A Long-Term Joint Dispatch Method for Virtual Power Plants

By dynamically adjusting the scheduling priorities of wind, solar, energy storage, and flexible loads in the virtual power plant, and combining the equivalent losses and cumulative compensation costs throughout the entire life cycle, the resource allocation sequence is optimized, solving the problem of unreasonable resource use in the long-term scheduling scheme of the virtual power plant, and achieving cost reduction and improved market adaptability.

CN123073798APending Publication Date: 2026-11-10GUONENG (SHAANXI) ENERGY SALES CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611549328.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-08
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

The existing long-term dispatch schemes for virtual power plants fail to effectively distinguish between wind and solar deviations, energy storage life loss, flexible load compensation, and resource allocation sequence, resulting in distorted total dispatch cost estimates, unreasonable resource utilization, accelerated energy storage life decay, and decreased flexible load coordination, making them unable to adapt to the needs of multiple market tasks.

Method used

By acquiring monthly bilateral energy contract curves, wind and solar power output forecasts, energy storage health status, and flexible load information, the scheduling priorities of wind and solar power, energy storage, and flexible loads are dynamically adjusted. Combined with the equivalent loss cost and cumulative compensation cost throughout the entire life cycle, the resource allocation order is optimized, and state switching penalties are set to adapt to market changes.

Benefits of technology

It enables the rational scheduling of wind, solar, energy storage, and flexible loads, reduces the total scheduling cost of the virtual power plant throughout its entire lifecycle, protects the lifespan of energy storage, improves the coordination of flexible loads, adapts to the needs of multiple market transactions, and enhances the economy and adaptability of scheduling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123073798A_ABST
    Figure CN123073798A_ABST
Patent Text Reader

Abstract

This invention relates to the field of joint dispatching technology, specifically to a medium- and long-term joint dispatching method for virtual power plants. The method includes the following steps: acquiring monthly bilateral energy contract curves, wind and solar power output forecast information, SOH values, and cumulative flexible load execution information; calculating the expected settlement cost of wind and solar power output deviating from the contract, the equivalent loss cost of energy storage over its entire life cycle, and the cumulative compensation cost; classifying market operation states into three categories based on the level of supply guarantee instructions, determining the corresponding resource dispatching priority coefficients, adding hysteresis intervals and switching penalties to suppress frequent switching, and finally solving the joint dispatching plan with the goal of minimizing the total dispatching cost while satisfying two types of constraints. This invention quantifies the implicit cost differences of different regulatory resources, dynamically adjusts resource dispatching priorities according to market and operational scenarios, adapts to the needs of virtual power plants participating in multiple types of electricity market transactions, reduces the overall medium- and long-term dispatching cost, and avoids the blind use of high-cost resources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of joint dispatching technology, and more specifically, to a long-term joint dispatching method for a virtual power plant. Background Technology

[0002] Virtual power plants are distributed resources that aggregate distributed new energy sources, energy storage systems, and flexible loads, relying on information and communication technologies and software control systems, and participate as a whole in grid dispatch and electricity market transactions. In recent years, the installed capacity of new energy sources in China has continued to increase, and the reform of the electricity market has been gradually promoted. Virtual power plants need to fulfill multiple types of transaction contracts and coordinate the operation rhythm of various adjustable resources within the medium and long term. Medium and long term joint dispatch has become the core technical link of virtual power plant operation and management. Current mainstream long-term dispatch schemes for virtual power plants mostly aim to minimize overall operating costs or maximize renewable energy consumption. They follow a fixed-priority dispatch logic, allocating deviation adjustment tasks according to the rule of prioritizing wind and solar power, followed by energy storage, and finally covering flexible loads. This priority does not change with market conditions and resource adjustment costs. In terms of adjustment cost accounting, wind and solar deviations are mostly estimated based on a unified spot price, energy storage losses are mostly taken as a fixed unit electricity cost, and flexible load compensation is mostly calculated based on a unified compensation price. In practice, problems with this type of solution gradually emerge in several specific aspects. First, the settlement of wind and solar power deviations does not distinguish between over-generation and under-generation. In the actual spot market, over-generation is settled at the over-generation settlement price, while under-generation is purchased at the under-generation purchase price. The different price standards mean that the cost calculated using a unified price does not match the actual settlement cost. The larger the deviation, the more obvious the distortion, which is directly passed to subsequent optimization solutions. Second, the energy storage lifespan loss does not distinguish between depth of discharge. Discharging 1 kilowatt-hour has a significantly different impact on battery lifespan depending on whether it's shallow or deep discharge. A fixed unit cost prevents the optimization model from recognizing the cost difference between shallow and deep cycles, leading to a bias towards selective discharge during arbitrage or peak periods. Concentrated deep deployment during high-price periods yields short-term gains but accelerates battery cycle life depletion; once battery health declines, damage at the same depth is greater and more pronounced; thirdly, flexible load compensation does not differentiate between quota limits and the number of calls; in actual agreements, monthly cumulative reductions within the quota are charged at the base price, while exceeding the limit is charged at a higher tiered price, and repeated calls also generate fatigue-based price increases; the existing solution does not distinguish these differences, and the optimization model treats flexible load as a resource with a fixed price, calling it whenever reduction is needed, leading to overuse, decreased user willingness to cooperate, and inability to call it when truly needed; fourthly, the fixed resource call order cannot adapt to multi-market tasks. Virtual power plants simultaneously face three types of tasks: energy market arbitrage, ancillary service capacity sales, and grid supply guarantee orders. These three tasks have different resource requirements, and fixed priorities cannot reflect these differences. This can easily lead to situations where energy storage is still arbitrageurized according to energy market logic when ancillary service revenue is high, and flexible loads fail to step up in time when supply guarantee orders are upgraded. Fifth, the order of dispatch is disconnected from the cost of regulation. The marginal cost of one wind and solar dispatch, the lifespan loss of one energy storage discharge, and the compensation expenditure for one flexible load reduction are inconsistent in terms of dimensions and scope. There is a lack of a unified benchmark, and the optimization model can only rely on artificially set priority coefficients for ranking. However, these coefficients do not change with the market environment and resource status, so the order of dispatch does not match the actual economics. The combination of these problems leads to two main issues: firstly, it distorts the overall scheduling cost estimate, causing the scheduling plan derived from the optimization model to exceed the expected cost during actual execution; secondly, it results in unreasonable resource utilization, accelerated degradation of energy storage lifespan, decreased coordination of flexible loads, and failure of energy storage and flexible loads to switch roles in ancillary services and supply guarantee tasks in a timely manner as required by the tasks, thus failing to fully realize the benefits of multi-market participation; therefore, a long-term joint scheduling method for virtual power plants is urgently needed. Summary of the Invention

[0003] The purpose of this invention is to provide a long-term joint dispatch method for virtual power plants to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides a method for long-term joint dispatching of a virtual power plant, comprising the following steps: S1. Obtain monthly bilateral energy contract curves, monthly ancillary service capacity contract information, wind and solar power output forecast information, spot market price forecast information, SOH value, and cumulative execution information of flexible load; S2. Based on the monthly bilateral energy contract curve and wind and solar power output forecast information, determine the expected settlement cost of wind and solar power output deviating from the contract curve in each period of the scheduling cycle. S3. Determine the equivalent loss cost of the entire life cycle of energy storage within the scheduling period based on the energy storage discharge depth and energy storage battery health status information for each time period within the scheduling period. S4. Determine the cumulative compensation cost of the flexible load based on the cumulative reduction amount and cumulative number of calls in the cumulative execution information of the flexible load; S5. Based on the ancillary service market price, peak-valley price difference forecast, and grid dispatch instruction level, identify the current market operation status of the virtual power plant, and determine the dispatch priority coefficients of wind and solar power, energy storage, and flexible loads under the current market operation status. S6. Taking the minimum total scheduling cost within the scheduling cycle as the optimization objective, and combining the expected settlement cost of wind and solar power, the equivalent loss cost of energy storage throughout its entire life cycle, the cumulative compensation cost, and the scheduling priority coefficient, under the premise of satisfying the cross-time coupling constraint of energy storage charge state and the continuous call constraint of flexible load, the joint scheduling plan of wind and solar power, energy storage and flexible load within the future scheduling cycle is obtained.

