Optimized scheduling method and system for new energy power system

Abstract

The invention relates to the technical field of large-scale power system safety guarantee and intelligent scheduling, and discloses a new energy power system optimization scheduling method and system, and the method comprises the steps: carrying out the internal processing steps before economic scheduling, and carrying out the self-adaptive correction of a frequency change rate safety threshold value based on the overall completeness of a power defense system, real-time data closed-loop verification is carried out based on a wide-area measurement system, and an inertia contribution constant of a started unit is selected; and then generating a minimum system inertia base line constraint based on the corrected frequency change rate safety threshold and the inertia contribution constant selected after verification, and implanting the minimum system inertia base line constraint into the economic dispatching optimization model for execution. According to the method, a closed-loop information interaction path is established between power system safety analysis logic and an economic dispatching model, and verified physical safety requirements are preposed as constraint conditions which must be obeyed by the economic model.

Description

TECHNICAL FIELD

[0001] The present application relates to a new energy power system optimization scheduling method and system, belonging to the technical field of large-scale power system security and intelligent dispatch. BACKGROUND

[0002] In the current operation and control of large-scale power systems, maintaining the stability of the system frequency is the core prerequisite for ensuring the safe operation of the power grid, and the total inertia of the system, i.e. the physical ability of the power grid to resist frequency changes, is a key physical support for building the entire security and defense system. In a high-inertia system dominated by traditional synchronous generators, sufficient physical rotational inertia provides the necessary response time margin for various safety and defense measures of the dispatch system. However, with the large-scale grid connection of new energy sources such as wind power and photovoltaic power through power electronic inverters, their near-zero inertia contribution changes the physical basis of the power grid. At the same time, the economic dispatch logic aimed at minimizing operating costs will automatically replace the operation of traditional synchronous generators that provide high inertia when faced with new energy sources with extremely low marginal costs. This economically oriented dispatch execution process continuously weakens the necessary inertia support of the power grid at the physical level, leading to a continuous decline in the total inertia level of the system and reducing the frequency stability of the power grid.

[0003] To address this challenge, the existing technical ideas are mostly focused on post-response and remediation, such as deploying fast frequency response resources or relying on emergency control measures, but these solutions on the one hand rely on high auxiliary service resources, and on the other hand, they do not prevent the formation of low inertia state from the root; When the economic dispatch determines the operating mode to be lower than the physical critical value, once a regular disturbance is encountered, the frequency change rate will far exceed the design limit of the defense system, causing all these remedial measures to be ineffective in time, and the system still faces the risk of frequency collapse; Analysis shows that there is a fundamental lack of information exchange and mechanism coordination between the current power grid security protection system and the economic dispatch logic, that is, the economic dispatch model lacks consideration of the system physical inertia, a core security element in its decision-making basis; Currently, the optimization scheduling methods proposed in the industry to address the uncertainty of new energy are mostly focused on dealing with the randomness of its output prediction, but fail to address this fundamental coordination gap at the physical level, for example, the Chinese invention patent with publication number CN114050609A discloses a high-proportion new energy power system self-adaptive robust day-ahead optimization scheduling method, which aims to solve the influence of wind and light output uncertainty on scheduling, which generates an uncertain scenario set using Monte Carlo simulation and constructs a self-adaptive robust optimization model with the goal of minimizing total operating cost, however, the robustness of this approach is aimed at the output prediction error of new energy, and its essence is still an optimization method that deals with randomness at the economic level, it cannot identify and verify the real physical inertia required by the system from the perspective of power grid physical safety, and it cannot translate this physical safety requirement into a rigid constraint that the economic dispatch model must comply with, the scheduling result may still lead the system to a critical low inertia state at the physical level, and cannot avoid the risk of frequency collapse from the source.

[0004] Therefore, how to provide a new energy power system optimization scheduling method, which can dynamically calculate and impose a minimum inertia bottom line constraint to ensure system frequency safety before economic dispatch is executed, so that the economic optimization behavior is subject to the rigid constraints of physical safety from the source, has become a technical problem to be solved by the present application. SUMMARY

[0005] The present application provides a new energy power system optimization scheduling method and system, which mainly aims to solve the problem that the existing economic dispatch logic lacks consideration of the system physical inertia, resulting in frequency safety risks in the scheduling scheme.

[0006] To achieve the above-mentioned purpose, the new energy power system optimization scheduling method provided by the present application comprises an internal processing step for generating and verifying the safety constraints of the power system before the execution of the economic dispatch optimization model, the internal processing step comprises: Step a, adaptive correction step of constraint reference, real-time aggregation of state information of multiple defense resources in power system, including low frequency load shedding device and generator frequency controller; based on state information, a dynamic compensation index representing overall completeness of power defense system is calculated; and based on dynamic compensation index, a frequency rate safety threshold for constraint generation is dynamically adjusted, the rule is that when overall completeness decreases, frequency rate safety threshold is automatically tightened; Step b, closed loop verification step of constraint parameter, using real-time data of wide area measurement system, real inertia contribution of online unit is identified and verified; real inertia contribution is compared with a reference inertia contribution constant determined based on nameplate parameters or control mode of the unit, and the smaller value of real inertia contribution and reference inertia contribution constant is selected as inertia contribution constant for step c constraint implantation; Step c, constraint generation and implantation step, based on frequency rate safety threshold obtained after adaptive correction in step a, minimum system inertia bottom line is reversely deduced; and based on inertia contribution constant selected after verification in step b, constraint equation is constructed; constraint equation is implanted into economic dispatch optimization model as mandatory constraint condition; Step d, constraint optimization step, economic dispatch optimization model is executed to generate dispatch instructions under the restriction of mandatory constraint condition.

