A new energy source load storage collaborative prediction and dispatching system and method

Abstract

The application discloses a new energy source-load-storage collaborative prediction and dispatching system and method, and relates to the technical field of new energy power dispatching.The system forms a dispatching input data set through a data acquisition module, determines an effective available power upper limit based on new energy prediction data, equipment state data and capacity mapping data through an equipment state constraint correction module, and performs state constraint correction on a theoretical available power distribution; a scene construction module constructs a representative scene and a scene probability based on the corrected new energy available power distribution, price uncertainty data and load deviation uncertainty data; and a risk dispatching module solves a dispatching strategy in combination with a flexible load response amount, energy storage charging and discharging power, contract declaration execution power and a conditional risk value item.The application can improve the matching of dispatching scenes and actual operation states of equipment, and enhance the executability and risk adaptability of a new energy source-load-storage collaborative dispatching strategy.

Description

Technical Field

[0001] This invention relates to the field of new energy power dispatching technology, and in particular to a new energy source-load-storage coordinated prediction and dispatching system and method. Background Technology

[0002] As the proportion of renewable energy generation connected to the grid continues to increase, the volatility and randomness of renewable energy output, as well as changes in equipment operating status, are increasingly impacting power dispatch. In the scenario of coordinated operation of energy sources, loads, and energy storage, the dispatch process needs to simultaneously coordinate renewable energy generation, flexible loads, energy storage devices, and market transactions, and formulate an executable dispatch strategy within a unified dispatch cycle. Due to differences in data sources, time scales, and operational constraints among various entities, discrepancies can easily arise between dispatch inputs and actual operating conditions.

[0003] Existing source-load-storage coordinated scheduling methods typically optimize solutions based on renewable energy forecasts, load forecasts, electricity price information, and energy storage operational constraints. However, in actual operation, the availability and degradation of renewable energy generation equipment affect its generating capacity and output uncertainty. If the scheduling process does not adequately characterize equipment state changes, the constructed scheduling scenario may fail to accurately reflect the actual generating capacity of renewable energy generation equipment within the scheduling cycle.

[0004] When renewable energy, flexible loads, energy storage, and market transactions participate in dispatching, dispatching strategies also need to consider source-load balance, operational constraints, and uncertainties. Existing methods, when dealing with uncertainties such as renewable energy output, electricity price changes, and load deviations, typically focus more on conventional forecasting or scenario optimization, and there is still room for improvement in characterizing the linkage between equipment state constraints, load response, energy storage operation, and risk dispatching. Therefore, it is necessary to propose a renewable energy-load-storage collaborative forecasting and dispatching scheme that can form a consistent logical chain among dispatching inputs, scenario construction, load response, and risk dispatching. Summary of the Invention

[0005] The purpose of this invention is to provide a new energy source-load-storage coordinated prediction and scheduling system and method to address the problems in existing source-load-storage coordinated scheduling, such as insufficient characterization of new energy power generation equipment status changes and generating capacity, low matching degree between scheduling scenarios and actual operating conditions, and insufficient linkage between flexible load response, energy storage operation, and risk scheduling. By coordinating equipment status constraints, source-side uncertainties, load response, energy storage operation, and risk control in a unified scheduling chain, a scheduling strategy suitable for new energy source-load-storage coordinated operation is generated, improving the executability and risk adaptability of the scheduling strategy.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] A new energy source-load-storage coordinated prediction and scheduling system is used to coordinate the prediction and scheduling of new energy power generation equipment, flexible load clusters and energy storage equipment included in the scheduling scope, including:

[0008] The data acquisition module is used to acquire new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data, and energy storage operation data within the scheduling cycle, and form a scheduling input dataset;

[0009] The equipment state constraint correction module is used to determine the upper bound of theoretical available power and the distribution of theoretical available power based on new energy prediction data, determine the equipment state constraint vector including available capacity coefficient and state risk coefficient based on equipment state data, determine the upper bound of effective available power based on the equipment state constraint vector and capacity mapping data, truncate the high-power side probabilistic mass that exceeds the upper bound of effective available power, and transfer the truncated probabilistic mass to a power range not higher than the upper bound of effective available power according to the low-power transfer kernel determined by the state risk coefficient, so as to obtain the distribution of new energy available power after equipment state constraint correction.

[0010] The scenario construction module is used to construct representative scenarios and scenario probabilities based on the distribution of renewable energy power generation capacity after equipment state constraint correction, electricity price uncertainty data, and load deviation uncertainty data.

[0011] The load response module is used to determine the flexible load response amount based on the demand response incentive price and price response characteristics.

[0012] The risk scheduling module is used to determine the source-load balance deviation based on representative scenarios, scenario probabilities, flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power. It also combines energy storage operation constraints, demand response constraints, contract declaration constraints, and conditional risk value terms to solve the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy.

[0013] The policy publishing module is used to generate and send scheduling policy data packets.

[0014] As a preferred embodiment of the present invention, the data acquisition module includes:

[0015] The scheduling time reference unit is used to generate a unified time period sequence based on the scheduling cycle and scheduling time granularity, and to map new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data and energy storage operation data to the unified time period sequence;

[0016] The equipment data association unit is used to establish a corresponding relationship between the new energy forecast data, equipment status data and capacity mapping data of the same new energy power generation equipment under the same scheduling period, with the new energy power generation equipment as the association object.

[0017] The data quality discrimination unit is used to perform timeliness verification, boundary consistency verification, and time sequence continuity verification on the data after the completion of time period mapping and device data association, and generate data quality identifiers;

[0018] The scheduling input generation unit is used to generate a scheduling input dataset based on a unified time period sequence, correspondence, and data quality identifier, and then send the scheduling input dataset to the equipment status constraint correction module, scenario construction module, load response module, and risk scheduling module.

[0019] As a preferred embodiment of the present invention, the device state constraint correction module includes:

[0020] The state vector construction unit is used to construct the equipment state constraint vector based on the equipment state data and data quality identifier in the scheduling input dataset. The equipment state constraint vector includes the available capacity coefficient and state risk coefficient of the new energy power generation equipment during the scheduling period.

[0021] The available capacity coefficient is a dimensionless coefficient in the interval [0,1], used to reflect the available capacity ratio of new energy power generation equipment in the corresponding scheduling period; the state risk coefficient is a dimensionless coefficient in the interval [0,1], used to reflect the output uncertainty corresponding to the state decay of new energy power generation equipment.

[0022] The effective upper bound determination unit is used to determine the theoretical upper bound of the generated power based on new energy forecast data, the upper bound of the capacity-mapped power based on capacity mapping data, and the effective upper bound of the available power based on the available capacity coefficient, the theoretical upper bound of the generated power, and the upper bound of the capacity-mapped power.

[0023] ;

[0024] In the formula, Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates new energy power generation equipment During the scheduling period Available capacity factor, Indicates new energy power generation equipment During the scheduling period The theoretical upper limit of the power output, Indicates new energy power generation equipment During the scheduling period The capacity is mapped to the upper limit of power.

[0025] As a preferred embodiment of the present invention, the device state constraint correction module further includes:

[0026] The high-power probability truncation unit is used to truncate the probability mass of the theoretically available power distribution that exceeds the effective available power upper bound, based on the upper bound of the effective available power. The theoretically available power distribution uses the non-negative renewable energy available power as the supporting variable, and the probability density function of the theoretically available power distribution is: The cutoff probability mass satisfies the following formula:

[0027] ;

[0028] In the formula, Indicates new energy power generation equipment During the scheduling period The high-power side probability mass that was intercepted Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates the power output of new energy sources;

[0029] The low-power probability transfer unit is used to determine the low-power transfer kernel based on the state risk coefficient, and to transfer the truncated probability mass to a power range no higher than the upper bound of the effective available power, thus obtaining the probability density function of the renewable energy power distribution after equipment state constraint correction:

[0030] ;

[0031] In the formula, This represents the probability density function of the renewable energy power distribution after equipment state constraint correction. Represents the characteristic function, Indicates a low-power transfer core. Indicates new energy power generation equipment During the scheduling period State risk coefficient;

[0032] Low-power transfer core To be based on state risk system The normalized probability density function of the parameters, low-power transfer kernel The support range is Outside the support interval, the value is 0, and the following formula is satisfied:

[0033] ;

[0034] The probability density function of the renewable energy power distribution after equipment state constraint correction Send to the scenario building module to generate renewable energy power scenarios.

