Intelligent Grid Multi-Time Scale Resource Scheduling Optimization Method and System

Abstract

The present invention discloses an intelligent power grid multi-time scale resource scheduling optimization method and system, which relates to the technical field of multi-time scale resource scheduling. The method includes the following steps: obtaining a first constraint condition for the rolling update compliance rate and determining an initial rolling update compliance rate; obtaining a second constraint condition for the scheduling margin configuration ratio and quantifying the initial scheduling margin configuration ratio; performing multi-sample simulation on the initial parameters to generate a number of to-be-adjusted configuration arrays; mapping the to-be-adjusted configuration arrays into multi-time scale resource scheduling through scheduling arrangement and optimization drive, and performing collaborative evaluation; fitting a high-order collaborative response display surface based on the evaluation results and the to-be-adjusted configuration arrays; selecting a scheduling sensitive area of the surface according to actual requirements, screening the best configuration array, and driving multi-time scale resource scheduling. The present application solves the problem of unbalanced coordination in multi-time scale scheduling by analyzing, screening and regulating the scheduling margin configuration ratio and the rolling update compliance rate.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-time scale resource scheduling, and more specifically, to an intelligent power grid multi-time scale resource scheduling optimization method and system. Background Art

[0002] The resource scheduling optimization technology of the intelligent power grid aims to solve the problems of renewable energy output fluctuation, load demand uncertainty, and multi-agent collaborative scheduling, and improve the economy, reliability, and energy utilization efficiency of power grid operation. The current technical solutions mainly adopt a multi-time scale optimization strategy, combined with short-term (second-hour), medium-term (hour-day), and long-term (day-month) scheduling frameworks, and construct a distributed energy generation prediction model through deep learning algorithms to improve the power prediction accuracy. In addition, adjustable resources such as energy storage systems, demand-side response, and electric vehicles are introduced, and algorithms such as reinforcement learning and multi-objective optimization are used for dynamic scheduling decisions to achieve flexible and optimal allocation of resources and improve the overall stability and sustainability of the power grid.

[0003] The existing intelligent power grid multi-time scale resource scheduling process is mainly divided into day-ahead scheduling, intra-day scheduling, and real-time scheduling, corresponding to long-term, medium-term, and short-term optimizations respectively, to ensure the economy, stability, and new energy consumption capacity of the power grid. In the day-ahead scheduling stage, based on load prediction and renewable energy output prediction, an optimization algorithm (such as mixed integer linear programming, reinforcement learning) is used to formulate a unit commitment plan and optimize the charge and discharge strategies of the energy storage system. In the intra-day scheduling stage, combined with the latest meteorological and market data, through short-term load prediction, economic dispatch, and demand-side response technologies, the unit output and reserve capacity are optimized to cope with new energy fluctuations and load changes. In the real-time scheduling stage, relying on the SCADA system and synchronized phasor measurement (PMU) for power grid state monitoring, and through automatic generation control (AGC), event-driven scheduling, and power electronic device regulation, millisecond-level response is achieved to maintain the stability of the power grid frequency and voltage and ensure safe and efficient operation.

[0004] For example, a method for rapid tracking of power sources and loads in a distribution network based on intra-day and real-time rolling control announced in the invention patent with publication number CN109765787B. This invention relates to distribution network control technology, specifically to a method for rapid tracking of power sources and loads in a distribution network based on intra-day and real-time rolling control. A static load model with time-varying parameters is established for the loads in the distribution network, and the parameters of the load model are identified online using the least squares method with constraints. The state space equation of the load is derived using the established load model. With the lowest scheduling cost, global optimization within a 15-minute time scale is performed for intra-day operation to obtain the economically optimal power at the grid connection point. Based on the established load model and the models of photovoltaic and energy storage output, the state space equation of the distribution network system is established. Multivariable generalized predictive control is used to control the output on both the power source and load sides. Using the economically optimal power at the grid connection point as the reference sequence, the control command for the next moment is obtained, forming a rolling optimization process. The results after control are evaluated using volatility indicators and economic indicators. This method can improve the control accuracy and suppress the influence of power source and load fluctuations.

[0005] For example, a load control method based on model prediction and multi-time scale robustness announced in the invention patent with publication number CN119024705A. This invention discloses a load control method based on model prediction and multi-time scale robustness, belonging to the technical field of load control. This method performs load prediction on each load device in the system to be controlled based on the multi-time scale load prediction model stored in the database to obtain the total load prediction result. A system dynamic model of the system to be controlled is established. Based on the system dynamic model and the load prediction result, a cost function is minimized to determine the load control input, which is the amount of load adjustment required for the system to be controlled. This invention can accurately adjust the load demand in the system, thereby optimizing energy distribution, reducing energy waste, and coping with the uncertainty and dynamic changes of energy demand, ensuring the economy and reliability of system operation. It can effectively solve the problem of load fluctuations in energy management and improve energy utilization efficiency and system stability.

[0006] In the above disclosed technical solutions, there are at least the following technical problems: In a smart grid, short-term scheduling (such as minute-level), medium-term scheduling (such as intra-day), and long-term scheduling (such as monthly / annual) are often driven by different mechanisms and models. Due to different focus targets, data accuracies, and response speeds at each scale, problems such as scheduling strategy conflicts, resource allocation fragmentation, and inconsistent feedback are likely to occur, resulting in a decline in the overall performance of the system.

[0007] There is a lack of effective coordination and feedback mechanisms between different time scales. Short-term scheduling is difficult to fully utilize the expected trends of long-term planning, and long-term planning is also difficult to quickly respond to short-term disturbances; especially in scenarios where multiple adjustable resources (such as renewable energy, energy storage, and load-side response) coexist.

[0008] When the rolling update compliance rate is relatively high, it can improve the response ability to short-term load changes, but it is easy to frequently disturb the medium- and long-term plans, increasing system instability and adjustment costs;

[0009] When the compliance rate is relatively low, the coordination burden is reduced, but the short-term scheduling deviation may accumulate due to untimely adjustment, affecting the immediacy and accuracy of system operation;

[0010] When the scheduling margin configuration ratio is set relatively high, it can enhance the system's bearing capacity for prediction errors and emergencies, but it will reduce resource utilization rate and bring higher operation costs;

[0011] When the configuration ratio is relatively low, the resource use efficiency and scheduling economy are improved, but the system's robustness in the face of uncertainty and load fluctuations will decline.

[0012] The scheduling margin configuration ratio and the rolling update compliance rate jointly affect the dynamic response ability and system robustness of multi-time scale scheduling; however, the existing technologies lack a dynamic optimization and adaptive adjustment mechanism between the rolling update compliance rate and the scheduling margin configuration ratio, resulting in the problem of unbalanced coordination of multi-time scale scheduling.

[0013] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0014] In order to overcome the above defects of the prior art, embodiments of the present invention provide an intelligent grid multi-time scale resource scheduling optimization method and system, which analyze, screen and regulate the scheduling margin configuration ratio and the rolling update compliance rate to solve the problem of unbalanced coordination of multi-time scale scheduling.

