Virtual power plant aggregation regulation method and device facing multiple uncertainties, equipment and medium

Abstract

This application discloses a method, device, equipment, and medium for aggregated control of a virtual power plant facing multiple uncertainties, relating to the field of smart grid technology. The method includes: constructing a resource state vector by collecting various resource data from the virtual power plant; generating and compressing a representative set of scenarios with multiple uncertainties; calculating response reliability by combining resource state and historical response data; deriving a reliable adjustable domain; solving the aggregated control plan by coupling the scenario set and the reliable domain with the goal of minimizing the overall cycle cost; generating and verifying the issuance and execution of equipment control commands in a hierarchical manner; collecting execution deviations to roll and correct the control plan; dynamically updating resource reliability; and forming a closed-loop control throughout the entire process. Through rolling correction and reliability updates, a closed-loop adaptive control is formed, balancing economy and reliability, and improving the virtual power plant's performance and control stability in complex scenarios.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to a method, apparatus, equipment and medium for virtual power plant aggregation control facing multiple uncertainties. Background Technology

[0002] Existing virtual power plant aggregation control technologies mainly include the following categories: First, deterministic optimization control methods, which directly use single-point predicted values ​​of renewable energy, load, and electricity prices to construct the scheduling model without additional consideration of various prediction deviations and equipment and communication disturbances; Second, basic stochastic scenario optimization methods, which generate a small number of renewable energy and load scenarios based on historical errors and solve for the optimal control plan through scenario expectations; Third, robust / partially robust optimization methods, which limit the fluctuation range of uncertain variables and generate conservative control schemes under the worst operating conditions; Fourth, cloud-edge layered coordinated control methods, which distribute computational and communication pressure through a three-layer architecture of platform, edge, and equipment.

[0003] While existing virtual power plant aggregation and control technologies can model, evaluate, and schedule some distributed resources, they still have the following shortcomings in complex operating scenarios with multiple uncertainties: Uncertainty factors are not comprehensively considered. Existing technologies typically focus on the uncertainty of renewable energy output or load forecasting errors, rarely modeling factors such as renewable energy output, load demand, market prices, user response behavior, equipment availability, communication latency, and execution deviations in a unified manner. In actual operation, these uncertainties often occur simultaneously and influence each other. Considering only a single uncertainty can easily lead to a mismatch between the virtual power plant aggregation and control plan and the actual executable capacity. Aggregation and control capabilities are mostly statically assessed, making it difficult to reflect changes in resource status. Regulation capacity assessment methods typically estimate the regulation capacity of a virtual power plant based on equipment capacity, historical response data, or fixed models. However, the internal resources of a virtual power plant are significantly time-varying. If static aggregation capacity assessment is used, it is easy to overestimate or underestimate the actual regulation capacity of the virtual power plant. Regulation commands lack executability verification. Some existing technologies focus on the aggregation output optimization at the virtual power plant level, but do not adequately consider the decomposition of resource-side commands, equipment constraints, user constraints, and communication execution constraints. The upward or downward adjustment targets obtained at the virtual power plant level may be feasible in the mathematical model, but they cannot be executed when decomposed to specific air conditioning loads, energy storage devices, or charging piles due to equipment power boundaries, ramping constraints, state of charge constraints, user comfort constraints, or response delays.

[0004] Therefore, how to accurately assess the reliability and adjustability of virtual power plants under multiple uncertainties and generate closed-loop aggregated control schemes that take into account economy, reliability and equipment feasibility has become an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, equipment, and medium for the aggregated control of virtual power plants in the face of multiple uncertainties, aiming to solve the technical problem of how to accurately assess the reliable and adjustable capabilities of virtual power plants and generate aggregated control schemes.

[0006] To achieve the above objectives, this application proposes a virtual power plant aggregation control method for multiple uncertainties, comprising: Acquire basic data, predictive data, operational data, historical response data, and communication status data of various distributed resources within the virtual power plant, and construct a resource status vector based on the basic data, predictive data, operational data, historical response data, and communication status data; Based on the predicted data, the operational data, the historical response data, and the communication status data, a set of multiple uncertainty scenarios is generated, and the set of multiple uncertainty scenarios is filtered and compressed to obtain a representative scenario set; The response reliability of each resource is calculated based on the resource state vector and the historical response data, and the reliable adjustable domain of the virtual power plant is calculated based on the response reliability and the resource state vector. Based on the aggregated regulation optimization objective function, the representative scenario set and the reliable adjustable domain are solved to obtain the aggregated regulation plan, wherein the aggregated regulation optimization objective function is to minimize the total comprehensive cost within the regulation period. Based on the aggregated control plan and the trusted adjustable domain, a resource group control instruction is generated. The resource group control instruction is converted into a device-side control instruction. The device-side control instruction is then validated for executability. Once the validation is successful, the instruction is sent to the corresponding device so that the corresponding device can perform the corresponding control operation and provide feedback on the execution result. The execution results fed back by the device are collected, the control deviation is calculated based on the execution results, and the aggregate control plan for subsequent periods is rolled out based on the control deviation. At the same time, the response credibility is updated based on the execution results.

[0007] Furthermore, to achieve the above objectives, this application also proposes a virtual power plant aggregation control device oriented towards multiple uncertainties, the virtual power plant aggregation control device oriented towards multiple uncertainties comprising: The data acquisition module is used to acquire basic data, predictive data, operational data, historical response data, and communication status data of various distributed resources within the virtual power plant, and to construct a resource status vector based on the basic data, predictive data, operational data, historical response data, and communication status data. The scenario construction module is used to generate a set of multiple uncertain scenarios based on the prediction data, the running data, the historical response data and the communication status data, and to filter and compress the set of multiple uncertain scenarios to obtain a representative scenario set. The credibility calculation module is used to calculate the response credibility of each resource based on the resource state vector and the historical response data, and to calculate the credibility adjustable domain of the virtual power plant based on the response credibility and the resource state vector. The optimization and control module is used to solve the representative scenario set and the reliable adjustable domain based on the aggregated control optimization objective function to obtain the aggregated control plan, wherein the aggregated control optimization objective function is to minimize the total comprehensive cost within the control period. The instruction sending module is used to generate resource group control instructions according to the aggregated control plan and the trusted adjustable domain, convert the resource group control instructions into device-side control instructions, perform executability verification on the device-side control instructions, and send them to the corresponding device after the verification is passed, so that the corresponding device can perform the corresponding control operation and feed back the execution result. The feedback correction module is used to collect the execution results fed back by the device, calculate the control deviation based on the execution results, and make rolling corrections to the aggregate control plan for subsequent periods based on the control deviation, while updating the response credibility based on the execution results.

[0008] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the virtual power plant aggregation control method for multiple uncertainties described above.

[0009] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the virtual power plant aggregation control method for multiple uncertainties described above.

[0010] This application constructs a resource state vector by collecting various resource data from a virtual power plant, generating and compressing a representative set of scenarios with multiple uncertainties. It calculates response reliability by combining resource state and historical response data, deriving a reliable adjustable domain. With the goal of minimizing the overall cycle cost, it couples the scenario set and the reliable domain to solve for an aggregated control plan. It generates and verifies the issuance and execution of equipment control commands in a layered manner, collects execution deviations to continuously correct the control plan, and dynamically updates resource reliability, forming a closed-loop control process. Dynamic correction of response reliability avoids inflated aggregation capabilities, and the reliable adjustable domain constraint ensures the feasibility of the solution. Feasibility verification improves engineering feasibility, and rolling correction and reliability updates form a closed-loop adaptive control, balancing economy and reliability, and enhancing the virtual power plant's performance and control stability in complex scenarios. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the first embodiment of the virtual power plant aggregation control method for multiple uncertainties in this application. Figure 2 This is a flowchart illustrating the second embodiment of the virtual power plant aggregation control method for multiple uncertainties in this application. Figure 3 This is a schematic diagram of the module structure of the virtual power plant aggregation control device for multiple uncertainties in this application; Figure 4 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the virtual power plant aggregation control method for multiple uncertainties in the embodiments of this application.

