A multi-element flexible load cooperative scheduling method and system for coping with wind and light uncertainty

Abstract

The application discloses a kind of multi-element flexible load cooperative scheduling method and system to deal with wind light uncertainty, method includes: according to the wind light load forecast power and flexible load baseline power generated by multi-source historical operation data, establish virtual power plant resource level set and various resource operation constraints;With the minimum running cost and wind light risk cost as the goal, the day-ahead robust power balance plan is constructed, the wind light uncertainty is described by typical scene and adaptive KL probability fuzzy set, and standby capacity is configured;Intra-day stage executes rehearsal according to day-ahead plan, calculates system comprehensive deviation and decomposes into source side, load side and flexible load side partial deviation;Establish upper layer rolling optimization model to determine the adjustment budget and call boundary of each resource;Under the budget constraint, dynamic weight is used to improve greedy algorithm to allocate real-time adjustment amount.Improve the power balance ability of virtual power plant and the cooperative regulation efficiency of multi-element flexible resource under the condition of wind light uncertainty.

Description

Technical Field

[0001] This invention belongs to the field of power system optimization and dispatching technology, and in particular relates to a method and system for collaborative dispatching of multiple flexible loads to cope with the uncertainties of wind and solar power. Background Technology

[0002] With the continuous advancement of new power system construction, the integration ratio of distributed renewable energy sources such as wind power and photovoltaics in virtual power plants is constantly increasing. Wind power output is affected by factors such as wind speed, wind direction, air density, and turbine operating status, while photovoltaic output is affected by factors such as solar irradiance, ambient temperature, module temperature, cloud cover, and weather conditions. Both exhibit strong randomness, volatility, and intermittency. When there is a deviation between the predicted and actual wind and solar power output, the original day-ahead dispatch plan of the virtual power plant is prone to supply-demand imbalance, leading to power shortages, surplus power, increased wind and solar curtailment, or increased pressure on the external power grid. Therefore, under conditions of high wind and solar integration, how to construct a robust day-ahead power balance plan and quickly correct source-load deviations during the day and in real-time is a crucial problem that needs to be solved in the field of virtual power plant optimized dispatch.

[0003] Existing research has yielded substantial results regarding the optimal scheduling of virtual power plants. Some studies develop day-ahead operating plans based on wind and solar power forecasts and load forecasts, achieving source-load balance through energy storage charging and discharging, grid power purchase and sale, or flexible load response. Other studies employ scenario-based methods, stochastic optimization, or interval optimization to describe the uncertainties in wind and solar power output, thereby improving the adaptability of scheduling plans to forecast errors. Still other methods utilize resources such as air conditioning, electric vehicles, or energy storage in real-time to correct system deviations and reduce intraday operational biases. While these studies have improved the economic efficiency of virtual power plant operation and the capacity for renewable energy integration to some extent, the following shortcomings remain:

[0004] (1) Insufficient consideration of the drift of wind and solar power output probability distribution. Existing methods mostly rely on central forecast values ​​or fixed probability scenarios to describe the uncertainty of wind and solar power, which is difficult to adapt to the situation where the actual probability distribution of wind and solar power output shifts relative to the nominal probability. When weather conditions change or forecast errors fluctuate, fixed probability scenarios are prone to insufficient backup capacity, increased remediation costs, or source-load imbalance.

[0005] (2) Insufficient connection between the day-ahead, intraday and real-time phases. Existing methods lack systematic design for information transmission and plan progression between phases. Intraday phases are often corrected only based on the total system deviation, lacking the identification of deviation sources and making it difficult to make targeted corrections based on the different deviations of the wind power source side, the base load side and the flexible load side.

[0006] (3) Real-time correction methods are difficult to comprehensively reflect the resource scheduling potential. Existing real-time correction methods mostly adopt fixed call order or single cost index, which makes it difficult to comprehensively consider the resource adjustability margin, response reliability, economy and subsequent availability. For resources with energy constraints and time constraints, such as energy storage, electric vehicles and drones, excessive call in the current period will weaken the adjustment capability in the subsequent period. Fixed call order is prone to excessive resource call and reduced scheduling effect.

[0007] To address the above problems, this invention proposes a multi-element flexible load collaborative scheduling method and system for coping with wind and solar uncertainties. Summary of the Invention

[0008] This invention provides a multi-variable flexible load collaborative scheduling method and system to address the uncertainty of wind and solar power. It addresses the problem of deviation between the day-ahead plan and the actual intraday operation status of virtual power plants caused by random fluctuations in wind and solar power output, and constructs a closed-loop scheduling process that coordinates day-ahead, intraday, and real-time operations.

[0009] In a first aspect, the present invention provides a method for coordinated scheduling of multiple flexible loads to address the uncertainties of wind and solar power, comprising:

[0010] Acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic power day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints;

[0011] Based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, a day-ahead robust power balance plan is constructed to obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is equipped with reserve capacity.

[0012] During the daytime phase, updated wind and solar power, load power, and resource status data are collected. An execution rehearsal is performed according to the daytime robust power balance plan. The daytime system comprehensive deviation is calculated, and the daytime system comprehensive deviation is decomposed into sub-deviations from different sources.

[0013] Based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, the real-time adjustable resources are divided into multi-level resource pools, and an upper-level hierarchical rolling optimization model is established to determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window.

[0014] Under the constraints of the adjustment budget and the current allowed call boundary, the available adjustment capacity of each resource is determined according to the operational constraints of various resources. A greedy algorithm is used to allocate the specific adjustment amount for the current time period in descending order of the overall resource priority to obtain the real-time scheduling result.

