A method and system for flexible adjustment of resource rolling joint clearing

By establishing a resource-constraint coupling graph and using machine learning models to assist optimization in the power system, the problem of insufficient flexible resource allocation under high-proportion renewable energy access was solved. This achieved coordinated optimization of power market clearing and system regulation capacity, improving the executability of clearing results and online solution efficiency.

CN123073799APending Publication Date: 2026-11-10STATE GRID JIANGXI ELECTRIC POWER CO LTD RES INST
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611577982.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-13
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In power systems with a high proportion of renewable energy integration, existing clearing methods are difficult to effectively allocate flexible adjustment resources, resulting in insufficient adjustment capacity when the system net load changes rapidly. Furthermore, the lack of a unified joint clearing mechanism makes it difficult to execute clearing results and results in excessively long calculation times.

Method used

By acquiring and time-synchronizing power grid operation data, upward and downward ramp demand sequences and network security margin sequences are generated. A resource-constraint coupling diagram is established, candidate regulation resources are screened, and machine learning models are used to assist in optimization solutions. A day-ahead base joint clearing status and a real-time rolling joint clearing mechanism are constructed to achieve unified optimization of power, frequency regulation reserve, balance reserve and flexible ramp products.

Benefits of technology

It improves the flexibility and efficiency of resource utilization in the new energy power system, reduces the risk of insufficient ramp-up capability, enhances the adaptability of the power market to uncertainty, shortens the calculation time for real-time rolling clearing, and improves the executability of clearing results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123073799A_ABST
    Figure CN123073799A_ABST
Patent Text Reader

Abstract

This invention discloses a flexible resource rolling joint clearing method and system. The method includes: acquiring and time-synchronizing grid operation data to generate upward and downward ramp demand sequences and network security margin sequences for each scheduling period; establishing a resource-constraint coupling diagram and screening candidate adjustment resources; using the operating cost of power generation resources, the risk of net load ramp tail, capacity-ramp coupling conflict, and the adjustment margin at the end of the scheduling cycle as a composite optimization objective to obtain the day-ahead basic joint clearing state; executing real-time rolling joint clearing assisted by the predicted state to obtain the unit active power output control instructions and upward and downward flexible capacity reservation instructions for the current rolling window; executing the instructions for the first time period and updating the subsequent rolling process according to the actual execution deviation. This invention can reduce the risk of insufficient ramp capacity and line over-limit under new energy fluctuation conditions, and shorten the online solution time for real-time scheduling.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of power system optimization operation technology, and in particular relates to a flexible adjustment resource rolling joint clearing method and system. Background Technology

[0002] With the continuous advancement of new power system construction, the installed capacity of new energy power generation such as wind power and photovoltaic power is growing rapidly, and the proportion of new energy in the power system is constantly increasing. Compared with traditional conventional units, the output of new energy is significantly affected by meteorological conditions, exhibiting randomness, volatility, and intermittency, resulting in stronger uncertainty and rapid changes in the system's net load over short time scales. When the output of new energy suddenly decreases or the system load rises rapidly, the power system needs to mobilize generating units, energy storage, or other flexible resources to compensate within a short period of time; when the output of new energy suddenly increases or the load drops rapidly, the system needs to quickly reduce the output of conventional units or increase the capacity to absorb new energy. Therefore, the power system not only needs to meet the power balance at a certain moment, but also needs to have the ability to quickly adjust across time periods.

[0003] In a spot electricity market environment, the electricity market primarily serves to achieve a balance between electricity supply and demand, while ancillary services such as frequency regulation reserves and balancing reserves are mainly used to address short-term disturbances and forecasting errors. However, with the increasing proportion of renewable energy, relying solely on traditional reserve products is insufficient to fully characterize the system's ramp-up demand between adjacent dispatch periods. Flexible ramp-up products, by reserving the upward and downward adjustment capabilities of generator units or other flexible resources in advance, provide support for rapid changes in net load that may occur within a short timescale, thereby reducing the risk of insufficient ramp-up capacity during real-time operation.

[0004] Existing electricity spot market clearing methods often model or process power generation, reserve capacity, and flexible ramping products separately, neglecting to consider the capacity occupancy relationships, ramping capacity coupling relationships, and day-ahead-real-time connection relationships among different products. When generating units simultaneously provide power generation, reserve capacity, and flexible ramping products, these products share the unit's output space and ramping margin. Without a unified joint clearing mechanism, the market clearing results may become unenforceable in actual operation. Furthermore, real-time markets typically clear on a rolling basis at fixed time intervals, requiring repeated solving of optimization models that include power generation adjustments, reserve configurations, flexible ramping product configurations, and operational constraints within a short period. Traditional methods relying entirely on optimization solvers may suffer from long computation times and insufficient online response capabilities, making it difficult to meet the engineering application requirements of real-time rolling clearing.

[0005] Therefore, there is an urgent need for a flexible rolling joint clearing method for new energy power systems, which combines the solution state migration of adjacent rolling windows, candidate active constraint prediction, hierarchical physical feasibility correction and credibility control. This method enables the prediction results to not only serve as numerical initial points, but also to determine the constraint screening and solution state reuse range, thereby improving the physical executability of the joint clearing results and the efficiency of online solution. Summary of the Invention

[0006] This invention aims to address the problem that traditional clearing methods struggle to allocate and flexibly adjust resources in a timely manner when the randomness and volatility of renewable energy output, such as wind and solar power, cause rapid changes in system net load under conditions of high-proportion renewable energy access. It also addresses the issues of redundant capacity allocation, conflicting adjustment capabilities, and difficulty in executing clearing results when power, frequency regulation reserves, balance reserves, and flexible ramping products share unit output space and ramping capacity due to a lack of unified coordination. Furthermore, it resolves the problem that existing rolling joint clearing methods fail to fully utilize the overlapping information of adjacent rolling windows, leading to drastic cross-cycle changes in unit output and flexible ramping capacity configuration, and lengthy repeated model solving times.