[0013] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0015] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0016] Existing virtual power plant aggregation control technologies mainly include aggregation control methods based on deterministic prediction, stochastic optimization methods based on scenario generation, aggregation control methods based on robust optimization, and control methods based on multi-level coordination or cloud-edge collaboration. However, these existing technologies generally suffer from the following shortcomings: they do not comprehensively consider uncertainties, making it difficult to uniformly model various uncertainties such as renewable energy output, load demand, market prices, user response behavior, equipment availability, and communication latency; aggregation control capabilities are mostly statically assessed, failing to fully consider the time-varying characteristics of resources and historical performance, which easily leads to overestimation or underestimation of the actual control capabilities of the virtual power plant.

[0017] Based on the above, this application also provides a virtual power plant aggregation control method for multiple uncertainties, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the virtual power plant aggregation control method for multiple uncertainties in this application.

[0018] In this embodiment, the virtual power plant aggregation control method for multiple uncertainties includes steps S10~S60: Step S10: Obtain basic data, prediction data, operation data, historical response data, and communication status data of various distributed resources within the virtual power plant, and construct resource status vectors based on the basic data, prediction data, operation data, historical response data, and communication status data.

[0019] It should be noted that the resource state vector refers to a multi-dimensional data structure used to uniformly describe the current operating status and adjustability characteristics of each controllable resource within a virtual power plant. This includes current power, energy status, operating constraint status, equipment availability status, communication latency status, and historical response performance. Basic data refers to static parameters such as resource type, rated power, capacity, and access nodes. Forecast data refers to estimated data for future periods, such as photovoltaic power output, wind power output, load forecasts, and electricity price forecasts. Operational data refers to dynamic data reflecting the current operating status of resources, such as real-time power, energy storage state of charge, air conditioning temperature, and charging pile status. Historical response data refers to data reflecting past resource response performance, such as historical output error, historical load error, user response records, and equipment fault records. Communication status data refers to data reflecting communication link quality, such as equipment online status, command arrival time, execution latency, and execution success rate.

[0020] Further, step S10 includes: First, through the data acquisition interface of the virtual power plant operation platform, obtaining the rated capacity, predicted output, historical prediction error, and grid connection point location of distributed photovoltaic and wind power resources; obtaining the rated capacity, maximum charging and discharging power, current state of charge, and charging and discharging efficiency of the energy storage system; obtaining the current power, indoor temperature, upper and lower temperature limits, and user comfort constraints of the air conditioning load; obtaining the production plan, interruptible power, interruptible duration, and advance notice time of the industrial load; obtaining the vehicle access status, current power, target power, expected departure time, and charging power limit of the electric vehicle charging pile; and simultaneously obtaining grid dispatch instructions, demand response prices, ancillary service prices, market electricity prices, deviation assessment rules, communication link status, equipment online status, instruction execution success rate, and historical execution deviations.

[0021] Secondly, based on the resource type, rated power, capacity, and access node information in the basic data, as well as the real-time power, energy storage state of charge, air conditioning temperature, and charging pile status in the operational data, a basic physical parameter mapping for each resource is established. Specifically, distributed resources are divided into photovoltaic resources, wind power resources, energy storage resources, air conditioning load resources, industrial load resources, and charging pile resources according to resource type. A basic physical parameter mapping is established for each type of resource. The basic physical parameter mapping includes a static parameter layer and a dynamic parameter layer. The static parameter layer records the rated power, capacity, and access node information, while the dynamic parameter layer records the real-time power, energy storage state of charge, air conditioning temperature, and charging pile status.

[0022] Next, based on the predicted output of photovoltaic power, wind power, load, and electricity price from the forecast data, and the historical output error, historical load error, user response records, and equipment failure records from the historical response data, multi-timescale uncertainty characteristics of each resource are constructed. Specifically, first, the deviation sequence between the predicted value and the historical actual value of each resource at the day-ahead time scale is calculated to obtain the day-ahead uncertainty characteristics; second, the deviation sequence between the predicted value and the historical actual value of each resource at the intraday time scale is calculated to obtain the intraday uncertainty characteristics; then, the deviation sequence between the predicted value and the historical actual value of each resource at the real-time time scale is calculated to obtain the real-time uncertainty characteristics; finally, the day-ahead uncertainty characteristics, intraday uncertainty characteristics, and real-time uncertainty characteristics are combined to form the multi-timescale uncertainty characteristics of each resource.

[0023] Next, based on the device online status, command arrival time, execution latency, and execution success rate in the communication status data, the communication link quality index for each resource is calculated. In this example, the communication link quality index is calculated as follows: Communication Link Quality Index = Preset Online Weight × Device Online Status Score + Preset Latency Weight × Command Transmission Latency Score + Preset Success Rate Weight × Execution Success Rate Score. The device online status score is determined based on the percentage of online time the device has spent in the most recent preset statistical period (e.g., 24 hours). The command transmission latency score is determined based on the ratio of the difference between the command arrival time and the command issuance time to a preset latency threshold (e.g., 500ms). The execution success rate score is determined based on the ratio of the number of successful executions to the total number of commands issued in the most recent preset statistical period.

[0024] Then, based on the mapping of fundamental physical parameters, multi-timescale uncertainty characteristics, and communication link quality indicators, a multi-dimensional resource state vector is constructed. It's important to understand that this multi-dimensional resource state vector includes current power, energy state, operational constraint state, equipment availability state, communication delay state, historical response performance, uncertainty fluctuation amplitude, and communication link quality indicators. Current power is taken from the dynamic parameter layer of the fundamental physical parameter mapping; energy state is calculated based on the energy storage state of charge or air conditioning temperature; operational constraint state is determined based on the physical constraints corresponding to the resource type; equipment availability state is determined based on equipment fault records; communication delay state is determined based on command transmission delay; historical response performance is determined based on user response records; uncertainty fluctuation amplitude is determined based on the standard deviation in the multi-timescale uncertainty characteristics; and communication link quality indicators are derived from the above calculation results.

[0025] Finally, based on the coupling relationships between the dimensions in the multi-dimensional resource state vector, a resource state correlation matrix is ​​established. The multi-dimensional resource state vector is then subjected to consistency verification and dynamic updates based on this correlation matrix to obtain the resource state vector. In this example, the resource state correlation matrix is ​​established as follows: the correlation coefficient between current power and energy status is calculated, the correlation coefficient between equipment availability status and communication link quality indicators is calculated, and the correlation coefficient between historical response performance and uncertainty fluctuation amplitude is calculated. These correlation coefficients are then combined to form the resource state correlation matrix. Consistency verification of the multi-dimensional resource state vector is performed based on the resource state correlation matrix: if the correlation coefficient between two dimensions in the resource state correlation matrix exceeds a preset strong coupling threshold (0.8), it is determined whether there is a numerical contradiction between the two dimensions. If a contradiction exists, it is marked as an abnormal state and data re-sampling is triggered; if no contradiction exists, the consistency verification is passed. The multi-dimensional resource state vector that passes the consistency verification is dynamically updated: with a preset update cycle (15 minutes) as the interval, operating data and communication status data are re-collected to update the current power, energy status, equipment availability status, communication delay status, and communication link quality indicators in the multi-dimensional resource state vector, resulting in the resource state vector. The specific formula is as follows: in This represents the resource state vector of resource i during time period t; This represents the current power of resource i in time period t; This represents the resource energy state of resource i during time period t; This indicates the resource operation constraint state of resource i in time period t; This indicates the device availability status of resource i during time period t; This indicates the communication delay status of resource i during time period t; This represents the historical response performance of resource i during time period t; i represents the i-th resource within the virtual power plant, and t represents the control period.

[0026] Step S20: Based on the prediction data, operation data, historical response data and communication status data, generate a set of multiple uncertainty scenarios, and filter and compress the set of multiple uncertainty scenarios to obtain a representative scenario set.

[0027] It should be noted that the multiple uncertainty scenario set refers to a set of scenarios generated based on historical forecast errors, weather data, load data, market price data, user response records, and equipment availability records. These scenarios describe possible combinations of variables such as renewable energy output, load demand, electricity price, response rate, equipment status, and communication latency in future periods. The representative scenario set refers to a simplified set of scenarios obtained by filtering and compressing the initial multiple uncertainty scenario set, which can reduce computational scale while retaining key risk information.