[0035] As a preferred embodiment of the present invention, the scene construction module includes:

[0036] The joint sampling unit is used to generate an original joint scenario by taking the distribution of renewable energy power generation capacity after equipment state constraint correction, electricity price uncertainty data and load deviation uncertainty data as joint sampling objects; the original joint scenario includes the renewable energy power generation capacity sequence, electricity price sequence and load deviation sequence within the same scheduling cycle;

[0037] The scenario reduction unit is used to determine the joint scenario distance based on the renewable energy power generation sequence, electricity price sequence, and load deviation sequence, and to reduce the original joint scenario based on the joint scenario distance to obtain a representative scenario; the joint scenario distance is determined based on the normalized renewable energy power generation sequence, electricity price sequence, and load deviation sequence.

[0038] The probability determination unit is used to determine the scenario probability corresponding to the representative scenario based on the attribution relationship between the original joint scenario and the representative scenario, and send the representative scenario and scenario probability to the risk scheduling module.

[0039] As a preferred embodiment of the present invention, the load response module includes:

[0040] The response calculation unit is used to calculate flexible load clusters based on demand response incentive prices, price response characteristics, baseline responsiveness, response constraint boundaries, and the average response of the previous round of clusters. During the scheduling period The flexible load response within the system, the number of flexible load clusters is Flexible load cluster number The flexible load response is located within the response constraint boundary;

[0041] The group average update unit is used to update the group average response based on the flexible load response of each flexible load cluster in the same scheduling period, and to feed back the updated group average response to the response calculation unit.

[0042] The convergence discrimination unit is used to determine whether the preset convergence condition is met based on the difference between the average response of two adjacent rounds of the group, and when the preset convergence condition is met, the converged flexible load response is sent to the risk scheduling module.

[0043] The preset convergence condition includes that the difference between the average response of two adjacent rounds is less than or equal to the preset convergence threshold, and the flexible load response after convergence is the response of the scheduling period corresponding to the demand response incentive price.

[0044] As a preferred embodiment of the present invention, the risk scheduling module includes:

[0045] The joint decision-making setting unit is used to set contract application strategy, demand response incentive strategy and energy storage charging and discharging strategy as joint decision-making objects;

[0046] The source-load balance deviation calculation unit is used to calculate the representative scenario based on the representative scenario, the converged flexible load response, the energy storage charging power, the energy storage discharging power, and the contract declaration execution power, according to the following relationship. Next scheduling period Source load balance deviation:

[0047] ;

[0048] In the formula, Indicates the representative scenario Next scheduling period Source load balance deviation, Indicates the representative scenario Next scheduling period Baseline load, Indicates the representative scenario Next scheduling period Load deviation, Indicates the representative scenario Lower flexible load cluster During the scheduling period The convergent response, Indicates the number of flexible load clusters. This represents a representative scenario generated by the scenario construction module based on the distribution of renewable energy power generation after correction of equipment state constraints. Next scheduling period The renewable energy generation capacity, Indicates the scheduling period The energy storage discharge power, Indicates the scheduling period Energy storage charging power, Indicates the scheduling period Contract declaration execution power;

[0049] The constraint processing unit is used to construct a constraint set for the joint decision-making object based on energy storage operation constraints, demand response constraints, and contract declaration constraints. The constraint set includes boundary constraints for energy storage charging and discharging power, energy storage state of charge constraints, demand response quantity boundary constraints, and contract declaration execution power boundary constraints.

[0050] The risk solution unit is used to solve the joint decision object based on the source load balance deviation, constraint set, and conditional risk value term.

[0051] In a preferred embodiment of the present invention, the risk solving unit uses the representative scenarios and scenario probabilities output by the scenario construction module as random inputs, and employs the following stochastic optimization model to solve the joint decision object:

[0052] ;

[0053] The constraints of the stochastic optimization model include energy storage operation constraints, demand response constraints, contract declaration constraints, and the following conditional value-at-risk auxiliary constraints:

[0054] ;

[0055] In the formula, , , , These represent the contract declaration execution power sequence, demand response incentive price sequence, energy storage charging power sequence, and energy storage discharging power sequence within the scheduling period, respectively. This represents a secondary variable indicating the conditional value at risk. Indicates the representative scenario Conditional value at risk auxiliary slack variables, Represents the set of scheduling periods. This indicates the number of scenes. Indicates the representative scenario The probability of the scenario, and , , Indicates the representative scenario The scene scheduling benefit evaluation value is as follows. Show representative scene The scene loss caused by the source-load balance deviation is as follows. Indicates risk weight. Indicates the constraint penalty weight, Indicates the scheduling period Comprehensive constraint and penalty items;

[0056] Scene loss The source-load balance deviation is determined by the deviation cost, demand response incentive cost, and energy storage operation cost. The source-load balance deviation is calculated based on a representative scenario generated from the distribution of renewable energy power generation capacity after equipment state constraint correction. , and The evaluation coefficients are converted to the same evaluation dimension.

[0057] As a preferred embodiment of the present invention, the strategy publishing module includes:

[0058] The strategy data generation unit is used to generate a scheduling strategy data package based on the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy obtained by the risk solving unit. The scheduling strategy data package includes a scheduling cycle identifier, a scheduling period identifier, contract declaration execution power, demand response incentive price, energy storage charging power, and energy storage discharging power.

[0059] The strategy distribution unit is used to send scheduling strategy data packets to the transaction reporting system, the demand response execution system, and the energy storage management system.

[0060] The execution feedback unit is used to receive execution deviation data returned by the transaction reporting system, the demand response execution system, and the energy storage energy management system, and send the execution deviation data to the data acquisition module.

[0061] This invention also provides a method for coordinated prediction and scheduling of new energy sources, loads, and storage, applied to a new energy source, load, and storage coordinated prediction and scheduling system as described above, comprising the following steps:

[0062] S1. Collect new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data and energy storage operation data within the scheduling cycle, and form a scheduling input dataset;

[0063] S2. Determine the upper bound of theoretically available power and the distribution of theoretically available power based on the new energy prediction data in the scheduling input dataset. Determine the equipment status constraint vector including the available capacity coefficient and the status risk coefficient based on the equipment status data in the scheduling input dataset. Determine the upper bound of effective available power based on the equipment status constraint vector and the capacity mapping data.

[0064] S3. Based on the upper bound of the effective available power, the high-power probabilistic mass in the theoretical power distribution that exceeds the upper bound of the effective available power is truncated, and the truncated probabilistic mass is transferred to a power range not higher than the upper bound of the effective available power according to the low-power transfer kernel determined by the state risk coefficient, so as to obtain the new energy power distribution after the equipment state constraint correction.

[0065] S4. Based on the distribution of renewable energy generation capacity after equipment state constraint correction, electricity price uncertainty data and load deviation uncertainty data, construct scenarios to obtain representative scenarios and scenario probabilities.

[0066] S5. Determine the flexible load response amount based on the demand response incentive price and price response characteristics;

[0067] S6. Based on representative scenarios, scenario probabilities, flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power, determine the source-load balance deviation, and combine energy storage operation constraints, demand response constraints, contract declaration constraints, and conditional risk value terms to solve the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy.

[0068] S7. Generate a scheduling strategy data packet based on the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy, and send the scheduling strategy data packet.

[0069] The beneficial effects of this invention are as follows: By introducing equipment state constraint correction processing during the coordinated scheduling of new energy sources, loads, and storage, this invention uses new energy forecast data, equipment state data, and capacity mapping data together to characterize the renewable energy generation capacity. This allows the distribution of renewable energy generation power to reflect the available capacity status and state risk of renewable energy power generation equipment during the scheduling cycle. Compared to methods that construct scheduling inputs solely based on conventional forecast results, this invention can reduce the deviation between the scheduling scenario and the actual renewable energy generation capacity of the equipment, providing a more suitable source-side uncertainty input for subsequent coordinated scheduling of sources, loads, and storage.