[0007] In a first aspect, the present invention provides a flexible adjustment method for rolling joint clearing of resources, comprising: The grid operation data is acquired and time-aligned, and then converted into discrete scheduling time periods. Based on the net load change and the quantile of the probability distribution of net load prediction error, the upward ramp demand sequence and the downward ramp demand sequence for each scheduling time period are generated. Based on the remaining transmission capacity of key transmission sections, a network security margin sequence is formed. Based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the power grid operation data, calculate the available capacity of each unit, the remaining capacity for upward ramp adjustment, the remaining capacity for downward ramp adjustment, and the occupancy of capacity and ramp rate by different regulation services. Establish a resource-constraint coupling diagram between units, scheduling periods, transmission sections, and operation constraints, and screen candidate regulation resources that are sensitive to ramp gaps and line over-limits based on the resource-constraint coupling diagram. Based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and end-of-scheduling cycle adjustment margin constitute a composite optimization objective. The unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve are jointly determined to obtain the day-ahead basic joint clearing status that meets the unit dynamic constraints and network security constraints. Based on the aforementioned day-ahead basic joint clearing status, a predictive state-assisted real-time rolling joint clearing is executed to obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The predictive state-assisted real-time rolling joint clearing includes using a machine learning model to predict the candidate joint clearing results within the current rolling optimization window. Based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window. Use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three-level feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

[0008] Secondly, the present invention provides a flexible resource rolling joint clearing system, comprising: The acquisition module is configured to acquire and time-align power grid operation data, convert the power grid operation data to discrete scheduling time periods, generate upward ramp demand sequences and downward ramp demand sequences for each scheduling time period based on the net load change and the quantile of the probability distribution of net load prediction error, and form a network security margin sequence based on the remaining transmission capacity of key transmission sections. The filtering module is configured to calculate the available capacity of each unit, the remaining capacity for upward and downward ramping, and the occupancy of capacity and ramping rate by different regulation services based on the upward ramping demand sequence, the downward ramping demand sequence, the network security margin sequence, and the power grid operation data. It then establishes a resource-constraint coupling diagram between the unit, scheduling period, transmission section, and operation constraints, and filters out candidate regulation resources that are sensitive to ramping gaps and line over-limits based on the resource-constraint coupling diagram. The construction module is configured to, based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, use the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and scheduling cycle end adjustment margin as composite optimization objectives to jointly determine the unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve, thereby obtaining a day-ahead basic joint clearing status that satisfies the unit dynamic constraints and network security constraints; The execution module is configured to perform real-time rolling joint clearing assisted by prediction status based on the day-ahead basic joint clearing status, and obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The real-time rolling joint clearing assisted by prediction status includes using a machine learning model to predict the candidate joint clearing results in the current rolling optimization window. The output module is configured to, based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window, use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three levels of feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

[0009] 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 flexible adjustment resource rolling joint clearing method according to any embodiment of the present invention.

[0010] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the processor performs the steps of the flexible adjustment resource rolling joint clearing method according to any embodiment of the present invention.

[0011] The flexible adjustment resource rolling joint clearing method and system of this application incorporates electricity, frequency regulation reserve, balance reserve, and flexible ramping products into a unified optimization framework, constructing a rolling joint clearing mechanism that coordinates the day-ahead market and the real-time market. This achieves coordinated optimization of electricity market clearing and system flexible adjustment capacity configuration under conditions of high proportion of renewable energy access. By dynamically determining upward and downward flexible ramping demands based on net load changes and prediction errors, it can more accurately reflect the real-time adjustment demands caused by renewable energy fluctuations and load changes, reducing the risk of insufficient ramping capacity. By simultaneously considering unit capacity boundaries, reserve capacity, ramping capacity, and flexible ramping product reservation constraints in the joint clearing process, it can avoid capacity occupancy conflicts among electricity, reserve, and flexible ramping products, improving the executability of the clearing results. The introduction of a machine learning model to assist the optimization solution mechanism, based on historical clearing samples, predicts the initial values ​​of real-time unit electricity adjustment and the initial configuration results of flexible ramping products, and after three levels of feasibility correction, serves as the warm-start initial solution for the optimization solver, thereby improving the online solution efficiency of real-time rolling joint clearing.

[0012] It can continuously adjust the clearing results based on the latest load, renewable energy output, and market boundary conditions, thereby ensuring the safe and stable operation of the power system, improving the efficiency of flexible resource utilization, promoting renewable energy consumption, reducing system operating costs, and enhancing the adaptability of the electricity spot market to uncertainties and rapid ramp-up demand. Attached Figure Description

[0013] 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.

[0014] Figure 1 A flowchart of a flexible resource rolling joint clearing method provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a flexible adjustment resource rolling joint clearing system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0015] 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.

[0016] Please see Figure 1 The diagram illustrates a flowchart of a flexible resource rolling joint clearing method according to this application.

[0017] like Figure 1 As shown, the flexible adjustment of the resource rolling joint clearing method specifically includes the following steps: Step S101: Obtain and time-align the power grid operation data, convert the power grid operation data to discrete scheduling time periods, generate the upward ramp demand sequence and downward ramp demand sequence for each scheduling time period based on the net load change and the quantile of the probability distribution of net load prediction error, and form a network security margin sequence based on the remaining transmission capacity of key transmission sections.

[0018] In this step, the scheduling cycle is divided into several equally spaced time periods, with a scheduling time interval of [missing information]. The net system load for the t-th time period is: , In the formula, Let t be the system net load. System net load refers to the remaining load after deducting the output of new energy sources from the total load demand of the power system. Let be the system load forecast value for the t-th time period. Let be the predicted wind power output value for the t-th time period. Let be the predicted photovoltaic power output for the t-th time period; Calculate the net load change in the t-th time period relative to the previous scheduling time period. , This represents the system net load for the (t-1)th time period; The upward error margin and the downward error margin are determined based on the probability distribution of the net load forecast error, wherein the upward error margin is the upper quantile of the probability distribution of the net load forecast error at the confidence level, and the downward error margin is the lower quantile of the probability distribution of the net load forecast error at the confidence level. Based on the upward error margin and the downward error margin, the upward flexible ramp requirement and the downward flexible ramp requirement are determined as follows: , , In the formula, Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period, This is the upper quantile of the probability distribution of net load forecast error at the confidence level, used to characterize the safety margin of uncertainty regarding net load increases. It is the lower quantile of the probability distribution of net load forecast error at the confidence level, used to characterize the safety margin of the uncertainty of net load decline.

[0019] Step S102: Based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the power grid operation data, calculate the available capacity of each unit, the remaining capacity for upward ramp adjustment, the remaining capacity for downward ramp adjustment, and the occupancy of capacity and ramp rate by different regulation services. Establish a resource-constraint coupling diagram between units, scheduling periods, transmission sections, and operation constraints. Based on the resource-constraint coupling diagram, select candidate regulation resources that are sensitive to ramp gaps and line over-limits.

[0020] In this step, a resource-constraint coupling graph is constructed, using generating units, scheduling periods, transmission sections, and operational constraints as nodes, and the relationships between nodes as edges. Specifically: For each unit g, establish an edge between the unit node and node i based on its access node i; For each transmission section, establish the edges between the transmission section and the relevant nodes based on the transmission lines it contains; For each operational constraint, based on the generating unit, time period, and transmission section involved, establish the edges between the operational constraint and the corresponding generating unit node, scheduling time period node, and transmission section node.