[0028] Further, step S20 includes: First, constructing a multi-source uncertainty coupling generation model based on the historical prediction error distribution characteristics in the prediction data, the weather change patterns in the operational data, the user response behavior transfer patterns in the historical response data, and the equipment failure association patterns in the communication status data. In this example, the multi-source uncertainty coupling generation model refers to a scenario generation model constructed based on the statistical characteristics and association patterns of multiple uncertain factors in the prediction data, operational data, historical response data, and communication status data, used to generate scenarios that simultaneously contain the joint values ​​of multiple uncertainty deviations. The construction method of the multi-source uncertainty coupling generation model is as follows: kernel density estimation is performed on the historical prediction error distribution characteristics to obtain the probability density functions of new energy output deviation and load prediction deviation; Markov chain modeling is performed on the weather change patterns to obtain the weather state transition probability matrix; state transition probability calculation is performed on the user response behavior transfer patterns to obtain the state transition matrix of user response rate deviation; association rule mining is performed on the equipment failure association patterns to obtain the joint occurrence probability of equipment availability state deviation and communication delay deviation. By coupling the above probability density function, state transition probability matrix, state transition matrix and joint occurrence probability, a multi-source uncertainty coupled generation model is constructed. When generating a single scenario, this model simultaneously samples the deviations in new energy output, load forecasting, market price, user response rate, communication delay, and equipment availability status according to the coupling relationship, ensuring that the correlation between the deviations conforms to historical statistical patterns.

[0029] Secondly, an initial set of multiple uncertainty scenarios is generated based on the multi-source uncertainty coupling generation model. Using Monte Carlo sampling, a preset number of initial scenarios (1000) are extracted from the multi-source uncertainty coupling generation model. Each sample constitutes a scenario, and each scenario includes the combined values ​​of new energy output deviation (including photovoltaic output prediction deviation and wind power output prediction deviation), load prediction deviation, market price deviation, user response rate deviation, communication delay deviation, and equipment availability status deviation. The generated initial set of multiple uncertainty scenarios is then validated for completeness, removing abnormal scenarios containing missing values ​​or exceeding physically reasonable ranges. The validated initial set of multiple uncertainty scenarios is obtained, expressed by the following formula: in This represents the s-th uncertainty scenario within time period t; This indicates the prediction deviation of photovoltaic power output in the s-th scenario; This indicates the wind power output prediction deviation in the s-th scenario; This represents the load forecast deviation in the s-th scenario; This represents the market price deviation in the s-th scenario; This represents the deviation in user response rate in the s-th scenario; This represents the communication or execution delay deviation in the s-th scenario; This represents the deviation in device availability status under the s-th scenario; This represents the initial set of multiple uncertainties. This represents the s-th scene; This represents the probability of the s-th scenario occurring; Indicates the total number of scenes.

[0030] Next, based on the initial set of multiple uncertainty scenarios, a multi-dimensional risk impact index for each scenario is calculated. The multi-dimensional risk impact index is an indicator that quantifies the comprehensive impact of a single uncertainty scenario on the overall aggregation and regulation capability of the virtual power plant. The multi-dimensional risk impact index is determined based on the comprehensive impact of each scenario on the aggregation and regulation capability of the virtual power plant.

[0031] Next, hierarchical clustering is performed on the initial set of multiple uncertainties scenarios based on the multi-dimensional risk impact index. In this example, the K-means clustering algorithm is used for hierarchical clustering, and the number of clusters is determined based on the size of the initial set of multiple uncertainties scenarios and computational resource constraints, set to a preset number of 10 clusters. Scenarios with similar risk impact indices are grouped into the same scenario cluster, and the risk impact index of the cluster center for each scenario cluster is calculated. Within each scenario cluster, the scenario with the risk impact index closest to the cluster center and covering the maximum uncertainty coupling pattern is selected as the first representative scenario. The first representative scenario refers to the scenario selected from the regular risk scenario clusters through hierarchical clustering that can represent the typical uncertainty characteristics of the cluster. The maximum uncertainty coupling pattern indicates that the scenario simultaneously includes the largest number of types of new energy output deviation, load demand deviation, market price deviation, user response rate deviation, communication delay deviation, and equipment availability status deviation. If multiple scenarios simultaneously meet the conditions, the scenario with the risk impact index closest to the cluster center is selected.

[0032] Then, scenarios whose multi-dimensional risk impact index exceeds the preset high-risk threshold are screened to obtain the second representative scenarios. The second representative scenarios refer to those scenarios that are forcibly retained from the high-risk scenarios and have a critical impact on the safety of virtual power plant regulation. The preset high-risk threshold (0.7) is determined based on the following: when the multi-dimensional risk impact index exceeds this threshold, the negative impact of the scenario on the aggregation and regulation capability of the virtual power plant reaches a level that requires special attention. This threshold is determined by statistical analysis of the risk impact index of historical extreme scenarios, and the preset quantile (75%) of the distribution of the risk impact index of historical extreme scenarios is used as a reference.

[0033] Finally, the first and second representative scenarios are merged into a candidate scenario set, and the probabilities of the scenarios in the candidate scenario set are adaptively weighted and normalized to obtain the representative scenario set. Specifically, first, the original occurrence probabilities of the first and second representative scenarios in the initial multi-uncertainty scenario set are calculated; then, the probability of the second representative scenario is amplified by a preset high-risk weighting coefficient, and the probability of the first representative scenario is reduced by a preset regular weighting coefficient; finally, the weighted probabilities are normalized so that the sum of the probabilities of all scenarios in the candidate scenario set is 1, thus obtaining the representative scenario set.

[0034] Step S30: Calculate the response reliability of each resource based on the resource state vector and historical response data, and calculate the reliable adjustable domain of the virtual power plant based on the response reliability and resource state vector.

[0035] It should be noted that response reliability refers to the degree of reliability with which a resource completes a response according to the control command, and is calculated comprehensively based on historical performance, response speed, communication status, and equipment health status. The reliable adjustable domain refers to the range of upward or downward power that a virtual power plant can reliably provide within a certain period, considering equipment physical constraints, user energy consumption constraints, response reliability, and multiple uncertainties.

[0036] Further, step S30 includes: First, constructing a multi-dimensional response reliability evaluation matrix for each resource based on the historical performance score, response speed score, communication status score, and device health status score in the historical response data, as well as the historical response performance and communication latency status in the resource status vector. In this example, the multi-dimensional response reliability evaluation matrix is ​​constructed as follows: the historical performance score, response speed score, communication status score, and device health status score are used as row vectors of the matrix, and the values ​​of each resource at different time periods are used as column vectors to form the multi-dimensional response reliability evaluation matrix. The historical performance score is determined based on the ratio of the number of times the resource completes the response according to the instruction within the most recent preset statistical period (30 days) to the total number of times the instruction is issued; the response speed score is determined based on the ratio of the average time from receiving the instruction to starting to execute the response to the preset baseline response time (1 minute); the communication status score is determined based on the resource's communication link quality index; and the device health status score is determined based on the ratio of the fault-free running time within the most recent preset statistical period (7 days) in the device fault record to the total running time.

[0037] Secondly, based on the temporal correlation characteristics between the dimensions in the multi-dimensional response credibility evaluation matrix, the credibility decay coefficient of each resource in different time periods is calculated. The credibility decay coefficient is determined by the number of consecutive unresponsive periods and the cumulative magnitude of response deviation. The number of consecutive unresponsive periods is determined by the number of periods between the resource's most recent successful response and the current period. The cumulative magnitude of response deviation is determined by the sum of the absolute values ​​of the differences between the actual response amount and the target response amount in each period within the most recent preset statistical period (7 days). When the credibility decay coefficient exceeds the preset decay limit (2.0), it is truncated to the preset decay limit (2.0) to prevent excessive decay.

[0038] Next, the multi-dimensional response credibility evaluation matrix is ​​dynamically weighted and corrected based on the credibility decay coefficient to obtain the dynamic response credibility of each resource at different time periods. In this example, the dynamic weighting correction is performed as follows: the score of each dimension in the multi-dimensional response credibility evaluation matrix is ​​divided by the credibility decay coefficient to obtain the corrected score of each dimension, and then the corrected scores of each dimension are weighted and summed to obtain the dynamic response credibility of each resource at different time periods. The dynamic response credibility ranges from 0 to 1. When the dynamic response credibility is lower than the preset minimum credibility threshold (0.3), the resource is marked as a low-credibility resource, and its priority is reduced in subsequent control task allocation.