[0015] Secondly, the present invention provides a multi-element flexible load collaborative scheduling system for coping with wind and solar uncertainties, comprising:

[0016] The acquisition module is configured to acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints;

[0017] The construction module is configured to construct a day-ahead robust power balance plan based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power and the various resource operation constraints, and obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is configured with reserve capacity.

[0018] The calculation module is configured to collect updated wind and solar power, load power and resource status data during the day, perform a pre-execution simulation according to the day-ahead robust power balance plan, calculate the day-ahead system comprehensive deviation, and decompose the day-ahead system comprehensive deviation into sub-item deviations from different sources.

[0019] The partitioning module is configured to divide real-time adjustable resources into multi-level resource pools based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, establish an upper-level hierarchical rolling optimization model, and determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window;

[0020] The output module is configured to determine the available adjustment capacity of each resource based on the constraints of the adjustment budget and the current allowed call boundary, and to allocate the specific adjustment amount for the current time period according to the overall priority of the resources from high to low using a greedy algorithm to obtain the real-time scheduling result.

[0021] Thirdly, an electronic device is provided, comprising: 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, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the multi-various flexible load collaborative scheduling method for coping with wind and solar uncertainties according to any embodiment of the present invention.

[0022] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the steps of the multi-various flexible load collaborative scheduling method for coping with wind and solar uncertainties according to any embodiment of the present invention.

[0023] This application presents a method and system for coordinated scheduling of multiple flexible loads to address wind and solar uncertainties. It generates predicted wind and solar load power and baseline power for flexible loads based on multi-source historical operating data, establishes a virtual power plant resource hierarchy set and various resource operating constraints, constructs a day-ahead robust power balance plan with the goal of minimizing operating costs and wind and solar risk costs, describes wind and solar uncertainties through typical scenarios and adaptive KL probabilistic fuzzy sets, and configures reserve capacity. During the intraday phase, a rehearsal is performed according to the day-ahead plan, calculating the overall system deviation and decomposing it into source-side, load-side, and flexible load-side sub-deviations. A hierarchical rolling optimization model is established to determine the adjustment budget and call boundaries for each resource. Under budget constraints, a dynamic weighted improved greedy algorithm is used to allocate real-time adjustment quantities. This improves the power balance capability of the virtual power plant and the efficiency of coordinated scheduling of multiple flexible resources under wind and solar uncertainty conditions. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating a multi-element flexible load collaborative scheduling method for addressing wind and solar uncertainties, provided in an embodiment of the present invention;

[0026] Figure 2 This is a structural block diagram of a multi-element flexible load collaborative scheduling system for coping with wind and solar uncertainties, provided in an embodiment of the present invention.

[0027] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figure 1 The diagram shows a flowchart of a multi-variable flexible load collaborative scheduling method for addressing wind and solar uncertainties, as proposed in this application.

[0030] like Figure 1 As shown, a multi-element flexible load collaborative scheduling method for addressing wind and solar uncertainties specifically includes the following steps:

[0031] Step S101: Obtain multi-source historical operating data and day-ahead external input data; generate wind power day-ahead predicted power, photovoltaic power day-ahead predicted power, base load day-ahead predicted power, and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data; and establish a virtual power plant resource hierarchy set and various resource operation constraints.

[0032] In this step, wind power forecasting input feature vectors are extracted from the multi-source historical operating data and the day-ahead external input data. The expression is:

[0033] ,

[0034] in, The input feature vector for wind power prediction in time period t. Let be the wind speed vector for time period t. Let t be the wind direction. Let be the air density during time period t. For the historical wind power output during time period t, Let t represent the available capacity of the wind turbines during time period t. The operating status of the wind turbine during time period t;

[0035] The photovoltaic prediction input feature vector is extracted from the multi-source historical operating data and the current-day external input data, and the expression is:

[0036] ,

[0037] In the formula, The input feature vector for photovoltaic prediction in time period t, Let be the solar irradiance during time period t. Let t be the ambient temperature. Let be the temperature of the photovoltaic module during time period t. Let be the cloud cover in time period t. The humidity at time t. For the historical photovoltaic output of time period t, Let t be the available photovoltaic installed capacity during time period t;

[0038] The base load forecast input feature vector is extracted from the multi-source historical operating data and the current-day external input data, and the expression is:

[0039] ,

[0040] In the formula, The input feature vector for the basic load forecast in time period t is... The historical base load power for time period t. For time period t, the day type is... Let t be the time period number. This is the holiday identifier for time period t. The seasonal type for time period t;

[0041] The wind power prediction input feature vector, the photovoltaic prediction input feature vector, and the base load prediction input feature vector are respectively input into the LSTM prediction model. The LSTM prediction model outputs the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, and the day-ahead predicted power of base load, respectively, as expressed in the following expressions:

[0042] ,

[0043] ,

[0044] ,

[0045] In the formula, The predicted wind power output for time period t is the day-ahead forecast. For the day-ahead forecast of photovoltaic power generation in time period t, For wind power day-ahead forecasting models, This is a day-ahead forecasting model for photovoltaics. Based on the day-ahead forecasting model for basic load;

[0046] Extract the state feature vector of the flexible load from the multi-source historical operating data and the current daytime external input data. For any flexible load , , For air conditioning load, For electric vehicle load, For drone load;

[0047] The flexible load State feature vector Input a baseline prediction model combining clustering and LightGBM regression, and output the flexible load. Baseline power when not participating in scheduling during time period t , For baseline prediction models;

[0048] The expression for the state characteristic vector of the air conditioning load AC is:

[0049] ,

[0050] In the formula, This is the state feature vector of the air conditioning load (AC). Let be the indoor temperature during time period t. Set the temperature for users in time period t. The historical operating power of the air conditioner during time period t. The historical operating power of the air conditioner during time period t. The user occupancy status for time period t;

[0051] The expression for the state feature vector of the electric vehicle load (EV) is:

[0052] ,

[0053] In the formula, This represents the state feature vector of the electric vehicle load (EV). Let be the time when the j-th electric vehicle enters the system. Let be the time when the j-th electric vehicle leaves the network. Let j be the current state of charge of the j-th electric vehicle in time period t. Let j be the minimum state of charge required for the j-th electric vehicle to disconnect from the grid. The maximum charging power of the j-th electric vehicle;

[0054] The expression for the state feature vector of the UAV (Unmanned Aerial Vehicle) is:

[0055] ,

[0056] In the formula, This represents the state feature vector of the UAV (Unmanned Aerial Vehicle) payload. Let u be the current remaining energy of the u-th drone in time period t. The mission deadline for the uth drone. This represents the minimum energy required before the end of the uth drone mission. Let u be the energy consumption of the uth drone during time period t. The maximum charging power for the uth drone;

[0057] Based on the aforementioned flexible load resource set and energy storage systems Define the internal flexible resource set Based on the aforementioned internal flexible resource set Interacting with external power grid resources Define a set of real-time adjustable resources to obtain the set of resource levels of the virtual power plant;

[0058] For the aforementioned air conditioning load AC, establish air conditioning comfort constraints. and air conditioning power constraints ,in, Let be the indoor temperature of the i-th air-conditioned user during time period t. and Let $\begin{cases}$ be the lower and upper limits of the acceptable indoor temperature for the $i$-th air conditioning user, respectively. Let be the operating power of the i-th air conditioner during time period t. The maximum operating power of the i-th air conditioner;

[0059] For the aforementioned electric vehicle load (EV), establish electric vehicle charging power constraints. Off-grid power constraints ,in, Let j be the charging power of the j-th electric vehicle during time period t. For the j-th electric vehicle at the off-grid time The state of charge;

[0060] For the aforementioned UAV (Unmanned Aerial Vehicle) payload, establish UAV energy boundary constraints. and task deadline constraints ,in, and These are the lower and upper energy limits for the u-th drone, respectively. For the uth drone at the mission deadline The remaining energy;

[0061] For the energy storage system ESS, establish energy storage boundary constraints. and energy storage charging and discharging power constraints , and charging / discharging mutual exclusion constraints ,in, The energy state of energy storage during time period t. and These are the lower and upper limits of energy storage capacity, respectively. and These represent the charging power and discharging power of the energy storage during time period t, respectively. and These are the maximum charging power and maximum discharging power of the energy storage, respectively. and These represent the charging state variables and discharging state variables of energy storage during time period t, respectively.

[0062] Step S102: Based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, a day-ahead robust power balance plan is constructed to obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is equipped with reserve capacity.

[0063] In this step, the predicted wind power daytime power is used. The photovoltaic day-ahead predicted power The day-ahead forecast power of the base load Based on the baseline power of each flexible load, a day-ahead power balance relationship is established, expressed as:

[0064] ,

[0065] In the formula, The day-ahead discharge power of the energy storage during time period t, The power purchased by the external power grid on the day-ahead of time period t. , , These represent the day-ahead planned power outputs of air conditioners, electric vehicles, and drones, respectively, during time period t. The day-ahead charging power of energy storage during time period t. The daytime electricity sales power of the external power grid during time period t;

[0066] Based on historical landscape data, day-ahead center prediction results, and prediction residuals, K typical landscape scenes are generated using residual resampling and scene clustering methods. KL probabilistic fuzzy sets are introduced to describe the shift in the true probability distribution of the landscape scenes relative to their nominal probabilities. The expression is:

[0067] ,

[0068] In the formula, Let be the probability of the s-th typical scenario under the true probability distribution. Let be the nominal probability of a typical scenario s. Let the radius of the KL fuzzy set be ;

[0069] The KL fuzzy set radius adopts an adaptive KL radius. The expression is:

[0070] ,

[0071] In the formula, Based on the KL radius, To predict the influence coefficient of error fluctuation intensity on the KL radius, Let be the intensity of the wind and solar forecast error fluctuation over time period t. The influence coefficient of weather disturbance intensity on the radius of KL. The intensity of the weather disturbance during time period t;

[0072] For the s-th typical scenario, the source-side deviation at time t Represented as:

[0073] ,

[0074] In the formula, For the s-th typical scenario, the wind power output during time period t is... For the s-th typical scenario, the photovoltaic output during time period t;

[0075] The objective function for constructing the current robust power balance plan is expressed as follows:

[0076] ,

[0077] ,

[0078] ,

[0079] ,

[0080] In the formula, The base operating cost is the current day. Cost of configuring standby capacity, To account for the uncertainties and risks associated with wind and solar power, For the set of scheduling periods, The duration of a single scheduling period. The electricity purchase price for time period t. The electricity price for time period t. The day-ahead adjustment cost of the air conditioner during time period t. For the day-ahead adjustment cost of electric vehicles in time period t, The day-ahead adjustment cost of the drone in time period t. The day-ahead operating cost of energy storage during period t, Configuration cost coefficient for upward backup The configuration cost factor for backward backup. For time period t, the upward reserve capacity is... This represents the downward reserve capacity for time period t. This represents the risk cost of the day-ahead plan under the s-th typical scenario;

[0081] The objective function is solved under the constraints of the various resource operations to obtain the day-ahead scheduling and allocation results for each resource.