[0015] To achieve the above object, the present invention provides the following technical solutions:

[0016] An intelligent grid multi-time scale resource scheduling optimization method, comprising the following steps: obtaining a first constraint condition of the rolling update compliance rate to determine an initial rolling update compliance rate; obtaining a second constraint condition of the scheduling margin configuration ratio to quantify the initial scheduling margin configuration ratio; performing multi-sample simulation on the initial parameters to generate a number of to-be-adjusted configuration arrays; mapping the to-be-adjusted configuration arrays into multi-time scale resource scheduling through scheduling arrangement and optimization drive, and performing collaborative evaluation; fitting a high-order collaborative response display surface based on the evaluation results and the to-be-adjusted configuration arrays; selecting a scheduling sensitive area of the surface according to actual requirements, screening the best configuration array, and driving multi-time scale resource scheduling.

[0017] In a preferred embodiment, the first constraint condition for obtaining the rolling update compliance rate and determining the initial rolling update compliance rate is as follows: The first constraint condition includes the maximum offset frequency of the scheduling plan rolling adjustment and the response coupling ability of each time-scale scheduling model; obtain the planned offset frequency and amplitude of the multi-time-scale model within the period, and construct a scheduling offset sample set; combine the response coupling ability of the time-scale scheduling model, and analyze the impact of the response delay of each scheduling model after the rolling adjustment trigger on the scheduling consistency; according to the maximum acceptable offset frequency, evaluate the coordination stability of the system at different frequencies; obtain the historical rolling update compliance rate data set and its stability and consistency loss functions, and screen the compliance rate in combination with the first constraint condition; based on the minimum deviation matching from the screened historical rolling update compliance rates, select the rolling update compliance rate with the smallest deviation from the current system as the initial rolling update compliance rate.

[0018] In a preferred embodiment, the second constraint condition for obtaining the scheduling margin configuration ratio and quantifying the initial scheduling margin configuration ratio is as follows: The second constraint condition includes the load prediction error distribution, the fluctuation range of renewable energy output, and the minimum guarantee ability of scheduling resources; the scheduling margin configuration ratio is the ratio of the resource redundancy rate in the typical operation scenario to the accident response demand, and the accident response demand is obtained by analyzing the sudden increase in load and the fluctuation of renewable energy; the second constraint condition is used to constrain the scheduling margin configuration ratios of several multi-scale scheduling systems within the historical preset period, and the average value is obtained as the initial scheduling margin configuration ratio.

[0019] In a preferred embodiment, the initial parameters are simulated with multiple samples to generate several arrays of configurations to be adjusted, which is specifically as follows: The initial parameters include the initial rolling update compliance rate and the initial scheduling margin configuration ratio; based on the first constraint condition and the second constraint condition, determine the value ranges of the rolling update compliance rate and the scheduling margin configuration ratio; based on Latin hypercube sampling, generate several arrays of configurations to be adjusted within the value ranges, and the configuration arrays contain different combinations of the rolling update compliance rate and the scheduling margin configuration ratio.

[0020] In a preferred embodiment, the collaborative evaluation is specifically as follows: The collaborative evaluation obtains the characteristic data for evaluating the adaptability of the configuration data by analyzing the characteristics of the multi-time-scale scheduling system; and trains according to the obtained characteristic data for evaluating the adaptability of the configuration data in combination with the historical data to obtain a driving effect evaluation model, and the characteristic data includes the cross-time-scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate.

[0021] In a preferred embodiment, fitting a high-order collaborative response display surface based on the evaluation result and the configuration array to be adjusted specifically includes: constructing a multi-dimensional feature mapping space for characterizing the response law of a multi-time-scale scheduling system based on the cross-time-scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization; the dimensions of the multi-dimensional feature mapping space include cross-time-scale scheduling fluctuation characteristics, dynamic matching characteristics of resource utilization, and drive effect evaluation values; associating and mapping the characteristic parameters of each group of configuration arrays to be adjusted with their corresponding drive effect evaluation values to form a set of mapping sample points; based on the mapping sample points, fitting and constructing a high-order collaborative response surface covering the entire feature mapping domain to reflect the change trend of the scheduling collaborative performance of the system under different configuration combinations.

[0022] In a preferred embodiment, selecting the scheduling sensitive area of the surface according to actual needs, screening the best configuration array, and driving multi-time-scale resource scheduling specifically includes: determining several peak areas of the high-order collaborative response display surface, and expanding the range based on several peaks according to a preset dynamic regulation amplitude to construct several candidate areas for adjusting the configuration array; screening several candidate areas according to the actual needs of the scheduling target; if the requirement for short-term response ability is relatively high, selecting the area with a higher rolling update compliance rate in the candidate areas as the scheduling sensitive area; if the system stability requirement is stronger, selecting the area with a higher scheduling margin configuration in the candidate areas as the scheduling sensitive area; if the resource usage efficiency is the current priority target, selecting the area with the highest peak in the candidate areas as the scheduling sensitive area; using the configuration array corresponding to the peak in the scheduling sensitive area as the best configuration array to drive multi-time-scale resource scheduling.

[0023] In a preferred embodiment, the specific method for obtaining the cross-time-scale scheduling fluctuation characteristics is as follows: at different scheduling time scales, obtaining the processing time from request to execution completion as the response time; for the same time scale, calculating the average value of several response times; obtaining the resource scheduling amount and total duration at different time scales, and calculating the scheduling efficiency; setting an initial cross-time-scale fluctuation model according to the average response time and scheduling efficiency; setting response time points for performance testing and collecting the scheduling fluctuation amplitude data at each point; based on the fluctuation amplitude data, optimizing the parameters of the initial fluctuation model using the non-linear least squares method to obtain a scheduling fluctuation characteristic calculation model, and the parameters of the initial fluctuation model include the maximum fluctuation amplitude, change speed, and sensitivity characteristics.

[0024] In a preferred embodiment, the specific method for obtaining the dynamic matching feature of resource utilization rate is as follows: Obtain the resource utilization rate changes and scheduling fluctuations under different scheduling strategies and resource allocations, and construct a matching evaluation model through a scheduling algorithm; perform clustering processing on the matching evaluation model; based on the extracted features of the evaluation model after clustering, calculate the matching accuracy evaluation features based on resource fluctuations, scheduling errors, and contentions; simulate the resource utilization rate under different scheduling strategies through the evaluation model, establish a timeliness model, and obtain the resource allocation timeliness evaluation features, and combine the matching accuracy evaluation features to sum to obtain the dynamic matching feature of resource utilization rate.