[0005] As a further improvement to this technical solution, S2 includes the following method steps: The predicted power output of wind and solar power in each time period within the scheduling cycle is compared with the monthly bilateral energy contract curve for each time period to obtain the deviation power. When the deviation in electricity volume is greater than zero, the expected revenue loss corresponding to the excess electricity volume during that period shall be calculated based on the spot market over-generation settlement price. When the deviation in electricity volume is less than zero, the expected replenishment cost corresponding to the shortfall in electricity volume during that period is calculated based on the spot market purchase price for the shortfall. The expected revenue loss and expected repurchase cost for each period are added together to obtain the expected settlement cost of wind and solar power within the scheduling cycle.

[0006] As a further improvement to this technical solution, S3 includes the following method steps: The discharge depth for each time period is determined based on the energy storage discharge amount and rated energy storage capacity during each time period within the scheduling cycle. Based on the SOH value, the mapping relationship between the depth of discharge corresponding to the current health state and the cost per unit of energy loss is extracted from the battery cycle life characteristic curve. The discharge depth of each time period is converted into the equivalent lifetime loss cost of each time period according to the mapping relationship; the equivalent lifetime loss cost of each time period is accumulated and converted according to the battery replacement cost and the rated cycle number to obtain the equivalent loss cost of the energy storage life cycle within the scheduling cycle.

[0007] As a further improvement to this technical solution, the mapping relationship between the depth of discharge and the cost per unit of energy loss includes the following steps: The unit power lifetime loss cost increases nonlinearly with the increase of discharge depth. The nonlinear exponent in the mapping relationship is updated online as the SOH value decreases. The distribution ratio of the total energy of energy storage in the scheduling cycle between deep cycle discharge and shallow cycle discharge is used as the optimization variable. The optimization variable and the discharge depth of each time period jointly participate in the search for the minimum total scheduling cost. The division between deep cycle discharge and shallow cycle discharge is defined by the discharge depth threshold.

[0008] As a further improvement to this technical solution, S4 includes the following method steps: Compare the cumulative reduction of flexible load within the scheduling cycle with the monthly cumulative reduction limit agreed upon in the contract; When the cumulative reduction does not exceed the monthly cumulative reduction limit, the basic compensation fee corresponding to the electricity reduction in each time period is calculated based on the basic compensation price. When the cumulative reduction exceeds the monthly cumulative reduction limit, the excess portion will be compensated at a tiered compensation price that is higher than the basic compensation price. The fatigue surcharge coefficient is determined based on the cumulative number of times the flexible load is called within the scheduling cycle, and the compensation cost for each time period is adjusted according to the fatigue surcharge coefficient to obtain the adjusted compensation cost for each time period. The adjusted compensation costs for each time period are summed to obtain the cumulative compensation cost for the flexible load.

[0009] As a further improvement to this technical solution, S5 includes the following method steps: Calculate the first ratio between the quoted capacity of the ancillary services market and the expected return of energy market arbitrage; calculate the second ratio between the predicted peak-valley price difference and the threshold for initiating energy storage arbitrage; Obtain the supply guarantee instruction level issued by the power grid dispatching agency; based on the first ratio, the second ratio and the supply guarantee instruction level, classify the current operating status of the virtual power plant into one of the following: energy market-dominated status, ancillary service-dominated status and supply guarantee constraint-dominated status; pre-configure the corresponding resource scheduling priority coefficient combination for the market operating status.

[0010] As a further improvement to this technical solution, classifying market operation status based on the first ratio, the second ratio, and the supply guarantee order level includes the following methods and steps: When the first ratio rises from below the first lower limit to above the first upper limit and continuously meets the preset number of rolling cycles, the market operation status will be switched from the energy market-dominated state to the ancillary service-dominated state. When the first ratio drops from above the first upper limit to below the first lower limit and continuously meets the preset rolling cycle number, the market operation state will be switched from the ancillary service-dominated state back to the energy market-dominated state; the first upper limit is higher than the first lower limit, and the two constitute the hysteresis interval for state switching.

[0011] As a further improvement to this technical solution, the optimization objective in S6 includes a state switching penalty term, which is determined based on the change in scheduling priority coefficient between two adjacent rolling cycles.

[0012] As a further improvement to this technical solution, S6 also includes the following method steps: When the cumulative number of calls to the flexible load exceeds the preset call count threshold, the callable capacity of the flexible load is reduced by a reduction factor. The reduced callable capacity is then used as the upper limit of the continuous call constraint of the flexible load in the rolling optimization solution. The reduction factor decreases non-linearly with the increase of the cumulative number of calls, and the reduction factor is updated online with the cumulative execution information of the flexible load.

[0013] As a further improvement to this technical solution, the priority of resource allocation for each time period in S6 is determined by the medium-to-long-term adjustment cost from smallest to largest; the cross-time period coupling constraint of energy storage charge state and the continuous allocation constraint of flexible load are embedded as hard constraints in the optimization model to participate in global optimization; the marginal impact cost of contract assessment in the medium-to-long-term adjustment cost is determined by taking the partial derivative of the time period deviation power with respect to the cost calculation relationship related to the deviation power in the monthly bilateral energy contract assessment rules, and the partial derivative is a piecewise constant.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This method for long-term joint dispatch of virtual power plants eliminates the cumulative losses of wind and solar power over time, only calculating the marginal impact cost of contract assessments. Energy storage is calculated based on its current battery health status, combined with a nonlinear mapping of discharge depth for each time period, to determine the equivalent loss cost over its entire lifecycle, incorporating both direct losses and marginal impacts. Flexible loads are priced in tiers based on the cumulative number of calls, with a nonlinear reduction in callable capacity after exceeding a threshold, calculating the cumulative compensation cost over time. A state-switching penalty term is introduced to suppress frequent market state switching. Combined with the current market dispatch priority, the joint dispatch plan is solved while satisfying various constraints. The method quantifies the implicit adjustment cost differences of different resources, avoiding the blind use of high-cost resources and effectively reducing the total dispatch cost of the virtual power plant over its entire lifecycle. Compared to dispatch schemes that do not classify and calculate implicit costs, the economic efficiency is significantly improved. It also takes into account the lifecycle protection of energy storage and the production constraints of flexible load users, further optimizing dispatch economics while meeting supply guarantee contract requirements.

[0015] 2. This method for medium- and long-term joint dispatch of virtual power plants identifies the current market operating status by using the first ratio of ancillary service market bids to energy market arbitrage profits, the second ratio of peak-valley price differences to the activation threshold, and the grid supply guarantee command level. The status is divided into three categories: energy market-dominated, ancillary service-dominated, and supply guarantee constraint-dominated. A corresponding dispatch priority coefficient combination is pre-configured for each status. A hysteresis interval is set, and the status is switched only when the ratio exceeds the limit for multiple consecutive rolling cycles. A status switching penalty term based on priority changes is added to suppress frequent switching. Finally, the dispatch plan is solved by combining the priority coefficients with the goal of minimizing the total dispatch cost. This method abandons the fixed-priority dispatch logic and can dynamically switch resource dispatch priorities according to different market scenarios and supply guarantee needs. It adapts to the diverse trading needs of virtual power plants participating in the energy market, ancillary service market, and supply guarantee tasks simultaneously. This ensures maximum returns in the core scenario while avoiding the problem of fixed priorities failing to adapt to scenario changes, thus improving the adaptability of medium- and long-term dispatch of virtual power plants to multi-market and multi-objective scenarios. Attached Figure Description

[0016] Figure 1 This is a flowchart summarizing the specification of the present invention; Figure 2 This is a flowchart illustrating the overall method steps of the present invention. Detailed Implementation

[0017] The technical solutions in 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, and 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.