[0082] Step S103: Collect updated wind and solar power, load power and resource status data during the day, perform a pre-execution simulation according to the day-ahead robust power balance plan, calculate the day-ahead system comprehensive deviation, and decompose the day-ahead system comprehensive deviation into sub-item deviations from different sources.

[0083] In this step, the updated wind power output for the day is collected. Photovoltaic power Basic load power And the air conditioner's executable power before real-time correction. The executable power of electric vehicles before real-time correction. And the executable power of the drone before real-time correction. ;

[0084] The execution rehearsal of the aforementioned robust power balance plan was carried out, and the intraday rehearsal supply-side power was calculated. Demand-side power forecast for the day The expression is:

[0085] ,

[0086] ,

[0087] Calculate the intraday system comprehensive deviation The intraday system comprehensive deviation is decomposed into intraday wind source side deviation. Intraday basic load deviation Air conditioning execution deviation Electric vehicle execution deviation and drone execution deviation The expression is:

[0088] ,

[0089] ,

[0090] ,

[0091] ,

[0092] .

[0093] Step S104: Based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, the real-time adjustable resources are divided into multi-level resource pools, an upper-level hierarchical rolling optimization model is established, and the adjustment budget and current allowed call boundary of each resource in the current window are determined within the short-term prediction window.

[0094] In this step, set up a real-time scrolling window. Based on the latest wind and solar forecasts, load forecasts, deviation trends, and resource status, a forecast deviation sequence is formed within the future window. ,in, Optimize window length for scrolling;

[0095] The set of real-time adjustable resources in the virtual power plant resource hierarchy set Divided into flexible load resource pool Energy storage resource pool and external power grid backup resource pool ;

[0096] For any resource Let the resources be Time period within the scrolling window The predicted adjustment amount is Calculate the time period within the scrolling window Predicted residual bias The expression is:

[0097] ,

[0098] Based on the time period within the scrolling window The objective function of the upper-level hierarchical rolling optimization model is established based on the predicted residual bias, and its expression is:

[0099] ,

[0100] ,

[0101] ,

[0102] ,

[0103] In the formula, To predict the residual bias penalty coefficient, For resources Within the scrolling window The predicted adjustment costs For time period Layered call penalty cost, For time period Adjustment and smoothing costs, For time period The cost of excessive resource use The penalty weight for tiered calls, To adjust the penalty weights for smoothing, Penalty weights for excessive resource use The penalty coefficient for calling up energy storage resources. The penalty coefficient for calling external power grid resource pools. For resources Time period within the scrolling window The predicted adjustment amount, For resources Within the scrolling window Cumulative usage It is a very small positive number. For resources Within the scrolling window Maximum effective adjustable capacity;

[0104] The constraints of the upper-level hierarchical rolling optimization model include:

[0105] The resource adjustment boundary constraint is expressed as follows:

[0106] ,

[0107] The resource budget constraint for the scrolling window is expressed as follows:

[0108] ,

[0109] In the formula, For resources The adjustment budget allowed within the current scrolling window.

[0110] Furthermore, based on the source-side bias contribution... The contribution of the basic load deviation The contribution of the flexible load deviation and the rate of change of deviation A resource budget adjustment model is established, and the adjusted budget for resource r after adjustment in time period t is calculated using the resource budget adjustment model. The expression is:

[0111] ,

[0112] In the formula, The basic budget for resource r, Let r be the sensitivity coefficient of resource r to source-side deviation. Let r be the sensitivity coefficient of resource r to the deviation of base load. Let r be the sensitivity coefficient of resource r to flexible load deviation. Let r be the sensitivity coefficient of resource r to the trend of deviation change. The deviation rate over time period t. To predict the maximum rate of change of deviation within the window, It is a very small positive number;

[0113] Substitute the calculated adjustment budget into the rolling window resource budget constraint of the upper-level hierarchical rolling optimization model, solve the upper-level hierarchical rolling optimization model, and obtain the current allowed call capacity of resource r in time period t.

[0114] Step S105: Under the constraints of the adjustment budget and the current allowed call boundary, determine the available adjustment capacity of each resource according to the operation constraints of various resources, and use a greedy algorithm to allocate the specific adjustment amount for the current time period from high to low according to the overall priority of resources to obtain the real-time scheduling result.

[0115] In this step, for any resource Based on the physical adjustable capacity of resources The upper-level scrolling optimization provides the current allowed call capacity. and the aforementioned adjustment budget Calculate the available adjustment capacity of resource r that can be used by lower-level scheduling in the current time period. :

[0116] ,

[0117] Calculate the maximum available capacity among the current candidate resources. and minimum available capacity The available adjustability capacity of resource r is normalized to obtain the adjustability margin index. :

[0118] ,

[0119] Calculate the unit adjustment cost of resource r :

[0120] ,

[0121] in, Let r be the adjustment cost function of resource r in time period t. The adjustment power of resource r in time period t;

[0122] Calculate the maximum marginal adjustment cost among the current candidate resources. and minimum marginal adjustment cost The unit adjustment cost is then reverse-normalized to obtain the economic indicators. :

[0123] ,

[0124] The response reliability index of computing resource r in time period t :

[0125] ,

[0126] in, For the historical response level of resource r in time period t, Let r be the suitability of the current state of resource r in time period t. For resource r, the willingness to respond in the current time period of time t. Let be the weight coefficient, and satisfy... ;

[0127] Computing resource response fulfillment rate in the previous period :

[0128] ,

[0129] in, Let r be the power of the scheduling instructions received by resource r in the previous time period t−1. This represents the actual response power of resource r in the previous time period t−1;

[0130] Historical response levels are updated using exponential smoothing. :

[0131] ,

[0132] in, Historical memory coefficient;