[0007] Preferably, the step of reversely deducing the minimum system inertia bottom line in step c further comprises: step a-1, real-time obtaining topology structure of current power system; step a-2, based on real-time topology structure, dynamically determining a list of one or more current trusted disturbances; step a-3, for each current trusted disturbance in the list, respectively calculating a disturbance specific inertia demand required by it; step a-4, determining the maximum value in the multiple disturbance specific inertia demands as the minimum system inertia bottom line.

[0008] Preferably, step a and step c further comprise: dividing the power system into at least two dynamic inertia partitions based on dynamic physical characteristics thereof; for each of the at least two dynamic inertia partitions, respectively calculating and determining its corresponding minimum local inertia bottom line; in step c, each minimum local inertia bottom line is implanted into the economic dispatch optimization model as mandatory constraint condition, to ensure that the total inertia contribution in each dynamic inertia partition meets its respective bottom line.

[0009] Preferably, the closed-loop verification step b of the constraint parameters further includes: step b-1 monitoring the disturbance richness in the real-time data of the wide-area measurement system to determine the data credibility of the true inertia contribution; step b-2 when the data credibility is determined to be lower than a credibility threshold set according to the historical statistical characteristics of the real-time data of the wide-area measurement system, actively injecting a standardized power perturbation signal of a specific frequency into the power system through a power system fast adjustment resource; step b-3 synchronously extracting the global frequency response amplitude corresponding to the specific frequency of the power perturbation signal from the real-time data of the wide-area measurement system. and the power response amplitude of each unit Step b-4 is based on the global frequency response amplitude. With the power response amplitude of each unit The physical relationship between them was re-identified and verified to determine the true contribution of inertia. Its identification rules follow: ,in, It is a calibration coefficient characterizing the physical properties of the power system; and by using the re-identified and verified real inertia contribution, the comparison and selection rules defined in step b of the closed-loop verification of the constraint parameters are executed.

[0010] Preferably, the method further includes: reusing the inertia contribution constant selected after verification in step b, wherein the inertia contribution constant represents the total actual inertia of the operating units of the power system; sending the inertia contribution constant as a power system security status parameter to the low-frequency load shedding defense device in the power system; and the low-frequency load shedding defense device automatically selecting and activating a set of action setting values ​​that matches the current inertia level from multiple preset action setting value tables based on the received power system security status parameter.

[0011] Preferably, in step a, the step of calculating the dynamic compensation index includes: using a weighted summation logic based on power system expert rules to aggregate the discrete state information of multiple defense resources into a single normalized value as the dynamic compensation index.

[0012] Preferably, the step of reverse deducing the minimum system inertia baseline in step c includes: step c1 establishing a dynamic response relationship model between the total system inertia and the frequency change rate when the power system encounters a preset disturbance; step c2 based on the frequency change rate safety threshold obtained after adaptive correction in step a, and reverse deducing the minimum system inertia baseline through the dynamic response relationship model.

[0013] Preferably, the constraint equation in step c is in the form of a sum of the product of the generator set start-up state decision variable and the inertia contribution constant selected after verification in step b, and the sum of the product is limited to be greater than or equal to the minimum system inertia baseline derived in reverse in step c.

[0014] Preferably, the economic dispatch optimization model is an optimization model with the objective function of minimizing total operating cost. The method further includes: during or after the constraint optimization in step d, extracting the shadow price corresponding to the mandatory constraint condition of the minimum system inertia baseline from the economic dispatch optimization model; using the shadow price as an economic cost indicator of power system security constraints for post-verification; and triggering an alarm or prompting for manual review of the constraint generation in steps a and c when the economic cost indicator exceeds an economic rationality threshold set by the power system dispatch rules.