[0018] Example This invention provides a method for medium- and long-term joint dispatch of virtual power plants, the overall method steps of which are as follows: Figure 2 As shown; S1. Obtain monthly bilateral energy contract curves, monthly ancillary service capacity contract information, wind and solar power output forecast information, spot market price forecast information, SOH value, and cumulative flexible load execution information; S2. Based on the monthly bilateral energy contract curves and wind and solar power output forecast information, determine the expected settlement cost of wind and solar power output deviating from the contract curve for each period within the scheduling cycle; S3. Based on the energy storage discharge depth and energy storage battery health status information for each period within the scheduling cycle, determine the equivalent loss cost of the energy storage life cycle within the scheduling cycle; S4. Based on the cumulative reduction amount and cumulative call count in the cumulative flexible load execution information, determine the cumulative... S5. Based on the ancillary service market price, peak-valley price difference forecast, and grid dispatch instruction level, identify the current market operation status of the virtual power plant and determine the dispatch priority coefficients of wind and solar, energy storage, and flexible loads under the current market operation status; S6. With the goal of minimizing the total dispatch cost within the dispatch cycle, and combining the expected settlement cost of wind and solar, the equivalent loss cost of energy storage throughout its life cycle, the cumulative compensation cost, and the dispatch priority coefficients, under the premise of satisfying the cross-time coupling constraints of energy storage load status and the continuous call constraints of flexible loads, solve for the joint dispatch plan of wind and solar, energy storage, and flexible loads within the future dispatch cycle.

[0019] The following is a detailed explanation of steps S1 to S6 above; specifically, it includes the following method steps: In step S1, the contract and prediction information acquisition operation is performed; specifically, it includes the following sub-steps: Sub-step S1.1: Establish a connection with the contract management system and read the monthly bilateral energy contract curve and monthly ancillary service capacity contract information; the monthly bilateral energy contract curve refers to the bilateral energy trading contract signed between the virtual power plant and the power purchaser or seller in the medium- and long-term market. The contract is decomposed into a sequence of agreed power values ​​by scheduling period, and then the agreed power value for each scheduling period is extracted from the agreed power value sequence. subscript This refers to the English meaning of the contract and is used only for distinction; among them, Indicates the scheduling period index. , The total number of time periods within the scheduling cycle; monthly ancillary service capacity contract information refers to the peak-shaving and frequency regulation ancillary service capacity committed by the virtual power plant and its corresponding contract price; that is, the overall capacity and price values ​​categorized by ancillary service type. , where subscript , These are abbreviations for capacity and price values, used only for distinction; this information is used to characterize the capacity and price level committed by virtual power plants in the ancillary services market. The output data of this sub-step is the sum of the agreed-upon power value, capacity value, and price value; wherein, the agreed-upon power value is used in the subsequent step S2 to compare the predicted wind and solar power output values ​​on a time-by-time basis to determine the wind and solar deviation cost; the sum of the capacity value and price value is used in the subsequent step S5 to calculate the first ratio of the ancillary service market capacity quotation to the expected return of energy market arbitrage, in order to identify the market operating status; Sub-step S1.2: By establishing connections with the meteorological forecasting module and the power forecasting module respectively, obtain wind and solar power output forecast information within the scheduling cycle, which refers to the predicted output values ​​of wind power and photovoltaic power generation for each time period within the future scheduling cycle; specifically: the meteorological forecasting module outputs a meteorological forecast sequence of wind speed and irradiance for the future scheduling cycle based on numerical weather forecast data; the power forecasting module outputs the predicted wind and solar power output values ​​for each time period based on the meteorological forecast sequence, historical output data, and unit operating status. ; Subscript The abbreviations for predicted values ​​are used only for distinction; historical output data is read from the historical operation database of the virtual power plant, and unit operating status is read from the monitoring terminals of each power generation unit. Simultaneously, historical spot price data is read from the electricity market trading platform. A neural network prediction method is used to output the over-generation settlement price and the shortfall purchase price for each time period based on this historical spot price data, thus forming the spot market price prediction information. Therefore, the spot market price prediction information includes the over-generation settlement price for each time period. Price of the shortfall ; Subscript and These are abbreviations for "over-issuance" and "under-issuance," used only to distinguish between them; the over-issuance settlement price refers to the predicted power output value of wind and solar power. When the generated electricity exceeds the agreed-upon value, the settlement price for selling the excess electricity in the spot market; the shortage purchase price refers to the price at which the shortfall in wind and solar power output is purchased in the spot market when the predicted output is less than the agreed-upon value; the predicted wind and solar power output is based on the predicted output value and the duration of the time period. The calculations show that the output data of this sub-step includes the predicted wind and solar power output, the over-generation settlement price, and the shortfall purchase price. The predicted wind and solar power output is used in subsequent step S2 to compare with the agreed-upon power output value over time periods to calculate the deviation power. The sign of the deviation in electricity volume determines whether the over-generation settlement price or the deficit purchase price will be used for settlement in the future; the over-generation settlement price and the deficit purchase price are used in the subsequent step S2 to calculate the wind and solar deviation cost in segments. Sub-step S1.3: Obtain the current SOH value of the energy storage battery by establishing a connection with the battery management system. The State of Harm (SOH) value is used to characterize the ratio of the battery's current capacity to its rated capacity. The battery management system estimates the SOH online using the ampere-hour integral method or the open-circuit voltage method. This estimation method is a conventional technique in the field and will not be elaborated here. Simultaneously, it establishes a connection with the load control terminal or demand response platform to read the cumulative execution information of flexible loads. This information includes the cumulative reduction of flexible loads within the scheduling cycle. With cumulative call count ; Subscript The abbreviation for "cumulative" is used only for distinction; the output data of this sub-step are the SOH value, cumulative reduction amount, and cumulative number of calls; the SOH value is used in the subsequent step S3 to extract the mapping relationship from the battery cycle life characteristic curve between the depth of discharge corresponding to the current health state and the cost per unit of energy loss, and is used to update the life loss nonlinearity index in this mapping relationship online. The cumulative reduction amount will be used in subsequent step S4 to match the monthly cumulative reduction limit agreed upon in the contract. A comparison is made to determine whether the tiered compensation price is triggered; subscript The English meaning of "quota" is used only for distinction; the cumulative number of calls will be used in the subsequent step S4 to determine the fatigue surcharge coefficient. And used to determine whether callable capacity reduction is triggered; subscript The English meaning of fatigue is used only for distinction.