[0070] This invention is based on the distribution of renewable energy generation capacity after equipment state constraints are corrected. It constructs representative scenarios and their probabilities by combining electricity price uncertainty data and load deviation uncertainty data. Simultaneously, it incorporates flexible load response, energy storage charging power, energy storage discharging power, and contract execution power into the source-load balance deviation calculation, thus establishing a unified scheduling solution relationship among renewable energy generation, flexible loads, energy storage equipment, and contract applications. By combining conditional risk value terms to solve the scheduling strategy while satisfying energy storage operation constraints, demand response constraints, and contract application constraints, this invention improves the executability, coordination, and risk adaptability of the scheduling strategy under uncertain operating scenarios. Attached Figure Description

[0071] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the modular structure of the system of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0072] The present invention will be further described below with reference to embodiments. It should be understood that the following embodiments are used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Without departing from the concept of the present invention, those skilled in the art can adaptively configure the relevant parameters, data interfaces and calculation methods according to the actual scheduling cycle, equipment type, data source and market operation rules.

[0073] like Figure 1As shown, this embodiment provides a new energy source-load-storage coordinated prediction and scheduling system for coordinated prediction and scheduling of new energy power generation equipment, flexible load clusters, and energy storage equipment included in the scheduling scope. The system includes a data acquisition module, an equipment state constraint correction module, a scenario construction module, a load response module, a risk scheduling module, and a strategy publishing module. Each module processes data in the order of data acquisition, equipment state constraint correction, scenario construction, load response, risk scheduling, and strategy publishing, and completes the generation of scheduling strategies within the same scheduling cycle and unified time period sequence.

[0074] (1) Data acquisition module

[0075] In this embodiment, the data acquisition module is used to acquire data related to new energy power generation equipment, flexible load clusters, energy storage equipment, and market operation within the same scheduling cycle, and organizes data from different sources, time scales, and formats into a scheduling input dataset. This scheduling input dataset serves as the unified data input for the equipment state constraint correction module, scenario construction module, load response module, and risk scheduling module.

[0076] New energy forecast data is used to characterize the predicted output of new energy power generation equipment during each scheduling period within the scheduling cycle; equipment status data is used to characterize the availability, operation, maintenance, derating, fault, or health status of new energy power generation equipment; capacity mapping data is used to characterize the correspondence between new energy power generation equipment and rated capacity, grid-connected capacity, line capacity, or scheduling access capacity; flexible load data is used to characterize the baseline load, responsive capacity, response boundary, response period, and price response characteristics of flexible load clusters; market transaction data is used to characterize electricity prices, contract declaration requirements, transaction periods, and settlement cycles; and energy storage operation data is used to characterize the state of charge, charge / discharge power boundary, capacity boundary, operation status, and availability status of energy storage equipment.

[0077] The scheduling time base unit generates a unified time period sequence based on the scheduling cycle and scheduling time granularity. This unified time period sequence serves as a common time index for various types of data. For data whose sampling frequency is inconsistent with the scheduling time granularity, the scheduling time base unit maps new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data, and energy storage operation data to the unified time period sequence by means of aggregation, interpolation, or maintaining the previous valid value.

[0078] The equipment data association unit uses new energy power generation equipment as the association object. Based on the equipment identifier and the scheduling period identifier, it establishes a correspondence between the new energy forecast data, equipment status data, and capacity mapping data of the same new energy power generation equipment under the same scheduling period, forming a data record for a single new energy power generation equipment and a single scheduling period. This data record is used by the subsequent equipment status constraint correction module to construct the equipment status constraint vector and determine the upper bound of the effective available power.

[0079] The data quality assessment unit performs timeliness, boundary consistency, and temporal continuity checks on the data after time period mapping and device data association, and generates data quality identifiers. Timeliness checks determine whether the data acquisition time, update time, or generation time meets the scheduling cycle requirements; boundary consistency checks determine whether data values ​​conform to device capacity, operating boundaries, and business rules; and temporal continuity checks determine whether there are missing, abnormal jumps, or time breakpoints for the same data object between adjacent scheduling time periods. Data quality identifiers characterize the usability of the corresponding data in subsequent calculations. Data quality identifiers include normal, missing, out-of-bounds, timeout, and imputation identifiers. For data with missing, out-of-bounds, timeout, or imputation identifiers, subsequent modules remove, correct, downgrade, or call backup data according to preset data processing rules.

[0080] The dispatch input generation unit generates a dispatch input dataset based on a unified time period sequence, correspondence, and data quality identifier. The dispatch input dataset includes dispatch time period identifiers, equipment identifiers, data types, data values, and data quality identifiers. For new energy power generation equipment, the dispatch input dataset includes new energy forecast data, equipment status data, and capacity mapping data. For flexible load clusters, the dispatch input dataset includes baseline load, responsive capacity, price response characteristics, and response constraint boundaries. For energy storage equipment, the dispatch input dataset includes state of charge, charge / discharge power boundaries, capacity parameters, and operating status. For market trading objects, the dispatch input dataset includes electricity price data, contract declaration-related data, and dispatch time period identifiers.

[0081] The scheduling input generation unit sends the scheduling input dataset to the equipment state constraint correction module, scenario construction module, load response module, and risk scheduling module. Specifically, the equipment state constraint correction module receives renewable energy forecast data, equipment state data, capacity mapping data, and data quality identifiers; the scenario construction module receives data related to renewable energy forecasts, market transactions, and load deviations, with market transaction data serving as the data basis for generating electricity price uncertainty data, and flexible load data serving as the data basis for generating load deviation uncertainty data; the load response module receives flexible load data, demand response incentive price-related data, and price response characteristics; and the risk scheduling module receives market transaction data, energy storage operation data, flexible load response-related data, and scheduling constraint-related data.

[0082] (2) Equipment state constraint correction module

[0083] In this embodiment, the equipment state constraint correction module receives the scheduling input dataset output by the data acquisition module and, based on the new energy prediction data, equipment state data, capacity mapping data, and data quality identifiers in the scheduling input dataset, corrects the state constraints on the distribution of renewable energy power generation capacity of the renewable energy power generation equipment within the scheduling cycle. The processing object of the equipment state constraint correction module is the renewable energy power generation equipment included in the scheduling scope, the processing granularity is the scheduling period, and the output result is the renewable energy power generation capacity distribution after equipment state constraint correction. This output result is sent to the scenario construction module to generate a renewable energy power generation capacity scenario.

[0084] Let the number of the new energy power generation equipment be The scheduling period is The equipment state constraint correction module reads the new energy power generation equipment from the scheduling input dataset. During the scheduling period The data includes renewable energy forecasting data, equipment status data, capacity mapping data, and data quality identifiers. Renewable energy forecasting data is used to determine the upper limit and distribution of theoretically available power. Equipment status data reflects the availability, operation, maintenance, derated, fault, or health status of renewable energy power generation equipment during the corresponding dispatch period. Capacity mapping data determines the upper limit of the capacity-mapped power of renewable energy power generation equipment during the corresponding dispatch period, limited by rated capacity, grid-connected capacity, line capacity, or dispatch access capacity. Data quality identifiers determine whether the corresponding data is directly used in the calculation and whether correction, deweighting, or the use of backup data is required.

[0085] The state vector construction unit is used to construct equipment state constraint vectors based on equipment state data and data quality identifiers in the scheduling input dataset. These equipment state constraint vectors include those for new energy power generation equipment. During the scheduling period Available capacity coefficient within and state risk coefficient Available capacity factor A dimensionless coefficient in the interval [0,1], used to characterize the capacity availability ratio of new energy power generation equipment during the corresponding dispatch period; state risk coefficient. is a dimensionless coefficient in the interval [0,1], used to characterize the output uncertainty corresponding to the state decay of new energy power generation equipment. The available capacity coefficient and the state risk coefficient characterize the capacity availability ratio and the output uncertainty, respectively, and are used to determine the upper bound of effective available power and handle low power probability transfer.

[0086] In one implementation, the available capacity coefficient is determined based on the ratio between the available capacity and rated capacity of the new energy power generation equipment during the dispatch period, and is limited to the range [0,1] through a limiting process. The available capacity on the equipment side is determined by the operating status, maintenance status, fault status, derating ratio, and available capacity records in the equipment status data. When the new energy power generation equipment is in normal operation, the available capacity coefficient is close to 1; when the new energy power generation equipment is in derating operation, partial fault, or partial shutdown state, the available capacity coefficient is determined based on the ratio between the available capacity and rated capacity on the equipment side; when the new energy power generation equipment is in an unavailable state, the available capacity coefficient is 0 or close to 0.

