeling-decision-execution-optimization.

[0128] Based on the foregoing embodiments, this application provides an energy management device for hydropower coupling control. The device includes various units and modules included in each unit, which can be implemented by a processor in a computer device; of course, it can also be implemented by specific logic circuits. In the implementation process, the processor can be a central processing unit (CPU), a microprocessor unit (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA), etc.

[0129] Figure 3 A schematic diagram of the composition structure of an energy management device for water-energy coupling control provided in an embodiment of this application is shown below. Figure 3 As shown, the energy management device 20 for water-energy coupling control includes: The processing module 201 is used to acquire watershed hydrological data, energy data and ecological data, construct a water-energy coupled sensing dataset, and establish a digital twin of watershed hydrology and energy coupling based on the dataset, quantify the spatiotemporal coupling correlation between runoff, water level changes and energy output and load demand, and generate a water-energy coupled dynamic response model. Based on the aforementioned hydropower coupling dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed, and a hierarchical control strategy including medium- and long-term scheduling plans and real-time adjustment strategies is formulated. Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the game between ecological protection objectives and energy efficiency objectives, and a hydropower coupling scheduling strategy is generated. The execution module 202 is used to perform water energy bidirectional conversion buffer regulation. Based on the dynamic changes in new energy output and grid load, it controls the pumped storage unit and hydropower unit to perform bidirectional conversion of electrical energy and water, so as to smooth out energy system fluctuations. Historical scheduling data and real-time operational feedback are collected, and the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy are iteratively optimized through a reinforcement learning agent to form a closed-loop energy management system.

[0130] Optionally, the processing module 201 is further configured to: Deploy hydrological monitoring stations to collect flow data, water level data, and rainfall data; deploy hydropower unit status sensors to collect unit operating parameters; deploy wind and solar power station output acquisition terminals to collect photovoltaic output data and wind power output data; deploy energy storage battery management systems to collect energy storage charging and discharging information; and deploy power grid load monitoring nodes to collect load data and electricity price data. The flow rate data, water level data, rainfall data, generator operating parameters, photovoltaic output data, wind power output data, energy storage charging and discharging information, load data, and electricity price data are collected on millisecond, minute, and hourly time scales, respectively. The collected data is time-synchronized, missing value imputation is performed, and multi-source fusion is carried out by the edge computing gateway to obtain the water energy coupled sensing dataset; Based on watershed topographic data, hydrogeological data, turbine output characteristic models, wind and solar power prediction data, energy storage charging and discharging information, and power grid flow data, a spatiotemporal coupling correlation model for runoff flow, water level changes, power generation output, and load demand is established. Based on the spatiotemporal coupling correlation model, watershed-level digital twins, station-level digital twins, and equipment-level digital twins are generated to map the operating status of the physical system in real time.

[0131] Optionally, the processing module 201 is further configured to: Configure the data cache queue of the edge computing gateway and the time synchronization reference clock; The flow rate data, photovoltaic output data, wind power output data, and energy storage charging and discharging information collected on a millisecond time scale, the unit operating parameters, load data, and electricity price data collected on a minute time scale, and the water level data and rainfall data collected on an hourly time scale are respectively stored in the data cache queue. All data in the data cache queue are time-aligned according to the unified timestamp of the time synchronization reference clock; Perform outlier detection on the time-aligned data and remove data points that exceed the preset reasonable range; The missing data resulting from the removal process is filled in using linear interpolation to generate complete time series data; After time synchronization and missing value completion, the hydrological, energy, and ecological data in the time series data are fused from multiple sources according to preset data fusion rules to generate a water-energy coupled sensing dataset containing hydrological feature fields, energy feature fields, and ecological feature fields.

[0132] Optionally, the processing module 201 is further configured to: At the basin-level overall control level, the hydrological forecast data, wind and solar power output prediction data and annual grid absorption target within a preset time period are used as inputs, and the basin ecological base flow, reservoir flood control capacity and grid security constraints are used as constraints to formulate medium and long-term scheduling plans for cascade reservoirs and overall control strategies for hydropower conversion. At the station-level coordination layer, the system receives the medium- and long-term scheduling plans for the cascade reservoirs and the overall control strategy for hydropower conversion from the basin-level overall control layer. Combining real-time hydrological data, real-time wind and solar power output data, and real-time load data, the system dynamically allocates the output of each hydropower unit, the charging and discharging power of the energy storage system, and the grid connection ratio of new energy sources. The system also uses a rolling optimization algorithm to correct the medium- and long-term scheduling plans to adapt to intraday meteorological and load fluctuations. At the equipment-level execution layer, the dynamic allocation commands generated by the station-level coordination layer are converted into turbine governor control signals, energy storage converter control signals, and wind and solar inverter control signals, and millisecond-level fine-tuning is performed. The operating status parameters of each device are collected and fed back to the station-level coordination layer and the watershed-level overall control layer, forming a hierarchical closed-loop control architecture.