[0025] The intelligent power grid multi-time scale resource scheduling optimization system includes an initial parameter acquisition module, a configuration array acquisition module, a drive evaluation module, and a configuration array adjustment module; the initial parameter acquisition module is used to obtain the first constraint condition of the rolling update compliance rate and determine the initial rolling update compliance rate; obtain the second constraint condition of the scheduling margin configuration ratio and quantify the initial scheduling margin configuration ratio; the configuration array acquisition module is used to perform multi-sample simulation on the initial parameters to generate several configuration arrays to be adjusted; the drive evaluation module is used to map the configuration arrays to be adjusted into the multi-time scale resource scheduling through scheduling arrangement and optimization drive, and perform collaborative evaluation; the configuration array adjustment module is used to fit a high-order collaborative response display surface based on the evaluation results and the configuration arrays to be adjusted; select the scheduling sensitive area of the surface according to actual needs, screen the best configuration array, and drive the multi-time scale resource scheduling.

[0026] The technical effects and advantages of the intelligent power grid multi-time scale resource scheduling optimization method of the present invention:

[0027] 1. By constructing a refined initial calculation mechanism for the rolling update compliance rate and combining historical scheduling offset data with multi-level scheduling response coupling characteristics, the present invention can effectively identify the rolling adjustment strategy with the strongest current system adaptability, thereby improving scheduling consistency and system stability. At the same time, by combining factors such as load prediction error, renewable energy fluctuations, and minimum guarantee capacity, a quantification model for the scheduling margin configuration ratio is established, realizing the dynamic balance between scheduling resource redundancy and risk response. Using Latin hypercube sampling to generate diverse parameter configurations further enhances the simulation coverage of the scheme and the integrity of the optimization space. The overall scheme has both real-time adaptability and interpretability, improving the coordinated control ability of the multi-time scale scheduling system in complex operating environments, and providing effective support for building a new power system with strong resilience, fast response, and reasonable resource allocation.

[0028] 2. By mapping the configuration array to be adjusted to a multi-time-scale resource scheduling model, the present invention combines the scheduling objectives and characteristics of short cycles, medium cycles, and long cycles, and conducts refined management from the perspectives of real-time performance, resource buffering, and global balance respectively, significantly enhancing the coordination and flexibility of the scheduling system at different time scales. On this basis, the system introduces a collaborative evaluation mechanism, and quantitatively analyzes the impact of each configuration array on the system scheduling behavior and resource efficiency around two key indicators: the cross-time-scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate, effectively solving the problems of inconsistent inter-period scheduling and unbalanced resource allocation existing in traditional scheduling systems. On the one hand, the introduction of cross-time-scale scheduling fluctuation characteristics helps to accurately capture the interference degree of the configuration array in short-term decisions on long-term scheduling objectives, and realizes the stability optimization of scheduling behavior; on the other hand, the dynamic matching characteristics of resource utilization rate focus on the supply-demand adaptation situation in the resource allocation process, improving the response ability and utilization efficiency of the scheduling system to resource state changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a schematic flow chart of the intelligent power grid multi-time-scale resource scheduling optimization method of the present invention;

[0030] Figure 2 is a schematic structural diagram of the intelligent power grid multi-time-scale resource scheduling optimization system of the present invention;

[0031] Figure 3 is a display diagram of the multi-dimensional feature mapping space based on the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0033] Embodiment 1 Figure 1 provides the intelligent power grid multi-time-scale resource scheduling optimization method of the present invention, including the following steps:

[0034] S1. Obtain the first constraint condition of the rolling update compliance rate, and combine the historical scheduling execution results to determine the initial rolling update compliance rate.

[0035] In this embodiment, the first constraint condition includes the maximum offset frequency of the rolling adjustment of the scheduling plan and the response coupling ability of each time-scale scheduling model;

[0036] In this embodiment, combining the historical scheduling execution results to determine the initial rolling update compliance rate is specifically as follows:

[0037] Obtain the plan offset frequency and amplitude change of the multi-time scale scheduling model during the rolling adjustment within the preset period, and construct a scheduling offset sample set;

[0038] And combine the response coupling capabilities of each time scale scheduling model to calculate the impact of the response delay of each scheduling model after the rolling adjustment trigger on the scheduling consistency;

[0039] Evaluate the coordination stability of the scheduling system at different rolling frequencies according to the maximum acceptable offset frequency of the rolling adjustment;

[0040] Based on the scheduling stability and historical offset sample data, determine the initial rolling update compliance rate adapted to the current scheduling system through the method of minimum deviation matching.

[0041] Furthermore, the response coupling capabilities between each time scale are specifically quantified as:

[0042] C i→j = f(Δt i→j , η i→j , ρ i→j )

[0043] In the formula, C i→j is the response coupling capability from time scale i to j, Δt i→j is the average response delay from time scale i to j, η i→j is the response intensity amplitude, and ρ i→j is the response direction consistency characteristic.

[0044] Furthermore, the impact on the scheduling consistency is quantified by defining a consistency loss function, specifically as:

[0045]

[0046] In the formula, L is the consistency loss function, W ij is the weight of the scheduling task in the scale i→j, ΔP ij and ΔT ij are the offset difference and response delay in the scale i→j, ε is a decimal number to prevent the denominator from being zero (such as 1e-6); α, β are weight characteristics;

[0047] Furthermore, the specific method for obtaining the coordination stability of the scheduling system at different rolling frequencies is:

[0048]

[0049] In the formula, Stab(h) is the coordination stability at the rolling frequency h, Var(ΔP(h)) is the scheduling offset variance at the rolling frequency, Fail(h) is the conflict and failure probability, φ, is a weight factor;

[0050] Furthermore, based on scheduling stability and historical offset sample data, an initial rolling update compliance rate suitable for the current scheduling system is determined through the method of minimum deviation matching, specifically as follows:

[0051] Obtain a historical rolling update compliance rate data set, obtain the coordination stability and consistency loss function of each rolling update compliance rate, and combine with a preset first constraint condition to screen the historical rolling update compliance rates; select the rolling update compliance rate with the smallest deviation from the current system based on minimum deviation matching from the screened historical rolling update compliance rates as the initial rolling update compliance rate.

[0052] In this embodiment, the definition of the historical rolling update compliance rate data set is: R = {r1, r2,..., r N} where R is the historical rolling update compliance rate data set, and r N is the rolling update compliance rate in the Nth scheduling period. In this embodiment, the first constraint condition is specifically:

[0053] R′ = {r i ∈R | S(r i ) ≥ S min , L(r i ) ≤ L max};

[0054] In the formula, R′ is the screened historical rolling update compliance rate data set, r i is the screened historical rolling update compliance rate, S(r i ) is the coordination stability of the ith historical rolling update compliance rate, S min is the lowest threshold of coordination stability, L(r i ) is the consistency loss function of the ith historical rolling update compliance rate, and L max is the highest threshold of consistency loss;

[0055] In this embodiment, the initial rolling update compliance rate is specifically:

[0056]

[0057] d(r i ) = |S(r i ) - S target | + μ * |L(r i ) - L target |

[0058] In the formula, r * is the initial rolling update compliance rate, d(r i ) is the deviation matching degree, and S targetFor the target coordination stability of the current system, L target is the target consistency loss of the current system, and μ is the control factor for adjusting the influence weight of the consistency loss.