[0133] Calculate the maximum reliability index among the current candidate resources. and minimum reliability index Regarding the response reliability index Normalization is performed to obtain the response reliability index. :

[0134] ,

[0135] The forward retention value of computing resource r in time period t :

[0136] , |

[0137] in, Before the current invocation of resource r, resource r in the future time period Adjustable capacity, After the current invocation of resource r, resource r will be available in a future time period. Adjustable capacity, Optimize window length for scrolling;

[0138] For electric vehicle load, calculate the off-grid urgency of the j-th electric vehicle in time period t. :

[0139] ,

[0140] in, The charging efficiency of the j-th electric vehicle. Let be the battery capacity of the j-th electric vehicle;

[0141] Calculate the capacity correction factor for the j-th electric vehicle. :

[0142] ,

[0143] Corrected effective adjustable capacity of electric vehicles for:

[0144] ,

[0145] in, Let be the original adjustable capacity of the j-th electric vehicle;

[0146] For the drone load, calculate the task urgency of the u-th drone in time period t. :

[0147] ,

[0148] in, The charging efficiency of the uth drone;

[0149] Calculate the capacity correction factor for the u-th UAV. :

[0150] ,

[0151] Corrected effective adjustable capacity of drones for:

[0152] ,

[0153] in, The original adjustable capacity of the uth drone;

[0154] Build the overall priority of resource r in time period t :

[0155] ,

[0156] in, For dynamic weights with adjustable margins, For economic dynamic weights, In response to the dynamic weights of reliability, Dynamic weights that retain value for forward-looking purposes, and satisfy the following conditions: ;

[0157] Calculate the emergency factor of the deviation :

[0158] ,

[0159] in, The maximum deviation magnitude within the historical or forecast window;

[0160] Candidate resources are ranked according to the comprehensive priority. Sort the resources from largest to smallest to obtain an ordered set. , where n is the total number of candidate resources;

[0161] Let the magnitude of the deviation to be eliminated in the current time period be... The initial residual deviation is For the k-th resource after sorting Distribute adjustment amount :

[0162] ,

[0163] in, For the k-th resource Available adjustable capacity, The remaining deviation after allocating the (k-1)th resource;

[0164] Call resources Then, update the remaining deviation. ,like If the current time-period deviation has been completely eliminated, the scheduling process stops. If a residual deviation still exists after all internal flexible resources have been utilized, the external power grid is invoked as a backup, ultimately yielding the real-time adjustment amounts of each resource for the current time period. .

[0165] In summary, the method in this application generates predicted wind and solar load power and flexible load baseline power based on multi-source historical operating data, establishes a virtual power plant resource hierarchy set and various resource operation constraints; constructs a day-ahead robust power balance plan with the goal of minimizing operating costs and wind and solar risk costs, describes wind and solar uncertainties through typical scenarios and adaptive KL probabilistic fuzzy sets, and configures reserve capacity; performs a rehearsal according to the day-ahead plan during the intraday phase, calculates the system comprehensive deviation and decomposes it into source-side, load-side, and flexible load-side sub-deviations; establishes an upper-level hierarchical rolling optimization model to determine the adjustment budget and call boundary of each resource; uses a dynamic weighted improved greedy algorithm to allocate real-time adjustment amounts under budget constraints; and finally performs anti-reverse adjustment verification and closed-loop feedback to the next time period; thus improving the power balance capability of the virtual power plant and the collaborative adjustment efficiency of multiple flexible resources under wind and solar uncertainty conditions.

[0166] Please see Figure 2 The diagram shows a structural block diagram of a multi-element flexible load collaborative scheduling system for dealing with wind and solar uncertainties, as proposed in this application.

[0167] like Figure 2 As shown, the multi-functional flexible load collaborative scheduling system 200 includes an acquisition module 210, a construction module 220, a calculation module 230, a partitioning module 240, and an output module 250.

[0168] The acquisition module 210 is configured to acquire multi-source historical operating data and day-ahead external input data, generate day-ahead predicted power for wind power, photovoltaic power, base load, and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operating constraints; the construction module 220 is configured to construct a day-ahead robust power balance plan based on the day-ahead predicted power for wind power, photovoltaic power, base load, flexible load baseline power, and various resource operating constraints, and obtain the day-ahead scheduling and allocation results for each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes wind and solar prediction uncertainties through typical scenarios and probabilistic fuzzy sets, and is configured with reserve capacity; the calculation module 230… The system is configured to collect updated wind and solar power, load power, and resource status data during the daytime phase, perform a pre-execution simulation according to the daytime robust power balance plan, calculate the daytime system comprehensive deviation, and decompose the daytime system comprehensive deviation into sub-item deviations from different sources; the partitioning module 240 is configured to divide real-time adjustable resources into multi-level resource pools based on the daytime system comprehensive deviation and the sub-item deviations, and according to the virtual power plant resource hierarchy set, establish an upper-level hierarchical rolling optimization model, and determine the adjustment budget and current allowable call boundary of each resource within the short-term prediction window; the output module 250 is configured to determine the available adjustment capacity of each resource under the constraints of the adjustment budget and the current allowable call boundary, and allocate the specific adjustment amount for the current time period according to the resource comprehensive priority from high to low using a greedy algorithm to obtain the real-time scheduling result.

[0169] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0170] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the multi-variable flexible load collaborative scheduling method for coping with wind and solar uncertainties in any of the above method embodiments.

[0171] In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows:

[0172] Acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic power day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints;

[0173] Based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, a day-ahead robust power balance plan is constructed to obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is equipped with reserve capacity.

