Intensive intelligent management and control method and system for new energy power station group

Abstract

The invention discloses an intensive intelligent management and control method and system for a new energy power station group, and relates to the technical field of operation monitoring of new energy power stations, and the method is executed by the intensive intelligent management and control system deployed in a cloud-side collaborative architecture, and comprises the steps: obtaining and fusing operation data from a plurality of heterogeneous new energy power stations, forming a unified data view of a power station group level; based on the unified data view, generating prediction information of future power generation power of the power station group and evaluation information of the health state of key equipment; and based on the prediction information and the evaluation information, generating a collaborative optimization control strategy for the power station group, and distributing a control instruction to each power station for execution. According to the invention, cross-power-station cooperative management and control can be realized, global operation optimization is guaranteed, operation and maintenance cost is effectively reduced, and large-scale intensive operation requirements of the new energy power station group are met.

Description

Technical Field

[0001] This invention relates to the field of new energy power plant operation monitoring technology, and in particular to an intensive intelligent management and control method and system for new energy power plant clusters. Background Technology

[0002] New energy sources, represented by wind power and photovoltaics, have entered a new stage of large-scale and clustered development. The planned installed capacity of wind and photovoltaic power generation in various regions is constantly increasing. While this growth is driving the transformation of the energy structure, it also poses unprecedented challenges to the operation and management of power plants. The traditional decentralized management model of one plant, one control unit leads to high operation and maintenance costs and a lack of cluster collaboration capabilities, which has become a key bottleneck restricting the improvement of industry quality and efficiency and high-quality development.

[0003] Currently, the operation and management of new energy power plant clusters mainly face the challenges of decentralized operation modes and siloed decision-making. Specifically, each new energy power plant operates independently, with its monitoring system having limited functionality, resulting in multiple information silos. This prevents operators from obtaining a unified, real-time, and accurate panoramic view of the power plant cluster's operation. All analysis and decision-making are based solely on local data and limited experience from individual power plants. This model cannot perform cross-power plant collaborative power prediction, nor can it formulate collaborative control strategies that consider global optimization, leading to low overall power system efficiency and failing to meet the needs of intensive operation of new energy power plant clusters. Summary of the Invention

[0004] To address the problem that existing technologies using independent operation modes for new energy power plant clusters cannot perform cross-station collaborative power prediction, nor can they formulate globally optimal collaborative control strategies, leading to low overall power system efficiency, this invention provides an intensive intelligent management and control method and system for new energy power plant clusters. This method enables cross-station collaborative management and control, ensures optimal global operation, effectively reduces operation and maintenance costs, and meets the needs of large-scale intensive operation of new energy power plant clusters. The specific technical solution is as follows: This invention provides an intensive intelligent management and control method for a new energy power plant cluster. The method is executed by an intensive intelligent management and control system deployed in a cloud-edge collaborative architecture, and includes the following steps: Acquire and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant cluster level; Based on the unified data view, predictive information on the future power generation of the power plant group and assessment information on the health status of key equipment are generated. Based on the predicted information and the evaluated information, a collaborative optimization control strategy for the power plant group is generated, and control commands are distributed to each power plant for execution.

[0005] Preferably, the acquisition and fusion of operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant cluster level includes: By deploying edge computing gateways in various power plants, the raw data collected based on different communication protocols is parsed and converted into different formats. The parsed and formatted data is uploaded to the cloud platform, and data mapping and association are performed based on a predefined unified information model to form the unified data view.

[0006] Preferably, after forming the unified data view, the method further includes: Based on the unified data view, establish and update the digital twin model corresponding to the power plant group; In the digital twin model, for missing or delayed actual collected data, virtual sensing and state inference are performed through a mechanism model or a data-driven model to fill in the blind spots or delays of the actual collected data, thereby obtaining the complete unified data view.

[0007] Preferably, based on the unified data view, generating predictive information on the future power generation of the power plant cluster includes: The meteorological data, historical power data and spatial correlation features in the unified data view are input into the multi-timescale power prediction model, and the multi-timescale power prediction model simultaneously outputs at least three types of prediction results with different time resolutions and prediction durations. Among them, the first type of prediction result corresponds to the first prediction duration with high temporal resolution, which is used for real-time power control; The second type of prediction result corresponds to a second prediction duration with medium time resolution, which is longer than the first prediction duration, and is used for day-ahead scheduling. The third type of prediction result corresponds to a third prediction duration with low temporal resolution. The third prediction duration is longer than the second prediction duration and is used for medium- and long-term operation planning.