[0059] It should be noted that the lowest threshold of coordination stability and the highest threshold of consistency loss are obtained by statistically analyzing the historical rolling update execution data and combining with the empirical thresholds of system stability and consistency requirements, and can be set according to actual needs.

[0060] It should be noted that the control factor for adjusting the influence weight of the consistency loss is obtained by fitting the historical scheduling data and is optimized and adjusted in combination with the scheduling offset samples and the target consistency loss threshold.

[0061] It should be noted that the target coordination stability and the target consistency loss of the current system are obtained according to the statistical characteristics of the historical scheduling data and in combination with the thresholds set according to the scheduling strategy requirements.

[0062] It should be noted that the rolling update compliance rate with the smallest deviation is the rolling update compliance rate that is closest to the current scheduling system state among the screened historical rolling update compliance rates.

[0063] It should be noted that the larger the scheduling offset variance, the more unstable it is, and the larger the coordination stability at the rolling frequency h, the stronger the coordination.

[0064] It should be noted that the amplitude of the response intensity is the proportion of the response layer reaching the upper layer instruction amplitude, and the quantization result of the response direction consistency feature is 1 for the same direction, -1 for the opposite direction, and 0 for partial misalignment.

[0065] It should be noted that the planned offset frequency refers to the statistical frequency of the number of rolling adjustments (i.e., modifying the original plan) of each scheduling plan in the multi-time scale scheduling model within a given time window (such as 1 day, 1 week, 1 month); for example: the original plan of the intraday scheduling model is updated every hour. If there are 6 times of re-rolling the plan due to external disturbances on a certain day, it is regarded as the "intraday scheduling offset frequency" of that day is 6 times / day; if the annual scheduling model has 2 times of plan reconstruction within a month, the offset frequency is 2 times / month; the planned offset frequency is used to reflect the activity degree of the rolling mechanism in actual operation.

[0066] It should be noted that the amplitude change refers to the degree of change in the key resource scheduling variables between the rolling plan and the original plan; it can be further quantified as: the amplitude of the load difference (e.g., the change amplitude of the planned power output or the load curve); the time window offset (e.g., how much the start or end time of the scheduling is advanced / retreated); the resource call change rate (e.g., originally planned to call device A, and after rolling, it is replaced by device B or a combination of multiple devices); for example: the original planned scheduling is to supply 50 MW of power from 9:00 to 10:00, and after rolling adjustment, it is changed to supply 60 MW of power from 9:30 to 10:30, then the offset amplitude includes a time offset of 30 minutes and a power increase of 10 MW; the amplitude change constitutes a quantitative index of the rolling offset, reflecting the intensity of the scheduling response and the flexibility boundary of resource allocation.

[0067] It should be noted that the response coupling ability between each time scale refers to the degree of response linkage and timeliness of different scheduling levels (such as annual / quarterly / monthly scheduling, daily scheduling, intra-day scheduling, real-time scheduling) to the same rolling adjustment event in a multi-time scale scheduling system, and it is a measurement index of the multi-level scheduling linkage feedback efficiency, reflecting: whether the upper layer (such as daily scheduling) can quickly transmit the changed plan to the lower layer (such as real-time scheduling) and adapt in a timely manner; whether the lower layer feedback can reverse-correct the upper layer plan; whether the middle layer (such as intra-day scheduling) is "mismatched" or "overly late in response".

[0068] It should be noted that the rolling update compliance rate refers to the degree of maintaining the previous scheduling plan by the scheduling system under continuous rolling adjustment, which can be understood as the "inheritance of the old plan";

[0069] The maximum acceptable offset frequency refers to the maximum rolling adjustment frequency allowed by the system to maintain scheduling coordination;

[0070] The historical offset sample data set refers to the paired data of the planned offset frequency and the planned offset amplitude extracted from the actual operation.

[0071] S2. Obtain the second constraint condition of the scheduling margin configuration ratio, and quantify the initial scheduling margin configuration ratio based on the resource redundancy rate and accident response requirements under typical operating scenarios.

[0072] The second constraint condition of the scheduling margin configuration ratio includes the load prediction error distribution, the fluctuation range of renewable energy output, and the minimum guarantee ability of scheduling resources;

[0073] In this embodiment, obtaining the second constraint condition of the scheduling margin configuration ratio is specifically:

[0074] The second constraint condition of the scheduling margin configuration ratio is specifically:

[0075] That is, max(m2, m3) ≤ M ≤ m1

[0076] Wherein, M is the scheduling margin configuration ratio, and σ D is the standard deviation of the load forecasting error, ω D is the mean value of the load forecasting error, σ P is the standard deviation of the renewable energy output fluctuation, ω P is the mean value of the renewable energy output fluctuation, P min is the minimum guaranteed scheduling power of the system;

[0077] In this embodiment, based on the resource redundancy rate and accident response requirements under typical operating scenarios, the initial scheduling margin configuration ratio is quantified, specifically:

[0078] The scheduling margin configuration ratio is the ratio of the resource redundancy rate to the accident response requirements under typical operating scenarios; the accident response requirements are obtained by analyzing the sudden increase in load and renewable energy fluctuations.

[0079] The scheduling margin configuration ratios of several multi-scale scheduling systems within a preset period of history are constrained by the second constraint condition of the scheduling margin configuration ratio, and the average value is obtained to get the initial scheduling margin configuration ratio;

[0080] The resource redundancy rate under the typical operating scenario is specifically:

[0081]

[0082] The accident response requirements are specifically:

[0083]

[0084] The scheduling margin configuration ratio is specifically:

[0085]

[0086] Wherein, R s is the resource redundancy rate, P A is the total available power that can be called by the current system, D N is the predicted load demand under the typical scenario, ΔD max is the maximum load disturbance range, ΔP R is the renewable energy output fluctuation range.

[0087] S3. Perform multi-sample simulations on the initial parameters to generate several arrays of configurations to be adjusted.

[0088] The initial parameters include the initial rolling update compliance rate and the initial scheduling margin configuration ratio;

[0089] Determine the value ranges of the rolling update compliance rate and the scheduling margin configuration ratio based on the first constraint condition and the second constraint condition;

[0090] Based on Latin hypercube sampling, a number of to-be-adjusted configuration arrays are generated within the value range, and each configuration array contains different combinations of rolling update compliance rates and scheduling margin configuration ratios.

[0091] It should be noted that the first constraint condition limits the range of the rolling update compliance rate based on historical rolling adjustment offset samples and scheduling consistency impacts, so as to avoid scheduling instability or resource waste caused by excessive or too low adjustment frequencies.

[0092] The second constraint condition combines the resource redundancy rate and accident response requirements in typical operation scenarios to ensure that the scheduling margin configuration ratio fluctuates within a reasonable range, avoiding situations of insufficient or excessive redundant scheduling resources.