[0133] Optionally, the processing module 201 is further configured to: Define a set of multi-objective functions, which includes the objectives of maximizing the compliance rate of watershed ecological flow, maximizing the comprehensive utilization efficiency of hydropower, maximizing the absorption rate of new energy sources, and maximizing the reliability of power supply. Define a set of constraints, which includes reservoir flood control limit water level constraints, hydropower unit output limit constraints, power grid cross-sectional transmission capacity constraints, and ecological base current hard constraints. An adaptive weighted multi-objective game algorithm is embedded to dynamically adjust the first weight coefficient corresponding to the ecological protection objective and the second weight coefficient corresponding to the energy efficiency objective. A non-dominated sorting genetic algorithm is used to iteratively optimize the set of multi-objective functions and the set of constraints to generate a Pareto optimal solution set. Obtain power grid dispatch demand data and ecological assessment results data, and select the optimal hydropower coupling dispatch strategy from the Pareto optimal solution set based on the power grid dispatch demand data and the ecological assessment results data.

[0134] Optionally, the execution module 202 is further configured to: Real-time monitoring of wind and solar power output data and grid load data to determine whether there is a situation where the output of new energy sources is excessive and the grid load is lower than the preset lower load threshold. When it is determined that there is an overcapacity of new energy output, the pumped storage units are started to convert the excess electrical energy into hydropower and store it in the upstream reservoir. The reservoir capacity regulation capacity of the cascade reservoirs is used to smooth out the fluctuations in new energy output. When it is determined that the output of new energy sources is insufficient or the grid load is higher than the preset load limit threshold, the cascade hydropower stations will be dispatched to generate more water energy and convert it into electricity to supplement the power supply gap of the grid. Based on the real-time perception of the physical system's operating status by the digital twin, the water energy bidirectional conversion buffer control strategy is updated according to a preset rolling time interval. By combining runoff forecast data, the water storage and release schedules of upstream reservoirs can be planned in advance, enabling the water energy conversion schedule to be coordinated with the fluctuation curves of new energy output and the changes in grid load in a timely and spatial manner.

[0135] Optionally, the execution module 202 is further configured to: A deep reinforcement learning agent was constructed, using historical hydropower coupled scheduling data, real-time operation status data, and regulation effect evaluation data as training sample sets. The training sample set is classified and labeled according to the hydrological scenarios of the high water season, the normal water season, the low water season, the peak-valley load energy scenario, and the new energy output fluctuation scenario. The deep reinforcement learning agent is trained using the training sample set after classification and labeling, so that the deep reinforcement learning agent learns the optimal control strategy under different hydrological scenarios and different energy scenarios. Based on the daily scheduling and execution feedback data, the coupling matching parameters of the hydropower coupling dynamic response model, the dynamic weight parameters of the multi-objective game algorithm, and the logical parameters of the hydropower bidirectional conversion buffer regulation are iteratively updated by the deep reinforcement learning agent. The model parameters of the deep reinforcement learning agent are fully updated according to a preset quarterly cycle, so as to realize the self-adaptation and self-optimization of the scheduling strategy.

[0136] This application embodiment improves the global allocation efficiency of water and energy resources, enhances the capacity for new energy absorption, ensures the ecological security of the basin, and improves the robustness of energy management in complex scenarios by constructing a digital twin and quantifying the coupling relationship between water and energy, deploying a hierarchical collaborative control architecture, adopting ecological and benefit multi-objective game optimization, performing two-way conversion buffer control of water and energy, and introducing reinforcement learning to achieve self-learning iteration.

[0137] The descriptions of the apparatus embodiments above are similar to those of the method embodiments above, and have similar beneficial effects. In some embodiments, the functions or modules included in the apparatus provided in this application can be used to perform the methods described in the method embodiments above. For technical details not disclosed in the apparatus embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0138] It should be noted that, in the embodiments of this application, if the above-described energy management method for hydropower coupling control is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware, software, or firmware, or any combination of hardware, software, and firmware.