[0008] Preferably, based on the unified data view, generating assessment information on the health status of key equipment includes: Based on the real-time operating parameters of the devices in the unified data view, calculate the index values ​​that characterize their health status; The indicator value is compared with a preset health benchmark threshold. When the comparison results meet the preset warning conditions, evaluation information including failure probability, component location, and remaining life prediction is generated.

[0009] Preferably, based on the predicted information and the evaluated information, a collaborative optimization control strategy for the power plant group is generated, and control commands are distributed to each power plant for execution, including: With the goal of minimizing the overall operating cost of the power plant group, an optimization decision-making model is constructed, which includes electricity market prices, equipment health, grid operation safety, and equipment health constraints. The power forecast value and electricity market price forecast in the forecast information, as well as the equipment health index in the evaluation information, are used as key parameters to input into the optimization decision model and solve it to obtain the optimized power setpoint for each power station. The optimized power setting value is converted into equipment control commands adapted to the local control systems of each power station and then issued for execution.

[0010] Preferably, the objective function of the optimization decision model is expressed as: in, To optimize the total number of time periods in the cycle, The total number of power stations, For power station During the period The planning has contributed its efforts. For the power generation cost function, To be related to output and equipment health index The relevant maintenance cost function, For time period Electricity market forecasts electricity prices.

[0011] Preferably, a method for intensive intelligent management and control of a new energy power plant cluster further includes: When the grid frequency is detected to deviate from the preset rated value, based on the adjustment rate characteristics of each unit in the power plant group, at least one first type of unit is selected to adjust its output within the first time window. The first type of unit includes energy storage systems and / or generator sets; After the power grid frequency recovers to the preset rated value, the output setting value of each unit in the subsequent time window is recalculated based on the optimization decision model, so as to enable the power plant group to smoothly transition to the cost-optimal operating state.

[0012] This invention also provides an intensive intelligent management and control system, which applies the aforementioned intensive intelligent management and control method for a new energy power plant cluster, including: The edge perception layer consists of edge computing gateways deployed in each power station, used to acquire operational data from multiple heterogeneous new energy power stations; The cloud-based intelligent layer is used to receive and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant group level; based on the unified data view, it generates prediction information on the future power generation of the power plant group and assessment information on the health status of key equipment; based on the prediction information and the assessment information, it generates a collaborative optimization control strategy for the power plant group and distributes control commands to each power plant for execution. The data transmission layer is used to connect the edge perception layer and the cloud intelligence layer.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides an intensive intelligent management and control method for new energy power plant clusters. By fusing and aggregating operational data from multiple heterogeneous power plants, it offers operators a unified, real-time, and accurate panoramic operational view of the power plant cluster, solving the problem of the inability to comprehensively control operational status under traditional models. Cluster-level power generation prediction and equipment health assessment information generated based on panoramic data replace the traditional experience-driven model. By generating and distributing globally collaborative optimization control strategies, it achieves cross-power plant collaborative management and control, ensuring optimal global operation, effectively reducing operation and maintenance costs, and meeting the intensive operation needs of large-scale new energy power plant clusters. This invention effectively solves the problems of high operation and maintenance costs and lack of collaborative capabilities in new energy power plant clusters under the traditional decentralized management model, improving the overall efficiency and intensive operation level of the power system. Attached Figure Description

[0014] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0015] Figure 1 This is a flowchart of the method of the present invention.

[0016] Figure 2 This is a system schematic diagram of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be understood that, when used in this specification, the terms “comprising” and “including” indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0019] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0020] It should also be further understood that the term "and / or" as used in this specification refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes such combinations.

[0021] Please see Figure 2 This invention provides an integrated intelligent management and control system, comprising: The edge perception layer consists of edge computing gateways deployed in each power station, used to acquire operational data from multiple heterogeneous new energy power stations; The edge perception layer consists of a cluster of intelligent gateways deployed at each new energy power plant site. In the communication room or equipment layer of each new energy power plant, one or more industrial intelligent gateways are deployed as edge computing nodes. Based on their configuration, the gateways invoke the corresponding drivers to communicate with equipment such as wind turbines, inverters, and substation monitoring devices within the plant, parsing raw binary messages into structured data. Simultaneously, the intelligent gateways also perform data filtering, compression, and simple statistical calculations to optimize uplink bandwidth.