[0093] The Latin hypercube sampling method is used to uniformly sample within the value range that meets the above constraint conditions, so as to ensure the representativeness of the configuration array, improve the coverage of the simulation samples, and enable the subsequent optimization analysis to comprehensively reflect the possible impacts of different scheduling strategies.

[0094] In this embodiment, by constructing a refined initial calculation mechanism for the rolling update compliance rate, combining historical scheduling offset data with multi-level scheduling response coupling characteristics, the rolling adjustment strategy with the strongest current system adaptability can be effectively identified, thereby improving scheduling consistency and system stability. At the same time, by combining elements such as load prediction error, renewable energy fluctuations, and minimum guarantee capacity, a quantitative model for the scheduling margin configuration ratio is established, achieving a dynamic balance between redundant scheduling resources and risk response. Using Latin hypercube sampling to generate diverse parameter configurations further enhances the simulation coverage of the scheme and the integrity of the optimization space. The overall scheme has both real-time adaptability and interpretability, improving the coordinated control ability of the multi-time scale scheduling system in complex operating environments and providing effective support for building a new type of power system with strong resilience, fast response, and reasonable resource allocation.

[0095] S4. Map the to-be-adjusted configuration arrays into multi-time scale resource scheduling through scheduling orchestration and optimization drive, and conduct collaborative evaluation;

[0096] In this embodiment, mapping the to-be-adjusted configuration arrays into multi-time scale resource scheduling through scheduling orchestration and optimization drive specifically means:

[0097] According to the rolling update compliance rate and scheduling margin configuration ratio in the configuration array, combined with the multi-time scale structure of the scheduling system, determine the parameter setting schemes for each time scale scheduling model.

[0098] Based on the scheduling orchestration strategy, map different configuration arrays to short-cycle scheduling, medium-cycle scheduling, and long-cycle scheduling models to ensure that the scheduling schemes for each time scale can operate coordinately;

[0099] Among them, the short-cycle scheduling model focuses on real-time adjustment and is mainly dynamically optimized based on the rolling update compliance rate;

[0100] The mid-cycle scheduling model combines the scheduling margin configuration ratio to optimize the resource allocation strategy;

[0101] The long-cycle scheduling model comprehensively considers the rolling update compliance rate and the scheduling margin configuration ratio to jointly drive the scheduling model.

[0102] It should be noted that the existing technology has mature applications for mapping the configuration array to be adjusted to multi-time scale resource scheduling through scheduling orchestration and optimization driving, which will not be described here.

[0103] In this implementation, collaborative assessment is conducted, specifically:

[0104] The collaborative evaluation performs feature analysis on the multi-time scale scheduling system to obtain feature data for evaluating the degree of adaptation of the configuration data, and performs training based on the feature data for evaluating the degree of adaptation of the configuration data combined with historical data to obtain a driving effect evaluation model.

[0105] The characteristic data includes cross-time scale scheduling fluctuation characteristics and resource utilization dynamic matching characteristics.

[0106] The cross-timescale scheduling fluctuation feature is an indicator used to measure the fluctuation amplitude of scheduling decisions between different time scales, reflecting the impact of short-term scheduling changes on long-term scheduling goals. Specifically, this feature describes the coordination between short-term and long-term scheduling decisions of different configuration arrays. If the rolling update compliance rate or scheduling margin configuration ratio in the configuration array is not properly selected, it may lead to frequent fluctuations in short-term scheduling, thereby affecting the stability of long-term scheduling goals. By evaluating scheduling fluctuations across time scales, the impact of these fluctuations on the overall performance of the system can be quantified, helping to select configuration arrays that are more stable and adaptable at different time scales.

[0107] In practical applications, different configuration arrays (including rolling update compliance rate and scheduling margin configuration ratio) will directly affect the scheduling fluctuation across time scales. For example, a higher rolling update compliance rate may lead to frequent adjustments in the short term, thereby increasing the scheduling fluctuation characteristics; while a scheduling margin configuration ratio that is too high or too low may also cause incoordination between scheduling at different time scales and increase system fluctuations. By simulating the application of different configuration arrays, its impact on scheduling fluctuations across time scales can be evaluated, so as to select the most appropriate configuration array and optimize the stability and adaptability of the scheduling system.

[0108] Analyzing the fluctuation characteristics of scheduling across time scales has the following advantages for analyzing the application effects of different configuration arrays and solving the problem of coordinated imbalance in multi-time scale scheduling:

[0109] Improve scheduling efficiency: By identifying the characteristics of scheduling fluctuations at different time scales, the configuration array can be adjusted accordingly to reduce the volatility of resource scheduling, thereby optimizing the scheduling efficiency of the multi-time scale system. A relatively stable scheduling fluctuation helps to improve the overall efficiency of the system and reduce the frequency and cost of excessive system adjustments.

[0110] Balance resource allocation: The analysis of cross-time scale scheduling fluctuation characteristics can reveal the impact of different configuration arrays on resource allocation during the scheduling process, helping to optimize resource configuration, avoid excessive or insufficient resource scheduling on certain time scales, and thus achieve reasonable and efficient utilization of resources.

[0111] Improve scheduling consistency: By analyzing cross-time scale scheduling fluctuation characteristics, the collaborative relationship and potential imbalance points between scheduling strategies at different time scales can be identified, and then the configuration array can be optimized to reduce the scheduling imbalance problem between different time scales. Controlling scheduling fluctuations helps to improve scheduling consistency and reduce unnecessary adjustments.

[0112] Enhance system stability: The analysis of cross-time scale scheduling fluctuation characteristics helps the system to better adapt to external environmental changes, timely adjust the configuration array, reduce system instability caused by fluctuations, and ensure continuous and stable operation in a multi-time scale scheduling environment.

[0113] Optimize multi-time scale scheduling coordination: In a multi-time scale scheduling system, the coordination between time scales is crucial. By analyzing cross-time scale scheduling fluctuation characteristics, the coordination imbalance problem between scheduling at different time scales can be identified and solved, so as to achieve optimized coordination between time scales, enabling the system to flexibly respond to complex scheduling requirements.

[0114] The specific method for obtaining the cross-time scale scheduling fluctuation characteristics is as follows:

[0115] At different scheduling time scales, obtain the processing time required from the initiation of a scheduling request to the completion of scheduling execution and the generation of a scheduling result, as the scheduling response time;

[0116] For the same scheduling time scale, perform several calculations of the response time and calculate the average response time;

[0117] Obtain the resource scheduling volume and total duration of the system at different scheduling time scales, and calculate the system scheduling efficiency;

[0118] Set an initial cross-time scale scheduling fluctuation model based on the average response time and system scheduling efficiency;

[0119] Set several response time points for system performance testing. At each response time point, collect the fluctuation amplitude data of the scheduling execution results;

[0120] Based on the amplitude fluctuation data, the undetermined parameters of the initial cross-time-scale scheduling fluctuation model are optimized using the non-linear least squares method to obtain a calculation model for the cross-time-scale scheduling fluctuation characteristics. The undetermined parameters include the maximum value of the amplitude fluctuation, the change rate of the amplitude fluctuation with the increase of the response time, and the sensitivity characteristics of the response time to the change of the amplitude fluctuation.