[0020] In step S2, based on the monthly bilateral energy contract curve and wind and solar power output forecast information obtained in step S1, the expected settlement costs of wind and solar power output deviating from the contract curve for each period within the scheduling cycle are determined. ; Subscript The English meaning of "cost" is used only for distinction; it specifically includes the following sub-steps: Sub-step S2.1: Compare the predicted wind and solar power output with the agreed-upon power consumption value time by time. Since the predicted wind and solar power output is power (in kilowatts or megawatts) and the agreed-upon power consumption value is electricity (in kilowatt-hours or megawatt-hours), the units are different and cannot be directly subtracted. Therefore, first multiply the predicted wind and solar power output value by the duration of a single dispatch period. Converted into the predicted power output of wind and solar power during that period. The duration of each time slot is determined based on the scheduling period. For example, in this embodiment, the scheduling period is 15 minutes, meaning each time slot is 0.25 minutes long; or the scheduling period is 30 minutes, meaning each time slot is 0.5 minutes long. The deviation in power output is calculated based on the predicted and agreed power output values ​​from wind and solar power. Then, the deviation in electricity consumption is used to determine whether the period constitutes an over-issuance or under-issuance: Scenario 1: When the deviation in power output is greater than 0, it indicates that the predicted power output of wind and solar power during that period is higher than the agreed power output for that period, which is considered an over-generation situation. Since the over-generated power is sold in the spot market at the over-generation settlement price, the contract price is extracted from the monthly ancillary service capacity contract information. Therefore, the difference between the over-generation settlement price and the contract price is used to determine whether the over-generation is settled at the spot price. If the over-generation settlement price is lower than the contract price, then the over-generation was originally settled at the contract price, but is now determined to be settled at the spot price. The difference between the over-generation settlement price and the contract price is the revenue loss per unit of electricity. The calculation is based on the deviation in electricity volume, combined with the over-generation settlement price and the contract electricity price. The specific formula for calculating the expected revenue loss corresponding to excess power generation during a given period is as follows: ; Subscript The English meaning of "loss" is used only for distinction; among them, This indicates that the value within the parentheses is the larger of the value and zero. The formula means that when the over-generation settlement price is lower than the contracted electricity price, the revenue loss is greater than 0; the more over-generation, the greater the revenue loss. When the over-generation settlement price is not lower than the contracted electricity price, the revenue loss is less than or equal to 0. Taking zero indicates that the excess power generation, settled at the spot price, does not result in any revenue loss. Since this case only calculates the loss and not the additional revenue, it is used... Cut off; Scenario 2: When the deviation in electricity volume is less than 0, it indicates that the volume is lower than the agreed-upon value for that period, which is considered a shortfall. Since the shortfall in electricity volume needs to be purchased in the spot market at the shortfall purchase price, the shortfall purchase price and the contract price are used to determine whether the shortfall in electricity volume is purchased at the spot price. If the shortfall purchase price is higher than the contract price, then this part of the shortfall in electricity volume, which was originally obtained at the contract price, now needs to be purchased at the spot price. The difference between the shortfall purchase price and the contract price is the additional cost per unit of electricity volume. By using the absolute value of the deviation in electricity volume, combined with the purchase price of the shortfall and the contract electricity price, The expected replenishment cost corresponding to the power shortage during a given period is calculated using the following formula: ; Subscript This is an abbreviation for "supplementary purchase"; since the deviation electricity price is negative, the absolute value represents the size of the shortfall. The formula means that when the purchase price is higher than the contract price, the additional cost is greater than 0, and the greater the shortfall, the higher the expected replenishment cost; when the spot purchase price is not higher than the contract price, the additional cost is less than or equal to 0. Taking zero indicates that the missing electricity can be purchased at the spot price without incurring additional costs; similarly, using... Cut off; Case 3: When the deviation power is equal to 0, it means that the predicted power output of wind and solar power in this period is equal to the agreed power output in this period. Sub-step S2.2: Add the expected revenue loss and the expected repurchase cost for each time period, and then sum them over the entire scheduling cycle to obtain the expected settlement cost of wind and solar power. The specific algorithm formula is as follows: The unit is yuan; it should be noted that for the same period of time Since the deviation in electricity volume cannot be both positive and negative simultaneously, at most one of the expected revenue loss and the expected replenishment cost will be non-zero, while the other will be zero. Adding them together in each time period will not cause double counting. The purpose of this step is to make the estimation of the expected settlement cost of wind and solar power closer to the actual market settlement rules by distinguishing between the two situations of over-issuance and shortage and using the spot market over-issuance settlement price and shortage purchase price for segmented settlement. If it is not segmented and only settled by a single price, it will not be able to reflect the difference in price caliber between over-issuance and shortage in the spot market, which will lead to the distortion of the estimation of the expected settlement cost of wind and solar power, and thus affect the optimization results in the subsequent step S6.

[0021] In step S3, based on the energy storage discharge depth at each time period within the scheduling cycle and the energy storage battery health status information from step S1, the equivalent loss cost over the entire life cycle of the energy storage within the scheduling cycle is determined. ; Subscript The abbreviation for lifespan includes the following sub-steps: Sub-step S3.1: Read the energy storage discharge amount for each time period from the energy storage management system. and rated energy storage capacity subscript This is an abbreviation for energy storage, used only for distinction; subscript The term "rated" is used only for distinction; it is used to calculate the depth of discharge for different time periods. The value of the depth of discharge ranges from 0 to 1. The meaning of this formula is: the depth of discharge is equal to the proportion of the energy storage discharge amount to the rated capacity of the energy storage during that period. When the depth of discharge is 0, it means that the energy storage did not discharge during that period. When the depth of discharge is 1, it means that the energy storage released all the rated capacity of the energy storage during that period. The closer the depth of discharge is to 1, the deeper the discharge during that period. Once the depth of the injection point is known, the nonlinear exponent of lifetime loss is calculated based on the SOH value. The specific algorithm formula is as follows: ;in, This is the initial lifetime loss nonlinear index, i.e., the lifetime loss nonlinear index when the battery is in a brand new state. The health status impact coefficient characterizes the amplification effect of SOH decline on the nonlinear exponent of battery life loss; the SOH value ranges from 0 to 1; both the initial nonlinear exponent of battery life loss and the health status impact coefficient can be obtained by fitting the cycle life data provided by the battery manufacturer; the meaning of this formula is: as the battery health status declines, the SOH value decreases, As the lifespan loss nonlinearity increases, the lifespan loss nonlinearity also increases. The larger the lifespan loss nonlinearity, the more significant the impact of the discharge depth on the unit power loss cost, and the higher the relative cost of deep cycling, thus guiding the scheduling plan of subsequent step S6 to favor shallow cycle discharge. When the nonlinear exponent of life loss is obtained, the battery cycle life characteristic curve is acquired. The battery cycle life characteristic curve describes the relationship between the depth of discharge and the number of cycles. The deeper the discharge depth, the fewer the number of cycles; the shallower the discharge depth, the more the number of cycles, but the single discharge amount of a shallow cycle is also less. Considering these two aspects, the life loss cost per unit of electricity increases nonlinearly with the increase of the depth of discharge. The battery cycle life characteristic curve is determined by the cycle life data provided by the battery manufacturer. In this embodiment, the number of cycles of this model of battery at different discharge depths is read from the energy storage management system parameter library to form a correspondence between the discharge depth and the number of cycles, and is pre-stored in the mapping relationship library. If the battery manufacturer does not directly provide curve data, it can also be obtained by fitting historical charge and discharge data recorded in the battery management system. The corresponding mapping curve is selected from the mapping relationship library based on the SOH value. This mapping relationship satisfies the following: the unit charge lifetime loss cost increases non-linearly with the increase of discharge depth. The specific algorithm formula is as follows: ;in, The cost per unit of charge lifetime is given a depth of discharge and a healthy state, expressed in yuan per kilowatt-hour. The benchmark unit energy loss cost is the unit energy loss cost when the depth of discharge is equal to the benchmark depth of discharge. It is calculated based on the cycle life data and replacement cost provided by the battery manufacturer. The English meaning of "benchmark" is used only for distinction. The baseline depth of discharge is taken from the standard depth of discharge used by the battery manufacturer for cycle life testing and is used for normalization. It is a nonlinear exponent of life loss, which is updated online as the SOH value decreases; Sub-step S3.2: Substitute the depth of discharge into the mapping relationship, and then multiply by the energy storage discharge amount to obtain the first... Equivalent lifetime loss cost over time period The specific algorithm formula is as follows: ; Subscript The abbreviation "equivalent" is used only for distinction; the equivalent lifespan loss cost of each time period is then added together, and then discounted according to battery replacement cost and rated cycle number to obtain the equivalent loss cost of the energy storage's entire lifespan within the dispatch cycle. The specific algorithm formula is as follows: ;in, This refers to the battery replacement cost, expressed in yuan, which is the cost required to replace a battery with a new one when it reaches the end of its lifespan. The battery replacement cost is retrieved from the equipment ledger or purchase contract of the energy storage management system. If the replacement has not yet occurred, the current market purchase price of the same model of battery is used. For reference replacement cost, used for normalization, the reference replacement cost is the battery replacement cost itself, with the subscript... The abbreviation for "reference" in English; The rated cycle life refers to the number of times a battery can be used at a reference depth of discharge. The rated cycle life is obtained from the battery manufacturer's specifications. The meaning of this formula is: It is the lifetime loss cost calculated based on the depth of discharge during this scheduling cycle, multiplied by... This conversion factor maps it to the full life cycle scale. The purpose of the conversion is to unify the dimensions of the equivalent loss cost of energy storage over its entire life cycle with the expected settlement cost of wind and solar power obtained in step S2 and the cumulative compensation cost in the subsequent step S4, so that they can be compared and summed in the same optimization objective function in the subsequent step S6. This step uses a nonlinear mapping from the depth of discharge to the unit energy loss cost and incorporates the SOH value into the online update of the mapping relationship, so that the estimation of the equivalent loss cost of energy storage over its entire life cycle is more in line with the battery aging characteristics. At the same time, the allocation ratio of deep cycle and shallow cycle is used as an optimization variable to provide a more refined adjustment means for global optimization in the subsequent step S6.