[0022] The cloud-based intelligent layer is used to receive and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant group level; based on the unified data view, it generates prediction information on the future power generation of the power plant group and assessment information on the health status of key equipment; based on the prediction information and the assessment information, it generates a collaborative optimization control strategy for the power plant group and distributes control commands to each power plant for execution. The cloud-based intelligence layer is deployed on enterprise private clouds or industry cloud platforms. It mainly consists of the following software subsystems: Data Platform: Responsible for data fusion. It uses a stream processing engine to clean and verify the incoming data in real time, and performs automated mapping and association based on a unified information model to build a standardized unified data view in the virtual space that fully corresponds to the physical power plant group.

[0023] AI Algorithm Platform: Responsible for intelligent analysis. Provides the training and inference environment for machine learning models. Deploys algorithm models for power prediction, equipment health assessment, and lifespan prediction. These models are invoked as microservices, generating predictive and evaluation information based on a unified data view.

[0024] The optimization decision-making and control center is responsible for collaborative decision-making. Based on market signals, scheduling instructions, and analysis results from the AI ​​platform, the optimization decision engine solves a multi-objective constrained optimization model to generate the globally optimal power setpoint. The control distribution service is responsible for converting the setpoint into specific device instructions and securely distributing them through the data transmission layer.

[0025] The data transmission layer is used to connect the edge perception layer and the cloud intelligence layer.

[0026] The data transmission layer is a logical channel comprised of intelligent gateways deployed at the edge computing edge, communication software stacks at the cloud entry point, and network security hardware. It ensures low-latency and reliable transmission of massive amounts of real-time data and control commands.

[0027] In practice, the cloud leverages its powerful computing and storage capabilities to execute complex tasks that are impossible at the edge. The data platform achieves data semantic consistency through a unified model, eliminating information silos. The AI ​​algorithm platform models based on panoramic rather than localized data, giving predictions and assessments a cluster perspective. The optimization decision-making and control center can coordinate multi-dimensional objectives such as economy, safety, and equipment status across the entire power station cluster, achieving global optimization that a single power station cannot accomplish.

[0028] The principle behind this system lies in its cloud-edge collaborative architecture. The edge perception layer standardizes heterogeneous data in real time, solving the problem of information silos; the cloud intelligence layer utilizes a unified data view and AI algorithm platform to achieve power prediction and equipment health assessment; finally, the optimization decision-making and control center generates and executes globally optimal control strategies based on multi-dimensional objectives such as economy, safety, and equipment status. This results in reduced operating costs.

[0029] Please see Figure 1 This invention also provides a method for intensive intelligent management and control of a new energy power plant cluster. The method is executed by the intensive intelligent management and control system provided in the foregoing embodiments and includes the following steps: Step S1: Acquire and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant group level; Data from power plants scattered across different geographical locations and from different vendors is integrated into a unified, standardized data center. This is achieved through the collaborative efforts of edge computing network management across each power plant and a cloud-based data platform, resulting in a unified data view of the power plant cluster that eliminates protocol differences, aligns timestamps, and verifies data quality.

[0030] Specifically, the acquisition and fusion of operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant cluster level includes: By deploying edge computing gateways in various power plants, the raw data collected based on different communication protocols is parsed and converted into different formats. The parsed and formatted data is uploaded to the cloud platform, and data mapping and association are performed based on a predefined unified information model to form the unified data view.

[0031] In the communication room or equipment layer of each new energy power plant, one or more smart gateways are deployed as edge computing nodes. Each gateway has a built-in protocol driver library that supports mainstream industry standard protocols as well as proprietary protocols from common equipment manufacturers. Configuration tools are used to assign corresponding drivers to each downstream device, such as the wind turbine main controller, photovoltaic inverter, and substation integrated automation system.

[0032] The intelligent gateway continuously receives data packets from downstream devices. Protocol-driven decoding extracts measurement points, their values, timestamps, and quality codes. The decoded data is then converted into a unified intermediate data format within the gateway. Through unified modeling, standardized integration of data from different manufacturers and device types is achieved, reducing system integration cycle time and costs.

[0033] Step S2: Based on the unified data view, generate prediction information on the future power generation of the power plant group and assessment information on the health status of key equipment; AI algorithm platforms deployed in the cloud. For example, using LSTM (Long Short-Term Memory) models for power prediction, and using Isolation Forest or deep autoencoders for device anomaly detection.