[0015] A new energy power system optimization dispatching system includes: The constraint benchmark adaptive correction unit is configured to: collect the status information of multiple defense resources in the power system in real time, including low-frequency load shedding devices and generator frequency controllers; calculate a dynamic compensation index characterizing the overall integrity of the power defense system based on the status information; and dynamically adjust a frequency change rate safety threshold used for constraint generation based on the dynamic compensation index, with the rule that the frequency change rate safety threshold is automatically tightened when the overall integrity decreases. The constraint parameter closed-loop verification unit is configured to: use real-time data from the wide-area measurement system to identify and verify the actual inertia contribution of the already started unit online; compare the actual inertia contribution with a reference inertia contribution constant determined based on the unit's nameplate parameters or control mode, and select the smaller value between the actual inertia contribution and the reference inertia contribution constant as the output of the inertia contribution constant required for subsequent constraint implantation; The constraint generation and implantation unit is configured to: derive the minimum system inertia baseline by reverse-engineering the frequency change rate safety threshold output by the constraint benchmark adaptive correction unit; construct constraint equations based on the inertia contribution constant output by the constraint parameter closed-loop verification unit; and implant the constraint equations as mandatory constraints into an economic scheduling optimization model. The constrained optimization unit is configured to execute an economic scheduling optimization model to generate scheduling instructions under the constraints of mandatory constraints.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. A two-way closed-loop information exchange path is established between the security analysis logic and the economic dispatch model of the power system. First, the system inertia requirement, which represents the physical bottom line and originates from the security analysis, is translated and implanted into the economic dispatch model, so that when seeking the optimal cost, the unit combination that meets physical safety must be retained. Furthermore, this method also utilizes the economic cost information generated by the economic model when executing this constraint, extracts this information in reverse and presents it to the security dispatch side, so that the makers of security strategies can, for the first time, quantitatively and clearly grasp the operating costs corresponding to their physical security decisions. In this way, the security logic and economic logic are transformed from being mutually separated and information blind to mutually constraining and transparent, providing a quantifiable and arbitrable decision-making basis for the coordinated dispatch of large-scale power grids between security and economy.

[0017] 2. The system inertia baseline used in this method is not a fixed static threshold, but rather an adaptive dynamic constraint generation mechanism that accurately and dynamically matches the real physical risks of the power grid that are constantly changing in time and space. Before calculating constraints, this mechanism first obtains the current topology of the power grid in real time and uses this as a basis to dynamically determine the list of credible disturbances that are truly threatening. At the same time, this mechanism also combines the analysis of the physical spatial characteristics of the large-scale power grid to automatically identify dynamic inertia partitions that may have local instability risks on key sections. Finally, what is implanted into economic dispatch is a multi-dimensional set of constraints, which not only ensures the overall inertia baseline for dealing with the maximum global risk, but also ensures the inertia baseline of each partition to prevent local frequency collapse. In this way, the problems of over-constraint or under-constraint caused by traditional fixed thresholds when the power grid operating conditions change are avoided.

[0018] 3. This method establishes a self-verifying closed-loop collaborative defense system based on physical reality. It does not rely on the theoretical commitment value of the inertia contribution of generating units (especially new energy sources). Instead, it uses an inertia verification and correction step to identify and verify the real inertia contribution of each online unit online using real-time data from a wide-area measurement system. This verified real inertia information is used, on the one hand, to dynamically correct and refresh the safety constraint boundary embedded in economic dispatch, ensuring that the decision-making of the first line of defense, optimized dispatch, is always based on physical reality. On the other hand, this verified situational parameter, which characterizes the real vulnerability of the power grid, is reused and sent to defense devices such as low-frequency load shedding in the power grid. The defense devices then automatically adapt their action strategies based on this information, ensuring that the response of the last line of defense, emergency control, always matches the real risk level of the power grid. This method of dynamically coordinating prevention and control through verified physical information constructs a logically unified and information-connected defense-in-depth system. Attached Figure Description

[0019] Fig. 1 This is a flowchart illustrating the closed-loop control process for the coordinated scheduling of safety and economy in this invention. Fig. 2 This is a comparison graph showing the frequency response performance of the method of the present invention and the control group under system disturbance; Fig. 3 The diagram shows the multi-layered functional modules and information interaction architecture of the optimized scheduling system of this invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the scope of protection of the invention.

[0021] This invention provides a method and system for optimizing the dispatch of new energy power systems. Its architecture involves pre-implanting an internal processing step for generating and verifying power system security constraints before the execution of an economic dispatch optimization model aimed at minimizing operating costs. This internal processing step constructs a closed-loop information interaction and constraint path from power system physical security analysis to the execution of the economic dispatch model. It first dynamically and adaptively generates a definite system inertia baseline constraint based on the physical requirements of (large-scale power grid security assurance). This constraint is then implanted into the economic dispatch optimization model of the (intelligent dispatch system), ensuring that economic optimization behavior, while seeking cost optimization, must meet the physical security boundaries of security assurance. This avoids large-scale power grid frequency security risks caused by inertia deficiencies from the source, before dispatch instructions are generated. Specifically, the internal processing step, as a security pre-module running before the main process of economic dispatch, such as Unit Commitment (UC) or Economic Dispatch (ED), has the following internal operating procedures and information flow configuration.