[0039] Next, based on the dynamic response reliability and the device availability state, power boundary, ramp constraint, and energy state in the resource state vector, the reliable up-adjustment capability and reliable down-adjustment capability of each resource at different time periods are calculated. The calculation method for the reliable up-adjustment capability and reliable down-adjustment capability is as follows: in This indicates the credible upscaling capability of resource i in time period t; This indicates the credible downscaling capability of resource i during time period t; Indicates the reliability of dynamic response; This indicates the availability of resources; a value of 1 indicates that the device is available, and a value of 0 indicates that it is unavailable. Indicates the maximum equivalent power of the resource; Indicates the minimum equivalent power of the resource; Indicates the baseline operating power of resources; This indicates the maximum ramp-up capability of resources; This indicates the maximum reduction in climbing capacity of resources; Indicates the time interval for regulation.

[0040] Then, based on the reliable up-adjustment and reliable down-adjustment capabilities of each resource, as well as the complementary adjustment characteristics between resources, a resource group collaborative adjustment matrix is ​​constructed. In this example, the resource group collaborative adjustment matrix is ​​constructed as follows: the complementary adjustment coefficient between energy storage resources and photovoltaic resources is calculated (energy storage discharge can compensate for the decline in photovoltaic output, so a positive value is taken); the complementary adjustment coefficient between air conditioning load resources and industrial load resources is calculated (the adjustment periods of the two types of loads can be staggered and complementary, so a positive value is taken); the complementary adjustment coefficient between charging pile resources and energy storage resources is calculated (the grid capacity released by the power reduction of charging piles can be utilized by energy storage charging, so a positive value is taken); and the above complementary adjustment coefficients are combined into the resource group collaborative adjustment matrix.

[0041] Finally, the trusted upward and trusted downward adjustment capabilities are aggregated and corrected based on the resource group collaborative adjustment matrix to obtain the trusted adjustable domain of the virtual power plant. In this example, the aggregation and correction method is as follows: for resources belonging to the same resource group and having complementary adjustment relationships, their trusted upward or trusted downward adjustment capabilities are multiplied by 1 and summed with the complementary adjustment coefficient to obtain the corrected trusted upward or trusted downward adjustment capabilities; for resources without complementary adjustment relationships, their trusted upward or trusted downward adjustment capabilities remain unchanged. The corrected trusted upward adjustment capabilities of each resource are summed to obtain the aggregated trusted upward adjustment capability of the virtual power plant; the corrected trusted downward adjustment capabilities of each resource are summed to obtain the aggregated trusted downward adjustment capability of the virtual power plant.

[0042] To further account for the risks arising from multiple uncertainties, a safety margin adjustment is made to the aggregate trust up-adjustment capability and aggregate trust down-adjustment capability: in This indicates the reliability of the security correction. This indicates the reliable downscaling capability after security modifications; This indicates an upward adjustment of the capacity safety margin; This indicates a reduction in the capacity safety margin.

[0043] Step S40: Solve the representative scenario set and the reliable adjustable domain based on the aggregated regulation optimization objective function to obtain the aggregated regulation plan.

[0044] It should be noted that the aggregated control plan refers to the overall output plan, regulation plan, and reserve plan formed by the virtual power plant at the day-ahead, intraday, or real-time stages. The aggregated control optimization objective function is a mathematical model that aims to minimize the sum of operating costs, user compensation costs, equipment loss costs, deviation assessment costs, and multiple uncertainty risk penalty costs within the control period. Specifically, the aggregated control optimization objective function aims to minimize the total comprehensive cost within the control period. The total comprehensive cost includes operating costs, user compensation costs, equipment loss costs, deviation assessment costs, and multiple uncertainty risk penalty costs.

[0045] Further, step S40 includes: First, based on the occurrence probability and multi-dimensional risk impact index of each scenario in the representative scenario set, a scenario adaptive weight allocation mechanism is constructed. The scenario adaptive weight allocation mechanism assigns a first weight value to high-risk scenarios and a second weight value to normal scenarios. The first weight value is greater than the second weight value. The first weight value is a preset multiple (greater than 1) of the second weight value. This multiple is determined by statistical analysis of the difference in losses from control failures in historical extreme scenarios and normal scenarios, so that the contribution of high-risk scenarios to the optimization objective matches their potential risk losses.

[0046] Secondly, based on the scenario-adaptive weight allocation mechanism, an aggregated control optimization objective function is established with the goal of minimizing the total comprehensive cost within the control period. The total comprehensive cost includes operating costs, user compensation costs, equipment depreciation costs, deviation assessment costs, and penalties for multiple uncertainties. The specific formula is as follows: Where J represents the objective function value; T represents the number of time periods within the control cycle; This represents the operating cost for time period t; This represents the cost of user compensation; This indicates the cost of energy storage or equipment wear and tear. Indicates the cost of deviation assessment; This represents the risk penalty cost caused by multiple uncertainties.

[0047] Next, a hierarchical constraint system is established, using the trustworthy and adjustable domain as a hard constraint and the upper limit of the aggregated adjustment deviation for each scenario in the representative scenario set as a soft constraint. In this example, the hard constraint is: in This represents the aggregate adjustment amount of the virtual power plant during time period t. This constraint is a hard constraint, meaning that the aggregate adjustment amount cannot exceed the reliable adjustable range under any circumstances. The soft constraint is: in This represents the actual adjustment amount in time period t of the s-th scenario in the representative scenario set; This represents the upper limit of the allowable aggregate adjustment deviation for time period t (based on the preset quantile (95%) of historical deviation statistics). This constraint is a soft constraint, meaning that a moderate excess is allowed in the optimization solution, but the excess is included in the deviation assessment cost.

[0048] Then, based on the hierarchical constraint system and the aggregated regulation optimization objective function, a rolling time-domain optimization strategy is used to solve the regulation cycle in segments, obtaining the aggregated regulation sub-plans for each time period. In this example, the implementation of the rolling time-domain optimization strategy is as follows: the regulation cycle is divided into multiple solution periods with a preset rolling window length (4h). Within each rolling window, the aggregated regulation sub-plan within that window is solved using the resource status at the start of the current window and the latest forecast data as initial conditions. Only the regulation plan for the first time period (15min) within the current window is executed, and then the window is rolled forward and solved again in the next time period. This strategy can incorporate the latest resource status information and forecast updates at each solution time, reducing the impact of long-cycle forecast errors on the regulation plan.

[0049] Finally, based on the temporal coupling relationship between the aggregated control sub-plans of each time period, the boundary smoothing of adjacent time periods is performed. In this example, the temporal coupling relationship is mainly reflected in the continuity constraint of the energy storage state of charge: in This indicates the state of charge of the energy storage system during time period t+1. This indicates the state of charge of the energy storage during time period t; Indicates charging efficiency; This represents the charging power during time period t; Indicates discharge efficiency; This represents the discharge power during time period t. The method for smoothing the boundary between adjacent time periods is as follows: check whether the energy storage charge state of the aggregated control sub-plans obtained from the solutions of two adjacent rolling windows is consistent at the boundary time period. If they are inconsistent, the boundary energy storage charge state of the first solution window is used as the benchmark to correct the initial energy storage charge state of the subsequent solution window, and the aggregated control sub-plan of the subsequent solution window is resolved until the boundary energy storage charge state of adjacent windows is consistent, thus obtaining the aggregated control plan.

[0050] Step S50: Generate resource group control instructions based on the aggregated control plan and the trusted adjustable domain, convert the resource group control instructions into device-side control instructions, perform executability verification on the device-side control instructions, and send them to the corresponding devices after the verification is passed, so that the corresponding devices can perform the corresponding control operations and provide feedback on the execution results.

[0051] It should be noted that resource group control instructions refer to the adjustment targets allocated to energy storage resource groups, air conditioning load resource groups, industrial load resource groups, charging pile resource groups, and distributed power resource groups. Equipment-side control instructions refer to the charging / discharging, load reduction, peak shifting, start / stop, or power adjustment instructions issued to specific equipment or control terminals.