[0034] Specifically, based on the unified data view, generating predictive information on the future power generation of the power plant cluster includes: The meteorological data, historical power data and spatial correlation features in the unified data view are input into the multi-timescale power prediction model, and the multi-timescale power prediction model simultaneously outputs at least three types of prediction results with different time resolutions and prediction durations. The meteorological data includes high-precision numerical weather forecast data for future periods, as well as real-time satellite cloud images, radar precipitation and other observation data, which are spatially interpolated and corrected to the coordinates of each power station; the historical power data is the continuous historical active power output time series data of each station in the power station group; the spatial correlation characteristics are topological feature vectors that characterize the output synchronicity or complementarity between different stations by analyzing the geographical distribution of the power station group and the historical output sequence.

[0035] Among them, the first type of prediction result corresponds to the first prediction duration with high temporal resolution, which is used for real-time power control; The initial forecast period covers the next 0 to 4 hours, with a time resolution typically of 15 minutes or 5 minutes. A Temporal Convolutional Network (TCN) is used to perform rolling updates and forecasts based on the latest real-time meteorological data and very short historical power data. The output power curve is used for real-time power control.

[0036] The second type of prediction result corresponds to a second prediction duration with medium time resolution, which is longer than the first prediction duration, and is used for day-ahead scheduling. The second forecast period covers the next 72 hours, with a time resolution typically of 15 minutes or 1 hour. Deep neural networks (DNNs) are usually fed with numerical weather prediction (NWP) data as input. They are run 1-2 times daily, outputting detailed power output plans for the next three days. The results are used to develop day-ahead dispatch plans, submit day-ahead electricity market transaction requests, and schedule unit start-ups, shutdowns, and maintenance.

[0037] The third type of prediction result corresponds to a third prediction duration with low temporal resolution. The third prediction duration is longer than the second prediction duration and is used for medium- and long-term operation planning.

[0038] The third forecast period covers the next 7 days to 1 month, with a time resolution typically daily or weekly. It employs a statistical learning model, focusing on analyzing long-term meteorological trends and historical statistical patterns. The output is the expected daily or weekly average power generation for the next few weeks or months. It is primarily used for medium- to long-term power generation revenue assessment, fuel and operation and maintenance resource procurement planning, and monthly power generation plan reporting to the grid.

[0039] Specifically, based on the unified data view, the generation of assessment information on the health status of key equipment includes: Based on the real-time operating parameters of the devices in the unified data view, calculate the index values ​​that characterize their health status; For equipment In time Its comprehensive health index By its The results are obtained by fusing and calculating key operating parameters. One implementation method is a model based on weighted normalization and nonlinear transformation: in, Indicates equipment No. The parameters in time Measured values, such as vibration amplitude, temperature, and current harmonic distortion rate; and These are the historical mean and standard deviation of this parameter under the equipment's health baseline condition, used for normalization; It is aimed at the first A nonlinear mapping function with one parameter is used to capture the nonlinear relationship between parameter deviation and health decline. It is the first The weights of each parameter satisfy... This can be determined through expert experience or machine learning; It is a Sigmoid class function used to map the composite score to... interval, Represents perfect health. This indicates a complete failure.

[0040] The indicator value is compared with a preset health benchmark threshold. The early warning mechanism in this embodiment is based on the deviation of the health index from a dynamic benchmark. A time window is defined. The health index series within is A dynamic benchmark model can be established using moving average or exponential smoothing models. : in, This is a smoothing factor.

[0041] When the comparison results meet the preset warning conditions, evaluation information including failure probability, component location, and remaining life prediction is generated.

[0042] Calculate the standardized deviation of the current health index from the dynamic benchmark. : in, It is an estimate of the standard deviation of the baseline model under historical health conditions; when ( When the threshold is set to a preset value, an early warning is triggered.

[0043] Based on the trajectory of health index decline, models such as the Weibull proportional hazards model can be used to estimate future health outcomes. Conditional probability of failure occurring within a time period .

[0044] By fitting the degradation curve of the health index, the system can predict when it will first reach the failure threshold. The time, i.e., remaining lifespan. .

[0045] Step S3: Based on the predicted information and the evaluation information, generate a collaborative optimization control strategy for the power plant group, and distribute control commands to each power plant for execution.

[0046] Based on analysis and forecasting, operational strategies are developed to minimize the operating costs or maximize the benefits of the entire power plant group, and control commands are issued.