[0022] The internal processing step first executes an adaptive correction step for the constraint benchmark, i.e., step a. This step is set because the overall health status or completeness of the defense system is dynamically changing, and a fixed rate of change of frequency (RoCoF) safety threshold may be insufficient when the defense system is compromised. Therefore, the constraint benchmark adaptive correction unit in this scheme is configured to collect the status information of key defense resources in the defense system in real time and periodically (e.g., every 5 minutes) through the EMS (Energy Management System) or SCADA (Supervisory Control and Data Acquisition) network. The defense resources specifically include, but are not limited to, the online / offline / maintenance status of low-frequency load shedding devices (UFLS) deployed in various substations, and the commissioning / decommissioning status of various generator frequency controllers (such as primary frequency regulation). When executing the adaptive correction step a for the constraint benchmark, in order to improve computational efficiency and ensure that this step matches the physical characteristics of a large-scale power grid, multiple data points in the real-time aggregated power system are used. The status information of defense resources specifically refers to the status of resources in a pre-determined list of key defense resources. The generation procedure for this list is as follows: during the offline calibration phase, a sensitivity analysis calculation is performed individually for each defense resource in the power grid benchmark topology using a power system simulation platform. This calculation quantifies the percentage of deterioration impact on the maximum frequency change rate RoCoF of the system under a preset N-1 disturbance by simulating the single resource's shutdown. All defense resources whose percentage of deterioration exceeds a predetermined engineering threshold, such as 2%, are selected and solidified into the list of key defense resources. The online calculation in step a only needs to aggregate the status of resources within this list to ensure that the dynamic compensation index accurately characterizes the completeness of the defense system that has a substantial impact on global frequency security. After obtaining this discrete status information, the unit executes a weighted summation logic based on security assurance operation rules to calculate a normalized dynamic compensation index characterizing the overall completeness of the power defense system. For example, its calculation procedure can be determined as follows: The weight and (For example The calibration coefficients are preset in the security rule base; subsequently, the unit calculates the dynamic compensation index (e.g., by using the dynamic compensation index). ), dynamically adjust a safety threshold for the rate of change of frequency used for subsequent constraint generation ( Its adjustment procedure was determined as follows: ,in, (For example This is a baseline threshold set by security assurance rules; thus, when the completeness of the defense system... Descending to The safety threshold for the rate of change of frequency is automatically tightened to This threshold will lead to a corresponding increase in the inertia requirement calculated subsequently, thereby using the increased inertia margin to compensate for the temporary reduction in downstream defense capabilities.

[0023] In parallel, the internal processing steps perform a closed-loop verification step for constraint parameters, namely step b. This step is set because the nameplate parameters or theoretical control values ​​(i.e., the reference inertia contribution constant) of the generator set may not be fully realized in the actual operation of a large-scale power grid due to control mode switching, abnormal equipment status, or communication delays. If the dispatching system trusts these theoretical values, the calculated total system inertia will be too high, posing a hidden danger to safety. Therefore, the constraint parameter closed-loop verification unit in this scheme is configured to utilize real-time data from a wide-area measurement system (WAMS) (e.g., from a PMU (Phasor Measurement Unit), refreshed at millisecond-level frequency, power angle, and active power of the unit). Phasor data), online identification and verification of the actual inertia contribution of the units already in operation ( In one implementation, the verification unit simultaneously captures the global frequency change rate when it detects a natural, minor disturbance in the power grid. ) and the power response of each unit ( ), and based on the swing equation of the power system ( Reverse identification ,in The rated frequency of the power grid (e.g.) In another preferred embodiment, the verification unit first performs step b-1, which involves monitoring the perturbation richness in the WAMS data (e.g., calculating the global frequency signal in...). The signal energy within the frequency band), when it is determined that the energy is below a confidence threshold set based on historical statistical characteristics (i.e., the power grid is operating smoothly, resulting in insufficient disturbance signals for passive identification), in addition, in the preferred embodiment of the active injection of power disturbance signals involved in the closed-loop verification step b of the constraint parameters, to ensure the safety margin of the verification operation itself for the large-scale power grid, this step also includes a pre-implementation online safety verification sub-step before injection. This sub-step uses real-time data from the wide-area measurement system to identify the real-time damping ratio of all current key low-frequency oscillation modes of the power grid through an online modal analysis algorithm. The constraint parameter closed-loop verification unit is configured to authorize the subsequent power disturbance signal injection operation only when the real-time damping ratio of all identified key oscillation modes is higher than a preset safety margin threshold, such as 5%. Furthermore, the specific frequency of the injected signal, for example... It is set to be far away from all identified low-damped oscillation mode frequency bands in the frequency domain, for example to The frequency band, thus achieving accurate parameter identification while avoiding the potential adverse effects of the verification process on the dynamic stability of the power grid, triggers step b-2, actively injecting a standardized, specific-frequency power perturbation signal (e.g., a...) into the large-scale power grid through a power system's rapid adjustment resources (e.g., a large HVDC (High Voltage Direct Current) transmission system or a BESS (Battery Energy Storage System) power station). , (a sine wave signal); synchronously, in step b-3, the verification unit uses a lock-in amplification or fast Fourier transform (FFT) algorithm to extract the signal from the WAMS data that is related to the sine wave signal. The global frequency response amplitude corresponding to a specific frequency ( ) and the power response amplitude of each unit ( Finally, in step b-4, based on the physical relationship between the response amplitudes (this physical relationship is characterized as follows: the larger the inertia of the unit, the larger its power response amplitude), Relative global frequency response amplitude (The larger), through identification rules Re-identify and verify the true inertia contribution ,in This is a coefficient pre-calibrated based on the physical characteristics of the power system; the verification unit obtains the true inertia contribution. (For example, After that, it is compared with the reference inertia contribution constant of the unit. (For example, The two values ​​are compared, and the smaller value is strictly selected (i.e., ), as the inertia contribution constant required for constraint implantation in subsequent step c ( Output.