[0087] In one implementation, the state risk coefficient is determined based on a health decay index, an alarm level normalization index, and a historical output deviation normalization index. These indices are derived from equipment status data or historical operating data associated with the equipment status data, and are normalized to a range of [0,1]. The state risk coefficient can be obtained by combining the above normalization indices according to preset weights, and then limited to a range of [0,1] through amplitude limiting. When the new energy power generation equipment is in a stable state, has a low alarm level, and a small historical output deviation, the state risk coefficient is low; when the new energy power generation equipment experiences state decay, increased alarm levels, abnormal operation, or increased historical output deviation, the state risk coefficient is high.

[0088] When the data quality identifier is normal, the state vector construction unit directly uses the corresponding equipment status data to calculate the available capacity coefficient and state risk coefficient. When the data quality identifier is missing, out-of-bounds, timeout, or imputation, the state vector construction unit processes the data according to preset data processing rules. These preset data processing rules include calling the equipment status data from the previous effective scheduling period, calling the status reference data of similar new energy power generation equipment, using the latest effective status from the operation and maintenance system, performing boundary correction on outliers, or applying conservative corrections to coefficients obtained from imputation data.

[0089] The effective upper bound determination unit is used to determine the theoretical upper bound of the generated power based on new energy forecast data, the upper limit of the capacity-mapped power based on capacity mapping data, and the effective available power upper bound based on the available capacity coefficient, the theoretical upper bound of the generated power, and the upper limit of the capacity-mapped power. Theoretical upper bound of generated power. This refers to the new energy power generation equipment obtained from new energy forecast data without considering equipment degradation. During the scheduling period The theoretical maximum achievable power output. The distribution of theoretical achievable power output is supported by the achievable power output of non-negative renewable energy sources; in one implementation, the theoretical achievable power output distribution is within... The probability density outside is 0. Capacity-mapped power upper limit. This indicates the new energy power generation equipment obtained based on capacity mapping data. During the scheduling period The upper limit of power is limited by rated capacity, grid-connected capacity, line capacity, or dispatch access capacity.

[0090] Upper limit of effective available power Determined according to the following relationship:

[0091] ;

[0092] In the formula, Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates new energy power generation equipment During the scheduling period Available capacity factor, Indicates new energy power generation equipment During the scheduling period The theoretical upper limit of the power output, Indicates new energy power generation equipment During the scheduling period The capacity is mapped to the upper limit of power.

[0093] Each power parameter takes a non-negative value, and the available capacity coefficient is located in the range of [0,1]. Therefore, the upper limit of the effective available power does not exceed the smaller value between the upper limit of the theoretically available power and the upper limit of the capacity-mapped power.

[0094] When the upper bound of effective available power is 0, the equipment state constraint correction module no longer calculates the low-power transfer core and sets the available power of new energy power generation equipment to 0 during the corresponding scheduling period; under discrete probability representation, the probability mass is concentrated on The corresponding power grid.

[0095] The device state constraint correction module also includes:

[0096] The high-power probability truncation unit is used to truncate the probability mass of the theoretically available power distribution that exceeds the effective available power upper bound, based on the upper bound of the effective available power. The theoretically available power distribution uses the non-negative renewable energy available power as the supporting variable, and the probability density function of the theoretically available power distribution is: The cutoff probability mass satisfies the following formula:

[0097] ;

[0098] In the formula, Indicates new energy power generation equipment During the scheduling period The high-power side probability mass that was intercepted Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates the power output of new energy sources;

[0099] The low-power probability transfer unit is used to determine the low-power transfer kernel based on the state risk coefficient, and to transfer the truncated probability mass to a power range no higher than the upper bound of the effective available power, thus obtaining the probability density function of the renewable energy power distribution after equipment state constraint correction:

[0100] ;

[0101] In the formula, This represents the probability density function of the renewable energy power distribution after equipment state constraint correction. Represents the characteristic function, Indicates a low-power transfer core. Indicates new energy power generation equipment During the scheduling period State risk coefficient;

[0102] Low-power transfer core To be based on state risk system The normalized probability density function of the parameters, low-power transfer kernel The support range is Outside the support interval, the value is 0, and the following formula is satisfied:

[0103] ;

[0104] The low-power transfer kernel also satisfies the non-negativity requirement. The state risk coefficient is used to adjust the distribution of the low-power transfer kernel within the support interval; when the state risk coefficient is small, the low-power transfer kernel can distribute the cutoff probability mass more widely in the interval close to the upper bound of the effective available power; when the state risk coefficient is large, the low-power transfer kernel can distribute the cutoff probability mass towards the lower power interval. In one embodiment, the low-power transfer kernel uses a truncated exponential normalized kernel, a Beta normalized kernel, a piecewise normalized kernel, or other probability density functions that satisfy the support interval, non-negativity, and integral normalization conditions, with the state risk coefficient as a parameter. The specific kernel function type mentioned above does not limit the unique implementation of the low-power transfer kernel.

[0105] The low-power probability transfer unit obtains the probability density function of the renewable energy power distribution after equipment state constraint correction according to the following formula:

[0106] ;

[0107] In the formula, This represents the probability density function of the renewable energy power distribution after equipment state constraint correction. Represents the characteristic function, The probability density function representing the theoretically achievable power distribution. This indicates the high-power side probabilistic mass that was truncated. This represents a low-power transfer kernel. The indicator function is used to preserve the portion of the theoretically available power distribution that does not exceed the upper bound of the effective available power. Used to represent the redistribution of the intercepted probability quality within a power range not exceeding the upper bound of the effective available power.

[0108] Since the theoretically available power distribution is a normalized probability distribution, and the low-power transfer kernel is a normalized probability density function, the sum of the probability mass retained within the upper bound of the effective available power in the corrected distribution and the cutoff probability mass redistributed by the low-power transfer kernel remains 1. Therefore, the renewable energy power distribution corrected by the equipment state constraints can serve as the probability input for the scenario construction module. The equipment state constraint correction module performs the following checks before output: A probability consistency check is performed, which includes nonnegativity verification and normalization verification. When the total probability mass deviates from 1 due to numerical discrepancies, calculation inaccuracies, or input anomalies, it is adjusted according to the total probability mass. Perform normalization correction.

[0109] In the discrete computation method, the equipment state constraint correction module divides the renewable energy power generation range into multiple power grids and represents the theoretical renewable power distribution as a discrete probability vector. The high-power probability cutoff unit sums the probability elements corresponding to power grids above the effective available power upper bound to obtain the cutoff probability mass. The low-power probability transfer unit generates normalized transfer weights corresponding to power grids no higher than the effective available power upper bound based on the state risk coefficient, and allocates the cutoff probability mass to the corresponding power grids according to these transfer weights; for power grids above the effective available power upper bound, the transfer weight is 0. The probability vector obtained from the discrete computation is the discrete form of the renewable energy power generation distribution after equipment state constraint correction.

[0110] The equipment state constraint correction module also handles boundary conditions. When the truncation probability quality is 0, it indicates that there is no probability quality in the theoretically available power distribution that exceeds the upper bound of the effective available power. The equipment state constraint correction module retains the theoretically available power distribution within the upper bound of the effective available power and performs normalization verification. When key data contains missing, out-of-bounds, or timeout indicators, the equipment state constraint correction module determines the available capacity coefficient and state risk coefficient according to preset data processing rules. When the upper bound of the effective available power is lower than the preset minimum available power threshold, the new energy power generation equipment is processed as a low-availability or unavailable state during the corresponding scheduling period.

[0111] The equipment state constraint correction module ultimately outputs the renewable energy power distribution after correcting the equipment state constraints. For a single renewable energy power generation device, this output is: Or the corresponding discrete probability distribution. For multiple new energy power generation devices, the equipment state constraint correction module generates the new energy power distribution after equipment state constraint correction for each new energy power generation device in each scheduling period. The scenario construction module performs independent sampling, related sampling, or summary processing based on the corrected distribution of each new energy power generation device to form a new energy power distribution scenario within the scheduling period, and further constructs representative scenarios and scenario probabilities.

[0112] (3) Scene construction module

[0113] In this embodiment, the scenario construction module receives the renewable energy power distribution after equipment state constraint correction from the equipment state constraint correction module, and combines it with electricity price uncertainty data and load deviation uncertainty data to construct representative scenarios and scenario probabilities. The scenario construction module converts source-side output uncertainty, market price uncertainty, and load deviation uncertainty into scenario data that can be used for calculation by the risk scheduling module within the same scheduling cycle.