[0022] In step S4, based on the cumulative reduction amount in step S1 With cumulative call count Determining the cumulative compensation cost of flexible loads includes the following sub-steps: Sub-step S4.1: Read the monthly cumulative reduction amount from the monthly ancillary service capacity contract information, and compare the cumulative reduction amount with the monthly cumulative reduction amount. The monthly cumulative reduction amount is the monthly cumulative reduction amount agreed upon in the contract, in kilowatt-hours or megawatt-hours. It is the upper limit of electricity that can be reduced free of charge or at the base price each month, as agreed between the virtual power plant and the flexible load user. Compare whether the basic compensation fee is calculated based on the base compensation price or the tiered compensation fee is calculated based on the tiered compensation price. When the cumulative reduction is less than or equal to the monthly cumulative reduction limit, the calculation is based on the base compensation price. The basic compensation fee for the time period is calculated using the following formula: ;in, For the first The flexible load reduction power during the time period, in kilowatt-hours or megawatt-hours, is determined by the optimization model in subsequent step S6 after a scheduling plan is given. This is the base compensation price, the bid. The term "basic" refers to the basic compensation price in English, and the distinction is only necessary. When a virtual power plant signs a dispatch agreement with a flexible load user, both parties negotiate and determine the basic compensation price based on factors such as the type of load the user can reduce, the duration of the reduction, the advance notice, and the electricity rights relinquished by the user. This price is then written into the agreement, hence the basic compensation price is extracted from the agreement. The meaning of this formula is: when the monthly cumulative reduction limit is not exceeded, the electricity reduction by the flexible load is settled according to the basic compensation price; the amount reduced is the same as the compensation, and the unit price remains unchanged. When the cumulative reduction exceeds the monthly cumulative reduction limit, a tiered compensation price will be applied. (subscript) (Referring to the English meaning of "stairs"; only the distinction is needed) Calculate the first... Tiered compensation fee for different time periods The specific algorithm formula is as follows: ;in, The tiered compensation price is negotiated and determined by the virtual power plant and flexible load users when signing the dispatch agreement, and is written into the agreement. Therefore, the tiered compensation price is extracted from the agreement; and the tiered compensation price is greater than the basic compensation price; the meaning of this formula is: It is up to the The cumulative reduction amount exceeding the monthly cumulative reduction limit will be used... Zero is taken when the limit is not exceeded; the excess portion is compensated at a tiered compensation price higher than the basic compensation price. The more the limit is exceeded, the higher the tiered compensation cost. The specific value of the tiered compensation price can be determined according to the contract or actual negotiation. Its purpose is to suppress the excessive use of flexible loads and prevent virtual power plants from using flexible loads as low-cost resources without restriction. Sub-step S4.2: Repeated use of flexible loads can be inconvenient for users and increase equipment wear. The more times the load is used, the more pronounced this fatigue effect becomes. Therefore, based on the cumulative number of uses, combined with a fatigue surcharge coefficient... Calculate the fatigue markup coefficient The specific algorithm formula is as follows: Among them, the fatigue premium coefficient is dimensionless, and the value of the fatigue premium coefficient is not less than 1. This is a reference call count used for normalization and can be set according to contractual agreements or historical call frequency. The fatigue surcharge nonlinear index is dimensionless; the fatigue surcharge magnitude coefficient and the fatigue surcharge nonlinear index are set according to the flexible load type and user comfort requirements; the meaning of this formula is: the greater the cumulative number of calls, the higher the fatigue surcharge rate. The larger the value, the greater the price increase after being amplified by the nonlinear exponent of fatigue pricing; the fatigue pricing magnitude coefficient controls the overall price increase, and the fatigue pricing nonlinear exponent controls how fast the price increases with the number of calls; when the fatigue pricing nonlinear exponent is greater than 1, the price increases faster with the number of calls, which has a stronger inhibitory effect on high-frequency calls. Add the basic compensation cost to the tiered compensation cost, and then multiply by the fatigue surcharge factor to obtain the corrected compensation cost for each time period. The unit is yuan, and the subscript is... "Revision" is an abbreviation for "correction"; the specific calculation formula for the compensation fees for each time period after the revision is as follows: It should be noted that, for the same time period If the period falls within the normal range, the tiered compensation cost is 0; if it falls within the excess range, both the basic compensation cost and the tiered compensation cost exist, and the entire compensation cost is uniformly adjusted using a fatigue surcharge coefficient. The adjusted compensation costs for each period are then summed to obtain the cumulative compensation cost for the flexible load. ; Subscript The English meaning of "flexible" is used only for distinction; the role of step S4 is to ensure that the compensation cost of flexible loads reflects their actual call intensity and user comfort loss through tiered compensation and fatigue pricing; and to prevent flexible loads from being over-reliant in rolling scheduling by reducing callable capacity, thus ensuring their sustainable participation capability.