[0121] Through these steps, we can effectively quantify the cross-time-scale scheduling fluctuation characteristics, providing a basis for further scheduling optimization.

[0122] The specific calculation formula for the average response time is as follows:

[0123]

[0124] The specific calculation formula for the adjustment step size of the cross-time-scale scheduling fluctuation model is as follows:

[0125]

[0126] The specific calculation formula for the cross-time-scale scheduling fluctuation characteristics is as follows:

[0127]

[0128] In the formula, is the average response time, T g，i is the response time of the i-th scheduling, n is the number of times to measure the scheduling response time at the same time scale, ε is the adjustment step size of the initial cross-time-scale scheduling fluctuation model, m is the number of fluctuation characteristic data of the data processing results for optimization, j is the label of the fluctuation characteristic data of the data processing results for optimization, E d,j is the cross-time-scale scheduling fluctuation data value of the data processing result, and δ is the cross-time-scale scheduling fluctuation characteristic.

[0129] The dynamic matching characteristic of resource utilization rate is an index used to measure the impact of different configuration arrays on the resource utilization efficiency during the resource allocation process, reflecting the matching degree of resource utilization by the scheduling system at different time scales. Specifically, the dynamic matching characteristic of resource utilization rate can reveal the optimization effects of different values of the rolling update compliance rate and the scheduling margin configuration ratio. This characteristic quantifies whether the configuration array can effectively balance the resource demand and the available resources during the dynamic resource scheduling process by comparing the difference between the actual resource utilization situation and the predetermined resource demand, thus ensuring the maximization and efficiency of resource utilization.

[0130] In practical applications, different configuration arrays directly affect the dynamic matching characteristics of resource utilization. For example, a higher rolling update compliance rate may lead to frequent resource reconfiguration and adjustment, causing over-scheduling or idleness of resources in the short term, thus affecting the overall resource utilization. Changes in the scheduling margin configuration ratio may also affect over-allocation or under-allocation of resources, especially in cases where the load demand changes drastically. By simulating the applications of different configuration arrays, their impacts on resource utilization can be evaluated, and the configuration array that best dynamically matches resource demand and supply can be selected, thereby optimizing resource utilization efficiency.

[0131] Analyzing the dynamic matching characteristics of resource utilization has the following advantages for analyzing the application effects of different configuration arrays and solving the problem of unbalanced multi-time scale scheduling coordination:

[0132] Improve resource utilization efficiency: By dynamically matching resource utilization, the resource requirements of each configuration array at different time scales can be accurately grasped, resource allocation can be optimized, and resource waste or over-scheduling can be avoided. This helps to improve the overall resource utilization of the system and ensure that resources are optimally configured.

[0133] Optimize multi-time scale scheduling coordination: The dynamic matching characteristics of resource utilization can reveal the changing trends of resource requirements at different time scales, thus providing an accurate basis for adjusting the configuration array, reducing conflicts and unbalanced coordination between time scales. Dynamic matching can be flexibly adjusted within different scheduling cycles to optimize scheduling coordination.

[0134] Reduce scheduling delay and fluctuations: The dynamic matching characteristics help to identify and adjust fluctuations in resource utilization, reducing system fluctuations caused by uneven resource allocation or scheduling delay. This helps to improve the response speed of the scheduling system and reduce unnecessary fluctuations caused by mismatched configuration arrays.

[0135] Enhance the adaptability and flexibility of the system: With the application of different time scale scheduling models, the system can quickly adjust through dynamic matching characteristics in the face of various external changes to adapt to fluctuations in resource requirements. Ensure the adaptability and flexibility of different configuration arrays in resource utilization in practical applications.

[0136] Improve system stability and reliability: By accurately matching resource utilization, over-scheduling or resource shortages can be avoided, thus reducing the instability of the system in complex scheduling scenarios. Through dynamic adjustment of the configuration array, the system can maintain a high degree of stability and reliability at different time scales.

[0137] Reduce system operating costs: Dynamically matching resource utilization can reduce over-scheduling caused by unreasonable resource configuration, thus saving energy and scheduling costs. By optimizing the configuration array and scheduling strategy, the system can allocate resources more effectively and reduce operating costs.

[0138] The selection of the cross - time - scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate as the directions for analyzing the application effects of different configuration arrays is mainly because these two aspects directly affect the stability and resource efficiency of the multi - time - scale scheduling system. First of all, the cross - time - scale scheduling fluctuation characteristics can reveal the fluctuation behavior of the system and the changes in resource requirements at different time scales, helping to evaluate the impact of the configuration array on the system stability, especially in solving the scheduling imbalance problem between multiple time scales. This is very important for ensuring the coordination of resource requirements within different scheduling cycles. The dynamic matching characteristics of resource utilization rate can accurately optimize the resource allocation, avoid resource waste or shortage, thereby improving the overall resource utilization efficiency of the system and reducing unnecessary delays and costs. Compared with other possible analysis directions, such as the scheduling fluctuation of a single time scale or the simple analysis of resource utilization rate, these two directions can capture the dynamic changes and resource utilization efficiency in the multi - time - scale scheduling collaboration more comprehensively, directly affecting the optimization effect and long - term stable operation of the system. Therefore, focusing on these two characteristics has higher practicality and pertinence.

[0139] The specific method for obtaining the dynamic matching characteristics of resource utilization rate is as follows:

[0140] Obtain the changes in the system resource utilization rate and the distribution of scheduling fluctuations under different scheduling strategies and resource allocation methods, and construct a resource utilization rate matching evaluation model through the resource scheduling algorithm;

[0141] Perform clustering processing on the resource utilization rate matching evaluation model;

[0142] Extract features from the processed resource utilization rate matching evaluation model, and calculate the matching accuracy evaluation features based on the resource utilization rate fluctuation, scheduling error, and resource contention situation;

[0143] Simulate the resource utilization rate under different scheduling strategies and resource allocation methods through the resource utilization rate matching evaluation model, establish a resource allocation timeliness model, and obtain the resource allocation timeliness evaluation features;

[0144] Calculate the dynamic matching characteristics of resource utilization rate by combining the matching accuracy evaluation features.