[0114] The inputs to the scenario construction module include the renewable energy power distribution after equipment state constraint correction, electricity price uncertainty data, and load deviation uncertainty data. The renewable energy power distribution after equipment state constraint correction is used to characterize the uncertainty of renewable energy power generation equipment in each dispatch period; the electricity price uncertainty data is obtained from market transaction data, historical electricity price data, electricity price forecast results, or transaction rule data; the load deviation uncertainty data is obtained from flexible load data, historical load deviation data, or load forecast error data.

[0115] The joint sampling unit is used to generate original joint scenarios by using the above three types of data as joint sampling objects. Each original joint scenario includes the renewable energy generation capacity sequence, electricity price sequence, and load deviation sequence within the same scheduling period. The joint sampling unit samples according to a unified time period sequence, so that the renewable energy generation capacity, electricity price, and load deviation in the same original joint scenario correspond one-to-one with the scheduling time period. The sampling method can be Monte Carlo sampling, Latin hypercube sampling, quantile sampling, or resampling based on historical samples. If there is a correlation between renewable energy generation capacity, electricity price, and load deviation, the joint sampling unit performs correlation sampling based on historical correlation relationships, correlation coefficient matrices, or joint distribution models; if the correlation is weak or correlation data is lacking, the data is sampled separately and then combined according to the scheduling time period to form the original joint scenario.

[0116] For multiple renewable energy power generation devices, the joint sampling unit can first generate device-level renewable energy power samples based on the renewable energy power distribution after the device state constraints are corrected, and then summarize them into a renewable energy power sequence according to the scheduling period. If the correlation between devices needs to be considered, the joint sampling unit will perform relevant sampling based on the spatial location of the devices, meteorological conditions, historical output correlation, or station grouping relationship; if the correlation between devices is not considered, each device will be sampled independently and then summarized.

[0117] The scenario reduction unit is used to determine the joint scenario distance based on the renewable energy power generation sequence, electricity price sequence, and load deviation sequence, and then reduces the original joint scenario based on the joint scenario distance to obtain a representative scenario. The joint scenario distance is determined based on the normalized renewable energy power generation sequence, electricity price sequence, and load deviation sequence. The normalization process is used to eliminate the dimensional differences between power, electricity price, and load deviation, and the normalization scale can be determined based on the historical sample range, prediction interval range, mean and standard deviation, or a preset scale.

[0118] In one implementation, the joint scenario distance is determined jointly based on the normalized differences in the renewable energy generation capacity sequence, the electricity price sequence, and the load deviation sequence. The differences in each dimension can be calculated using Euclidean distance, Manhattan distance, weighted distance, or other sequence distance methods. The weights can be set based on scheduling experience, the impact of historical deviations, market risk sensitivity, or the requirements of risk scheduling models.

[0119] Scene reduction units can reduce the original joint scene using methods such as clustering, forward selection, backward reduction, nearest-neighbor assignment, or probabilistic distance approximation. Taking clustering as an example, the scene reduction unit divides the original joint scene into several scene clusters according to the distance between the joint scenes, and selects one scene from each scene cluster as a representative scene; the representative scene can be the scene corresponding to the center of the scene cluster, or the original joint scene with the smallest average distance to other scenes in the scene cluster.

[0120] In one implementation, the scenario reduction unit may also retain joint risk scenarios during the reduction process. Joint risk scenarios can be determined based on the threshold relationship between renewable energy generation capacity, electricity price, and load deviation. For example, scenarios where renewable energy generation capacity is lower than a preset first risk threshold, electricity price is higher than a preset second risk threshold, and load deviation is higher than a preset third risk threshold. Risk thresholds can be determined by historical sample quantiles, scheduling safety boundaries, market price ranges, or operational experience. This approach illustrates feasible forms of risk scenario retention and does not limit the scenario reduction unit to a single implementation method.

[0121] The probability determination unit determines the scenario probability corresponding to a representative scenario based on the attribution relationship between the original joint scenarios and the representative scenario. When the original joint scenarios have the same initial probability, the representative scenario probability is determined by the proportion of the number of original joint scenarios belonging to that representative scenario to the total number of original joint scenarios; when the original joint scenarios have different initial probabilities, the representative scenario probability is determined by the sum of the initial probabilities of all original joint scenarios belonging to that representative scenario. The probability determination unit also performs a consistency check on the scenario probabilities, ensuring that the probabilities of each representative scenario are non-negative and the sum of the scenario probabilities is 1; if the probability sum deviates from 1 due to rounding, normalization is performed.

[0122] The representative scenarios output by the scenario construction module include a sequence of renewable energy generation capacity, a sequence of electricity prices, and a sequence of load deviations. The scenario probability represents the probability weight of the corresponding representative scenario. After receiving the representative scenarios and scenario probabilities, the risk scheduling module calls up the renewable energy generation capacity, electricity price, and load deviation data under each representative scenario. It then combines the flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power to calculate the source-load balance deviation, and then performs risk scheduling solutions.

[0123] (4) Load response module

[0124] In this embodiment, the load response module determines the flexible load response amount based on the demand response incentive price and price response characteristics, and sends the converged flexible load response amount to the risk scheduling module. The load response module processes flexible load clusters included in the scheduling scope, with a processing granularity of scheduling time periods, and outputs the flexible load response amount of each flexible load cluster within each scheduling time period. This output is used for subsequent source-load balance deviation calculation and participates in the joint solution of contract application strategy, demand response incentive strategy, and energy storage charging and discharging strategy.

[0125] The response calculation unit is used to calculate flexible load clusters based on demand response incentive prices, price response characteristics, baseline responsiveness, response constraint boundaries, and the average response of the previous round of clusters. During the scheduling period The flexible load response within the system, the number of flexible load clusters is Flexible load cluster number The flexible load response is located within the response constraint boundary;

[0126] The group average update unit is used to update the group average response based on the flexible load response of each flexible load cluster in the same scheduling period, and to feed back the updated group average response to the response calculation unit.

[0127] The convergence discrimination unit is used to determine whether the preset convergence condition is met based on the difference between the average response of two adjacent rounds of the group, and when the preset convergence condition is met, the converged flexible load response is sent to the risk scheduling module.

[0128] The preset convergence condition includes that the difference between the average response of two adjacent rounds is less than or equal to the preset convergence threshold, and the flexible load response after convergence is the response of the scheduling period corresponding to the demand response incentive price.

[0129] (5) Risk scheduling module

[0130] In this embodiment, the risk scheduling module receives representative scenarios and scenario probabilities output by the scenario construction module, and combines them with the converged flexible load response, energy storage operation data, and market transaction data output by the load response module to calculate the source-load balance deviation under each representative scenario. This allows for the solution of contract application strategies, demand response incentive strategies, and energy storage charging and discharging strategies. The risk scheduling module processes the joint source-load-storage decision-making process within the same scheduling cycle, and its output is a scheduling strategy that can be sent to the strategy publishing module.

[0131] The inputs to the risk scheduling module include representative scenarios, scenario probabilities, flexible load response quantities, energy storage charging power boundaries, energy storage discharging power boundaries, energy storage state of charge boundaries, contract declaration execution power boundaries, demand response constraint parameters, and evaluation parameters related to revenue, losses, and constraint penalties. Representative scenarios include renewable energy generation power sequences, electricity price sequences, and load deviation sequences; scenario probabilities represent the probability weight of the corresponding representative scenario in the original set of joint scenarios; flexible load response quantities are determined by the load response module based on demand response incentive prices and price response characteristics; energy storage operation data is used to determine the charging and discharging capabilities and state of charge boundaries of energy storage devices; and market transaction data is used to determine contract declaration execution power, revenue evaluation, and deviation cost-related parameters.

[0132] The joint decision-making setting unit is used to set the contract application strategy, demand response incentive strategy, and energy storage charging and discharging strategy as joint decision-making objects. The contract application strategy includes the contract application execution power sequence for each scheduling period within the scheduling cycle; the demand response incentive strategy includes the demand response incentive price sequence for each scheduling period within the scheduling cycle; and the energy storage charging and discharging strategy includes the energy storage charging power sequence and energy storage discharging power sequence for each scheduling period within the scheduling cycle.