[0047] Specifically, based on the predicted information and the evaluated information, a collaborative optimization control strategy for the power plant group is generated, and control commands are distributed to each power plant for execution, including: With the goal of minimizing the overall operating cost of the power plant group, an optimization decision-making model is constructed, which includes electricity market prices, equipment health, grid operation safety, and equipment health constraints. The objective function is to minimize the total cost. in, To optimize the total number of time periods in the cycle, The total number of power stations, For power station During the period The planning has contributed its efforts. For the power generation cost function, To be related to output and equipment health index The relevant maintenance cost function, For time period Electricity market forecasts electricity prices.

[0048] The set of constraints includes: Power balance constraints: in, For power plant groups during time periods The overall dispatching instructions; Equipment operating constraints: in, To be related to health index The relevant output limit function, For power station The slope rate limit; Network topology security constraints: in, The power transfer distribution factor matrix, For time period The output vectors of each power station This is the limit vector for line transmission; The power forecast value and electricity market price forecast in the forecast information, as well as the equipment health index in the evaluation information, are used as key parameters to input into the optimization decision model and solve it to obtain the optimized power setpoint for each power station. The power prediction value provided by the prediction information As a constraint for equipment operation One of the input criteria, and the equipment health index provided by the evaluation information. As a cost function and output limit function Using the input parameters, the optimization decision model is solved to obtain the optimal power setpoint for each power station. .

[0049] The optimization decision-making model, aimed at minimizing the overall operating cost of a power plant group, essentially transforms complex real-world operational problems into computable mathematical optimization problems and applies efficient algorithms to find global or near-optimal solutions. The solution process fundamentally involves searching for the combination of decision variables that minimizes the objective function within the feasible region comprised of all constraints. A core algorithm typically combines the branch-and-bound method with interior-point or simplex methods to find the optimal solution within a finite time frame. Details are as follows: The nonlinear terms that may exist in the model are transformed into the standard form of mixed integer linear programming (MILP) or quadratic programming (QP) by piecewise linearization or quadratic term processing. For ultra-large-scale power plant groups, a decomposition-coordination strategy is adopted. The global problem is decomposed into multiple subproblems that are easier to solve by space or time. Coordination variables are introduced through Lagrange relaxation or objective cascading methods, and iterative updates are performed between the main subproblems until convergence to the global optimum or an acceptable suboptimal solution.

[0050] The optimization iteration is not completed in one go. The system performs a complete optimization calculation at fixed intervals, but only implements the decision instructions for the current period. In the next cycle, the system re-solves the optimization problem based on the latest measured data and updated prediction information.

[0051] The power setting value is converted into specific equipment control commands and then issued.

[0052] The optimal power setting value The commands are converted into control instructions that conform to the communication protocols of each power station's equipment and then issued.

[0053] This invention provides an intensive intelligent management and control method for new energy power plant clusters. By fusing and aggregating operational data from multiple heterogeneous power plants, it offers operators a unified, real-time, and accurate panoramic operational view of the power plant cluster, solving the problem of the inability to comprehensively control operational status under traditional models. Cluster-level power generation prediction and equipment health assessment information generated based on panoramic data replace the traditional experience-driven model. By generating and distributing globally collaborative optimization control strategies, it achieves cross-power plant collaborative management and control, ensuring optimal global operation, effectively reducing operation and maintenance costs, and meeting the intensive operation needs of large-scale new energy power plant clusters. This invention effectively solves the problems of high operation and maintenance costs and lack of collaborative capabilities in new energy power plant clusters under the traditional decentralized management model, improving the overall efficiency and intensive operation level of the power system.

[0054] In other embodiments, the goal is to maximize the overall benefits of the power plant group. The process of generating a collaborative optimization control strategy for the power plant group based on the predicted information and the evaluation information further includes: With the goal of maximizing the overall economic benefits of the power plant cluster, an optimization decision-making model is constructed, which includes electricity market prices, equipment health constraints, and network topology. By defining the optimization period as Each time period, such as 96 15-minute intervals. Let... For the number of power plants, the decision variable is... Indicates power station During the period They contributed to the plan.

[0055] The objective function for maximizing benefits is expressed as: in, It is a time period The current market electricity price; It refers to the length of the time period; It is a time period No. The price of various ancillary services; It is based on the power output vector of the power plant group The provided first This type of ancillary service volume is The function; It is a power station The operation and maintenance cost function is related to the current output. and equipment health index Negative correlation, meaning that the lower the health level, the higher the equivalent maintenance cost for the same output.