[0023] In step S5, based on the ancillary service market price, peak-valley price difference forecast, and grid dispatch instruction level, the current market operation status of the virtual power plant is identified, and the dispatch priority coefficients of the wind and solar power output forecast, energy storage discharge, discharge depth, and flexible load reduction are determined under the current market operation status; specifically, this includes the following sub-steps: Sub-step S5.1: By establishing a connection with the electricity market trading platform, historical price data is obtained from the platform. Then, the load forecasting module outputs load forecast values ​​for each time period within the future dispatch cycle based on historical load data, meteorological data (temperature, humidity), and user electricity consumption patterns. Peak-valley price difference forecast values ​​are obtained using historical price data, load forecast values, and wind and solar power output forecast values. Specifically: the dispatch cycle is divided into peak and off-peak periods; using historical price data, load forecast values, and wind and solar power output forecast values ​​from the electricity market trading platform as input features, the price forecast model outputs electricity price forecast values ​​for each time period; peak-hour electricity price forecast values ​​are extracted from the electricity price forecast values ​​for each time period. Electricity price forecast during off-peak hours subscript and These terms refer to "high" and "low" in English and are used only to distinguish between them; the peak-valley price difference is calculated based on the peak-hour electricity price forecast and the off-peak electricity price forecast. The unit is yuan / kWh; the price prediction model is established as follows: historical spot market clearing price series are read from the electricity market trading platform, and historical load data, load forecast values, and wind and solar power output forecast values ​​for the corresponding period are read. The three are aligned by time to form a training sample set; regression prediction or neural network prediction methods are used to fit the mapping relationship between the load forecast values, wind and solar power output forecast values ​​and historical prices and clearing prices of the same period using the training sample set. After training, the model parameters are fixed; during online prediction, the load forecast values ​​and wind and solar power output forecast values ​​output by the load forecast module are input into the model, and the electricity price forecast values ​​for each future period are output; Read the current state of charge from the battery management system. Minimum state of charge and highest state of charge According to the rated capacity of energy storage Calculate the actual usable capacity using SOH value. ; Subscript This refers to the actual English meaning and is only used for distinction; further, it depends on the actual available capacity and the current state of charge. and minimum state of charge Calculate the available energy storage capacity for the current rolling cycle. The unit is kilowatt-hour, and the subscript is... The English meaning of "rolling" is used only for distinction; Once the rolling cycle of available energy storage capacity and peak-valley price difference forecasts are obtained, the expected return from energy market arbitrage is calculated. ; Subscript The English meaning of "for profit" is used only for distinction; among them, It should be converted on a monthly basis, with the unit being yuan / kilowatt-month; The index represents the number of arbitrage cycles that can be executed within a month. "Monthly" is the English term used only for distinction; it refers to extracting ancillary service market capacity quotes from monthly ancillary service capacity contract information. subscript The abbreviation for "ancillary services" is used only for distinction; its unit is the same as the expected return from energy market arbitrage. The first ratio is calculated using the expected return from energy market arbitrage and the market capacity quotation for ancillary services. The first ratio is dimensionless; it reflects the level of ancillary service market capacity pricing relative to the expected arbitrage return in the energy market. When the first ratio is obtained, the predicted value based on the peak-to-valley price difference is based on the second ratio. The second ratio is dimensionless; among which, The first ratio is the threshold for initiating energy storage arbitrage, expressed in yuan / kWh. It can be set based on the energy storage cycle cost and the historical market price difference. The energy storage cycle cost can be calculated from the unit power lifespan loss cost and the charging and discharging efficiency. The historical market price difference can be obtained from historical price data statistics from the power market trading platform. The second ratio reflects the current peak-valley price difference prediction value relative to the energy storage arbitrage initiation threshold. The meaning of this formula is: the larger the peak-valley price difference prediction value, the larger the second ratio, indicating that the space for energy market arbitrage is larger, and the more motivated energy storage is to participate in arbitrage. The second ratio does not participate in the state switching judgment, but is used to assign values ​​to the subsequent dispatch priority coefficient. Sub-step S5.2: Set the supply guarantee command level subscript The English term "level" is used only for distinction. The power supply guarantee order level refers to the level of dispatch order issued by the power grid dispatching agency to ensure the security of power supply. The power supply guarantee order level is usually divided into several levels, such as normal, attention, early warning, and emergency. The power supply guarantee order level obtains the level information from the power grid dispatching agency through the dispatching data network. Based on the first ratio, the second ratio, and the supply guarantee order level, the current operating status of the virtual power plant is classified into one of the following: energy market-dominated, ancillary service-dominated, or supply guarantee constraint-dominated status; specifically, the classification rules are as follows: First, when the supply guarantee order level is warning or emergency, it is determined to be a state dominated by supply guarantee constraints. At this time, the tension of the power grid supply reaches the warning or emergency level, and virtual power plants should prioritize responding to the supply guarantee order, with market benefits taking a back seat. Second, when the supply guarantee order level is normal or under attention, the market operation status is determined according to the hysteresis switching rule between the energy market-dominated state and the ancillary services-dominated state; the specific hysteresis switching rule is as follows: The first ratio is lower than the first lower limit value. Rise above the first upper limit And continuously meet the preset number of rolling cycles When the first ratio drops from above the first upper limit to below the first lower limit and continuously meets the preset rolling cycle number, the market operation status will be switched from the ancillary service-dominated state back to the energy market-dominated state; wherein, The hysteresis interval for state switching is based on the first upper limit and the first lower limit. The first upper limit and the first lower limit can be determined statistically based on historical market data. The preset rolling cycle number indicates how many rolling cycles the first ratio must continuously meet the hysteresis switching condition before state switching is allowed. The preset rolling cycle number is set by the operators during the configuration phase, and the value is based on statistical analysis of historical market data and actual operating experience: statistically analyzing the distribution of the length of continuous periods in historical spot market and ancillary service market price data where the first ratio is higher than the first upper limit or lower than the first lower limit. If most real market state transitions last for more than 3 rolling cycles, while short-term noise mostly falls back within 2 rolling cycles, then a preset rolling cycle number of 3 can effectively filter noise. At the same time, the rolling step length is considered to match the preset rolling cycle number with the rolling step length, and the filtering window covers the duration of typical noise. This is used to avoid erroneous switching due to a single fluctuation. The purpose of setting the hysteresis interval and the number of consecutive rolling cycles is as follows: If only a single threshold is used for judgment, when the first ratio fluctuates slightly around the threshold, the state will switch frequently, causing the scheduling priority coefficient to change repeatedly, which in turn causes scheduling plan oscillations; after adopting the hysteresis interval, the first ratio must rise from below the first lower limit to above the first upper limit before it can switch to the auxiliary service dominant state, and the reverse switching is the same, and the original state remains unchanged within the intermediate interval; in addition, the requirement of a number of consecutive preset rolling cycles is added to further filter out short-term fluctuations; It should be noted that the second ratio is not involved in the above state transition judgment, but is used to assign values ​​to the subsequent scheduling priority coefficient. The reason for this is that the criterion for state transition should focus on the relative attractiveness between ancillary services and the energy market, i.e., the first ratio; while the second ratio reflects the absolute level of the arbitrage space in the energy market. Pre-configure corresponding resource scheduling priority coefficient combinations for each market operation state; the resource scheduling priority coefficient combinations include scheduling priority coefficients for wind and solar power, energy storage, and flexible negative resources; In an energy market-dominated environment, energy storage discharge capacity and discharge depth resources are used for low-charge-high-discharge arbitrage. Flexible load power reduction resources with higher dispatch priority coefficients are used to adjust electricity consumption curves, with a dispatch priority coefficient of the next highest. Wind and solar power output forecasts are executed according to the forecast output, with the lowest dispatch priority coefficient. If the second ratio is greater than 1, it indicates that the peak-valley price difference forecast is higher than the energy storage arbitrage activation threshold, indicating ample arbitrage opportunities, and the dispatch priority coefficient for energy storage can be further increased. If the second ratio is less than or equal to 1, it indicates that the peak-valley price difference forecast has not exceeded the energy storage arbitrage activation threshold, indicating limited arbitrage opportunities, and the dispatch priority coefficient for energy storage can be decreased. In an ancillary service-dominated state, energy storage is used to provide peak-shaving and frequency regulation capacity, and has the highest scheduling priority coefficient; wind and solar power output fluctuations can be used as adjustment resources for ancillary services, and have the next highest scheduling priority coefficient; flexible load scheduling has the lowest priority coefficient. Under the dominant supply constraint, flexible loads can be used for rapid load reduction, with the highest scheduling priority coefficient; energy storage can be used for peak discharge, with the second highest scheduling priority coefficient; wind and solar power operate at maximum output, with the lowest scheduling priority coefficient; specific scheduling priority coefficients are set based on actual operating experience, and the sum of the scheduling priority coefficients of the predicted wind and solar power output, energy storage discharge amount, discharge depth, and flexible load power reduction resources under the same condition can be normalized to 1; the scheduling priority coefficient for the current rolling cycle is output. (No. The first rolling cycle (Scheduling priority coefficient of resource class), and the scheduling priority coefficient of the previous rolling cycle. The data is temporarily stored in the rolling cycle parameter storage area; Step S5 achieves dynamic identification of the market operation status through the first ratio, the supply guarantee instruction level and the hysteresis switching rule, avoids frequent switching of status near the threshold through the hysteresis interval and the number of consecutive rolling cycles, further suppresses switching at the optimization target level through the status switching penalty term, and enables the calling order of various resources to be adaptively adjusted according to the market environment by configuring the scheduling priority coefficient by status.