[0021] The method for calculating edge weights in a resource-constraint coupled graph is as follows: The weight of the edge between the unit node and the scheduling time period node is determined based on the available capacity and remaining ramp-up capacity of unit g in time period t: , In the formula, Let g be the weight of the edge between unit g and time period t. Let g be the available capacity of unit g during time period t. The maximum capacity of unit g is... and These represent the remaining ramp-up capacity and ramp-down capacity of unit g during time period t, respectively. Let g be the maximum ramp rate of unit g. and All are weighting coefficients, and .

[0022] The weight of the edge between the generating unit node and the transmission section node is determined according to the power transmission distribution factor of the generating unit g to the transmission section k: , In the formula, Let g be the power transmission distribution factor of unit g to transmission section k. For the unit The power transmission distribution factor for transmission section k; For edges between operational constraint nodes and unit nodes, scheduling time period nodes, and transmission section nodes, their weights are determined based on the constraint type and the penalty for exceeding the constraint limit. For example, for branch transmission capacity constraints, the edge weight is taken as... , to characterize the activity level of the constraint, where, For power transmission lines The active power flow during time period t For power transmission lines The maximum permissible transmission power.

[0023] The resource-constraint coupling diagram is used to characterize the impact path of unit output adjustment on power flow and operational constraints of transmission sections.

[0024] For each unit g, calculate its contribution index to the ramp-up gap: , For each unit g, calculate its power transmission distribution factor to the critical transmission section k based on the resource-constraint coupling diagram. And in conjunction with the network security margin of section k, calculate the line over-limit sensitivity index. , This is a set of key transmission sections. Let be the remaining transmission capacity of section k during time period t.

[0025] Unit g is selected as a candidate regulation resource when it meets the following conditions: , , , , , In the formula, The threshold for contributing ability to climbing slopes. For network security sensitive thresholds; and These represent the upward coupling occupancy index and the downward coupling occupancy index for unit g during time period t, respectively. and These are the safe thresholds for upward capacity-climbing coupling occupancy indicators and downward capacity-climbing coupling occupancy indicators, respectively.

[0026] Through the above screening, a set of candidate adjustment resources that are insensitive to ramp gaps and line overruns or have positive adjustment capabilities are obtained, which will serve as the decision objects for subsequent day-ahead joint clearing and real-time rolling joint clearing.

[0027] Step S103: Based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, a composite optimization objective is formed by the operating cost of power generation resources, the tail risk of net load ramp, the capacity-ramp coupling conflict, and the adjustment margin at the end of the scheduling cycle. The unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve are jointly determined to obtain the day-ahead basic joint clearing status that satisfies the unit dynamic constraints and network security constraints.

[0028] In this step, the expression for the composite optimization objective is: , In the formula, The overall objective function for the recent joint clearing of basic funds is: The basic operational costs reserved for unit operation, start-up and shutdown, standby capacity, and flexible ramp-up capability. This represents the tail risk of a gap in net load ramping capacity. The capacity-climbing capability coupling conflict For insufficient adjustment margin at the end of the scheduling window, The weights for the tail risk of the net load ramping capacity gap, the weights for capacity-ramping capacity coupling conflicts, and the weights for insufficient adjustment margin at the end of the scheduling window. The basic operational cost is: , In the formula, For time period sets, For generator sets, For the day-ahead active power output of unit g during time period t, The cost of the electrical output of unit g during time period t. and These are the startup and shutdown costs for unit g, respectively. and These are the startup and shutdown variables of unit g during time period t, respectively. and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward frequency regulation reserve quotes and downward frequency regulation reserve quotes provided by unit g during time period t. and These represent the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve price and the downward balancing reserve price, respectively. and These represent the reserved capacity for upward flexible ramping and downward flexible ramping capabilities provided by unit g during time period t. and These are the upward flexible ramp product quotes and the downward flexible ramp product quotes provided by unit g in time period t. The risk at the tail end of the net load ramp-up is: , In the formula, This is a set of net load forecasting scenarios. For confidence level, As an auxiliary variable for Value at Risk, and These represent the upward climbing requirement and the downward climbing requirement under scenario s, respectively. and These refer to the upward and downward flexible ramping capabilities configured for unit g in scenario s. This indicates taking the positive value. The capacity-ramp coupling conflict is as follows: , In the formula, and These represent the upward coupling occupancy index and the downward coupling occupancy index for unit g during time period t, respectively. and These are the safe thresholds for upward capacity-climbing coupling occupancy indicators and downward capacity-climbing coupling occupancy indicators, respectively. The adjustment margin at the end of the scheduling window is: , In the formula, This is the end of the current scheduling window. and For the expected upward and downward demand in the next period, and It provides flexible upward and downward ramping capabilities for the end of the window period.

[0029] Furthermore, the current basic joint clearing status satisfies the power system supply and demand balance constraints, unit operation boundary constraints, ancillary service capacity constraints, flexible ramp product capacity constraints, unit ramp constraints, unit start-up and shutdown logic constraints, and network security operation constraints. The power system supply and demand balance constraint is: , In the formula, For the day-ahead active power output of unit g during time period t, Let t be the total predicted output of new energy sources in the t-th time period. For generator sets, This represents the predicted system load for the t-th time period. The unit's operating boundary constraints are: , In the formula, Let g be the operating state variable of unit g during time period t. and These are the minimum and maximum technical outputs of unit g, respectively; The auxiliary service capacity constraint is: , , , , In the formula, and These are the power system's upward frequency regulation reserve requirements and downward frequency regulation reserve requirements, respectively. and These refer to balancing reserve demand upwards and balancing reserve demand downwards within the power system. and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g in time period t, respectively. The capacity constraint for the flexible ramping product is: , , In the formula, and These represent the reserved capacity for upward flexible ramping and downward flexible ramping capabilities provided by unit g during time period t. Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period; The unit's ramp-up constraint is: , In the formula, and These refer to the uphill and downhill climbing capabilities of unit g, respectively. Let g be the day-ahead active power output of unit g during time period t−1. This is the scheduling time interval.

[0030] Step S104: Based on the day-ahead basic joint clearing status, perform real-time rolling joint clearing assisted by prediction status to obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The real-time rolling joint clearing assisted by prediction status includes using a machine learning model to predict the candidate joint clearing results in the current rolling optimization window.