[0056] The constraints include: Power balancing and dispatching instructions: in, This is a superior dispatch instruction. This refers to the allowable deviation range.

[0057] Equipment operating limits: The maximum / minimum output limits of the equipment are its health index. The function's output limit may decrease when health declines.

[0058] Slope rate constraint: Cybersecurity constraints: Nonnegativity constraint: Using the predicted information and the evaluation information as input parameters, the optimization decision model is solved to obtain the power setpoint for each power station.

[0059] The generated forecast information (such as future electricity prices) Wind and solar power forecast As Part of) and assessment information (i.e., real-time health indices of each power station) Used for dynamic adjustment and , serving as the input parameters for the aforementioned model. This optimization problem is typically a mixed-integer linear programming (MILP) or nonlinear programming (NLP) problem, which can be efficiently solved using commercial solvers such as CPLEX and Gurobi, or algorithms such as the interior-point method, to obtain the optimal solution set { Ultimately, the optimal planned output for each time period will be determined. Convert them into specific device control commands and issue them.

[0060] This embodiment maximizes the overall economic benefits of the power plant cluster while managing asset risks. The model not only responds to dispatch commands but also proactively integrates multi-dimensional factors such as market prices and equipment health for dynamic optimization. Under the premise of ensuring grid security, it maximizes power generation revenue by participating in the electricity market and ancillary services market; by considering equipment health constraints, it automatically avoids overload operation of unhealthy equipment, achieving an optimal balance between power generation revenue and asset protection.

[0061] Specifically, in a preferred embodiment of this application, the construction of the unified data view further includes: Based on the unified data view, establish and update the digital twin model corresponding to the power plant group; In the digital twin model, for missing or delayed actual collected data, virtual sensing and state inference are performed through a mechanism model or a data-driven model to fill in the blind spots or delays of the actual collected data, thereby obtaining the complete unified data view.

[0062] In the cloud, a corresponding digital twin is created for each physical entity. This twin not only stores static parameters and real-time data but also runs a mechanistic model reflecting its physical characteristics. When real-time data from a sensor is lost due to communication interruption, the digital twin can deduce the possible numerical sequence of that measurement point within that time period based on the last received valid data, combined with the mechanistic model and data from nearby sensors. This generates virtual, reasonable data points in a unified data view, ensuring uninterrupted downstream analysis.

[0063] This preferred embodiment uses a mechanistic model and a data-driven model for virtual sensing, which can still provide high-fidelity state estimation when some data is missing, avoiding overall functional degradation due to single-point data failure. It is especially suitable for new energy power stations in mountainous areas with complex communication conditions.

[0064] Specifically, in a preferred embodiment of this application, a method for intensive intelligent management and control of a new energy power plant cluster further includes: When the grid frequency is detected to deviate from the preset rated value, based on the adjustment rate characteristics of each unit in the power plant group, at least one first type of unit is selected to adjust its output within the first time window. The first type of unit includes energy storage systems and / or generator sets with rapid adjustment capabilities; After the power grid frequency recovers to the preset rated value, the output setting value of each unit in the subsequent time window is recalculated based on the optimization decision model, so as to enable the power plant group to smoothly transition to the cost-optimal operating state.

[0065] Before the actual issuance of commands, the system uses the planned power setpoints and current equipment status as inputs to drive a rapid simulation of the digital twin model. The simulation simulates the dynamic response of the entire system under these commands, checking for risks such as electrical quantity exceeding limits, equipment overload, or control instability. If the simulation detects a risk, an alarm is triggered, and the risky command is sent back to the optimization decision-making process for recalculation, or intervention by operators is required.

[0066] This preferred embodiment avoids system risks caused by improper control strategies, enabling the control system to continuously learn and improve itself, becoming increasingly intelligent and safer over the long term.

[0067] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0068] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0069] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0070] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, 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 steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the specification of the present invention.

Claims

1. A method for intensive intelligent management and control of a new energy power plant cluster, characterized in that, The method is executed by an intensive intelligent management and control system deployed in a cloud-edge collaborative architecture, and includes the following steps: Acquire and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant cluster level; Based on the unified data view, predictive information on the future power generation of the power plant group and assessment information on the health status of key equipment are generated. Based on the predicted information and the evaluation information, a collaborative optimization control strategy for the power plant group is generated, and control commands are distributed to each power plant for execution.