[0133] The source-load balance deviation calculation unit is used to calculate the representative scenario based on the representative scenario, the converged flexible load response, the energy storage charging power, the energy storage discharging power, and the contract declaration execution power, according to the following relationship. Next scheduling period Source load balance deviation:

[0134] ;

[0135] In the formula, Indicates the representative scenario Next scheduling period Source load balance deviation, Indicates the representative scenario Next scheduling period Baseline load, Indicates the representative scenario Next scheduling period Load deviation, Indicates the representative scenario Lower flexible load cluster During the scheduling period The convergent response, Indicates the number of flexible load clusters. This represents a representative scenario generated by the scenario construction module based on the distribution of renewable energy power generation after correction of equipment state constraints. Next scheduling period The renewable energy generation capacity, Indicates the scheduling period The energy storage discharge power, Indicates the scheduling period Energy storage charging power, Indicates the scheduling period Contract declaration execution power;

[0136] In the above source-load balance deviation calculation relationship, the baseline load and load deviation constitute the load-side demand in the representative scenario. The flexible load response is used to represent the load amount reduced or transferred after demand response. The total renewable energy generation capacity, energy storage discharge capacity, and contract declaration execution capacity are included in the supply side, while energy storage charging capacity is included in the load-side consumption. All items are in the power dimension, so the calculation of source-load balance deviation is consistent in dimension.

[0137] The constraint processing unit is used to construct a constraint set for the joint decision-making object based on energy storage operation constraints, demand response constraints, and contract declaration constraints. Energy storage operation constraints include boundary constraints for energy storage charging power, energy storage discharging power, energy storage state of charge (SBC) boundary constraints, recursive SBC constraints, and constraints preventing simultaneous charging and discharging of energy storage devices within the same scheduling period. Demand response constraints include boundary constraints for flexible load response, scheduling period response constraints, and total response constraints. Contract declaration constraints include boundary constraints for contract declaration execution power, scheduling period declaration constraints, and declaration constraints corresponding to market trading rules. These constraint parameters can be derived from energy storage operation data, flexible load data, market trading data, and scheduling rule data.

[0138] In one implementation, the energy storage state of charge (SBC) is updated based on the SBC of the previous scheduling period, the energy storage charging power, energy storage discharging power, charging / discharging efficiency, and scheduling time length of the current scheduling period. The SBC remains between a preset minimum and maximum SBC throughout the entire scheduling cycle. The energy storage charging power and energy storage discharging power do not exceed their respective power boundaries. If the system uses net power control, control rules prevent simultaneous charging and discharging within the same scheduling period; if the system uses mixed-integer optimization, charging / discharging mutual exclusion is constrained by charging / discharging state variables.

[0139] The risk solving unit is used to solve for the joint decision object based on the source-load balance deviation, constraint set, and conditional risk value term. The risk solving unit takes the representative scenario and scenario probability output by the scenario construction module as random input and uses the following stochastic optimization model to solve for the joint decision object:

[0140] ;

[0141] The constraints of the stochastic optimization model include energy storage operation constraints, demand response constraints, contract declaration constraints, and the following conditional value-at-risk auxiliary constraints:

[0142] ;

[0143] In the formula, , , , These represent the contract declaration execution power sequence, demand response incentive price sequence, energy storage charging power sequence, and energy storage discharging power sequence within the scheduling period, respectively. This represents a secondary variable indicating the conditional value at risk. Indicates the representative scenario Conditional value at risk auxiliary slack variables, Represents the set of scheduling periods. This indicates the number of scenes. Indicates the representative scenario The probability of the scenario, and , , Indicates the representative scenario The scene scheduling benefit evaluation value is as follows. Show representative scene The scene loss caused by the source-load balance deviation is as follows. Indicates risk weight. Indicates the constraint penalty weight, Indicates the scheduling period Comprehensive constraint and penalty items;

[0144] Scene scheduling benefit evaluation value Losses in a given scenario can be determined based on contract execution revenue, energy storage operation revenue, demand response costs, and deviation settlement results. The cost is determined by the deviation cost corresponding to the source-load balance deviation, the demand response incentive cost, and the energy storage operation cost. (Comprehensive constraint penalty item) Used to characterize scheduling periods The degree of violation or penalty associated with constraints related to energy storage operation constraints, demand response constraints, and contract declaration constraints. In this embodiment, , and By converting the preset evaluation coefficients into the same evaluation dimension, the benefit terms, risk terms, and constraint penalty terms in the stochastic optimization model can be evaluated in a unified manner.

[0145] The conditional value at risk (VaR) is used to characterize the risk level of a scenario's loss within the high-loss range. In the model above, the VaR is expressed through auxiliary variables. and Represented as:

[0146] ;

[0147] In the formula, This indicates the confidence level of the conditional value of risk, and Risk weighting Used to adjust the influence of risk terms on the objective function, constraining penalty weights. Used to adjust the influence of the comprehensive constraint penalty term on the objective function.

[0148] The risk solution unit can use linear programming, quadratic programming, mixed integer programming, convex optimization algorithms, or heuristic optimization algorithms to solve the above stochastic optimization model. The specific solution method is determined based on the form of the revenue evaluation function, the scenario loss function, the energy storage operation constraints, and the demand response constraints. After the solution is completed, the risk solution unit outputs the contract application strategy, the demand response incentive strategy, and the energy storage charging and discharging strategy.

[0149] The risk scheduling module has a clear data connection with the aforementioned modules. The representative scenarios and scenario probabilities output by the scenario construction module serve as scenario inputs for the stochastic optimization model; the converged flexible load response output by the load response module participates in the source-load balance deviation calculation; the renewable energy power distribution corrected for equipment state constraints is transformed into a renewable energy power sequence in the representative scenario through the scenario construction module; the data acquisition module provides energy storage operation data, market transaction data, and scheduling constraint-related data. The contract application strategy, demand response incentive strategy, and energy storage charging and discharging strategy output by the risk scheduling module are sent to the strategy publishing module.

[0150] (6) Strategy Deployment Module

[0151] In this embodiment, the strategy publishing module receives the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy output by the risk scheduling module, and generates a scheduling strategy data packet. The scheduling strategy data packet is sent to the transaction declaration system, demand response execution system, and energy storage management system for the corresponding execution systems to execute the strategy; the execution deviation data returned by each execution system is sent to the data acquisition module as data input for subsequent scheduling cycles.

[0152] The strategy data generation unit generates a scheduling strategy data package based on the contract application strategy, demand response incentive strategy, and energy storage charging / discharging strategy obtained by the risk solving unit. The contract application strategy includes the contract application execution power for each scheduling period within the scheduling cycle; the demand response incentive strategy includes the demand response incentive price for each scheduling period within the scheduling cycle; and the energy storage charging / discharging strategy includes the energy storage charging power and energy storage discharging power for each scheduling period within the scheduling cycle. The scheduling strategy data package is organized according to the scheduling cycle and scheduling periods, ensuring a correspondence between the contract application execution power, demand response incentive price, energy storage charging power, and energy storage discharging power within the same scheduling period.

[0153] The scheduling strategy data package includes a scheduling cycle identifier, a scheduling period identifier, contract application execution power, demand response incentive price, energy storage charging power, and energy storage discharging power. The scheduling cycle identifier indicates the applicable scheduling cycle, the scheduling period identifier indicates the execution period within the scheduling cycle, the contract application execution power is used by the trading application system to execute contract applications or contracts for the corresponding scheduling period, the demand response incentive price is used by the demand response execution system to issue response incentives, and the energy storage charging power and energy storage discharging power are used by the energy storage management system to control the charging and discharging behavior of energy storage devices.

[0154] In one implementation, the strategy data generation unit performs a consistency check on the scheduling strategy output by the risk scheduling module before generating the scheduling strategy data packet. The consistency check includes scheduling period integrity check, power boundary check, energy storage charging / discharging mutual exclusion check, and data format check. After passing the consistency check, the strategy data generation unit generates the scheduling strategy data packet.

[0155] The strategy distribution unit is used to send scheduling strategy data packets to the transaction application system, the demand response execution system, and the energy storage management system. Specifically, the strategy distribution unit sends the contract application execution power and the corresponding scheduling cycle identifier and scheduling period identifier to the transaction application system; it sends the demand response incentive price and the corresponding scheduling cycle identifier and scheduling period identifier to the demand response execution system; and it sends the energy storage charging power, energy storage discharging power, and the corresponding scheduling cycle identifier and scheduling period identifier to the energy storage management system. The strategy distribution unit can send scheduling strategy data packets through application programming interfaces (APIs), message queues, the scheduling master station communication interface, or file exchange.