[0031] In this step, the real-time rolling joint clearing adopts an adaptive rolling window, with a rolling window length of... The value is determined jointly by the net load forecast error, the maximum expected ramp rate, and the safety margin of critical transmission sections, and is expressed as follows: , In the formula, and These are the upper and lower limits of the scroll window length, respectively. The standard deviation of the net load forecast error. For the maximum expected climbing range, For scheduling time intervals, The remaining transmission capacity of the critical transmission section k in time period t. This is the baseline value for the remaining transmission capacity of the cross-section. Indicates rounding up; Let the current real-time scheduling time be... The length of the scroll window is The scrolling optimization window is then represented as: , In the formula, A real-time scrolling optimized window starting from [the specified point]. The real-time rolling joint clearing aims to minimize the weighted sum of control command adjustment costs, ramp gap risks, line over-limit risks, cross-cycle command changes, and window end adjustment margins. , In the formula, For The overall objective function for real-time rolling joint clearing, starting from [the target function], The cost of adjusting real-time active power output relative to the current baseline plan. To provide a real-time scrolling window for the climbing capacity gap. To mitigate the risk of line exceeding limits, The change in control command during the first period of an adjacent rolling cycle. The current scroll window's end adjustment capability is insufficient. These are the weighting coefficients for the climbing capacity gap within the real-time rolling window, the weighting coefficient for the risk of line exceeding limits, the weighting coefficient for the change in control commands at the beginning of the adjacent rolling cycle, and the weighting coefficient for the insufficient adjustment capacity at the end of the current rolling window. Confidence level Conditional Value at Risk (VaR); The cost of adjusting the active power output is: , In the formula, and These represent the upward and downward output adjustments of unit g during time period t, respectively. and These refer to the upward and downward adjustment of electricity prices for unit g during time period t, respectively. and These represent the upward and downward flexible ramp product capacities of unit g during time period t, respectively. and These are the upward and downward flexible ramp product quotes for unit g during time period t. and These represent the variables indicating insufficient upward adjustment of electrical energy and insufficient downward adjustment of electrical energy in time period t, respectively. and These represent the upward flexible ramp product shortage variable and the downward flexible ramp product shortage variable for time period t, respectively. These are the penalty coefficients for insufficient upward adjustment of power supply in time period t, insufficient downward adjustment of power supply in time period t, insufficient upward flexible ramp product in time period t, and insufficient downward flexible ramp product in time period t, respectively. The risk of the line exceeding the limit is as follows: , In the formula, For time The set of time periods contained in the scrolling optimization window starting from [the specified point]. This refers to the set of transmission lines in the power system that participate in network security verification. Numbering of transmission lines For power transmission lines The active power flow during time period t For power transmission lines Maximum permissible transmission power This represents the penalty amount for exceeding the limit when the absolute value of the active power flow of a transmission line exceeds the maximum permissible transmission power. The cross-cycle instruction change amount is: , In the formula, For generator sets, For unit i in the current rolling cycle, targeting the time period The resulting contribution For the previous rolling cycle, targeting the time period The resulting active power output guidance value, and For unit i, this refers to the time period in the current rolling cycle. The resulting ability to climb upwards and downwards flexibly. and Each period is for the previous rolling cycle. The resulting guideline values ​​for upward and downward flexible climbing capabilities. For unit i during the time period The weighting coefficient for the change in the instruction of flexible climbing ability.

[0032] The real-time rolling joint clearing process assisted by the predicted state includes sample construction, state feature extraction, prediction of candidate joint clearing variables, identification of candidate active constraints, three-level feasibility correction, prediction credibility calculation, and solving for state reuse. The sample construction includes: collecting upward flexible ramp demand, downward flexible ramp demand, unit available capacity, reserve occupancy capacity, and corresponding historical best clearing results within historical scheduling cycles to form a training sample set. , In the formula, For the training sample set, The number of historical scheduling samples. Let m be the input feature vector corresponding to the m-th historical scheduling sample. The historical best clearing result vector corresponding to the m-th historical scheduling sample; The state feature extraction includes: using the upward flexible ramp demand, downward flexible ramp demand, unit available capacity, and reserve occupancy capacity within the current rolling optimization window as input features for the machine learning model, and the current real-time scheduling time. The corresponding input feature vector is represented as: , In the formula, For the current real-time scheduling moment The corresponding input feature vector, To optimize the upward flexible ramping requirement vector within the scrolling optimization window. To optimize the downward flexible ramping requirement vector within the scrolling optimization window, This is the vector of available capacity for the generator units. This is a vector of available spare capacity. The candidate joint clearing variable prediction includes: using a trained machine learning model to predict the candidate clearing results within the current rolling optimization window. , In the formula, The candidate joint clearing results output by the machine learning model. For parameters The machine learning model is a support vector regression model, a random forest regression model, or a deep neural network regression model. In one specific embodiment, a deep neural network regression model is used, with the number of input layer nodes being... These correspond to the upward flexible ramp demand, downward flexible ramp demand, unit available capacity, and reserve capacity for each time period within the rolling optimization window; the hidden layer is set to 3 layers, with 256, 128, and 64 nodes in each layer, respectively, and the ReLU function is used for activation; the number of nodes in the output layer is... These correspond to the active power output adjustment and flexible ramp-up product capacity of each unit at different time periods. During training, supervised learning is performed using a historical scheduling sample set. The loss function is the mean squared error function, the optimizer is the Adam optimizer, the learning rate is set to 0.001, the batch size is set to 64, and the number of training rounds is set to 500. After training, the input feature vector of the current rolling optimization window is input into the deep neural network regression model, and the candidate joint clearing result is output.

[0033] The three-level feasibility correction includes: sequentially projecting the predicted candidate joint clearing results into the first-level basic feasible region, the second-level ramping and backup coupling feasible region, and the third-level network security feasible region; and correcting the prediction results that do not meet the constraints by projection, to obtain the candidate initial solution after the three-level feasibility correction. , In the formula, This is the candidate initial solution after three levels of feasibility correction. The vector of clearing decision variables to be corrected. The basic feasible region consists of supply and demand balance constraints, unit capacity constraints, unit ramping constraints, and flexible ramping product demand constraints. The prediction confidence calculation includes: calculating the weighted distance of the candidate joint clearing results before and after the three-level feasibility correction. And combine candidate active constraints to identify probabilities and the degree of line over-limit To obtain the prediction confidence level ,in, , , All are preset weighting coefficients; The solution state reuse includes: based on prediction confidence. The optimization solver selects at least one of the following: candidate joint clearing variables, active constraints, dual variables, and discrete variable fixed states. The solution operates within a complete set of physical constraints: power system supply and demand balance constraints, unit operation boundary constraints, ancillary service capacity constraints, flexible ramping product capacity constraints, unit ramping constraints, unit start-up and shutdown logic constraints, and network security operation constraints. The final rolling joint clearing result is obtained below: , In the formula, This is the final rolling joint clearing result of the real-time rolling joint clearing. For the real-time rolling joint clearing objective function, This is the initial solution for the warm-start of the optimization solver.