[0052] Further, step S50 includes: First, calculating the dynamic priority coefficient of each resource group based on the time-period adjustment demand direction and adjustment amount in the aggregated control plan, as well as the reliable upward and reliable downward adjustment capabilities of each resource in the reliable adjustable domain. In this example, the dynamic priority coefficient is calculated as follows: The dynamic priority coefficient is calculated as follows: Preset credibility weight × resource response credibility + preset speed weight × resource adjustment response speed score + preset cost weight × resource adjustment economic cost score. Here, preset credibility weight + preset speed weight + preset cost weight = 1. Resource response credibility, resource adjustment response speed score, and resource adjustment economic cost score are all dimensionless normalized scores, ranging from 0 to 1. Therefore, they have unified dimensions and can be directly weighted and summed. The resource adjustment response speed score is determined by the ratio of the average time from receiving the instruction to reaching the target adjustment amount to the preset baseline response time (1 minute). The smaller the ratio, the higher the score. The resource adjustment economic cost score is determined by the ratio of the sum of user compensation cost, equipment wear cost, and operating cost per unit of adjustment amount to the preset baseline cost (the historical average cost of each resource group). The smaller the ratio, the higher the score. The preset credibility weight (0.4), preset speed weight (0.3), and preset cost weight (0.3) are determined based on the following: resource response credibility has the most critical impact on the success of regulatory compliance, hence it is assigned the highest weight; response speed and economic cost have comparable impacts on regulatory efficiency and market competitiveness, hence they are assigned the same weight. When the direction of regulatory demand is upward, the dynamic priority coefficient is calculated using credible upward adjustment capability; when the direction of regulatory demand is downward, the dynamic priority coefficient is calculated using credible downward adjustment capability.

[0053] Secondly, the resource groups are sorted according to their dynamic priority coefficients, and the adjustment tasks in the aggregated control plan are allocated to each resource group according to the sorting results, thus obtaining the resource group control instructions. In this example, the adjustment task allocation method is as follows: the resource group with the highest dynamic priority coefficient is taken as the first priority allocation object, and adjustment tasks are allocated according to its reliable upward or downward adjustment capability. If the reliable adjustment capability of the resource group is sufficient to cover all adjustment needs, the allocation stops; if it is insufficient, the remaining adjustment tasks are allocated to the resource group with the second highest dynamic priority coefficient, and so on until all adjustment tasks are allocated. The amount of adjustment tasks obtained by each resource group is the amount of adjustment tasks allocated to it. The resource group control instructions include the resource group identifier, adjustment direction, adjustment amount, and execution period.

[0054] Then, based on the resource type characteristics of each resource group, the resource group control commands are mapped to equipment-side control commands. The equipment-side control commands corresponding to the energy storage resource group are energy storage charging and discharging power adjustment commands, including the target charging and discharging power value and charging / discharging direction; the equipment-side control commands corresponding to the air conditioning load resource group are air conditioning temperature setpoint adjustment commands, including the target temperature setpoint and adjustment duration; the equipment-side control commands corresponding to the industrial load resource group are industrial load interruption duration adjustment commands, including the interrupted load power value, interruption start time, and interruption duration; the equipment-side control commands corresponding to the charging pile resource group are charging pile charging power adjustment commands, including the target charging power value and power adjustment method (constant power adjustment or stepped power adjustment). The equipment-side control commands also include equipment number, execution period, response priority, and failure feedback mechanism.

[0055] Finally, multi-level executability verification is performed on the equipment-side control commands. This multi-level executability verification includes physical constraint layer verification, energy state layer verification, user contract layer verification, and communication link layer verification. Physical constraint layer verification verifies the equipment power boundary and ramp-up constraints; energy state layer verification verifies the energy storage charge state boundary and the electric vehicle target charge constraint; user contract layer verification verifies user comfort constraints and production process constraints; and communication link layer verification verifies the equipment online status and command transmission delay. Specifically, physical constraint layer verification verifies whether the target power value in the equipment-side control commands is within the equipment power boundary range and whether the power change rate meets the ramp-up constraint. If any verification fails, it is marked as a physical constraint layer verification failure. Energy state layer verification verifies whether, for energy storage devices, the energy storage charge state after executing the equipment-side control commands is between the preset charge state lower limit (10%) and the preset charge state upper limit (90%); for electric vehicle charging piles, it verifies whether the vehicle charge level reaches the target charge level after executing the equipment-side control commands. If any verification fails, it is marked as an energy state layer verification failure. User contract layer verification: For air conditioning load, verify whether the adjusted indoor temperature is within the user comfort constraint range; for industrial load, verify whether the interruption duration meets the production process constraint and the maximum interruption duration constraint. If any verification fails, it is marked as a user contract layer verification failure. Communication link layer verification: Verify whether the device is online and whether the command transmission delay is lower than the preset maximum allowable delay (5s). If any verification fails, it is marked as a communication link layer verification failure.

[0056] The system determines whether the device-side control command passes multi-level executability checks. If all four levels of checks pass, the device-side control command is sent to the corresponding device. The specific device-side control command is expressed as follows: ,in, This indicates the device control command for resource i during time period t; Indicates the device number; t indicates the execution period; Indicates the adjustment amount; Indicates the direction of adjustment; This indicates the execution duration, enabling the corresponding device to perform control operations and provide feedback on the execution results. If any level of verification fails, the device-side control command is returned to the resource group control command generation stage, the dynamic priority coefficient of the resource group is recalculated and downgraded, the control tasks originally assigned to the resource group are reassigned to other verified resource groups, and the device-side control command is regenerated for multi-level executability verification.

[0057] Step S60: Collect the execution results fed back by the equipment, calculate the control deviation based on the execution results, and make rolling corrections to the aggregate control plan for subsequent periods based on the control deviation. At the same time, update the response credibility based on the execution results.

[0058] It should be noted that control deviation refers to the difference between the actual aggregated output or actual regulation of the virtual power plant and the target aggregated output or target regulation. Rolling correction refers to the method by which the virtual power plant continuously updates forecast data, resource status, and execution deviations at the day-ahead, intraday, and real-time stages, and periodically resolves the control scheme.

[0059] Specifically, firstly, the system collects the execution results from the equipment feedback. This involves acquiring real-time data on actual power output, energy storage state of charge changes, air conditioning temperature changes, industrial load interruption duration, and charging power of each device after executing control commands through the data acquisition interface of the virtual power plant operation platform. Secondly, the system calculates the control deviation based on the execution results. This involves calculating the difference between the actual response data and the target response data, and determining the direction and magnitude of the control deviation. Then, the system performs rolling corrections on the aggregated control plan for subsequent periods based on the control deviation. When the control deviation exceeds a preset deviation threshold, a rolling correction mechanism is triggered. The control deviation for the current period is converted according to a preset rolling correction coefficient and then deducted or compensated from the aggregated control target for subsequent periods to obtain a corrected aggregated control plan. Simultaneously, the system adjusts the priority of resource group task allocation for subsequent periods based on the type of deviation source. For example, if the deviation originates from early departure of charging pile users, the dynamic priority coefficient of the charging pile resource group in subsequent periods is reduced, and the transferred task load is allocated to the energy storage resource group or the industrial load resource group. Finally, the response credibility is updated based on the execution results, and the updated response credibility is fed back to the trusted adjustable domain calculation stage. The purpose of this is to enable the virtual power plant to have continuous learning and adaptive control capabilities. Resources with continuous and stable responses are given higher priority in subsequent task allocation, while resources with insufficient responses are gradually downgraded, thereby improving the accuracy of resource management and control stability during long-term operation.

[0060] This embodiment constructs a resource state vector by collecting various resource data from a virtual power plant, generating and compressing a representative set of scenarios with multiple uncertainties. It calculates response reliability by combining resource state and historical response data, deriving a reliable adjustable domain. With the goal of minimizing the overall cycle cost, it couples the scenario set and the reliable domain to solve for an aggregated control plan. It generates and verifies the issuance and execution of equipment control commands in a layered manner, collects execution deviations to continuously correct the control plan, and dynamically updates resource reliability, forming a closed-loop control process. Dynamic correction of response reliability avoids inflated aggregation capabilities, and the reliable adjustable domain constraint ensures the feasibility of the solution. Feasibility verification improves engineering feasibility, and rolling correction and reliability updates form a closed-loop adaptive control, balancing economy and reliability, and enhancing the virtual power plant's performance and control stability in complex scenarios.