[0156] The execution feedback unit receives execution deviation data from the transaction declaration system, demand response execution system, and energy storage management system. Execution deviation data includes one or more of the following: contract declaration execution power deviation, actual flexible load response, actual energy storage charging power, actual energy storage discharging power, and energy storage state of charge change. The execution feedback unit organizes the execution deviation data according to the scheduling cycle identifier and scheduling period identifier, and sends the execution deviation data to the data acquisition module for data updates in subsequent scheduling cycles.

[0157] Based on the above system, such as Figure 2 As shown, this embodiment also provides a method for coordinated prediction and scheduling of new energy sources, loads, and storage. The method executes sequentially as follows:

[0158] S1. Collect new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data and energy storage operation data within the scheduling cycle, and form a scheduling input dataset;

[0159] S2. Determine the upper bound of theoretically available power and the distribution of theoretically available power based on the new energy prediction data in the scheduling input dataset. Determine the equipment status constraint vector including the available capacity coefficient and the status risk coefficient based on the equipment status data in the scheduling input dataset. Determine the upper bound of effective available power based on the equipment status constraint vector and the capacity mapping data.

[0160] S3. Based on the upper bound of the effective available power, the high-power probabilistic mass in the theoretical power distribution that exceeds the upper bound of the effective available power is truncated, and the truncated probabilistic mass is transferred to a power range not higher than the upper bound of the effective available power according to the low-power transfer kernel determined by the state risk coefficient, so as to obtain the new energy power distribution after the equipment state constraint correction.

[0161] S4. Based on the distribution of renewable energy generation capacity after equipment state constraint correction, electricity price uncertainty data and load deviation uncertainty data, construct scenarios to obtain representative scenarios and scenario probabilities.

[0162] S5. Determine the flexible load response amount based on the demand response incentive price and price response characteristics;

[0163] S6. Based on representative scenarios, scenario probabilities, flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power, determine the source-load balance deviation, and combine energy storage operation constraints, demand response constraints, contract declaration constraints, and conditional risk value terms to solve the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy.

[0164] S7. Generate a scheduling strategy data packet based on the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy, and send the scheduling strategy data packet.

[0165] This embodiment focuses on the collaborative prediction and scheduling process of new energy sources, loads, and storage, forming a continuous processing chain from multi-source data acquisition, equipment state constraint correction, scenario construction, load response, risk scheduling to strategy release. Compared with methods that rely solely on conventional new energy prediction results for scenario construction and scheduling solutions, this embodiment introduces equipment state data and capacity mapping data from new energy power generation equipment into the new energy renewable power distribution correction process. The renewable power distribution corrected for equipment state constraints is then used to represent scenario construction and source-load balance deviation calculation, ensuring a unified data continuity between equipment operating status, source-side uncertainties, flexible load response, energy storage operation, and risk scheduling within the same scheduling cycle. Those skilled in the art can implement the system and method described in this invention based on the processing procedures, parameter sources, formula relationships, and module connections disclosed in this embodiment.

[0166] In summary, this invention provides a complete implementation framework for the collaborative prediction and scheduling of new energy sources, loads, and energy storage, encompassing multi-source data access, equipment status constraint processing, scenario construction, load response, risk scheduling, and strategy deployment. Each stage has a clear data transfer relationship and processing order, providing a unified technical implementation path for the collaborative scheduling of new energy power generation equipment, flexible load clusters, and energy storage equipment within the same scheduling cycle. This system and method can be configured according to different scheduling cycles, equipment types, load organization forms, and market operating conditions, making it suitable for application scenarios such as new energy power plants, virtual power plants, integrated energy systems, and combined source-load-storage operations.

[0167] In the description of this specification, the terms "an embodiment," "an implementation," etc., indicate that a specific feature, structure, parameter, or process described in connection with that embodiment or implementation may be included in at least one embodiment of the present invention. Unless there is an obvious conflict between different implementations, those skilled in the art can reasonably combine the technical features of different implementations in this specification.

[0168] The present invention has been described above with reference to specific embodiments, but these embodiments are not intended to limit the scope of protection of the present invention. Without departing from the technical concept of the present invention, those skilled in the art can adaptively adjust the data processing method, parameter setting method, scenario construction method, response function form, optimization solution method, and strategy distribution method in the embodiments according to the actual scheduling cycle, equipment type, data source, market operation rules, and execution system interface. All such adjustments, without changing the data connection relationship between data acquisition, equipment state constraint correction, scenario construction, load response, risk scheduling, and strategy issuance described in the present invention, should fall within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A new energy source-load-storage coordinated prediction and scheduling system, used for coordinated prediction and scheduling of new energy power generation equipment, flexible load clusters and energy storage equipment included in the scheduling scope, characterized in that, include: The data acquisition module is used to acquire new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data, and energy storage operation data within the scheduling cycle, and form a scheduling input dataset; The equipment state constraint correction module is used to determine the upper bound of theoretical available power and the distribution of theoretical available power based on new energy prediction data, determine the equipment state constraint vector including available capacity coefficient and state risk coefficient based on equipment state data, determine the upper bound of effective available power based on the equipment state constraint vector and capacity mapping data, truncate the high-power side probabilistic mass that exceeds the upper bound of effective available power, and transfer the truncated probabilistic mass to a power range not higher than the upper bound of effective available power according to the low-power transfer kernel determined by the state risk coefficient, so as to obtain the distribution of new energy available power after equipment state constraint correction. The scenario construction module is used to construct representative scenarios and scenario probabilities based on the distribution of renewable energy power generation capacity after equipment state constraint correction, electricity price uncertainty data, and load deviation uncertainty data. The load response module is used to determine the flexible load response amount based on the demand response incentive price and price response characteristics. The risk scheduling module is used to determine the source-load balance deviation based on representative scenarios, scenario probabilities, flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power. It also combines energy storage operation constraints, demand response constraints, contract declaration constraints, and conditional risk value terms to solve the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy. The policy publishing module is used to generate and send scheduling policy data packets.

2. The new energy source-load-storage coordinated prediction and scheduling system according to claim 1, characterized in that, The data acquisition module includes: The scheduling time reference unit is used to generate a unified time period sequence based on the scheduling cycle and scheduling time granularity, and to map new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data and energy storage operation data to the unified time period sequence; The equipment data association unit is used to establish a corresponding relationship between the new energy forecast data, equipment status data and capacity mapping data of the same new energy power generation equipment under the same scheduling period, with the new energy power generation equipment as the association object. The data quality discrimination unit is used to perform timeliness verification, boundary consistency verification, and time sequence continuity verification on the data after the completion of time period mapping and device data association, and generate data quality identifiers; The scheduling input generation unit is used to generate a scheduling input dataset based on a unified time period sequence, correspondence, and data quality identifier, and then send the scheduling input dataset to the equipment status constraint correction module, scenario construction module, load response module, and risk scheduling module.

3. The new energy source-load-storage coordinated prediction and scheduling system according to claim 2, characterized in that, The device state constraint correction module includes: The state vector construction unit is used to construct the equipment state constraint vector based on the equipment state data and data quality identifier in the scheduling input dataset. The equipment state constraint vector includes the available capacity coefficient and state risk coefficient of the new energy power generation equipment during the scheduling period. The available capacity coefficient is a dimensionless coefficient in the interval [0,1], used to reflect the available capacity ratio of new energy power generation equipment in the corresponding scheduling period; the state risk coefficient is a dimensionless coefficient in the interval [0,1], used to reflect the output uncertainty corresponding to the state decay of new energy power generation equipment. The effective upper bound determination unit is used to determine the theoretical upper bound of the generated power based on new energy forecast data, the upper bound of the capacity-mapped power based on capacity mapping data, and the effective upper bound of the available power based on the available capacity coefficient, the theoretical upper bound of the generated power, and the upper bound of the capacity-mapped power. ； In the formula, Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates new energy power generation equipment During the scheduling period Available capacity factor, Indicates new energy power generation equipment During the scheduling period The theoretical upper limit of the power output, Indicates new energy power generation equipment During the scheduling period The capacity is mapped to the upper limit of power.