[0034] Real-time rolling joint clearing also includes constraints on flexible ramping product demand satisfaction, real-time unit capacity boundary constraints, and real-time ramping capability constraints. Among them, the requirements for flexible ramp products must meet the following constraints: , , In the formula, For generator sets, and These represent the upward and downward flexible ramp product capacities of unit g during time period t, respectively. Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period, and These are the upward flexible ramp product shortage variables and the downward flexible ramp product shortage variables for time period t, respectively. The real-time capacity boundary constraints of the unit are: , , In the formula, Let g be the operating state variable of unit g during time period t. and These are the minimum and maximum technical outputs of unit g, respectively. This refers to the actual output of unit g in the real-time market at time t. The plan is to contribute to the foundation established in the market recently. and These represent the upward and downward output adjustments of unit g during time period t, respectively. The real-time capacity reservation constraint for the generating unit is: , In the formula, and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g in time period t, respectively. The real-time climbing ability constraint is: , In the formula, For scheduling time intervals, and These refer to the uphill and downhill climbing capabilities of unit g, respectively. This refers to the actual output of unit g in the real-time market during time period t-1.

[0035] The real-time rolling joint clearing also considers the power system network security operation constraints, which include node power balance constraints, branch power flow calculation constraints, and branch transmission capacity constraints. The node power balance constraint is: , In the formula, Let i be the set of generator sets connected to node i. The actual output of unit g in the real-time market during time period t. For node i to contribute energy to new energy sources during time period t, Let i be the load demand of node i in time period t. Let i be the set of nodes connected to node i. The line susceptance between node i and node j and These are the voltage phase angles of nodes i and j during time period t, respectively. The constraints for branch power flow calculation are: , The tributary transmission capacity constraint is: , In the formula, Let be the transmission power of branch ij during scheduling period t; This represents the maximum allowed transmission capacity of branch ij.

[0036] Step S105: Based on the unit active power output control instructions and the instructions to increase and decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window. Use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three-level feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

[0037] In this step, after each rolling optimization calculation is completed, only the active power output control command and flexible capacity reservation command corresponding to the first scheduling period within the rolling optimization window are executed; The results of the remaining scheduling periods within the rolling optimization window, excluding the first scheduling period, will be used as subsequent guiding plans and will not be executed directly. When entering the next scheduling cycle, the power grid operation data is updated to obtain the updated power grid operation data; The rolling optimization window is moved forward by one scheduling period, and the updated rolling optimization window is as follows: , In the formula, For the updated scrolling optimized window, This is the current real-time scheduling moment. This is the length of the scroll window; When the preset scheduling period ends, the final market joint clearing result is output.

[0038] In summary, the method of this application acquires and times-aligns grid operation data to generate upward and downward ramp demand sequences and network security margin sequences for each scheduling period; establishes a resource-constraint coupling graph and screens candidate regulation resources; uses the operating cost of power generation resources, the tail risk of net load ramp, the capacity-ramp coupling conflict, and the regulation margin at the end of the scheduling cycle as a composite optimization objective to obtain the day-ahead basic joint clearing state; executes real-time rolling joint clearing assisted by the prediction state to obtain the unit active power output control instructions and upward and downward adjustment flexibility reserve instructions for the current rolling window; executes the instructions for the first time period, updates the sample weights and prediction confidence parameters of the machine learning model according to the actual execution deviation, and updates the subsequent rolling process; it can reduce the insufficient ramping capacity and line over-limit risk under the condition of new energy fluctuations, and shorten the online solution time of real-time scheduling.

[0039] Please see Figure 2 The diagram shows a structural block diagram of a flexible adjustment resource rolling joint clearing system according to this application.

[0040] like Figure 2 As shown, the flexible adjustment resource rolling joint clearing system 200 includes an acquisition module 210, a filtering module 220, a construction module 230, an execution module 240, and an output module 250.

[0041] The acquisition module 210 is configured to acquire and time-align grid operation data, convert the grid operation data to discrete scheduling time periods, generate upward ramp demand sequences and downward ramp demand sequences for each scheduling time period based on the net load change and the quantile of the probability distribution of net load prediction error, and form a network security margin sequence based on the remaining transmission capacity of key transmission sections. The filtering module 220 is configured to calculate the available capacity of each unit, the remaining upward ramp demand sequence, the remaining downward ramp capacity, and the capacity of different regulation services based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the grid operation data. Based on the occupancy of the ramp rate, a resource-constraint coupling diagram is established between generating units, scheduling periods, transmission sections, and operational constraints. Candidate adjustment resources sensitive to ramp gaps and line over-limits are then selected based on this diagram. Module 230 is configured to, based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, use a composite optimization objective consisting of the operating cost of generation resources, the risk at the tail end of net load ramp, capacity-ramp coupling conflict, and the adjustment margin at the end of the scheduling cycle to jointly determine the unit start-up / shutdown status, active power output plan, reserve capacity, and upward and downward adjustment flexibility reserves. The system obtains a day-ahead basic joint clearing state that satisfies unit dynamic constraints and network security constraints; the execution module 240 is configured to execute a prediction-assisted real-time rolling joint clearing based on the day-ahead basic joint clearing state, obtaining the unit active power output control command and the instructions to increase or decrease flexible capacity reservation in the current rolling window, wherein the prediction-assisted real-time rolling joint clearing includes using a machine learning model to predict the candidate joint clearing results within the current rolling optimization window; the output module 250 is configured to, based on the unit active power output control command and the instructions to increase or decrease flexible capacity reservation in the current rolling window, obtain ... The system lowers the flexible capacity reservation instruction, executes the unit active power output control instruction and flexible capacity reservation instruction for the first scheduling period of the current rolling window, uses the results of the remaining periods as the guiding state, obtains the actual execution state, calculates the deviation between the candidate joint clearing result, the candidate initial solution after three-level feasibility correction, and the actual execution state, updates the sample weights and prediction confidence parameters of the machine learning model, moves the rolling window forward by one scheduling period, uses the moved rolling window as the new current rolling window, and returns to execute the real-time rolling joint clearing assisted by the prediction state until the preset scheduling cycle end point is reached, and outputs the final market joint clearing result.

[0042] 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 2The various modules in the document will not be described in detail here.

[0043] 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 flexible adjustment resource rolling joint clearing method in any of the above method embodiments. In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows: The grid operation data is acquired and time-aligned, and then converted into discrete scheduling time periods. Based on the net load change and the quantile of the probability distribution of net load prediction error, the upward ramp demand sequence and the downward ramp demand sequence for each scheduling time period are generated. Based on the remaining transmission capacity of key transmission sections, a network security margin sequence is formed. Based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the power grid operation data, calculate the available capacity of each unit, the remaining capacity for upward ramp adjustment, the remaining capacity for downward ramp adjustment, and the occupancy of capacity and ramp rate by different regulation services. Establish a resource-constraint coupling diagram between units, scheduling periods, transmission sections, and operation constraints, and screen candidate regulation resources that are sensitive to ramp gaps and line over-limits based on the resource-constraint coupling diagram. Based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and end-of-scheduling cycle adjustment margin constitute a composite optimization objective. The unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve are jointly determined to obtain the day-ahead basic joint clearing status that meets the unit dynamic constraints and network security constraints. Based on the aforementioned day-ahead basic joint clearing status, a predictive state-assisted real-time rolling joint clearing is executed to obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The predictive state-assisted real-time rolling joint clearing includes using a machine learning model to predict the candidate joint clearing results within the current rolling optimization window. Based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window. Use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three-level feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

[0044] 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 the flexible resource rolling joint clearing system, etc. 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 located relative to the processor, and this remote memory may be connected to the flexible resource rolling joint clearing system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0045] 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 realizing the flexible resource rolling joint clearing method described in the above 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 flexible resource rolling joint clearing system. The output device 340 may include a display screen or other display device.