[0061] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 The virtual power plant aggregation control method for multiple uncertainties, step S60, further includes steps S201 to S206: Step S201: Based on the execution results, obtain the actual response data of each device after executing the device-side control command, and calculate the control deviation for the current period based on the actual response data and the target response data.

[0062] Specifically, the calculation method for the control deviation is as follows: in This indicates the control deviation during time period t; This represents the actual amount of regulation completed by the virtual power plant during time period t; This indicates the target adjustment amount required by the power grid dispatching agency or market transactions. When the deviation exceeds the preset deviation threshold (which is a preset percentage (5%) of the target adjustment amount), it is determined that a rolling correction needs to be triggered.

[0063] Step S202: Identify the type of deviation source based on the direction and magnitude of the control deviation.

[0064] It should be noted that the types of deviation sources include new energy output fluctuation deviation, load change deviation, insufficient user response deviation, equipment failure deviation, and communication delay deviation. Specifically, if the control deviation is consistent with the direction and magnitude of the new energy output prediction deviation, it is determined to be a new energy output fluctuation deviation; if the control deviation is consistent with the direction and magnitude of the load prediction deviation, it is determined to be a load change deviation; if the control deviation is concentrated in a certain resource group and the actual response of that resource group is lower than the target response, it is determined to be an insufficient user response deviation; if the control deviation is concentrated in a certain device and the communication status of that device is offline or the device status is faulty, it is determined to be an equipment failure deviation; if the control deviation exhibits lag characteristics and the command transmission delay exceeds the preset maximum allowable delay (5s), it is determined to be a communication delay deviation.

[0065] Step S203: Determine the rolling correction strategy based on the type of deviation source.

[0066] It should be noted that the rolling correction strategy includes a rapid backup resource deployment strategy for fluctuations in new energy output, a load transfer and redistribution strategy for sudden load changes, a low-reliability resource degradation strategy for insufficient user response, a redundant resource switching strategy for equipment failure, and a forward-looking instruction pre-issuance strategy for communication delay. Specifically, the following strategies are employed: For fluctuations in new energy output, a rapid backup resource deployment strategy is implemented: the dynamic priority coefficients of energy storage and distributed photovoltaic resource groups are increased by a preset rapid deployment coefficient, prioritizing energy storage discharge or photovoltaic curtailment reduction for compensation. For load abrupt changes, a load transfer and redistribution strategy is implemented: the adjustment tasks of industrial load and air conditioning load resource groups are transferred between time periods, shifting peak-hour tasks to off-peak periods. For insufficient user response, a low-reliability resource degradation strategy is implemented: the dynamic priority coefficient of resource groups with insufficient response is reduced by a preset degradation coefficient, and the unfinished adjustment tasks of these resource groups are redistributed to high-reliability resource groups. For equipment failure, a redundant resource switching strategy is implemented: the adjustment tasks corresponding to the failed equipment are switched to redundant equipment of the same type; if no redundant equipment is available, they are switched to other types of resource groups. For communication delays, a forward-looking command pre-issuance strategy is implemented: the command issuance time for subsequent time periods is pre-set with a forward time (one time period) to offset the impact of communication delays.

[0067] Step S204: Modify the aggregate control plan for subsequent periods according to the rolling correction strategy to obtain the revised aggregate control plan, and adjust the priority of resource group task allocation for subsequent periods according to the actual response data.

[0068] Specifically, the revised aggregation control plan is as follows: in This indicates the revised control target for the next period; This indicates the control target for the next period before the revision; This represents the rolling correction coefficient, determined based on the type of deviation source. In this example, the deviation for fluctuations in new energy output is 0.8, for sudden load changes it is 0.7, for insufficient user response it is 0.9, for equipment failure it is 1.0, and for communication delay it is 0.6. The priority of resource group task allocation for subsequent periods is adjusted based on actual response data: the corrected dynamic priority coefficients of the resource groups are reordered, and adjustment tasks for subsequent periods are allocated according to the new order.

[0069] Step S205: Calculate the actual execution score of each resource in the current time period based on the actual response data, and calculate the confidence update magnitude based on the actual execution score and the type of deviation source.

[0070] Specifically, the actual performance score is calculated as follows: in This represents the actual execution score of resource i during time period t; This represents the actual response amount of resource i during time period t; This represents the target response quantity of resource i in time period t; when the actual response quantity equals the target response quantity, the actual execution score is 1; when the actual response quantity is 0, the actual execution score is 0. The confidence update magnitude is then calculated based on the actual execution score and the type of deviation source. The confidence update magnitude is calculated as follows: in This represents the confidence update magnitude of resource i in time period t. The confidence update magnitude is adjusted differently based on whether the source of the deviation belongs to the resource's own controllable factors. The resource's own controllable factors include user response insufficiency deviation and equipment failure deviation. External uncontrollable factors include new energy output fluctuation deviation, load change deviation and communication delay deviation. This represents the credibility update coefficient, which is determined based on the resource type. In this example, energy storage resources are assigned a coefficient of 0.3, air conditioning load resources are assigned a coefficient of 0.2, industrial load resources are assigned a coefficient of 0.3, and charging pile resources are assigned a coefficient of 0.4. This indicates the credibility of the current response of resource i in time period t; This indicates the correction factor for the type of deviation source.

[0071] Step S206: Dynamically update the response credibility of each resource according to the credibility update magnitude to obtain the updated response credibility, and feed the updated response credibility back to the credibility adjustable domain calculation link to form a closed-loop iterative control.

[0072] Specifically, the response credibility of each resource is dynamically updated based on the credibility update magnitude. The updated response credibility is: in Let represent the response credibility of resource i in the next time period t+1. The updated response credibility is fed back to the trusted adjustable domain calculation stage to form a closed-loop iterative control: the updated response credibility serves as the basis for calculating the trusted up-adjustment capability and trusted down-adjustment capability in the next time period. Input values ​​participate in the construction of the reliable adjustable domain for the next regulation cycle.

[0073] This embodiment calculates control deviations by collecting actual equipment response data, identifies deviation sources such as fluctuations in renewable energy output, sudden load changes, insufficient user response, equipment failures, and communication delays, and adopts targeted rolling correction strategies such as rapid deployment of backup resources, load transfer and reallocation, degradation of low-reliability resources, switching of redundant resources, and pre-issuance of forward-looking instructions. It then modifies the aggregated control plan for subsequent time periods and adjusts the priority of resource group task allocation. Simultaneously, it updates resource response reliability based on actual execution scores and deviation source types, feeding this information back to the reliable adjustable domain calculation stage, forming a closed-loop iterative control. This improves the dynamic adaptive control capability of the virtual power plant under multiple uncertainties, reduces the risk of accumulated control deviations, and achieves continuous optimization of resource reliability and rolling correction of control schemes.

[0074] Based on the first embodiment of this application, this application also provides a virtual power plant aggregation control device for multiple uncertainties, please refer to... Figure 3 The device includes: The data acquisition module 10 is used to acquire basic data, predictive data, operational data, historical response data, and communication status data of various distributed resources within the virtual power plant, and to construct resource status vectors based on the basic data, predictive data, operational data, historical response data, and communication status data.

[0075] The scenario construction module 20 is used to generate a set of multiple uncertain scenarios based on prediction data, running data, historical response data and communication status data, and to filter and compress the set of multiple uncertain scenarios to obtain a representative scenario set.

[0076] The credibility calculation module 30 is used to calculate the response credibility of each resource based on the resource state vector and historical response data, and to calculate the credibility adjustable domain of the virtual power plant based on the response credibility and resource state vector.

[0077] The optimization and control module 40 is used to solve the representative scenario set and the reliable adjustable domain based on the aggregated control optimization objective function to obtain the aggregated control plan. The aggregated control optimization objective function is to minimize the total comprehensive cost within the control period.

[0078] The instruction sending module 50 is used to generate resource group control instructions based on the aggregated control plan and the trusted adjustable domain, convert the resource group control instructions into equipment-side control instructions, perform executability verification on the equipment-side control instructions, and send them to the corresponding equipment after the verification is passed, so that the corresponding equipment can perform the corresponding control operation and provide feedback on the execution result.