[0139] This application provides a computer device including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements some or all of the steps in the above-described method.

[0140] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements some or all of the steps in the above-described method. The computer-readable storage medium can be transient or non-transient.

[0141] This application provides a computer program including computer-readable code, wherein when the computer-readable code is executed in a computer device, a processor in the computer device performs some or all of the steps in the above-described method.

[0142] This application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, it implements some or all of the steps in the above-described method. This computer program product can be implemented specifically through hardware, software, or a combination thereof. In some embodiments, the computer program product is specifically embodied as a computer storage medium; in other embodiments, the computer program product is specifically embodied as a software product, such as a software development kit (SDK), etc.

[0143] It should be noted that the descriptions of the various embodiments above tend to emphasize the differences between them, while their similarities or commonalities can be referred to interchangeably. The descriptions of the above embodiments of the device, storage medium, computer program, and computer program product are similar to the descriptions of the above method embodiments and have similar beneficial effects. For technical details not disclosed in the embodiments of the device, storage medium, computer program, and computer program product of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0144] It should be noted that, Figure 4 This is a schematic diagram of a hardware entity of a computer device in an embodiment of this application, such as... Figure 4 As shown, the hardware entity of the computer device 700 includes: one or more processors 701, a communication interface 702, and a memory 703, wherein: Processor 701 typically controls the overall operation of computer device 700.

[0145] Communication interface 702 enables computer devices to communicate with other terminals or servers over a network.

[0146] The memory 703 is configured to store instructions and applications executable by the processor 701, and can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data, and video communication data) in the processor 701 and various modules in the computer device 700. It can be implemented using flash memory or random access memory (RAM). Data transfer between the processor 701, the communication interface 702, and the memory 703 can be performed via bus 704. Only one processor is shown in the figure; each processor 700 includes one or more cores.

[0147] It should be noted that the computer device may include multiple processors 701, and each processor 701 can interact with each other through aggregated communication methods such as all-to-all, all-gather, or all-reduce. The processors 701 may be central processing units (CPUs), graphics processing units (GPUs), embedded neural network processing units (NPUs), tensor processing units (TPUs), data processing units (DPUs), accelerated processing units (APUs), floating-point processing units (FPUs), or application-specific integrated circuits (ASICs). The processors may also be single-core or multi-core processors. The processor may consist of a CPU and hardware chips. The hardware chips may be ASICs, PLDs, or combinations thereof. The PLDs may be complex programmable logic devices (CPLDs), FPGAs, generic array logic (GALs), or any combination thereof. The processor can also be implemented using logic devices with built-in processing logic, such as FPGAs or digital signal processors (DSPs).

[0148] The communication interface 702 can be a wired interface or a wireless interface, used to communicate with other modules or devices. The wired interface can be an Ethernet interface, a local interconnect network (LIN), etc., and the wireless interface can be a cellular network interface or a wireless LAN interface, etc.

[0149] Memory 703 can be non-volatile memory, such as read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Memory 703 can also be volatile memory, which can be random access memory (RAM) used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synclink dynamic random access memory (SLDRAM), and direct rambus RAM (DRRAM), direct rambus DRAM (DRDRAM), and rambus DRAM.

[0150] The 704 bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc.

[0151] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above steps / processes do not imply a sequential order of execution; the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0152] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0153] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0154] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0155] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0156] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0157] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, magnetic disks, or optical disks.

[0158] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An energy management method for hydropower coupling control, characterized in that, The method includes: Acquire watershed hydrological, energy, and ecological data, construct a water-energy coupled sensing dataset, and establish a digital twin of watershed hydrology and energy coupling based on the dataset. Quantify the spatiotemporal coupling correlation between runoff, water level changes, energy output, and load demand, and generate a water-energy coupled dynamic response model. Based on the aforementioned hydropower coupling dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed, and a hierarchical control strategy including medium- and long-term scheduling plans and real-time adjustment strategies is formulated. Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the game between ecological protection objectives and energy efficiency objectives, and a hydropower coupling scheduling strategy is generated. Implement water energy bidirectional conversion buffer regulation, and control pumped storage units and hydropower units to carry out bidirectional conversion of electricity and water according to the dynamic changes of new energy output and grid load, so as to smooth energy system fluctuations. Historical scheduling data and real-time operational feedback are collected, and the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy are iteratively optimized through a reinforcement learning agent to form a closed-loop energy management system.