[0024] Subsequently, the internal processing steps execute a constraint generation and implantation step, namely step c; the constraint generation and implantation unit in this scheme first receives the frequency change rate safety threshold obtained after adaptive correction in step a ( ,For example ), and inversely deduce the minimum system inertia baseline that must be met under the current operating conditions of a large-scale power grid ( ); This derivation step c1 first establishes the scenario where the power system encounters a preset disturbance (power imbalance is ); When ), the total inertia of the system ( ) and the rate of change of frequency ( A dynamic response relationship model between () and (), which can be calibrated as Step c2, based on the output of step a ( ), derived through reverse engineering of the model To ensure robustness of security, the derivation process is further refined: the unit first obtains the current power system topology from the EMS in real time; based on the real-time topology (e.g., knowing that a 750kV line A is under maintenance), it dynamically determines a list of one or more current credible disturbances (Nk) (e.g., disturbances involving line A are removed from the list, but a new, currently most serious disturbance, a single-phase ground fault on 750kV line B, is added). For each currently trusted perturbation in the list; Calculate the required perturbation-specific inertia for each. ; to have multiple The maximum value in the range is determined as the final minimum system inertia baseline. For example, if If it is the maximum value, then After obtaining Subsequently, the constraint generation and implantation unit inherits the inertia contribution constant selected after verification in step b. ), construct constraint equations; the form of the constraint equations is a decision variable based on the generator set's start-up state ( (an integer variable that is either 0 or 1) and Sum of the products, with the condition that the sum of the products is greater than or equal to 1. ,Right now: (For example, Finally, this unit incorporates the constraint equations as mandatory constraints into the economic scheduling optimization model of the (intelligent scheduling system).

[0025] In a preferred embodiment, to address the spatial distribution characteristics of large-scale power grids caused by long-distance AC transmission above 750kV, i.e., the risk of local inertia loss, steps a and c are further refined: the system utilizes dynamic physical characteristic analysis tools of the power system (such as modal analysis or electrical distance calculation) to divide the power grid into at least two dynamic inertia partitions (e.g., partition A and partition B connected by high-impedance tie lines) based on real-time topology; for each partition (e.g., partition A), it separately calculates and determines the most severe credible disturbance within the partition (e.g., the most severe credible disturbance within the partition). The minimum local inertia baseline corresponding to ) (For example, Accordingly, each minimum local inertia baseline is incorporated as a mandatory constraint into the economic scheduling optimization model. At this point, the incorporated constraint equation set takes the form of: (global constraint) (Local A constraint) (Local B constraint) This set of constraints ensures that the economic optimization results of the (intelligent dispatch system) not only satisfy the total global inertia, but also ensure that the inertia contribution of each critical local area in the large-scale power grid meets its respective physical baseline.

[0026] Finally, the internal processing steps execute a constraint optimization step, namely step d; the constraint optimization unit in this scheme (i.e., the economic scheduling (UC / ED) solver in the intelligent scheduling system) executes an economic scheduling optimization model with the objective function of minimizing total operating cost; this solver searches for the unit combination with the minimum cost ( When formulating power allocation schemes, the power balance constraints of the power grid, the start-up and shutdown constraints of generating units, and the mandatory constraints of one or more minimum systems and local inertia baselines embedded in this invention must be met simultaneously. If the lowest-cost scheme (e.g., extensive use of zero-inertia renewable energy sources) cannot meet these constraints... Constraints will force the optimization algorithm to automatically add ( ) to its decisions. One or more with high The synchronous generator sets are processed until all inertia constraints are satisfied, ultimately generating scheduling instructions that meet the physical safety baseline. In a preferred embodiment, this method also establishes a reverse information interaction path from the economic model to the safety analysis: in step e, during or after the constraint optimization in step d, the system automatically extracts the mandatory constraints related to the minimum system inertia baseline (e.g., from the solution results of the economic scheduling optimization model (i.e., the constraint optimization unit)). The corresponding shadow price () Step f, the shadow price ( As an economic cost indicator for security constraints, it is subject to post-verification. The physical meaning is to ensure that the power grid meets the requirement of increasing by 1 during the current scheduling period (e.g., 1 hour). The constraint of the inertia baseline, the marginal increase required to reduce the total system operating cost (unit: yuan / hour), its unit is... Step g, set an economic rationality threshold defined by (intelligent scheduling) rules (e.g., When economic cost indicators When the threshold is exceeded ( The dispatching and operation personnel who trigger alarms or prompt security assurances should manually review the constraint generation premises of step a (defense completeness assessment) and step c (trusted disturbance list) from a security assurance perspective to determine whether the current security constraints may be set too conservatively, resulting in unreasonable economic costs.