[0046] 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.

[0047] In one implementation, the above-described electronic device is applied in a flexible adjustment resource rolling joint clearing system for a client, 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: The grid operation data is acquired and time-aligned, and then converted into discrete scheduling time periods. Based on the net load change and the quantile of the probability distribution of net load prediction error, the upward ramp demand sequence and the downward ramp demand sequence for each scheduling time period are generated. Based on the remaining transmission capacity of key transmission sections, a network security margin sequence is formed. Based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the power grid operation data, calculate the available capacity of each unit, the remaining capacity for upward ramp adjustment, the remaining capacity for downward ramp adjustment, and the occupancy of capacity and ramp rate by different regulation services. Establish a resource-constraint coupling diagram between units, scheduling periods, transmission sections, and operation constraints, and screen candidate regulation resources that are sensitive to ramp gaps and line over-limits based on the resource-constraint coupling diagram. Based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and end-of-scheduling cycle adjustment margin constitute a composite optimization objective. The unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve are jointly determined to obtain the day-ahead basic joint clearing status that meets the unit dynamic constraints and network security constraints. Based on the aforementioned day-ahead basic joint clearing status, a predictive state-assisted real-time rolling joint clearing is executed to obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The predictive state-assisted real-time rolling joint clearing includes using a machine learning model to predict the candidate joint clearing results within the current rolling optimization window. Based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window. Use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three-level feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

[0048] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0049] 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 flexible resource rolling joint clearing method, characterized in that, include: The grid operation data is acquired and time-aligned, and then converted into discrete scheduling time periods. Based on the net load change and the quantile of the probability distribution of net load prediction error, the upward ramp demand sequence and the downward ramp demand sequence for each scheduling time period are generated. Based on the remaining transmission capacity of key transmission sections, a network security margin sequence is formed. Based on the upward ramp demand sequence, the downward ramp demand sequence, the network security margin sequence, and the power grid operation data, calculate the available capacity of each unit, the remaining capacity for upward ramp adjustment, the remaining capacity for downward ramp adjustment, and the occupancy of capacity and ramp rate by different regulation services. Establish a resource-constraint coupling diagram between units, scheduling periods, transmission sections, and operation constraints, and screen candidate regulation resources that are sensitive to ramp gaps and line over-limits based on the resource-constraint coupling diagram. Based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and end-of-scheduling cycle adjustment margin constitute a composite optimization objective. The unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve are jointly determined to obtain the day-ahead basic joint clearing status that meets the unit dynamic constraints and network security constraints. Based on the aforementioned day-ahead basic joint clearing status, a predictive state-assisted real-time rolling joint clearing is executed to obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The predictive state-assisted real-time rolling joint clearing includes using a machine learning model to predict the candidate joint clearing results within the current rolling optimization window. Based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window. Use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three-level feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.

2. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The process of generating the upward and downward ramp demand sequences for each scheduling period based on the net load change and the quantiles of the net load prediction error probability distribution includes: The scheduling cycle is divided into several equally spaced time periods, with the scheduling time interval being... The net system load for the t-th time period is: , In the formula, Let t be the system net load. System net load refers to the remaining load after deducting the output of new energy sources from the total load demand of the power system. Let be the system load forecast value for the t-th time period. Let be the predicted wind power output value for the t-th time period. Let be the predicted photovoltaic power output for the t-th time period; Calculate the net load change in the t-th time period relative to the previous scheduling time period. , This represents the system net load for the (t-1)th time period; The upward error margin and the downward error margin are determined based on the probability distribution of the net load forecast error, wherein the upward error margin is the upper quantile of the probability distribution of the net load forecast error at the confidence level, and the downward error margin is the lower quantile of the probability distribution of the net load forecast error at the confidence level. Based on the upward error margin and the downward error margin, the upward flexible ramp requirement and the downward flexible ramp requirement are determined as follows: , , In the formula, Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period, This is the upper quantile of the probability distribution of net load forecast error at the confidence level, used to characterize the safety margin of uncertainty regarding net load increases. It is the lower quantile of the probability distribution of net load forecast error at the confidence level, used to characterize the safety margin of the uncertainty of net load decline.

3. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The expression for the composite optimization objective is: , In the formula, The overall objective function for the recent joint clearing of basic funds is: The basic operational costs reserved for unit operation, start-up and shutdown, standby capacity, and flexible ramp-up capability. This represents the tail risk of a gap in net load ramping capacity. The capacity-climbing capability coupling conflict For insufficient adjustment margin at the end of the scheduling window, The weights for the tail risk of the net load ramping capacity gap, the weights for capacity-ramping capacity coupling conflicts, and the weights for insufficient adjustment margin at the end of the scheduling window. The basic operational cost is: , In the formula, For time period sets, For generator sets, For the day-ahead active power output of unit g during time period t, The cost of the electrical output of unit g during time period t. and These are the startup and shutdown costs for unit g, respectively. and These are the startup and shutdown variables of unit g during time period t, respectively. and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward frequency regulation reserve quotes and downward frequency regulation reserve quotes provided by unit g during time period t. and These represent the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve price and the downward balancing reserve price, respectively. and These represent the reserved capacity for upward flexible ramping and downward flexible ramping capabilities provided by unit g during time period t. and These are the upward flexible ramp product quotes and the downward flexible ramp product quotes provided by unit g in time period t. The risk at the tail end of the net load ramp-up is: , In the formula, This is a set of net load forecasting scenarios. For confidence level, As an auxiliary variable for Value at Risk, and These represent the upward climbing requirement and the downward climbing requirement under scenario s, respectively. and These refer to the upward and downward flexible ramping capabilities configured for unit g in scenario s. This indicates taking the positive value. The capacity-ramp coupling conflict is as follows: , In the formula, and These represent the upward coupling occupancy index and the downward coupling occupancy index for unit g during time period t, respectively. and These are the safe thresholds for upward capacity-climbing coupling occupancy indicators and downward capacity-climbing coupling occupancy indicators, respectively. The adjustment margin at the end of the scheduling window is: , In the formula, This is the end of the current scheduling window. and For the expected upward and downward demand in the next period, and It provides flexible upward and downward ramping capabilities for the end of the window period.

4. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The aforementioned day-ahead joint clearing status satisfies the power system supply and demand balance constraints, unit operation boundary constraints, ancillary service capacity constraints, flexible ramp product capacity constraints, unit ramp constraints, unit start-up and shutdown logic constraints, and network security operation constraints. The power system supply and demand balance constraint is: , In the formula, For the day-ahead active power output of unit g during time period t, Let t be the total predicted output of new energy sources in the t-th time period. For generator sets, This represents the predicted system load for the t-th time period. The unit's operating boundary constraints are: , In the formula, Let g be the operating state variable of unit g during time period t. and These are the minimum and maximum technical outputs of unit g, respectively; The auxiliary service capacity constraint is: , , , , In the formula, and These are the power system's upward frequency regulation reserve requirements and downward frequency regulation reserve requirements, respectively. and These refer to balancing reserve demand upwards and balancing reserve demand downwards within the power system. and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g in time period t, respectively. The capacity constraint for the flexible ramping product is: , , In the formula, and These represent the reserved capacity for upward flexible ramping and downward flexible ramping capabilities provided by unit g during time period t. Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period; The unit's ramp-up constraint is: , In the formula, and These refer to the uphill and downhill climbing capabilities of unit g, respectively. Let g be the day-ahead active power output of unit g during time period t−1. This is the scheduling time interval.

5. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The step of performing real-time rolling joint clearing with predictive status assistance based on the day-ahead basic joint clearing status to obtain the unit active power output control command, the upward adjustment of flexible capacity reservation command, and the downward adjustment of flexible capacity reservation command for the current rolling window includes: The real-time rolling joint clearing adopts an adaptive rolling window, with a rolling window length of... The value is determined jointly by the net load forecast error, the maximum expected ramp rate, and the safety margin of critical transmission sections, and is expressed as follows: , In the formula, and These are the upper and lower limits of the scroll window length, respectively. The standard deviation of the net load forecast error. For the maximum expected climbing range, For scheduling time intervals, The remaining transmission capacity of the critical transmission section k in time period t. This is the baseline value for the remaining transmission capacity of the cross-section. Indicates rounding up; Let the current real-time scheduling time be... The length of the scroll window is The scrolling optimization window is then represented as: , In the formula, A real-time scrolling optimized window starting from [the specified point]. The real-time rolling joint clearing aims to minimize the weighted sum of control command adjustment costs, ramp gap risks, line over-limit risks, cross-cycle command changes, and window end adjustment margins. , In the formula, For The overall objective function for real-time rolling joint clearing, starting from [the target function], The cost of adjusting real-time active power output relative to the current baseline plan. To provide a real-time scrolling window for the climbing capacity gap. To mitigate the risk of line exceeding limits, The change in control command during the first period of an adjacent rolling cycle. The current scroll window's end adjustment capability is insufficient. These are the weighting coefficients for the climbing capacity gap within the real-time rolling window, the weighting coefficient for the risk of line exceeding limits, the weighting coefficient for the change in control commands at the beginning of the adjacent rolling cycle, and the weighting coefficient for the insufficient adjustment capacity at the end of the current rolling window. Confidence level Conditional Value at Risk (VaR); The cost of adjusting the active power output is: , In the formula, and These represent the upward and downward output adjustments of unit g during time period t, respectively. and These refer to the upward and downward adjustment of electricity prices for unit g during time period t, respectively. and These represent the upward and downward flexible ramp product capacities of unit g during time period t, respectively. and These are the upward and downward flexible ramp product quotes for unit g during time period t. and These represent the variables indicating insufficient upward adjustment of electrical energy and insufficient downward adjustment of electrical energy in time period t, respectively. and These represent the upward flexible ramp product shortage variable and the downward flexible ramp product shortage variable for time period t, respectively. These are the penalty coefficients for insufficient upward adjustment of power supply in time period t, insufficient downward adjustment of power supply in time period t, insufficient upward flexible ramp product in time period t, and insufficient downward flexible ramp product in time period t, respectively. The risk of the line exceeding the limit is as follows: , In the formula, For time The set of time periods contained in the scrolling optimization window starting from [the specified point]. This refers to the set of transmission lines in the power system that participate in network security verification. Numbering of transmission lines For power transmission lines The active power flow during time period t For power transmission lines Maximum permissible transmission power This represents the penalty amount for exceeding the limit when the absolute value of the active power flow of a transmission line exceeds the maximum permissible transmission power. The cross-cycle instruction change amount is: , In the formula, For generator sets, For unit i in the current rolling cycle, targeting the time period The resulting contribution For the previous rolling cycle, targeting the time period The resulting active power output guidance value, and For unit i, this refers to the time period in the current rolling cycle. The resulting ability to climb upwards and downwards flexibly. and Each period is for the previous rolling cycle. The resulting guideline values ​​for upward and downward flexible climbing capabilities. For unit i during the time period The weighting coefficient for the change in the instruction of flexible climbing ability.

6. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The real-time rolling joint clearing with prediction state assistance includes sample construction, state feature extraction, candidate joint clearing variable prediction, candidate active constraint identification, three-level feasibility correction, prediction credibility calculation, and solution state reuse. The sample construction includes: collecting upward flexible ramp demand, downward flexible ramp demand, unit available capacity, reserve occupancy capacity, and corresponding historical best clearing results within historical scheduling cycles to form a training sample set. , In the formula, For the training sample set, The number of historical scheduling samples. Let m be the input feature vector corresponding to the m-th historical scheduling sample. The historical best clearing result vector corresponding to the m-th historical scheduling sample; The state feature extraction includes: using the upward flexible ramp demand, downward flexible ramp demand, unit available capacity, and reserve occupancy capacity within the current rolling optimization window as input features for the machine learning model, and the current real-time scheduling time. The corresponding input feature vector is represented as: , In the formula, For the current real-time scheduling moment The corresponding input feature vector, To optimize the upward flexible ramping requirement vector within the scrolling optimization window. To optimize the downward flexible ramping requirement vector within the scrolling optimization window, This is the vector of available capacity for the generator units. This is a vector of available spare capacity. The candidate joint clearing variable prediction includes: using a trained machine learning model to predict the candidate clearing results within the current rolling optimization window. , In the formula, The candidate joint clearing results output by the machine learning model. For parameters The machine learning model is a support vector regression model, a random forest regression model, or a deep neural network regression model. The three-level feasibility correction includes: sequentially projecting the predicted candidate joint clearing results into the first-level basic feasible region, the second-level ramping and backup coupling feasible region, and the third-level network security feasible region; and correcting the prediction results that do not meet the constraints by projection, to obtain the candidate initial solution after the three-level feasibility correction. , In the formula, This is the candidate initial solution after three levels of feasibility correction. The vector of clearing decision variables to be corrected. The basic feasible region consists of supply and demand balance constraints, unit capacity constraints, unit ramping constraints, and flexible ramping product demand constraints. The prediction confidence calculation includes: calculating the weighted distance of the candidate joint clearing results before and after the three-level feasibility correction. And combine candidate active constraints to identify probabilities and the degree of line over-limit To obtain the prediction confidence level ,in, , , All are preset weighting coefficients; The solution state reuse includes: based on prediction confidence. The optimization solver selects at least one of the following: candidate joint clearing variables, active constraints, dual variables, and discrete variable fixed states. The solution operates within a complete set of physical constraints: power system supply and demand balance constraints, unit operation boundary constraints, ancillary service capacity constraints, flexible ramping product capacity constraints, unit ramping constraints, unit start-up and shutdown logic constraints, and network security operation constraints. The final rolling joint clearing result is obtained below: , In the formula, This is the final rolling joint clearing result of the real-time rolling joint clearing. For the real-time rolling joint clearing objective function, This is the initial solution for the warm-start of the optimization solver.