[0145] The specific calculation formula for the matching accuracy evaluation features is as follows:

[0146]

[0147] The specific calculation formula for the resource allocation timeliness evaluation features is as follows:

[0148]

[0149] The specific calculation formula for the dynamic matching feature of resource utilization rate is as follows:

[0150]

[0151] In the formula, WS q is the matching accuracy evaluation feature, TF q is the resource allocation timeliness evaluation feature, is the dynamic matching feature of resource utilization rate, z is the deviation value, TJ is the energy consumption fluctuation under different load allocation and resource scheduling strategies, θ is the resource utilization coverage under different load allocations, md is the scheduling feedback rate, μ is the scheduling error, is the resource scheduling model, is the resource scheduling efficiency loss, u is the stability of the resource scheduling system, is the resource response time adjustment term.

[0152] In this embodiment, the drive effect evaluation model is specifically:

[0153] Establish an initial drive effect evaluation model, and according to the cross-time scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate in historical data, use a multi-layer perceptron (MLP) to supervise and train the proportional coefficients of the initial drive effect evaluation model to obtain the drive effect evaluation model;

[0154] The initial drive effect evaluation model is specifically:

[0155]

[0156] In the formula, H is the drive effect evaluation value, that is, the evaluation result, θ1 is the proportional coefficient to be trained for the cross-time scale scheduling fluctuation characteristics, and θ2 is the proportional coefficient to be trained for the dynamic matching characteristics of resource utilization rate.

[0157] It should be noted that the multi-layer perceptron is only a method used in this application, and any supervised training method with strong interpretability or high prediction accuracy can be applied to this application, such as gradient boosting decision tree (GBDT).

[0158] S5. Based on the evaluation result and the configuration array to be adjusted, fit a high-order collaborative response display surface.

[0159] In this embodiment, based on the evaluation result and the configuration array to be adjusted, fitting a high-order collaborative response display surface is specifically:

[0160] Based on the cross-time scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate, construct a multi-dimensional feature mapping space for characterizing the response law of the multi-time scale scheduling system;

[0161] The dimensions of the multi-dimensional feature mapping space include cross-time-scale scheduling fluctuation features, resource utilization dynamic matching features, and drive effect evaluation values;

[0162] Associate and map the characteristic parameters of each group of configuration arrays to be adjusted with their corresponding drive effect evaluation values to form a set of mapping sample points;

[0163] Based on the mapping sample points, fit and construct a high-order collaborative response surface covering the entire feature mapping domain to reflect the change trend of the scheduling collaborative performance of the system under different configuration combinations.

[0164] S6. Select the scheduling sensitive area of the surface according to the actual requirements, screen the optimal configuration array, and drive the multi-time-scale resource scheduling.

[0165] In this embodiment, selecting the scheduling sensitive area of the surface according to the actual requirements, screening the optimal configuration array, and driving the multi-time-scale resource scheduling specifically include:

[0166] Determine several peak regions of the high-order collaborative response display surface, and according to the preset dynamic regulation amplitude, expand the range based on several peaks to construct several candidate regions for adjusting the configuration array;

[0167] According to the actual requirements of the scheduling target, screen several candidate regions;

[0168] If the requirement for short-term response ability is relatively high, select the region with a higher rolling update compliance rate in the candidate region as the scheduling sensitive area;

[0169] If the system stability requirement is stronger, select the region with a high scheduling margin configuration in the candidate region as the scheduling sensitive area;

[0170] If the resource usage efficiency is the current priority target, select the region with the highest peak in the candidate region as the scheduling sensitive area;

[0171] Take the configuration array corresponding to the peak in the scheduling sensitive area as the optimal configuration array to drive the multi-time-scale resource scheduling.

[0172] In this embodiment, by mapping the configuration array to be adjusted to a multi-time-scale resource scheduling model and combining the scheduling objectives and characteristics of short cycles, medium cycles, and long cycles, refined management is carried out from the aspects of real-time performance, resource buffering, and global balance, significantly enhancing the coordination and flexibility of the scheduling system at different time scales. On this basis, the system introduces a collaborative evaluation mechanism, quantifies and analyzes the impact of each configuration array on the system scheduling behavior and resource efficiency around two key indicators: the cross-time-scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate, effectively solving problems such as inconsistent inter-cycle scheduling and unbalanced resource allocation existing in traditional scheduling systems. On the one hand, the introduction of cross-time-scale scheduling fluctuation characteristics helps to accurately capture the interference degree of the configuration array in short-term decisions on long-term scheduling objectives and optimize the stability of scheduling behavior; on the other hand, the dynamic matching characteristics of resource utilization rate focus on the supply-demand adaptation situation in the resource allocation process, improving the response ability and utilization efficiency of the scheduling system to resource state changes.

[0173] Figure 2 is a multi-time-scale resource scheduling optimization system for smart grids, including an initial parameter acquisition module, a configuration array acquisition module, a drive evaluation module, and a configuration array adjustment module;

[0174] The initial parameter acquisition module is used to obtain the first constraint condition of the rolling update compliance rate and determine the initial rolling update compliance rate; obtain the second constraint condition of the scheduling margin configuration ratio and quantify the initial scheduling margin configuration ratio;

[0175] The configuration array acquisition module is used to perform multi-sample simulation on the initial parameters to generate a number of configuration arrays to be adjusted;

[0176] The drive evaluation module is used to map the configuration array to be adjusted to multi-time-scale resource scheduling through scheduling orchestration and optimization drive and perform collaborative evaluation;

[0177] The configuration array adjustment module is used to fit a high-order collaborative response display surface based on the evaluation results and the configuration array to be adjusted; select the scheduling sensitive area of the surface according to actual needs, screen the best configuration array, and drive multi-time-scale resource scheduling.

[0178] Figure 3 is a display diagram of a multi-dimensional feature mapping space based on, where Resourc Usage is the dimension of the dynamic matching characteristics of the resource utilization rate, Sensitivity is the dimension of the cross-time-scale scheduling fluctuation characteristics, and Efficiency is the dimension of the drive effect evaluation value, Figure 3 The points in are the mapped sample points of the display.

[0179] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination.

[0180] As described above, this is only a specific implementation manner of the present application. However, the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims described above.

[0181] Finally: The above is only a preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent power grid multi-time scale resource scheduling optimization method, characterized in that, It includes the following steps: Obtain the first constraint condition of the rolling update compliance rate and determine the initial rolling update compliance rate; Obtain the second constraint condition of the scheduling margin configuration ratio and quantify the initial scheduling margin configuration ratio; The rolling update compliance rate refers to the degree of maintaining the previous scheduling plan by the scheduling system under continuous rolling adjustment; The scheduling margin configuration ratio is the ratio of the resource redundancy rate in a typical operation scenario to the accident response demand; Perform multi-sample simulation on the initial parameters to generate several configuration arrays to be adjusted; Map the configuration arrays to be adjusted into multi-time-scale resource scheduling through scheduling orchestration and optimization drive and conduct collaborative evaluation; The specific content of the collaborative evaluation is as follows: The collaborative evaluation obtains characteristic data for evaluating the adaptability of configuration data by analyzing the characteristics of the multi-time-scale scheduling system; And train according to the obtained characteristic data for evaluating the adaptability of configuration data combined with historical data to obtain a drive effect evaluation model. The characteristic data includes cross-time-scale scheduling fluctuation characteristics and dynamic matching characteristics of resource utilization rate; Fit a high-order collaborative response display surface based on the evaluation results and the configuration arrays to be adjusted; Select the scheduling sensitive area of the surface according to actual requirements, screen the best configuration array, and drive multi-time-scale resource scheduling.

2. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 1, wherein The specific content of obtaining the first constraint condition of the rolling update compliance rate and determining the initial rolling update compliance rate is as follows: The first constraint condition includes the maximum offset frequency of the rolling adjustment of the scheduling plan and the response coupling ability of each time-scale scheduling model; Obtain the planned offset frequency and amplitude of the multi-time-scale model within a period, and construct a scheduling offset sample set; Combined with the response coupling ability of the time-scale scheduling model, analyze the impact of the response delay of each scheduling model after the rolling adjustment trigger on the scheduling consistency; Evaluate the coordination stability of the system at different frequencies according to the maximum acceptable offset frequency; Obtain the historical rolling update compliance rate data set and its stability and consistency loss function, and screen the compliance rate in combination with the first constraint condition; Select the rolling update compliance rate with the smallest deviation from the current system as the initial rolling update compliance rate based on the minimum deviation matching from the screened historical rolling update compliance rates.

3. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 2, wherein The specific content of obtaining the second constraint condition of the scheduling margin configuration ratio and quantifying the initial scheduling margin configuration ratio is as follows: The second constraint condition includes the load prediction error distribution, the fluctuation range of renewable energy output, and the minimum guarantee ability of scheduling resources; The scheduling margin configuration ratio is the ratio of the resource redundancy rate in a typical operation scenario to the accident response demand, and the accident response demand is obtained by analyzing the sudden increase in load and the fluctuation of renewable energy; Constrain the scheduling margin configuration ratios of several multi-scale scheduling systems within the historical preset period through the second constraint condition and calculate the average to obtain the initial scheduling margin configuration ratio.

4. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 3, wherein The specific content of performing multi-sample simulation on the initial parameters to generate several configuration arrays to be adjusted is as follows: The initial parameters include the initial rolling update compliance rate and the initial scheduling margin configuration ratio; Determine the value ranges of the rolling update compliance rate and the scheduling margin configuration ratio based on the first constraint condition and the second constraint condition; Based on Latin hypercube sampling, several arrays of configurations to be adjusted are generated within the value range, and the configuration arrays contain different combinations of rolling update compliance rates and scheduling margin configuration ratios.

5. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 4, wherein Based on the evaluation results and the arrays of configurations to be adjusted, a high-order collaborative response display surface is fitted, specifically as follows: Based on the cross-time-scale scheduling fluctuation characteristics and the dynamic matching characteristics of resource utilization rate, a multi-dimensional feature mapping space is constructed to characterize the response law of the multi-time-scale scheduling system. The dimensions of the multi-dimensional feature mapping space include cross-time-scale scheduling fluctuation characteristics, dynamic matching characteristics of resource utilization rate, and driving effect evaluation values. The characteristic parameters of each array of configurations to be adjusted are associated and mapped with their corresponding driving effect evaluation values to form a set of mapping sample points. Based on the mapping sample points, a high-order collaborative response surface covering the entire feature mapping domain is fitted and constructed to reflect the change trend of the scheduling collaborative performance of the system under different configuration combinations.

6. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 5, characterized in that According to the actual requirements, the scheduling sensitive area of the surface is selected, the optimal configuration array is screened, and the multi-time-scale resource scheduling is driven, specifically as follows: Several peak regions of the high-order collaborative response display surface are determined, and according to the preset dynamic regulation amplitude, range expansion is carried out based on several peaks to construct several candidate regions for adjusting the configuration array. According to the actual requirements of the scheduling target, several candidate regions are screened. If the actual requirement is short-term response ability, the region with the highest rolling update compliance rate in the candidate regions is selected as the scheduling sensitive area. If the actual requirement is system stability, the region with the highest scheduling margin configuration in the candidate regions is selected as the scheduling sensitive area. If the actual requirement is resource usage efficiency, the region with the highest peak in the candidate regions is selected as the scheduling sensitive area. The configuration array corresponding to the peak in the scheduling sensitive area is used as the optimal configuration array to drive the multi-time-scale resource scheduling.

7. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 6, wherein The specific method for obtaining the cross-time-scale scheduling fluctuation characteristics is as follows: Under different scheduling time scales, the processing time from request to execution completion is obtained as the response time. For the same time scale, the average value of several response times is calculated. The resource scheduling volume and total duration under different time scales are obtained, and the scheduling efficiency is calculated. An initial cross-time-scale fluctuation model is set according to the average response time and scheduling efficiency. Response time points are set for performance testing, and the scheduling fluctuation amplitude data at each point are collected. According to the fluctuation amplitude data, the parameters of the initial fluctuation model are optimized based on the nonlinear least squares method to obtain a scheduling fluctuation characteristic calculation model, and the parameters of the initial fluctuation model include the maximum value of the fluctuation amplitude, the change speed, and the sensitivity characteristics.

8. The intelligent power grid multi-time scale resource scheduling optimization method according to claim 7, characterized in that The specific method for obtaining the dynamic matching characteristics of resource utilization rate is as follows: The changes in resource utilization rate and scheduling fluctuations under different scheduling strategies and resource allocations are obtained, and a matching evaluation model is constructed through a scheduling algorithm. Clustering processing is performed on the matching evaluation model. Based on the extracted characteristics of the evaluation model after clustering, the matching accuracy evaluation characteristics based on resource fluctuations, scheduling errors, and contentions are calculated. By evaluating the model to simulate the resource utilization under different scheduling strategies, an aging model is established and the aging evaluation features of resource allocation are obtained, and the dynamic matching features of resource utilization are obtained by summing the matching accuracy evaluation features.

9. A system using the intelligent grid multi-time scale resource scheduling optimization method according to any one of claims 1-8, characterized in that, It includes an initial parameter acquisition module, a configuration array acquisition module, a driving evaluation module, and a configuration array adjustment module; The initial parameter acquisition module is used to obtain the first constraint condition of the rolling update compliance rate and determine the initial rolling update compliance rate; obtain the second constraint condition of the scheduling margin configuration ratio and quantify the initial scheduling margin configuration ratio. The configuration array acquisition module is used to perform multi-sample simulation on the initial parameters to generate a number of configuration arrays to be adjusted. The driving evaluation module is used to map the configuration arrays to be adjusted into multi-time scale resource scheduling through scheduling orchestration and optimization driving, and perform collaborative evaluation. The configuration array adjustment module is used to fit the high-order collaborative response display surface based on the evaluation results and the configuration arrays to be adjusted; select the scheduling sensitive area of the surface according to actual requirements, screen the optimal configuration array, and drive the multi-time scale resource scheduling.

Images