[0027] In another preferred embodiment, this method also constructs a collaborative mechanism for (intelligent scheduling) pre-emptive prevention and post-event defense: step f, reuses the inertia contribution constant (representing the total actual inertia of the grid's operating units) selected after verification in step b. Step g), will As a power system security status parameter, it is sent to the Low Frequency Load Shedding (UFLS) defense device in the defense system through the communication network of the (intelligent dispatch system); in step h, the UFLS device in the defense system has multiple preset action setting tables (e.g., three sets A, B, and C), corresponding to high, medium, and low inertia levels respectively; the UFLS device receives the data... The safety status parameters automatically select and activate a set of action settings that match the current actual inertia level of the power grid; for example, when the received... At lower levels, UFLS automatically switches to the more aggressive C-set setting table (which operates more frequently and cuts off larger loads), thereby ensuring that the defense strategy of the defense system always matches the actual operational risk (inertia level) of the power grid. This invention also provides a new energy power system optimization dispatching system, which implements any of the above methods. Its hardware architecture can be built based on existing (intelligent dispatching system) EMS / SCADA platforms, WAMS wide-area measurement systems, and computing server clusters. Its core lies in the fact that the system, in terms of functional modules, includes at least: a constraint benchmark adaptive correction unit, configured to perform the function of step a above, namely, aggregating the resource status of the defense system, calculating the dynamic compensation index, and dynamically adjusting the frequency change rate safety threshold; a constraint parameter... A closed-loop verification unit is configured to perform the function of step b above, that is, to identify and verify the actual inertia contribution of the unit online using WAMS real-time data in parallel, and to select the smaller value between the actual value and the benchmark value as the output inertia contribution constant; a constraint generation and implantation unit is configured to perform the function of step c above, that is, to derive the minimum system (and local) inertia baseline based on the corrected safety threshold and the dynamic credible disturbance list, and to construct constraint equations based on the verified inertia contribution constant, and to implant the constraint equations as mandatory constraints into the economic dispatch optimization model; and a constraint optimization unit is configured to perform the function of step d above, that is, to execute the economic dispatch optimization model to generate the final dispatch instructions under the constraints of the mandatory constraints.

[0028] Example 1: In a typical nighttime off-peak electricity operation of a large-scale power grid, the total grid load is low, while the new energy (wind power) is predicted to reach a high level. Simultaneously, the EMS energy management system status monitoring of the intelligent dispatch system shows that the low-frequency load shedding devices (UFLS) in two areas of the grid are offline due to communication link maintenance. A newly connected large wind farm reports that it has virtual inertia response capability, but its actual contribution to safety assurance has not yet been verified in operation. Before the start of this dispatch cycle, the economic dispatch optimization model of the intelligent dispatch system, based on its function aimed at minimizing total operating cost, initially tends to maximize the acceptance of wind power with near-zero marginal cost and plans to disconnect multiple (e.g., 3) synchronous generator units with high operating costs from the grid. However, in the optimization dispatch method and system provided by this invention, the execution of this economic dispatch optimization model is constrained by a preceding internal processing step. First, the constraint benchmark adaptive correction unit step a is triggered. This unit gathers the offline status information of the UFLS devices in real time, and calculates the dynamic compensation index based on this information. Descending to Accordingly, the unit automatically uses the safety threshold for the frequency change rate used to constrain the generation. From the benchmark Tighten to .

[0029] In parallel, step b of the constraint parameter closed-loop verification unit is initiated. This unit utilizes real-time data from the WAMS wide-area measurement system to verify the inertia contribution of all operational units, including the newly connected wind farm. Given the insufficient disturbance richness of the power grid under the current off-peak electricity conditions, this unit, according to the procedures in steps b-2 to b-4, injects a power perturbation signal of a specific frequency through the HVDC transmission system of a power system and simultaneously identifies the response. The identification result confirms the true inertia contribution of the new wind farm. Only the reported reference inertia contribution constant of The verification unit then proceeded to verify the wind farm station's... Used for subsequent calculations to determine its baseline value The verification value is used, while the smaller reference value is used for other synchronous generators that have passed verification. Subsequently, the constraint generation and implantation unit step c begins to execute. This unit first determines the most severe credible disturbance of the current large-scale power grid according to the real-time topology of the power grid and the procedures in steps a-2 to a-4. For example, a fault N-1 on a 750kV line corresponds to a power loss of This unit is used to derive the minimum system inertia baseline in reverse engineering. At that time, the material tightened in step a was used. ( The calculation process is as follows: This unit then constructs constraint equations, the sum of inertia on the left-hand side of which... In the middle, the new wind farm station It has been determined to be lower in step b. Verification value; ultimately, this value, generated based on the tightened safety threshold and the verified inertia contribution, is... The minimum system inertia baseline constraint is incorporated into the economic scheduling optimization model.

[0030] The constrained optimization unit, step d, begins execution. The economic dispatch optimization model, in finding the minimum cost, discovers that its initial plan to shut down three synchronous generators will result in the total grid inertia, especially after considering the low verification value of the wind farm, being far lower than... The algorithm imposes mandatory boundary constraints. To satisfy these safety constraints, the optimization algorithm automatically modifies its unit combination decision, retaining two of the three units originally planned to be shut down in its final generated scheduling instructions (these two units satisfy the mandatory boundary constraints). The most economical combination required by the constraints was used, while fully incorporating renewable wind power. During the execution of this dispatch instruction, the large-scale power grid encountered a 750kV line N-1 fault. Because the power grid retained [the necessary components]... The constraints guarantee Physical inertia, the rate of change of the power grid frequency is suppressed within The following provides the necessary time margin for UFLS devices in the defense system that are under maintenance and have reduced responsiveness, so that the grid frequency can be restored to stability without triggering UFLS action, thus avoiding a potential large-scale grid frequency collapse event.