7. The flexible resource rolling joint clearing method according to claim 5, characterized in that, The real-time rolling joint clearing also includes flexible ramping product demand satisfaction constraints, unit real-time capacity boundary constraints, and real-time ramping capability constraints. Among them, the requirements for flexible ramp products must meet the following constraints: , , In the formula, For generator sets, and These represent the upward and downward flexible ramp product capacities of unit g during time period t, respectively. Let t represent the upward flexible ramping requirement in the t-th time period. For the downward flexible ramp requirement in the t-th time period, and These are the upward flexible ramp product shortage variables and the downward flexible ramp product shortage variables for time period t, respectively. The real-time capacity boundary constraints of the unit are: , , In the formula, Let g be the operating state variable of unit g during time period t. and These are the minimum and maximum technical outputs of unit g, respectively. This refers to the actual output of unit g in the real-time market at time t. The plan is to contribute to the foundation established in the market recently. and These represent the upward and downward output adjustments of unit g during time period t, respectively. The real-time capacity reservation constraint for the generating unit is: , In the formula, and These represent the upward frequency regulation reserve capacity and the downward frequency regulation reserve capacity provided by unit g during time period t, respectively. and These are the upward balancing reserve capacity and the downward balancing reserve capacity provided by unit g in time period t, respectively. The real-time climbing ability constraint is: , In the formula, For scheduling time intervals, and These refer to the uphill and downhill climbing capabilities of unit g, respectively. This refers to the actual output of unit g in the real-time market during time period t-1.

8. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The real-time rolling joint clearing also considers the power system network security operation constraints, which include node power balance constraints, branch power flow calculation constraints, and branch transmission capacity constraints. The node power balance constraint is: , In the formula, Let i be the set of generator sets connected to node i. The actual output of unit g in the real-time market during time period t. For node i to contribute energy to new energy sources during time period t, Let i be the load demand of node i in time period t. Let i be the set of nodes connected to node i. The line susceptance between node i and node j and These are the voltage phase angles of nodes i and j during time period t, respectively. The constraints for branch power flow calculation are: , The tributary transmission capacity constraint is: , In the formula, Let be the transmission power of branch ij during scheduling period t; This represents the maximum allowed transmission capacity of branch ij.

9. The flexible resource rolling joint clearing method according to claim 1, characterized in that, The step of moving the scrolling window forward by one scheduling period and using the moved scrolling window as the new current scrolling window includes: After each rolling optimization calculation is completed, only the active power output control command and flexible capacity reservation command corresponding to the first scheduling period within the rolling optimization window are executed; The results of the remaining scheduling periods within the rolling optimization window, excluding the first scheduling period, will be used as subsequent guiding plans and will not be executed directly. When entering the next scheduling cycle, the power grid operation data is updated to obtain the updated power grid operation data; The rolling optimization window is moved forward by one scheduling period, and the updated rolling optimization window is as follows: , In the formula, For the updated scrolling optimized window, This is the current real-time scheduling moment. This is the length of the scroll window; When the preset scheduling period ends, the final market joint clearing result is output.

10. A flexible resource rolling joint clearing system, characterized in that, include: The acquisition module is configured to acquire and time-align power grid operation data, convert the power grid operation data to discrete scheduling time periods, generate upward ramp demand sequences and downward ramp demand sequences for each scheduling time period based on the net load change and the quantile of the probability distribution of net load prediction error, and form a network security margin sequence based on the remaining transmission capacity of key transmission sections. The filtering module is configured to calculate the available capacity of each unit, the remaining capacity for upward and downward ramping, and the occupancy of capacity and ramping rate by different regulation services based on the upward ramping demand sequence, the downward ramping demand sequence, the network security margin sequence, and the power grid operation data. It then establishes a resource-constraint coupling diagram between the unit, scheduling period, transmission section, and operation constraints, and filters out candidate regulation resources that are sensitive to ramping gaps and line over-limits based on the resource-constraint coupling diagram. The construction module is configured to, based on the candidate adjustment resources, the upward ramp demand sequence, the downward ramp demand sequence, and the network security margin sequence, use the power generation resource operation cost, net load ramp tail risk, capacity-ramp coupling conflict, and scheduling cycle end adjustment margin as composite optimization objectives to jointly determine the unit start-up and shutdown status, active power output plan, reserve capacity, upward adjustment flexibility reserve, and downward adjustment flexibility reserve, thereby obtaining a day-ahead basic joint clearing status that satisfies the unit dynamic constraints and network security constraints; The execution module is configured to perform real-time rolling joint clearing assisted by prediction status based on the day-ahead basic joint clearing status, and obtain the unit active power output control command and the instructions to increase and decrease flexible capacity reservation in the current rolling window. The real-time rolling joint clearing assisted by prediction status includes using a machine learning model to predict the candidate joint clearing results in the current rolling optimization window. The output module is configured to, based on the unit active power output control instructions and the instructions to increase or decrease flexible capacity reservations in the current rolling window, execute the unit active power output control instructions and flexible capacity reservation instructions for the first scheduling period of the current rolling window, use the results of the remaining periods as the guiding state, obtain the actual execution state, calculate the deviation between the candidate joint clearing results, the candidate initial solutions after three levels of feasibility correction, and the actual execution state, update the sample weights and prediction confidence parameters of the machine learning model, move the rolling window forward by one scheduling period, use the moved rolling window as the new current rolling window, return to execute the real-time rolling joint clearing assisted by the prediction state, until the preset scheduling cycle endpoint is reached, and output the final market joint clearing result.