4. The new energy source-load-storage coordinated prediction and scheduling system according to claim 3, characterized in that, The device state constraint correction module also includes: The high-power probability truncation unit is used to truncate the probability mass of the theoretically available power distribution that exceeds the effective available power upper bound, based on the upper bound of the effective available power. The theoretically available power distribution uses the non-negative renewable energy available power as the supporting variable, and the probability density function of the theoretically available power distribution is: The cutoff probability mass satisfies the following formula: ； In the formula, Indicates new energy power generation equipment During the scheduling period The high-power side probability mass that was intercepted Indicates new energy power generation equipment During the scheduling period The upper limit of effective available power, Indicates the power output of new energy sources; The low-power probability transfer unit is used to determine the low-power transfer kernel based on the state risk coefficient, and to transfer the truncated probability mass to a power range no higher than the upper bound of the effective available power, thus obtaining the probability density function of the renewable energy power distribution after equipment state constraint correction: ； In the formula, This represents the probability density function of the renewable energy power distribution after equipment state constraint correction. Represents the characteristic function, Indicates a low-power transfer core. Indicates new energy power generation equipment During the scheduling period State risk coefficient; Low-power transfer core To be based on state risk system The normalized probability density function of the parameters, low-power transfer kernel The support range is Outside the support interval, the value is 0, and the following formula is satisfied: ； The probability density function of the renewable energy power distribution after equipment state constraint correction Send to the scenario building module to generate renewable energy power scenarios.

5. A new energy source-load-storage coordinated prediction and scheduling system according to claim 4, characterized in that, The scene construction module includes: The joint sampling unit is used to generate an original joint scenario by taking the distribution of renewable energy power generation capacity after equipment state constraint correction, electricity price uncertainty data and load deviation uncertainty data as joint sampling objects; the original joint scenario includes the renewable energy power generation capacity sequence, electricity price sequence and load deviation sequence within the same scheduling cycle; The scenario reduction unit is used to determine the joint scenario distance based on the renewable energy power generation sequence, electricity price sequence, and load deviation sequence, and to reduce the original joint scenario based on the joint scenario distance to obtain a representative scenario; the joint scenario distance is determined based on the normalized renewable energy power generation sequence, electricity price sequence, and load deviation sequence. The probability determination unit is used to determine the scenario probability corresponding to the representative scenario based on the attribution relationship between the original joint scenario and the representative scenario, and send the representative scenario and scenario probability to the risk scheduling module.

6. A new energy source-load-storage coordinated prediction and scheduling system according to claim 1, characterized in that, The load response module includes: The response calculation unit is used to calculate flexible load clusters based on demand response incentive prices, price response characteristics, baseline responsiveness, response constraint boundaries, and the average response of the previous round of clusters. During the scheduling period The flexible load response within the system, the number of flexible load clusters is Flexible load cluster number The flexible load response is located within the response constraint boundary; The group average update unit is used to update the group average response based on the flexible load response of each flexible load cluster in the same scheduling period, and to feed back the updated group average response to the response calculation unit. The convergence discrimination unit is used to determine whether the preset convergence condition is met based on the difference between the average response of two adjacent rounds of the group, and when the preset convergence condition is met, the converged flexible load response is sent to the risk scheduling module. The preset convergence condition includes that the difference between the average response of two adjacent rounds is less than or equal to the preset convergence threshold, and the flexible load response after convergence is the response of the scheduling period corresponding to the demand response incentive price.

7. A new energy source-load-storage coordinated prediction and scheduling system according to claim 6, characterized in that, The risk scheduling module includes: The joint decision-making setting unit is used to set contract application strategy, demand response incentive strategy and energy storage charging and discharging strategy as joint decision-making objects; The source-load balance deviation calculation unit is used to calculate the representative scenario based on the representative scenario, the converged flexible load response, the energy storage charging power, the energy storage discharging power, and the contract declaration execution power, according to the following relationship. Next scheduling period Source load balance deviation: ； In the formula, Indicates the representative scenario Next scheduling period Source load balance deviation, Indicates the representative scenario Next scheduling period Baseline load, Indicates the representative scenario Next scheduling period Load deviation, Indicates the representative scenario Lower flexible load cluster During the scheduling period The convergent response, Indicates the number of flexible load clusters. This represents a representative scenario generated by the scenario construction module based on the distribution of renewable energy power generation after correction of equipment state constraints. Next scheduling period The renewable energy generation capacity, Indicates the scheduling period The energy storage discharge power, Indicates the scheduling period Energy storage charging power, Indicates the scheduling period Contract declaration execution power; The constraint processing unit is used to construct a constraint set for the joint decision-making object based on energy storage operation constraints, demand response constraints, and contract declaration constraints. The constraint set includes boundary constraints for energy storage charging and discharging power, energy storage state of charge constraints, demand response quantity boundary constraints, and contract declaration execution power boundary constraints. The risk solution unit is used to solve the joint decision object based on the source load balance deviation, constraint set, and conditional risk value term.

8. A new energy source-load-storage coordinated prediction and scheduling system according to claim 7, characterized in that, The risk solving unit takes the representative scenarios and scenario probabilities output by the scenario construction module as random inputs, and uses the following stochastic optimization model to solve the joint decision object: ； The constraints of the stochastic optimization model include energy storage operation constraints, demand response constraints, contract declaration constraints, and the following conditional value-at-risk auxiliary constraints: ； In the formula, , , , These represent the contract declaration execution power sequence, demand response incentive price sequence, energy storage charging power sequence, and energy storage discharging power sequence within the scheduling period, respectively. This represents a secondary variable indicating the conditional value at risk. Indicates the representative scenario Conditional value at risk auxiliary slack variables, Represents the set of scheduling periods. This indicates the number of scenes. Indicates the representative scenario The probability of the scenario, and , , Indicates the representative scenario The scene scheduling benefit evaluation value is as follows. Show representative scene The scene loss caused by the source-load balance deviation is as follows. Indicates risk weight. Indicates the constraint penalty weight, Indicates the scheduling period Comprehensive constraint and penalty items; Scene loss The source-load balance deviation is determined by the deviation cost, demand response incentive cost, and energy storage operation cost. The source-load balance deviation is calculated based on a representative scenario generated from the distribution of renewable energy power generation capacity after equipment state constraint correction. , and The evaluation coefficients are converted to the same evaluation dimension.

9. A new energy source-load-storage coordinated prediction and scheduling system according to claim 1, characterized in that, The policy publishing module includes: The strategy data generation unit is used to generate a scheduling strategy data package based on the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy obtained by the risk solving unit. The scheduling strategy data package includes a scheduling cycle identifier, a scheduling period identifier, contract declaration execution power, demand response incentive price, energy storage charging power, and energy storage discharging power. The strategy distribution unit is used to send scheduling strategy data packets to the transaction reporting system, the demand response execution system, and the energy storage management system. The execution feedback unit is used to receive execution deviation data returned by the transaction reporting system, the demand response execution system, and the energy storage energy management system, and send the execution deviation data to the data acquisition module.

10. A method for coordinated prediction and scheduling of new energy sources, loads, and storage, applied to a coordinated prediction and scheduling system for new energy sources, loads, and storage as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. Collect new energy forecast data, equipment status data, capacity mapping data, flexible load data, market transaction data and energy storage operation data within the scheduling cycle, and form a scheduling input dataset; S2. Determine the upper bound of theoretically available power and the distribution of theoretically available power based on the new energy prediction data in the scheduling input dataset. Determine the equipment status constraint vector including the available capacity coefficient and the status risk coefficient based on the equipment status data in the scheduling input dataset. Determine the upper bound of effective available power based on the equipment status constraint vector and the capacity mapping data. S3. Based on the upper bound of the effective available power, the high-power probabilistic mass in the theoretical power distribution that exceeds the upper bound of the effective available power is truncated, and the truncated probabilistic mass is transferred to a power range not higher than the upper bound of the effective available power according to the low-power transfer kernel determined by the state risk coefficient, so as to obtain the new energy power distribution after the equipment state constraint correction. S4. Based on the distribution of renewable energy generation capacity after equipment state constraint correction, electricity price uncertainty data and load deviation uncertainty data, construct scenarios to obtain representative scenarios and scenario probabilities. S5. Determine the flexible load response amount based on the demand response incentive price and price response characteristics; S6. Based on representative scenarios, scenario probabilities, flexible load response, energy storage charging power, energy storage discharging power, and contract declaration execution power, determine the source-load balance deviation, and combine energy storage operation constraints, demand response constraints, contract declaration constraints, and conditional risk value terms to solve the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy. S7. Generate a scheduling strategy data packet based on the contract declaration strategy, demand response incentive strategy, and energy storage charging and discharging strategy, and send the scheduling strategy data packet.

Images