[0024] In step S6, with the goal of minimizing the total scheduling cost within the scheduling cycle, and combining the expected settlement cost of wind and solar power in step S2, the equivalent loss cost of energy storage over its entire life cycle in step S3, the cumulative compensation cost in step S4, and the scheduling priority coefficient in step S5, a joint scheduling plan for wind and solar power, energy storage, and flexible loads within the future scheduling cycle is determined, while satisfying the constraints of cross-time coupling of energy storage charge state and continuous dispatch of flexible loads. Specifically, this includes the following sub-steps: Sub-step S6.1: Calculate the state transition penalty term based on the scheduling priority coefficient of the current rolling cycle and the scheduling priority coefficient of the previous rolling cycle. ; Subscript The abbreviation for punishment is used only for distinction; among them, This refers to the penalty coefficient, expressed in yuan. The value is determined based on the following: First, the common range of the sum of absolute values ​​of the changes in scheduling priority coefficients between adjacent rolling cycles is calculated within a typical scheduling cycle without the application of a penalty term. Then, this common range is compared with the magnitude of the expected settlement cost of wind and solar power, the equivalent loss cost of energy storage over its entire life cycle, and the cumulative compensation cost. The penalty coefficient is chosen such that the cost incurred by the penalty term when a complete state switch of the scheduling priority coefficient occurs accounts for approximately 1% to 5% of the total scheduling cost. In this way, the penalty term is sufficient to suppress frequent switching but will not overwhelm other cost items. The specific value is determined based on actual operating experience. When a state transition penalty is incurred, the cost is settled based on the expected value of the scenery. Equivalent loss cost over the entire life cycle of energy storage Cumulative compensation costs and state switching penalty Combined with scheduling priority coefficient Medium- and long-term adjustment costs (subscript) (Referring to the English meaning of "adjustment," used only for distinction) and resource deviation power. Construct an optimization objective function that minimizes the total scheduling cost. The specific algorithm formula is as follows: ; in, For resource type indexing, three categories are selected: wind and solar, energy storage, and flexible load. Index for scheduling periods; For the first Class resources in the The medium- and long-term adjustment cost of the unit deviation electricity received during the time period, in yuan / kWh; For the first Class resources in the The resource deviation electricity received during the time period, in kilowatt-hours, is the decision variable for the optimization model. Regarding the definition of the two types of costs in the optimization objective function, the following needs to be explained: the expected settlement cost of wind and solar power, the equivalent loss cost of energy storage throughout its entire life cycle, and the cumulative compensation cost are the total costs calculated according to steps S2 to S4 under a given scheduling plan, that is, the expected settlement cost of wind and solar power, the equivalent loss cost of energy storage throughout its entire life cycle, and the cumulative compensation cost when executed according to the contract benchmark for the entire scheduling cycle. It is the marginal adjustment cost added by the additional acceptance of deviation adjustment of each resource. The medium and long-term adjustment cost is the marginal quantity, and the resource deviation electricity is the deviation electricity accepted. The two have different calibers. The former is the total amount, and the latter is the increment. They do not constitute double counting. The medium- to long-term adjustment costs consist of three parts: direct adjustment costs, cumulative adjustment losses over the cycle, and marginal impact costs of contract performance evaluations. These three types of resources have different composition methods, which are explained below: Wind and solar resources do not generate cumulative regulation loss costs over time. The medium- and long-term regulation costs consist of two parts: direct regulation costs and marginal impact costs of contract assessment. The direct regulation costs of wind and solar resources are zero, meaning that wind and solar power output is executed according to the forecast value without additional direct regulation costs. The marginal impact costs of contract assessment for wind and solar resources are determined by taking the partial derivative of the resource deviation power volume according to the monthly bilateral energy contract assessment rules. The medium- and long-term regulation costs of energy storage resources consist of three parts: direct regulation costs, cumulative regulation loss costs over the cycle, and marginal impact costs of contract assessments. Among them, direct regulation costs are the energy loss costs during the charging and discharging process, which are calculated based on the energy storage cycle efficiency. Cumulative regulation loss costs over the cycle are determined based on the nonlinear mapping relationship between the depth of discharge and the cost of loss per unit of electricity. Marginal impact costs of contract assessments are determined based on the marginal impact of energy storage acceptance deviations on monthly contract execution deviations. The medium- and long-term adjustment costs of flexible load resources also consist of three parts: direct adjustment costs, cumulative adjustment loss costs over the period, and marginal impact costs of contract assessment. Among them, the direct adjustment cost of flexible load resources is the basic compensation price; the cumulative adjustment loss cost over the period is calculated by applying the tiered compensation price and the fatigue surcharge coefficient; and the marginal impact cost of contract assessment is determined by the marginal impact of the flexible load acceptance deviation on the monthly contract execution deviation.

[0025] Regarding the definition of the cumulative regulation loss cost, the following explanation is required: The cumulative regulation loss cost refers to the cost corresponding to the proportion of the regulation electricity that the resource has accumulated during the scheduling cycle to its upper limit of available regulation electricity. For energy storage resources, this proportion is obtained by dividing the accumulated regulation electricity by the upper limit of available regulation electricity of energy storage, and then multiplying it by the unit electricity lifetime loss cost, which is the cumulative regulation loss cost. For flexible load resources, this proportion is obtained by dividing the portion of the monthly cumulative reduction that exceeds the contract amount by the upper limit of the contract's allowed excess reduction, and then multiplying it by the tiered compensation price and fatigue surcharge coefficient, which is the cumulative regulation loss cost. The scheduling priority of each resource in each time period is determined by its medium- and long-term adjustment cost from smallest to largest; that is, the resource with the lower the medium- and long-term adjustment cost, the higher the priority in accepting the resource deviation. Under the premise of satisfying various constraints, the optimization model will prioritize the use of resources with low medium- and long-term adjustment costs, and reserve resources with high medium- and long-term adjustment costs for use when necessary, thereby reducing the overall scheduling cost. The scheduling priority coefficient and the medium- and long-term adjustment cost are jointly incorporated into the optimal objective function. The scheduling priority coefficient reflects the impact of market operation status on the importance of resources, and the medium- and long-term adjustment cost reflects the economic cost of the resource itself in accepting the deviation. The product of the two determines the actual order of resource use, thus obtaining the scheduling priority of each resource in each time period. When the cumulative number of calls to the flexible load exceeds the preset call count threshold, the callable capacity of the flexible load is reduced by a reduction factor, and the reduced callable capacity is used as the upper limit of the continuous call constraint of the flexible load to participate in the rolling optimization solution. The reduction factor decreases nonlinearly with the increase of the cumulative number of calls, and the reduction factor is updated online with the cumulative execution information of the flexible load. In addition, the distribution ratio of the total energy of energy storage between deep cycle discharge and shallow cycle discharge is used as an optimization variable, and together with the discharge depth of each time period, it participates in the search for the minimum total scheduling cost. Among them, the division between deep cycle discharge and shallow cycle discharge is defined by the discharge depth threshold. When the current state of charge is not less than the discharge depth threshold, it is a deep cycle discharge; otherwise, it is a shallow cycle discharge. Sub-step S6.2: Construct constraints, including the cross-period coupling constraint of energy storage charge state and the continuous call constraint of flexible load. The two are embedded in the optimization model as hard constraints to participate in global optimization. Based on the current state of charge and the rated capacity of the energy storage, a cross-time-period coupling constraint for the energy storage state of charge is constructed. The specific algorithm formula is as follows: ; Subscript and These refer to the English meanings of charging and discharging, respectively; among them, the state of charge of energy storage is dimensionless and ranges from 0 to 1. For the first State of charge of energy storage during a given time period; Charging efficiency is dimensionless and is estimated online by the battery management system. Discharge efficiency is dimensionless and originates from the same source as charging efficiency. For the first The charging power during the time period, in kilowatts, is a decision variable for the optimization model and will be determined after subsequent solution. For the first The discharge power during the time period, in kilowatts, is a decision variable in the optimization model and will be determined after subsequent solving; it also satisfies The meaning of this formula is: the state of charge in the next time period is equal to the state of charge in the current time period plus the charging amount divided by the rated capacity, and then minus the discharging amount divided by the rated capacity; the charging amount is calculated according to the charging efficiency, and the discharging amount is calculated according to the discharging efficiency; this constraint couples the state of charge in adjacent time periods, so that the optimization model cannot only consider the single time period optimization, but must consider the charging and discharging arrangement of energy storage across time periods. When the cross-time coupling constraints of the energy storage state of charge are known, take the first time period as an example. Starting with the time period, then setting the length as... Sliding window and maximum number of calls allowed within the sliding window And construct constraints for continuous call of flexible loads, the specific formula is as follows: Among them, the number of time periods during which flexible loads are invoked shall not exceed the maximum number of times allowed within the sliding window; For the first The indicator variable for whether flexible loads are invoked during a given time period is a binary decision variable, with a value of 1 indicating invocation and 0 indicating non-invocation. This value is determined by the optimization model during the solution process. The purpose of this constraint is to prevent flexible loads from being repeatedly invoked in multiple consecutive time periods, ensuring user comfort. It should be noted that the indicator variable is related to the power reduction of flexible loads. When the indicator variable is 1, the power reduction of flexible loads can take a value greater than zero; when the indicator variable is 0, the power reduction of flexible loads is associated with zero. This relationship is achieved in the optimization model by constraining the power reduction of flexible loads to be less than or equal to the product of the indicator variable and the maximum power reduction of flexible loads in a single time period. The maximum power reduction of flexible loads in a single time period is the maximum power reduction of flexible loads in a single time period, which is read from contract information or load control terminal parameters. The objective function, priority rules, and constraints are combined into a complete optimization model, which is then solved using mixed-integer linear programming or mixed-integer nonlinear programming methods to obtain a joint scheduling plan for wind, solar, energy storage, and flexible loads within the future scheduling cycle. The solution results include the wind and solar power output plan for each time period, the energy storage charging and discharging plan (i.e., the values ​​of charging power and discharging power), the flexible load reduction plan (i.e., the flexible load reduction amount and the values ​​of indicator variables), and the resource deviation amount undertaken by each resource.