[0174] During the daytime phase, updated wind and solar power, load power, and resource status data are collected. An execution rehearsal is performed according to the daytime robust power balance plan. The daytime system comprehensive deviation is calculated, and the daytime system comprehensive deviation is decomposed into sub-deviations from different sources.

[0175] Based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, the real-time adjustable resources are divided into multi-level resource pools, and an upper-level hierarchical rolling optimization model is established to determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window.

[0176] Under the constraints of the adjustment budget and the current allowed call boundary, the available adjustment capacity of each resource is determined according to the operational constraints of various resources. A greedy algorithm is used to allocate the specific adjustment amount for the current time period in descending order of the overall resource priority to obtain the real-time scheduling result.

[0177] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of a multi-functional flexible load coordination scheduling system to address wind and solar uncertainties. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely configured relative to a processor, which can be connected via a network to the multi-functional flexible load coordination scheduling system to address wind and solar uncertainties. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0178] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3 As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the multi-functional flexible load collaborative scheduling method for addressing wind and solar uncertainties as described in the above-described method embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the multi-functional flexible load collaborative scheduling system for addressing wind and solar uncertainties. The output device 340 may include a display screen or other display device.

[0179] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0180] In one implementation, the aforementioned electronic device is applied to a multi-various flexible load collaborative scheduling system for addressing wind and solar uncertainties. As a client, it 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, which, when executed by the at least one processor, enable the at least one processor to:

[0181] Acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic power day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints;

[0182] Based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, a day-ahead robust power balance plan is constructed to obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is equipped with reserve capacity.

[0183] During the daytime phase, updated wind and solar power, load power, and resource status data are collected. An execution rehearsal is performed according to the daytime robust power balance plan. The daytime system comprehensive deviation is calculated, and the daytime system comprehensive deviation is decomposed into sub-deviations from different sources.

[0184] Based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, the real-time adjustable resources are divided into multi-level resource pools, and an upper-level hierarchical rolling optimization model is established to determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window.

[0185] Under the constraints of the adjustment budget and the current allowed call boundary, the available adjustment capacity of each resource is determined according to the operational constraints of various resources. A greedy algorithm is used to allocate the specific adjustment amount for the current time period in descending order of the overall resource priority to obtain the real-time scheduling result.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for coordinated scheduling of multiple flexible loads to address the uncertainties of wind and solar power, characterized in that, include: Acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic power day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints; Based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, a day-ahead robust power balance plan is constructed to obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is equipped with reserve capacity. During the intraday phase, updated wind and solar power, load power, and resource status data are collected. An execution rehearsal is performed according to the day-ahead robust power balance plan. The intraday system comprehensive deviation is calculated and decomposed into sub-deviations from different sources. Based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, the real-time adjustable resources are divided into multi-level resource pools, and an upper-level hierarchical rolling optimization model is established to determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window. Under the constraints of the adjustment budget and the current allowed call boundary, the available adjustment capacity of each resource is determined according to the various resource operation constraints. A greedy algorithm is used to allocate the specific adjustment amount for the current time period according to the overall resource priority from high to low, so as to obtain the real-time scheduling result.