2. The intensive intelligent management and control method for a new energy power plant cluster according to claim 1, characterized in that, The acquisition and fusion of operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant cluster level includes: By deploying edge computing gateways in various power plants, the raw data collected based on different communication protocols is parsed and converted into different formats. The parsed and formatted data is uploaded to the cloud platform, and data mapping and association are performed based on a predefined unified information model to form the unified data view.

3. The intensive intelligent management and control method for a new energy power plant cluster according to claim 2, characterized in that, After forming the unified data view, it also includes: Based on the unified data view, establish and update the digital twin model corresponding to the power plant group; In the digital twin model, for missing or delayed actual collected data, virtual sensing and state inference are performed through a mechanism model or a data-driven model to fill in the blind spots or delays of the actual collected data, thereby obtaining the complete unified data view.

4. The intensive intelligent management and control method for a new energy power plant cluster according to claim 1, characterized in that, Based on the unified data view, the generation of prediction information for the future power generation of the power plant group includes: The meteorological data, historical power data and spatial correlation features in the unified data view are input into the multi-timescale power prediction model, and the multi-timescale power prediction model simultaneously outputs at least three types of prediction results with different time resolutions and prediction durations. Among them, the first type of prediction result corresponds to the first prediction duration with high temporal resolution, which is used for real-time power control; The second type of prediction result corresponds to a second prediction duration with medium time resolution, which is longer than the first prediction duration, and is used for day-ahead scheduling. The third type of prediction result corresponds to a third prediction duration with low temporal resolution. The third prediction duration is longer than the second prediction duration and is used for medium- and long-term operation planning.

5. A method for intensive intelligent management and control of a new energy power plant cluster according to claim 1 or 4, characterized in that, Based on the unified data view, the assessment information for the health status of key equipment is generated, including: Based on the real-time operating parameters of the devices in the unified data view, calculate the index values ​​that characterize their health status; The indicator value is compared with a preset health benchmark threshold. When the comparison results meet the preset warning conditions, evaluation information including failure probability, component location, and remaining life prediction is generated.

6. The intensive intelligent management and control method for a new energy power plant cluster according to claim 5, characterized in that, Based on the predicted information and the evaluated information, a collaborative optimization control strategy for the power plant group is generated, and control commands are distributed to each power plant for execution, including: With the goal of minimizing the overall operating cost of the power plant group, an optimization decision-making model is constructed, which includes electricity market prices, equipment health, grid operation safety, and equipment health constraints. The power forecast value and electricity market price forecast in the forecast information, as well as the equipment health index in the evaluation information, are used as key parameters to input into the optimization decision model and solve it to obtain the optimized power setpoint for each power station. The optimized power setting value is converted into equipment control commands adapted to the local control systems of each power station and then issued for execution.

7. The intensive intelligent management and control method for a new energy power plant cluster according to claim 6, characterized in that, The objective function of the optimization decision model is expressed as: in, To optimize the total number of time periods in the cycle, The total number of power stations, For power station During the period The planning has contributed its efforts. For the power generation cost function, To be related to output and equipment health index The relevant maintenance cost function, For time period Electricity market forecasts electricity prices.

8. The intensive intelligent management and control method for a new energy power plant cluster according to claim 6, characterized in that, Also includes: When the grid frequency is detected to deviate from the preset rated value, based on the adjustment rate characteristics of each unit in the power plant group, at least one first type of unit is selected to adjust its output within the first time window. The first type of unit includes energy storage systems and / or generator sets; After the power grid frequency recovers to the preset rated value, the output setting value of each unit in the subsequent time window is recalculated based on the optimization decision model, so as to enable the power plant group to smoothly transition to the cost-optimal operating state.

9. An intensive intelligent control system, characterized in that, The method for intensive intelligent management and control of a new energy power plant cluster according to any one of claims 1-8 includes: The edge perception layer consists of edge computing gateways deployed in each power station, used to acquire operational data from multiple heterogeneous new energy power stations; The cloud-based intelligent layer is used to receive and integrate operational data from multiple heterogeneous new energy power plants to form a unified data view at the power plant group level; based on the unified data view, it generates prediction information on the future power generation of the power plant group and assessment information on the health status of key equipment; based on the prediction information and the assessment information, it generates a collaborative optimization control strategy for the power plant group and distributes control commands to each power plant for execution. The data transmission layer is used to connect the edge perception layer and the cloud intelligence layer.

Images