[0026] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A long-term joint dispatching method for virtual power plants, characterized in that: The methods and steps include the following: S1. Obtain monthly bilateral energy contract curves, monthly ancillary service capacity contract information, wind and solar power output forecast information, spot market price forecast information, SOH value, and cumulative execution information of flexible load; S2. Based on the monthly bilateral energy contract curve and wind and solar power output forecast information, determine the expected settlement cost of wind and solar power output deviating from the contract curve in each period of the scheduling cycle. S3. Determine the equivalent loss cost of the entire life cycle of energy storage within the scheduling period based on the energy storage discharge depth and energy storage battery health status information for each time period within the scheduling period. S4. Determine the cumulative compensation cost of the flexible load based on the cumulative reduction amount and cumulative number of calls in the cumulative execution information of the flexible load; S5. Based on the ancillary service market price, peak-valley price difference forecast, and grid dispatch instruction level, identify the current market operation status of the virtual power plant, and determine the dispatch priority coefficients of wind and solar power, energy storage, and flexible loads under the current market operation status. S6. Taking the minimum total scheduling cost within the scheduling cycle as the optimization objective, and combining the expected settlement cost of wind and solar power, the equivalent loss cost of energy storage throughout its entire life cycle, the cumulative compensation cost, and the scheduling priority coefficient, under the premise of satisfying the cross-time coupling constraint of energy storage charge state and the continuous call constraint of flexible load, the joint scheduling plan of wind and solar power, energy storage and flexible load within the future scheduling cycle is obtained.

2. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, S2 includes the following method steps: The predicted power output of wind and solar power in each time period within the scheduling cycle is compared with the monthly bilateral energy contract curve for each time period to obtain the deviation power. When the deviation in electricity volume is greater than zero, the expected revenue loss corresponding to the excess electricity volume during that period shall be calculated based on the spot market over-generation settlement price. When the deviation in electricity volume is less than zero, the expected replenishment cost corresponding to the shortfall in electricity volume during that period is calculated based on the spot market purchase price for the shortfall. The expected revenue loss and expected repurchase cost for each period are added together to obtain the expected settlement cost of wind and solar power within the scheduling cycle.

3. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, S3 includes the following method steps: The discharge depth for each time period is determined based on the energy storage discharge amount and rated energy storage capacity during each time period within the scheduling cycle. Based on the SOH value, the mapping relationship between the depth of discharge corresponding to the current health state and the cost per unit of energy loss is extracted from the battery cycle life characteristic curve. The discharge depth at each time period is converted into the equivalent lifetime loss cost at each time period according to the mapping relationship; The equivalent lifespan loss cost of each time period is summed up and then converted according to the battery replacement cost and the rated number of cycles to obtain the equivalent lifespan loss cost of energy storage within the scheduling cycle.

4. The long-term joint dispatch method for virtual power plants according to claim 3, characterized in that, The mapping relationship between depth of discharge and cost per unit of charge loss includes the following steps: The unit power lifetime loss cost increases nonlinearly with the increase of discharge depth. The nonlinear exponent in the mapping relationship is updated online as the SOH value decreases. The distribution ratio of the total energy of energy storage in the scheduling cycle between deep cycle discharge and shallow cycle discharge is used as the optimization variable. The optimization variable and the discharge depth of each time period jointly participate in the search for the minimum total scheduling cost. The division between deep cycle discharge and shallow cycle discharge is defined by the discharge depth threshold.

5. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, S4 includes the following method steps: Compare the cumulative reduction of flexible load within the scheduling cycle with the monthly cumulative reduction limit agreed upon in the contract; When the cumulative reduction does not exceed the monthly cumulative reduction limit, the basic compensation fee corresponding to the electricity reduction in each time period is calculated based on the basic compensation price. When the cumulative reduction exceeds the monthly cumulative reduction limit, the excess portion will be compensated at a tiered compensation price that is higher than the basic compensation price. The fatigue surcharge coefficient is determined based on the cumulative number of times the flexible load is called within the scheduling cycle, and the compensation cost for each time period is adjusted according to the fatigue surcharge coefficient to obtain the adjusted compensation cost for each time period. The adjusted compensation costs for each time period are summed to obtain the cumulative compensation cost for the flexible load.

6. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, S5 includes the following method steps: Calculate the first ratio between the quoted capacity of the ancillary services market and the expected return of energy market arbitrage; calculate the second ratio between the predicted peak-valley price difference and the threshold for initiating energy storage arbitrage; Obtain the supply guarantee instruction level issued by the power grid dispatching agency; based on the first ratio, the second ratio and the supply guarantee instruction level, classify the current operating status of the virtual power plant into one of the following: energy market-dominated status, ancillary service-dominated status and supply guarantee constraint-dominated status; pre-configure the corresponding resource scheduling priority coefficient combination for the market operating status.

7. The long-term joint dispatch method for virtual power plants according to claim 6, characterized in that, The market operation status is classified according to the first ratio, the second ratio, and the supply guarantee order level, including the following methods and steps: When the first ratio rises from below the first lower limit to above the first upper limit and continuously meets the preset number of rolling cycles, the market operation status will be switched from the energy market-dominated state to the ancillary service-dominated state. When the first ratio drops from above the first upper limit to below the first lower limit and continuously meets the preset rolling cycle number, the market operation state will be switched from the ancillary service-dominated state back to the energy market-dominated state; the first upper limit is higher than the first lower limit, and the two constitute the hysteresis interval for state switching.

8. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, The optimization objective in S6 includes a state transition penalty term, which is determined based on the change in scheduling priority coefficient between two adjacent rolling cycles.

9. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, S6 also includes the following method steps: When the cumulative number of calls to the flexible load exceeds the preset call count threshold, the callable capacity of the flexible load is reduced by a reduction factor. The reduced callable capacity is then used as the upper limit of the continuous call constraint of the flexible load in the rolling optimization solution. The reduction factor decreases non-linearly with the increase of the cumulative number of calls, and the reduction factor is updated online with the cumulative execution information of the flexible load.

10. The long-term joint dispatch method for virtual power plants according to claim 1, characterized in that, The constraints on cross-time coupling of energy storage state of charge and continuous call-up of flexible loads in S6 include the following methods and steps: Construct cross-time period coupling constraints for energy storage state of charge based on the current state of charge and the rated capacity of energy storage; Starting from the first time period, a length of [length to be specified] is then set. The sliding window and the maximum number of calls allowed within the sliding window are defined, and continuous call constraints with flexible load are constructed.