2. The method for coordinated scheduling of multiple flexible loads to address wind and solar uncertainties according to claim 1, characterized in that, The process of generating day-ahead predicted power for wind power, day-ahead predicted power for photovoltaic power, day-ahead predicted power for base load, and baseline power for flexible load based on the multi-source historical operating data and day-ahead external input data, and establishing a virtual power plant resource hierarchy set and various resource operating constraints, includes: Wind power forecasting input feature vectors are extracted from the multi-source historical operational data and day-ahead external input data. The expression is: ， in, The input feature vector for wind power prediction in time period t. Let be the wind speed vector for time period t. Let t be the wind direction. Let be the air density during time period t. For the historical wind power output during time period t, Let t represent the available capacity of the wind turbines during time period t. The operating status of the wind turbine during time period t; The photovoltaic prediction input feature vector is extracted from the multi-source historical operating data and the current-day external input data, and the expression is: ， In the formula, The input feature vector for photovoltaic prediction in time period t. Let be the solar irradiance during time period t. Let t be the ambient temperature. Let be the temperature of the photovoltaic module during time period t. Let be the cloud cover in time period t. The humidity at time t. For the historical photovoltaic output of time period t, Let t be the available photovoltaic installed capacity during time period t; The base load forecast input feature vector is extracted from the multi-source historical operating data and the current-day external input data, and the expression is: ， In the formula, The input feature vector for the basic load forecast in time period t is... The historical base load power for time period t. For time period t, the day type is... Let t be the time period number. This is the holiday identifier for time period t. The seasonal type for time period t; The wind power prediction input feature vector, the photovoltaic prediction input feature vector, and the base load prediction input feature vector are respectively input into the LSTM prediction model. The LSTM prediction model outputs the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, and the day-ahead predicted power of base load, respectively, as expressed in the following expressions: ， ， ， In the formula, The day-ahead forecast output of wind power for time period t. For the day-ahead forecast of photovoltaic power generation in time period t, For wind power day-ahead forecasting models, This is a day-ahead forecasting model for photovoltaics. Based on the day-ahead forecasting model for basic load; Extract the state feature vector of the flexible load from the multi-source historical operating data and the current daytime external input data. For any flexible load... , , For air conditioning load, For electric vehicle load, For drone load; The flexible load State feature vector Input a baseline prediction model combining clustering and LightGBM regression, and output the flexible load. Baseline power when not participating in scheduling during time period t , For baseline prediction models; The expression for the state characteristic vector of the air conditioning load AC is: ， In the formula, This is the state feature vector of the air conditioning load (AC). Let be the indoor temperature during time period t. Set the temperature for users in time period t. The historical operating power of the air conditioner during time period t. The historical operating power of the air conditioner during time period t. The user occupancy status for time period t; The expression for the state feature vector of the electric vehicle load (EV) is: ， In the formula, This represents the state feature vector of the electric vehicle load (EV). Let be the time when the j-th electric vehicle enters the system. Let be the time when the j-th electric vehicle leaves the network. Let j be the current state of charge of the j-th electric vehicle in time period t. Let j be the minimum state of charge required for the j-th electric vehicle to disconnect from the grid. The maximum charging power of the j-th electric vehicle; The expression for the state feature vector of the UAV (Unmanned Aerial Vehicle) is: ， In the formula, This represents the state feature vector of the UAV (Unmanned Aerial Vehicle) payload. Let u be the current remaining energy of the u-th drone in time period t. The mission deadline for the uth drone. This represents the minimum energy required before the end of the uth drone mission. Let u be the energy consumption of the uth drone during time period t. The maximum charging power for the uth drone; Based on the aforementioned flexible load resource set and energy storage systems Define the internal flexible resource set Based on the aforementioned internal flexible resource set Interacting with external power grid resources Define a set of real-time adjustable resources to obtain the set of resource levels of the virtual power plant; For the aforementioned air conditioning load AC, establish air conditioning comfort constraints. and air conditioning power constraints ,in, Let be the indoor temperature of the i-th air-conditioning user during time period t. and Let $\begin{cases}$ be the lower and upper limits of the acceptable indoor temperature for the $i$-th air conditioning user, respectively. Let be the operating power of the i-th air conditioner during time period t. The maximum operating power of the i-th air conditioner; For the electric vehicle load (EV), establish an electric vehicle charging power constraint. Off-grid power constraints ,in, Let j be the charging power of the j-th electric vehicle during time period t. For the j-th electric vehicle at the off-grid time The state of charge; For the aforementioned UAV (Unmanned Aerial Vehicle) payload, establish UAV energy boundary constraints. and task deadline constraints ,in, and These are the lower and upper limits of the energy of the u-th drone, respectively. For the uth drone at the mission deadline The remaining energy; For the energy storage system ESS, establish energy storage boundary constraints. and energy storage charging and discharging power constraints , and charging / discharging mutual exclusion constraints ,in, The energy state of energy storage during time period t. and These are the lower and upper limits of energy storage capacity, respectively. and These represent the charging power and discharging power of the energy storage system during time period t, respectively. and These are the maximum charging power and maximum discharging power of energy storage, respectively. and These represent the charging and discharging state variables of energy storage during time period t, respectively.

3. The method for coordinated scheduling of multiple flexible loads to address wind and solar uncertainties according to claim 1, characterized in that, The process of constructing a day-ahead robust power balance plan based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power, and the various resource operation constraints, and obtaining the day-ahead scheduling and allocation results for each resource, includes: Based on the day-ahead predicted wind power The photovoltaic day-ahead predicted power The day-ahead forecast power of the base load Based on the baseline power of each flexible load, a day-ahead power balance relationship is established, expressed as: ， In the formula, The day-ahead discharge power of the energy storage during time period t, The power purchased by the external power grid on the day of day t. , , These represent the day-ahead planned power outputs of air conditioners, electric vehicles, and drones, respectively, during time period t. The day-ahead charging power of energy storage during time period t. The daytime electricity sales power of the external power grid during time period t; Based on historical landscape data, day-ahead center prediction results, and prediction residuals, K typical landscape scenes are generated using residual resampling and scene clustering methods. KL probabilistic fuzzy sets are introduced to describe the shift in the true probability distribution of the landscape scenes relative to their nominal probabilities. The expression is: ， In the formula, Let be the probability of the s-th typical scenario under the true probability distribution. Let be the nominal probability of a typical scenario s. Let the radius of the KL fuzzy set be ; The KL fuzzy set radius adopts an adaptive KL radius. The expression is: ， In the formula, Based on the KL radius, To predict the influence coefficient of error fluctuation intensity on the KL radius, Let be the intensity of the wind and solar forecast error fluctuation over time period t. The influence coefficient of weather disturbance intensity on the radius of KL. The intensity of the weather disturbance during time period t; For the s-th typical scenario, the source-side deviation at time t Represented as: ， In the formula, For the s-th typical scenario, the wind power output during time period t is... For the s-th typical scenario, the photovoltaic output during time period t; The objective function for constructing the current robust power balance plan is expressed as follows: ， ， ， ， In the formula, The base operating cost is the current day. Cost of configuring standby capacity, To account for the uncertainties and risks associated with wind and solar power, For the set of scheduling periods, The duration of a single scheduling period. The electricity purchase price for time period t. The electricity price for time period t. The day-ahead adjustment cost of the air conditioner during time period t. For the day-ahead adjustment cost of electric vehicles in time period t, The day-ahead adjustment cost of the drone in time period t. The day-ahead operating cost of energy storage during period t, Configuration cost coefficient for upward backup The configuration cost factor for backward backup. For time period t, the upward reserve capacity is... This represents the downward reserve capacity for time period t. This represents the risk cost of the day-ahead plan under the s-th typical scenario; The objective function is solved under the constraints of the various resource operations to obtain the day-ahead scheduling and allocation results for each resource.