2. The method according to claim 1, characterized in that, The process involves establishing a digital twin of watershed hydrology and energy coupling based on the dataset, quantifying the spatiotemporal coupling correlation between runoff, water level changes, and energy output and load demand, and generating a hydropower coupling dynamic response model, including: Deploy hydrological monitoring stations to collect flow data, water level data, and rainfall data; deploy hydropower unit status sensors to collect unit operating parameters; deploy wind and solar power station output acquisition terminals to collect photovoltaic output data and wind power output data; deploy energy storage battery management systems to collect energy storage charging and discharging information; and deploy power grid load monitoring nodes to collect load data and electricity price data. The flow rate data, water level data, rainfall data, generator operating parameters, photovoltaic output data, wind power output data, energy storage charging and discharging information, load data, and electricity price data are collected on millisecond, minute, and hourly time scales, respectively. The collected data is time-synchronized, missing value imputation is performed, and multi-source fusion is carried out by the edge computing gateway to obtain the water energy coupled sensing dataset; Based on watershed topographic data, hydrogeological data, turbine output characteristic models, wind and solar power prediction data, energy storage charging and discharging information, and power grid flow data, a spatiotemporal coupling correlation model for runoff flow, water level changes, power generation output, and load demand is established. Based on the spatiotemporal coupling correlation model, watershed-level digital twins, station-level digital twins, and equipment-level digital twins are generated to map the operating status of the physical system in real time.

3. The method according to claim 2, characterized in that, The process of time synchronization, missing value imputation, and multi-source fusion of the collected data through an edge computing gateway to obtain a water energy coupled sensing dataset includes: Configure the data cache queue of the edge computing gateway and the time synchronization reference clock; The flow rate data, photovoltaic output data, wind power output data, and energy storage charging and discharging information collected on a millisecond time scale, the unit operating parameters, load data, and electricity price data collected on a minute time scale, and the water level data and rainfall data collected on an hourly time scale are respectively stored in the data cache queue. All data in the data cache queue are time-aligned according to the unified timestamp of the time synchronization reference clock; Perform outlier detection on the time-aligned data and remove data points that exceed the preset reasonable range; The missing data resulting from the removal process is filled in using linear interpolation to generate complete time series data; After time synchronization and missing value completion, the hydrological, energy, and ecological data in the time series data are fused from multiple sources according to preset data fusion rules to generate a water-energy coupled sensing dataset containing hydrological feature fields, energy feature fields, and ecological feature fields.

4. The method according to claim 1, characterized in that, Based on the aforementioned hydropower coupled dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed. A hierarchical control strategy, including medium- and long-term scheduling plans and real-time adjustment strategies, is formulated, comprising: At the basin-level overall control level, the hydrological forecast data, wind and solar power output prediction data and annual grid absorption target within a preset time period are used as inputs, and the basin ecological base flow, reservoir flood control capacity and grid security constraints are used as constraints to formulate medium and long-term scheduling plans for cascade reservoirs and overall control strategies for hydropower conversion. At the station-level coordination layer, the system receives the medium- and long-term scheduling plans for the cascade reservoirs and the overall control strategy for hydropower conversion from the basin-level overall control layer. Combining real-time hydrological data, real-time wind and solar power output data, and real-time load data, the system dynamically allocates the output of each hydropower unit, the charging and discharging power of the energy storage system, and the grid connection ratio of new energy sources. The system also uses a rolling optimization algorithm to correct the medium- and long-term scheduling plans to adapt to intraday meteorological and load fluctuations. At the equipment-level execution layer, the dynamic allocation commands generated by the station-level coordination layer are converted into turbine governor control signals, energy storage converter control signals, and wind and solar inverter control signals, and millisecond-level fine-tuning is performed. The operating status parameters of each device are collected and fed back to the station-level coordination layer and the watershed-level overall control layer, forming a hierarchical closed-loop control architecture.

5. The method according to claim 1, characterized in that, Under the hard constraint of ecological flow, a multi-objective game theory algorithm is used to optimize the ecological protection objective and the energy efficiency objective, generating a hydropower coupled scheduling strategy, including: Define a set of multi-objective functions, which includes the objectives of maximizing the compliance rate of watershed ecological flow, maximizing the comprehensive utilization efficiency of hydropower, maximizing the absorption rate of new energy sources, and maximizing the reliability of power supply. Define a set of constraints, which includes reservoir flood control limit water level constraints, hydropower unit output limit constraints, power grid cross-sectional transmission capacity constraints, and ecological base current hard constraints. An adaptive weighted multi-objective game algorithm is embedded to dynamically adjust the first weight coefficient corresponding to the ecological protection objective and the second weight coefficient corresponding to the energy efficiency objective. A non-dominated sorting genetic algorithm is used to iteratively optimize the set of multi-objective functions and the set of constraints to generate a Pareto optimal solution set. Obtain power grid dispatch demand data and ecological assessment results data, and select the optimal hydropower coupling dispatch strategy from the Pareto optimal solution set based on the power grid dispatch demand data and the ecological assessment results data.