[0079] The feedback correction module 60 is used to collect the execution results fed back by the equipment, calculate the control deviation based on the execution results, and make rolling corrections to the aggregate control plan for subsequent periods based on the control deviation. At the same time, it updates the response credibility based on the execution results.

[0080] The virtual power plant aggregation control device for multiple uncertainties provided in this application, employing the virtual power plant aggregation control method for multiple uncertainties in the above embodiments, can solve the technical problem of how to accurately assess the reliable and adjustable capabilities of a virtual power plant and generate an aggregation control scheme. Compared with the prior art, the beneficial effects of the virtual power plant aggregation control device for multiple uncertainties provided in this application are the same as those of the virtual power plant aggregation control method for multiple uncertainties provided in the above embodiments, and other technical features in the virtual power plant aggregation control device for multiple uncertainties are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0081] This application provides a virtual power plant aggregation control device for multiple uncertainties. The virtual power plant aggregation control device for multiple uncertainties includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the virtual power plant aggregation control method for multiple uncertainties in the above embodiment 1.

[0082] The following is for reference. Figure 4This document illustrates a structural schematic diagram of a virtual power plant aggregation and control device suitable for implementing embodiments of this application, addressing multiple uncertainties. The virtual power plant aggregation and control device for multiple uncertainties in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The virtual power plant aggregation control device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0083] like Figure 4 As shown, a virtual power plant aggregation control device accommodating multiple uncertainties may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 1002 or programs loaded from storage device 1003 into random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the virtual power plant aggregation control device accommodating multiple uncertainties. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the virtual power plant aggregation control equipment for multiple uncertainties to communicate wirelessly or wiredly with other devices to exchange data. Although various virtual power plant aggregation control devices for multiple uncertainties are shown in the figures, it should be understood that it is not required to implement or possess all of them. More or fewer may be implemented alternatively.

[0084] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0085] The virtual power plant aggregation control device for multiple uncertainties provided in this application, employing the virtual power plant aggregation control method for multiple uncertainties in the above embodiments, can solve the technical problem of how to accurately assess the reliable and adjustable capabilities of a virtual power plant and generate an aggregation control scheme. Compared with the prior art, the beneficial effects of the virtual power plant aggregation control device for multiple uncertainties provided in this application are the same as those of the virtual power plant aggregation control method for multiple uncertainties provided in the above embodiments, and other technical features in this virtual power plant aggregation control device for multiple uncertainties are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0086] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0087] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations 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. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0088] This application provides a computer-readable medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the virtual power plant aggregation control method for multiple uncertainties in the above embodiments.

[0089] The computer-readable medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor devices, or any combination thereof. More specific examples of computer-readable media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable medium may be any tangible medium containing or storing a program that can be executed by instructions, used by a device, or used in conjunction with it. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0090] The aforementioned computer-readable medium may be included in a virtual power plant aggregation control device for multiple uncertainties; or it may exist independently and not be assembled into a virtual power plant aggregation control device for multiple uncertainties.

[0091] The aforementioned computer-readable medium carries one or more programs that, when executed by a virtual power plant aggregation and control device oriented towards multiple uncertainties, enable the device to write computer program code for performing the operations of this application in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of this application. In this regard, all blocks in the flowcharts or block diagrams may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that all blocks in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using dedicated hardware-based implementations that perform the specified functions or operations, or using a combination of dedicated hardware and computer instructions.

[0093] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0094] The readable medium provided in this application is a computer-readable medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described virtual power plant aggregation control method for multiple uncertainties. This solves the technical problem of how to accurately assess the reliable adjustability of a virtual power plant and generate aggregation control schemes. Compared with the prior art, the beneficial effects of the computer-readable medium provided in this application are the same as those of the virtual power plant aggregation control method for multiple uncertainties provided in the above embodiments, and will not be repeated here.

[0095] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the virtual power plant aggregation control method for multiple uncertainties described above.

[0096] The computer program product provided in this application can solve the technical problem of how to accurately assess the reliable and adjustable capabilities of a virtual power plant and generate aggregated control schemes. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the virtual power plant aggregated control method for multiple uncertainties provided in the above embodiments, and will not be repeated here.

[0097] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A virtual power plant aggregation control method for multiple uncertainties, characterized in that, The method includes: Acquire basic data, predictive data, operational data, historical response data, and communication status data of various distributed resources within the virtual power plant, and construct a resource status vector based on the basic data, predictive data, operational data, historical response data, and communication status data; Based on the predicted data, the operational data, the historical response data, and the communication status data, a set of multiple uncertainty scenarios is generated, and the set of multiple uncertainty scenarios is filtered and compressed to obtain a representative scenario set; The response reliability of each resource is calculated based on the resource state vector and the historical response data, and the reliable adjustable domain of the virtual power plant is calculated based on the response reliability and the resource state vector. Based on the aggregated regulation optimization objective function, the representative scenario set and the reliable adjustable domain are solved to obtain the aggregated regulation plan, wherein the aggregated regulation optimization objective function is to minimize the total comprehensive cost within the regulation period. Based on the aggregated control plan and the trusted adjustable domain, a resource group control instruction is generated. The resource group control instruction is converted into a device-side control instruction. The device-side control instruction is then validated for executability. Once the validation is successful, the instruction is sent to the corresponding device so that the corresponding device can perform the corresponding control operation and provide feedback on the execution result. The execution results fed back by the device are collected, the control deviation is calculated based on the execution results, and the aggregate control plan for subsequent periods is rolled out based on the control deviation. At the same time, the response credibility is updated based on the execution results.

2. The method as described in claim 1, characterized in that, The step of constructing a resource state vector based on the basic data, the predicted data, the operational data, the historical response data, and the communication state data includes: Based on the resource type, rated power, capacity and access node information in the basic data, and the real-time power, energy storage state of charge, air conditioning temperature and charging pile status in the operation data, a basic physical parameter mapping for each resource is established. Based on the photovoltaic power output, wind power output, load forecast, and electricity price forecast in the forecast data, as well as the historical output error, historical load error, user response record, and equipment failure record in the historical response data, a multi-timescale uncertainty feature of each resource is constructed. The communication link quality indicators for each resource are calculated based on the device online status, command arrival time, execution delay, and execution success rate in the communication status data. Based on the basic physical parameter mapping, the multi-timescale uncertainty characteristics, and the communication link quality indicators, a multi-dimensional resource state vector is constructed, wherein the multi-dimensional resource state vector includes current power, energy state, operating constraint state, equipment availability state, communication delay state, historical response performance, uncertainty fluctuation amplitude, and communication link quality indicators. Based on the coupling relationship between the dimensions in the multi-dimensional resource state vector, a resource state association matrix is ​​established, and the multi-dimensional resource state vector is subjected to consistency verification and dynamic update based on the resource state association matrix to obtain the resource state vector.

3. The method as described in claim 1, characterized in that, The step of generating a set of multiple uncertainty scenarios based on the predicted data, the operational data, the historical response data, and the communication status data, and then filtering and compressing the set of multiple uncertainty scenarios to obtain a representative set of scenarios, includes: Based on the historical prediction error distribution characteristics in the prediction data, the weather change patterns in the operational data, the user response behavior transfer patterns in the historical response data, and the equipment fault association patterns in the communication status data, a multi-source uncertainty coupling generation model is constructed. An initial set of multiple uncertainty scenarios is generated based on the multi-source uncertainty coupling generation model, wherein each scenario in the initial set of multiple uncertainty scenarios includes the joint values ​​of new energy output deviation, load forecast deviation, market price deviation, user response rate deviation, communication delay deviation, and equipment availability status deviation. Based on the initial set of multiple uncertainty scenarios, a multi-dimensional risk impact index for each scenario is calculated, wherein the multi-dimensional risk impact index is determined according to the comprehensive impact of each scenario on the virtual power plant's aggregated regulation capability. Based on the multi-dimensional risk impact index, the initial set of multiple uncertainty scenarios is hierarchically clustered. Scenarios with similar risk impact indices are divided into the same scenario cluster. In each scenario cluster, the scenario with the risk impact index closest to the cluster center and covering the maximum uncertainty coupling mode is selected as the first representative scenario. The maximum uncertainty coupling mode indicates that the scenario simultaneously includes the largest number of types of new energy output deviation, load demand deviation, market price deviation, user response rate deviation, communication delay deviation, and equipment availability status deviation. Scenarios whose multi-dimensional risk impact index exceeds a preset high-risk threshold are screened to obtain a second representative scenario; The first representative scene and the second representative scene are merged into a candidate scene set, and the scene probabilities in the candidate scene set are adaptively weighted and normalized to obtain a representative scene set.