4. The multi-element flexible load collaborative scheduling method for coordinating wind and solar uncertainties according to claim 3, characterized in that, The process involves collecting updated wind and solar power, load power, and resource status data during the intraday phase, conducting a pre-execution simulation according to the day-ahead robust power balance plan, calculating the intraday system comprehensive deviation, and decomposing the intraday system comprehensive deviation into sub-deviations from different sources, including: Collect wind power updated within the day Photovoltaic power Basic load power And the air conditioner's executable power before real-time correction. The executable power of electric vehicles before real-time correction. And the executable power of the drone before real-time correction. ; The execution rehearsal of the aforementioned robust power balance plan was carried out, and the intraday rehearsal supply-side power was calculated. Demand-side power forecast for the day The expression is: ， ， Calculate the intraday system comprehensive deviation The intraday system comprehensive deviation is decomposed into intraday wind source side deviation. Intraday basic load deviation Air conditioning execution deviation Electric vehicle execution deviation and drone execution deviation The expression is: ， ， ， ， 。 5. A method for coordinated scheduling of multiple flexible loads to address wind and solar uncertainties according to claim 2, characterized in that, The step of establishing an upper-level hierarchical rolling optimization model by dividing real-time adjustable resources into multi-level resource pools based on the intraday system comprehensive deviation and the individual item deviations, and according to the virtual power plant resource hierarchy set, includes: Set up a real-time scrolling window Based on the latest wind and solar forecasts, load forecasts, deviation trends, and resource status, a forecast deviation sequence is formed for the future window. ,in, Optimize window length for scrolling; The set of real-time adjustable resources in the virtual power plant resource hierarchy set Divided into flexible load resource pool Energy storage resource pool and external power grid backup resource pool ; For any resource Let the resources be Time period within the scrolling window The predicted adjustment amount is Calculate the time period within the scrolling window Predicted residual bias The expression is: ， Based on the time period within the scrolling window The objective function of the upper-level hierarchical rolling optimization model is established based on the predicted residual bias, and its expression is: ， ， ， ， In the formula, To predict the residual bias penalty coefficient, For resources Within the scrolling window The predicted adjustment costs For time period Layered call penalty cost, For time period Adjustment and smoothing costs, For time period The cost of excessive resource use The penalty weight for tiered calls, To adjust the penalty weights for smoothing, Penalty weights for excessive resource use This refers to the penalty coefficient for accessing energy storage resource pools. The penalty coefficient for calling external power grid resource pools. For resources Time period within the scrolling window The predicted adjustment amount, For resources Within the scrolling window Cumulative usage It is a very small positive number. For resources Within the scrolling window Maximum effective adjustable capacity; The constraints of the upper-level hierarchical rolling optimization model include: The resource adjustment boundary constraint is expressed as follows: ， The resource budget constraint for the scrolling window is expressed as follows: ， In the formula, For resources The adjustment budget allowed within the current scrolling window.

6. A multi-element flexible load collaborative scheduling method for coordinating wind and solar uncertainties according to claim 5, characterized in that, The step of determining the adjustment budget and current allowed call boundaries for each resource within the short-term forecast window includes: Based on the source-side bias contribution The contribution of the basic load deviation The contribution of the flexible load deviation and the rate of change of deviation A resource budget adjustment model is established, and the adjusted budget for resource r after adjustment in time period t is calculated using the resource budget adjustment model. The expression is: ， In the formula, The basic budget for resource r, Let r be the sensitivity coefficient of resource r to source-side deviation. Let r be the sensitivity coefficient of resource r to the deviation of base load. Let r be the sensitivity coefficient of resource r to flexible load deviation. Let r be the sensitivity coefficient of resource r to the trend of deviation change. The deviation rate over time period t. To predict the maximum rate of change of deviation within the window, It is a very small positive number; Substitute the calculated adjustment budget into the rolling window resource budget constraint of the upper-level hierarchical rolling optimization model, solve the upper-level hierarchical rolling optimization model, and obtain the current allowed call capacity of resource r in time period t.

7. A multi-element flexible load collaborative scheduling system for coping with wind and solar uncertainties, characterized in that, include: The acquisition module is configured to acquire multi-source historical operating data and day-ahead external input data, generate wind power day-ahead predicted power, photovoltaic day-ahead predicted power, base load day-ahead predicted power and flexible load baseline power based on the multi-source historical operating data and day-ahead external input data, and establish a virtual power plant resource hierarchy set and various resource operation constraints; The construction module is configured to construct a day-ahead robust power balance plan based on the day-ahead predicted power of wind power, the day-ahead predicted power of photovoltaic power, the day-ahead predicted power of base load, the flexible load baseline power and the various resource operation constraints, and obtain the day-ahead scheduling and allocation results of each resource. The day-ahead robust power balance plan aims to minimize the sum of operating costs and wind and solar uncertainty risk costs, and describes the wind and solar prediction uncertainty through typical scenarios and probabilistic fuzzy sets, and is configured with reserve capacity. The calculation module is configured to collect updated wind and solar power, load power and resource status data during the day, perform a pre-execution simulation according to the day-ahead robust power balance plan, calculate the day-ahead system comprehensive deviation, and decompose the day-ahead system comprehensive deviation into sub-item deviations from different sources. The partitioning module is configured to divide real-time adjustable resources into multi-level resource pools based on the intraday system comprehensive deviation and the sub-item deviation, and according to the virtual power plant resource hierarchy set, establish an upper-level hierarchical rolling optimization model, and determine the adjustment budget and current allowed call boundary of each resource in the current window within the short-term prediction window; The output module is configured to determine the available adjustment capacity of each resource based on the constraints of the adjustment budget and the current allowed call boundary, and to allocate the specific adjustment amount for the current time period according to the overall priority of the resources from high to low using a greedy algorithm to obtain the real-time scheduling result.

8. An electronic device, characterized in that, include: 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 to enable the at least one processor to perform the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the method described in any one of claims 1 to 6.

Images