[0031] Example 2: To objectively verify the synergistic effectiveness of the internal processing steps in the new energy power system optimization scheduling method of the present invention, this example conducts a comparative experiment on a large-scale high-fidelity digital simulation platform for the power grid. This simulation platform is based on a simplified power grid model containing a 750kV transmission section. The model includes four large synchronous generator units (G1-G4, with different operating costs and inertia constants), two large new energy (wind power) power plants (WF1 and WF2, both equipped with virtual inertia control functions), and an economic scheduling optimization model (ED solver) for the intelligent scheduling system. The experiment used a WAMS (Wide Area Measurement System) simulator and a model of the Low Frequency Load Shedding (UFLS) device in the defense system. The test scenario was set as a typical high-risk operating condition of a power grid: large-scale power generation of new energy sources (WF1 and WF2 are both close to full capacity), and the system load was at a near-low point. To assess the robustness of the method under complex boundary conditions, two key safety adverse factors were introduced into the experiment: abnormal status of defense resources, with a total of 30% of the UFLS device capacity in the defense system set to a communication interruption / out-of-operation state; and distortion of constraint parameters, with the controller of the new energy power station WF1 reporting its reference inertia contribution constant. for However, its actual inertia contribution in the simulation physics engine Set as (That is, only 40% of the reported value); In the test, a safe N-1 reference disturbance was used, that is, 10 seconds after the dispatch command was executed, a large synchronous generator unit (G1, power 1100MW) in the power grid was simulated to trip.

[0032] To verify the synergistic effect of various technical features in the internal processing steps of this invention, the following three test groups were set up for comparison and operated under the same test scenario and disturbance: Control group A: using the existing pure economic scheduling method without applying any minimum system inertia baseline constraint; Control group B: using a simplified static constraint method, which applies a minimum system inertia baseline constraint, but the generation of the constraint uses a fixed reference frequency change rate safety threshold. The adaptive correction in step a is not performed; the reference inertia contribution constant reported by the unit is used. Without performing the closed-loop verification in step b, the sample group of this invention adopts the complete optimization scheduling method provided by this invention, that is, before the execution of the economic scheduling optimization model, the internal processing steps are fully executed, including the adaptive correction of the constraint benchmark in step a, the closed-loop verification of the constraint parameters in step b, the generation and implantation of constraints in step c, and the constraint optimization in step d. The scheduling process and key performance data of the response after disturbance in the three experimental groups are recorded in Table 1.

[0033] Table 1: Comparison of system operation status and security performance under different scheduling methods.

[0034] Analysis of the experimental data in Table 1 shows that: Control group A, lacking any safety-guaranteed inertia constraints, had its economic dispatch optimization model focused solely on cost minimization, shutting down all high-cost synchronous generators except G1. This resulted in the actual system inertia of the power grid being only [a certain value]. After encountering a 1100MW disturbance, the system frequency change rate achieve Lowest frequency Fall to This caused the UFLS devices in the defense system, including those still in operation, to activate, triggering load shedding; the static constraint method used in control group B, due to its Fixed as The minimum system inertia baseline it calculates Only More seriously, step b of the method incorrectly relies on the WF1 report. Inertia affects its scheduling scheme; for example, retaining G2 units plus trusted WF1 meets the requirements on paper. The constraints, but the actual system inertia of its power grid is only This includes WF1 The actual inertia differs from the book value. Therefore, after the disturbance, the system Reached It exceeded its Despite the design goals, the power grid is still on the verge of frequency instability.

[0035] When the sample of this invention performs internal processing steps, firstly, step a (constraint datum adaptive correction) detects 30% of UFLS offline and calculates... for Then Automatic tightening to Simultaneously, step b (constraint parameter closed-loop verification) identifies, through WAMS simulation data, that the actual inertia contribution of WF1 is only [amount missing]. And select this As WF1 Subsequently, step c (constraint generation and implantation) uses these two revised and verified parameters to derive a higher and more physically accurate minimum system inertia baseline. In the constrained optimization step d, the economic scheduling optimization model is forced to contribute only to WF1. In this case, make up The total inertia forced it to automatically retain the two synchronous generator units, G2 and G3, so that the actual system inertia of the power grid reached Although the total dispatch cost (119,000) was higher than the control group, after encountering the same 1,100 MW disturbance, the power grid's... Suppressed in (lower than) (Tightening threshold), lowest frequency Stay At high levels, large-scale power grids maintained frequency stability; the experimental data shows that there is a synergistic effect between step a (adaptive correction threshold) and step b (closed-loop verification parameters) in the method of this invention, and neither can be omitted; step a ensures the bottom line of safety assurance. The calculations are able to respond to changes in the completeness of the defense system, while step b ensures the availability of power grid resources to meet this baseline. It is real and reliable; the two work together to make the constraints ultimately embedded in the economic scheduling optimization model. It can accurately reflect the physical security requirements of large-scale power grids, thereby ensuring the frequency security of the power grid at the source of economic dispatch.