6. The method according to claim 1, characterized in that, The aforementioned implementation of hydropower bidirectional conversion buffer regulation, based on the dynamic changes in new energy output and grid load, controls pumped storage units and hydropower units to perform bidirectional conversion of electrical energy and water, in order to smooth energy system fluctuations, including: Real-time monitoring of wind and solar power output data and grid load data to determine whether there is a situation where the output of new energy sources is excessive and the grid load is lower than the preset lower load threshold. When it is determined that there is an overcapacity of new energy output, the pumped storage units are started to convert the excess electrical energy into hydropower and store it in the upstream reservoir. The reservoir capacity regulation capacity of the cascade reservoirs is used to smooth out the fluctuations in new energy output. When it is determined that the output of new energy sources is insufficient or the grid load is higher than the preset load limit threshold, the cascade hydropower stations will be dispatched to generate more water energy and convert it into electricity to supplement the power supply gap of the grid. Based on the real-time perception of the physical system's operating status by the digital twin, the water energy bidirectional conversion buffer control strategy is updated according to a preset rolling time interval. By combining runoff forecast data, the water storage and release schedules of upstream reservoirs can be planned in advance, enabling the water energy conversion schedule to be coordinated with the fluctuation curves of new energy output and the changes in grid load in a timely and spatial manner.

7. The method according to claim 1, characterized in that, The collection of historical scheduling data and real-time operational feedback, combined with the iterative optimization of the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy using a reinforcement learning agent, forms a closed-loop energy management system, including: A deep reinforcement learning agent was constructed, using historical hydropower coupled scheduling data, real-time operation status data, and regulation effect evaluation data as training sample sets. The training sample set is classified and labeled according to the hydrological scenarios of the high water season, the normal water season, the low water season, the peak-valley load energy scenario, and the new energy output fluctuation scenario. The deep reinforcement learning agent is trained using the training sample set after classification and labeling, so that the deep reinforcement learning agent learns the optimal control strategy under different hydrological scenarios and different energy scenarios. Based on the daily scheduling and execution feedback data, the coupling matching parameters of the hydropower coupling dynamic response model, the dynamic weight parameters of the multi-objective game algorithm, and the logical parameters of the hydropower bidirectional conversion buffer regulation are iteratively updated by the deep reinforcement learning agent. The model parameters of the deep reinforcement learning agent are fully updated according to a preset quarterly cycle, so as to realize the self-adaptation and self-optimization of the scheduling strategy.

8. An energy management device for hydroelectric coupling control, characterized in that, The device includes: The processing module is used to acquire watershed hydrological data, energy data and ecological data, construct a water-energy coupled sensing dataset, and establish a digital twin of watershed hydrology and energy coupling based on the dataset, quantify the spatiotemporal coupling correlation between runoff, water level changes and energy output and load demand, and generate a water-energy coupled dynamic response model. Based on the aforementioned hydropower coupling dynamic response model, a three-tiered collaborative decision-making and control architecture at the basin level, station level, and equipment level is deployed, and a hierarchical control strategy including medium- and long-term scheduling plans and real-time adjustment strategies is formulated. Under the hard constraint of ecological flow, a multi-objective game algorithm is used to optimize the game between ecological protection objectives and energy efficiency objectives, and a hydropower coupling scheduling strategy is generated. The execution module is used to perform the two-way conversion buffer regulation of hydropower. Based on the dynamic changes in the output of new energy sources and the load of the power grid, it controls the pumped storage units and hydropower units to carry out the two-way conversion of electrical energy and water, so as to smooth the fluctuations of the energy system. Historical scheduling data and real-time operational feedback are collected, and the hydropower coupling dynamic response model and the hydropower coupling scheduling strategy are iteratively optimized through a reinforcement learning agent to form a closed-loop energy management system.

9. A computer device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the energy management method for hydroelectric coupling control as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the energy management method for hydroelectric coupling control as described in any one of claims 1 to 7.

Images