4. The method as described in claim 1, characterized in that, The steps of calculating the response reliability of each resource based on the resource state vector and the historical response data, and calculating the reliable adjustable domain of the virtual power plant based on the response reliability and the resource state vector, include: Based on the historical performance score, response speed score, communication status score, and device health status score in the historical response data, as well as the historical response performance and communication delay status in the resource status vector, a multi-dimensional response credibility evaluation matrix for each resource is constructed. The reliability decay coefficient of each resource at different time periods is calculated based on the temporal correlation characteristics between the dimensions in the multi-dimensional response reliability evaluation matrix. The reliability decay coefficient is determined based on the number of consecutive unresponsive periods of the resource and the cumulative magnitude of the response deviation. The multi-dimensional response credibility evaluation matrix is ​​dynamically weighted and corrected based on the credibility decay coefficient to obtain the dynamic response credibility of each resource at different time periods. Based on the dynamic response confidence level, and the device availability status, power boundary, ramp constraint and energy status in the resource state vector, calculate the confidence up-adjustment capability and confidence down-adjustment capability of each resource at different time periods. Based on the reliable up-adjustment capability and reliable down-adjustment capability of each resource, as well as the complementary adjustment characteristics between resources, a resource group collaborative adjustment matrix is ​​constructed. The trusted up-adjustment capability and the trusted down-adjustment capability are aggregated and corrected based on the resource group collaborative adjustment matrix to obtain the trusted adjustable domain of the virtual power plant.

5. The method as described in claim 1, characterized in that, The step of solving the representative scenario set and the reliable adjustable domain based on the aggregated regulation optimization objective function to obtain the aggregated regulation plan includes: Based on the occurrence probability of each scenario in the representative scenario set and the multi-dimensional risk impact index, a scenario adaptive weight allocation mechanism is constructed, wherein the scenario adaptive weight allocation mechanism assigns a first weight value to high-risk scenarios and a second weight value to normal scenarios, and the first weight value is greater than the second weight value. Based on the scenario-adaptive weight allocation mechanism, an aggregated regulation optimization objective function is established with the goal of minimizing the total comprehensive cost within the regulation cycle. The total comprehensive cost includes operating costs, user compensation costs, equipment wear and tear costs, deviation assessment costs, and multiple uncertainty risk penalty costs. A hierarchical constraint system is established by taking the trustworthy and adjustable domain as a hard constraint and the upper limit of the aggregate adjustment deviation of each scenario under the representative scenario set as a soft constraint. Based on the hierarchical constraint system and the aggregated regulation optimization objective function, the regulation cycle is solved piecewise using a rolling time-domain optimization strategy to obtain the aggregated regulation sub-plan for each time period. Based on the temporal coupling relationship between the aggregated control sub-plans in each time period, the boundary smoothing of adjacent time periods is performed to obtain the aggregated control plan.

6. The method as described in claim 1, characterized in that, The steps of generating resource group control instructions based on the aggregation control plan and the trusted adjustable domain, converting the resource group control instructions into device-side control instructions, performing executability verification on the device-side control instructions, and issuing them to the corresponding devices after successful verification include: The dynamic priority coefficients of each resource group are calculated based on the time-period adjustment demand direction and adjustment amount in the aggregated control plan, as well as the reliable upward adjustment capability and reliable downward adjustment capability of each resource in the reliable adjustable domain. The resource groups are sorted according to the dynamic priority coefficient, and the adjustment tasks in the aggregated control plan are allocated to each resource group according to the sorting results to obtain the resource group control instructions. Based on the resource type characteristics of each resource group, the resource group control instructions are mapped to equipment-side control instructions, wherein the equipment-side control instructions include energy storage charging and discharging power adjustment instructions, air conditioning temperature setpoint adjustment instructions, industrial load interruption duration adjustment instructions, and charging pile charging power adjustment instructions. The device-side control command is subjected to multi-level executability verification. If it passes, the device-side control command is sent to the corresponding device so that the corresponding device can perform the control operation and provide feedback on the execution result. If it fails, the control task is returned to be reassigned.

7. The method as described in claim 1, characterized in that, The steps of collecting the execution results fed back by the device, calculating the control deviation based on the execution results, and making rolling corrections to the aggregate control plan for subsequent periods based on the control deviation, while updating the response reliability based on the execution results, include: Based on the execution results, the actual response data of each device after executing the device-side control command is obtained, and the control deviation for the current time period is calculated based on the actual response data and the target response data. The source type of deviation is identified based on the direction and magnitude of the control deviation, wherein the source type of deviation includes deviation due to fluctuation in new energy output, deviation due to sudden load changes, deviation due to insufficient user response, deviation due to equipment failure, and deviation due to communication delay. The rolling correction strategy is determined based on the type of deviation source. The rolling correction strategy includes a rapid backup resource call strategy for deviations caused by fluctuations in new energy output, a load transfer and redistribution strategy for deviations caused by sudden load changes, a low-reliability resource degradation strategy for deviations caused by insufficient user response, a redundant resource switching strategy for deviations caused by equipment failure, and a forward-looking instruction pre-issuance strategy for deviations caused by communication delays. The aggregation control plan for subsequent periods is revised according to the rolling correction strategy to obtain the revised aggregation control plan, and the priority of resource group task allocation for subsequent periods is adjusted according to the actual response data. The actual execution score of each resource in the current time period is calculated based on the actual response data, and the confidence update magnitude is calculated based on the actual execution score and the type of deviation source, wherein the confidence update magnitude is adjusted differently depending on whether the deviation source belongs to a factor controllable by the resource itself. The response credibility of each resource is dynamically updated based on the credibility update magnitude to obtain the updated response credibility. The updated response credibility is then fed back to the trusted adjustable domain calculation stage to form a closed-loop iterative control.

8. A virtual power plant aggregation control device for multiple uncertainties, characterized in that, The apparatus is applied to the virtual power plant aggregation control method for multiple uncertainties as described in any one of claims 1 to 7, and the apparatus comprises: The data acquisition module is used to acquire basic data, predictive data, operational data, historical response data, and communication status data of various distributed resources within the virtual power plant, and to construct a resource status vector based on the basic data, predictive data, operational data, historical response data, and communication status data. The scenario construction module is used to generate a set of multiple uncertain scenarios based on the prediction data, the running data, the historical response data and the communication status data, and to filter and compress the set of multiple uncertain scenarios to obtain a representative scenario set. The credibility calculation module is used to calculate the response credibility of each resource based on the resource state vector and the historical response data, and to calculate the credibility adjustable domain of the virtual power plant based on the response credibility and the resource state vector. The optimization and control module is used to solve the representative scenario set and the reliable adjustable domain based on the aggregated control optimization objective function to obtain the aggregated control plan, wherein the aggregated control optimization objective function is to minimize the total comprehensive cost within the control period. The instruction sending module is used to generate resource group control instructions according to the aggregated control plan and the trusted adjustable domain, convert the resource group control instructions into device-side control instructions, perform executability verification on the device-side control instructions, and send them to the corresponding device after the verification is passed, so that the corresponding device can perform the corresponding control operation and feed back the execution result. The feedback correction module is used to collect the execution results fed back by the device, calculate the control deviation based on the execution results, and make rolling corrections to the aggregate control plan for subsequent periods based on the control deviation, while updating the response credibility based on the execution results.

9. A virtual power plant aggregation and control device for multiple uncertainties, characterized in that, The device includes: a memory, a processor, and a virtual power plant aggregation control program for multiple uncertainties stored in the memory and running on the processor, the virtual power plant aggregation control program for multiple uncertainties being configured to implement the steps of the virtual power plant aggregation control method for multiple uncertainties as described in any one of claims 1-7.

10. A storage medium, characterized in that, The storage medium stores a virtual power plant aggregation control program for multiple uncertainties. When the virtual power plant aggregation control program for multiple uncertainties is executed by the processor, it implements the steps of the virtual power plant aggregation control method for multiple uncertainties as described in any one of claims 1-7.

Images