[0036] Example 3: This example combines Figs. 1 to 3 A description of an optimized dispatching method and system for a new energy power system, such as... Fig. 1 As shown, its core lies in a pre-processing step that precedes the economic dispatch step d. This pre-processing step executes steps a and b in parallel. Step a is the adaptive correction of the constraint benchmark, which calculates the dynamic compensation index and dynamically adjusts the frequency change rate safety threshold based on the aggregated power defense resource status information, such as the status of the low-frequency load shedding device and the generator frequency controller. Step b is the closed-loop verification of the constraint parameters, which uses real-time data from the Wide Area Measurement System (WAMS), such as millisecond-level phasor data from the PMU, to identify and verify the actual inertia contribution of the already started units online, and selects the smaller value between the actual value and the benchmark value. The actual inertia parameters verified in step b are sent to the low-frequency load shedding defense device U. FLS automatically selects and activates the matching action setpoint table. On the other hand, it is incorporated into step c along with the safety threshold corrected in step a. Step c is constraint generation and implantation. Based on the corrected threshold and verified inertia, it reverse-derives the minimum system inertia baseline, constructs constraint equations and implants them into the model. Subsequently, step d performs constraint optimization, that is, it executes an economic scheduling optimization model with the goal of minimizing total operating cost, while satisfying the implanted inertia baseline constraint. Finally, it generates a scheduling instruction that satisfies both safety and economic synergy. This method also includes a reverse verification path, which extracts the shadow price from step d as the economic cost indicator of the safety constraint. When the indicator exceeds the threshold, an alarm is triggered or a manual review is prompted; otherwise, the verification passes.

[0037] like Fig. 2 As shown in the figure, the horizontal axis represents time in seconds (s), and the vertical axis represents system frequency in hertz (Hz). The curves show that all groups were at 50.0 Hz before the disturbance. After the disturbance, the unconstrained control group A experienced the deepest frequency drop, reaching a minimum of 49.08 Hz at 5 seconds. The statically constrained control group B experienced the second deepest frequency drop, reaching a minimum of 49.34 Hz ​​at 5 seconds. The sample group of this invention, due to the implementation of the optimized scheduling method of this invention, experienced the shallowest frequency drop, with the minimum frequency only decreasing to 49.55 Hz at 5 seconds, and maintained the highest frequency throughout the entire response process. Fig. 3 As shown, the system architecture is logically divided into three layers: the data source and physical device layer, the core application service of the intelligent scheduling system, and the controlled object and collaboration layer. The data source and physical device layer includes a wide-area measurement system, which provides real-time phasor data, as well as power defense resources, such as low-frequency load shedding devices and generator frequency controllers. The intelligent scheduling system, as the core application service, includes four key units: a constraint benchmark adaptive correction unit, which aggregates the status information of power defense resources and receives data from the wide-area measurement system; a constraint parameter closed-loop verification unit, which also receives real-time phasor data from the wide-area measurement system; a constraint generation and implantation unit, which receives the outputs of the first two units; and a constraint optimization unit, which executes the economic scheduling model and generates the final constraint. The instructions, controlled objects, and coordination layer include already operational generating units (including synchronous units), new energy power plants, fast-regulating resources (including high-voltage DC), battery energy storage systems, and low-frequency load shedding defense devices. The information flow path (solid arrows) shows that the constraint benchmark adaptive correction unit obtains data from power defense resources and the wide-area measurement system, while the constraint parameter closed-loop verification unit obtains data from the wide-area measurement system. The processing results of both are passed to the constraint generation and implantation unit, and finally the constraint optimization unit generates scheduling instructions. At the same time, there is also a collaborative control path (dashed arrows) in the system. The constraint parameter closed-loop verification unit can control fast-regulating resources, such as for actively injecting perturbations, and send the verified security status parameters to the low-frequency load shedding defense device.

[0038] Example 4: In the engineering calibration phase before deploying a new energy power system optimization dispatching system and its method to a smart dispatching system of a large-scale power grid containing 750 kV AC transmission sections, to ensure the objectivity and reproducibility of the setting of key parameters in the method, an offline calibration procedure based on the high-fidelity digital simulation platform of the power grid is executed. This procedure aims to provide a systematic parameter tuning basis for the constraint benchmark adaptive correction logic in step a and the constraint parameter closed-loop verification logic in step b. The calibration procedure first targets the dynamic compensation index in step a. Calculation logic and safety threshold for reference frequency change rate The first step in calibration is to set the overall integrity of the power defense system to an ideal state on the simulation platform, that is... Subsequently, simulations were performed one by one on the list of credible disturbances dynamically determined in step a-2. For each disturbance, the total inertia of the system was repeatedly adjusted using the bisection method. This continues until a critical inertia value is found, at which the low-frequency load shedding device (UFLS) can just stabilize the system frequency without causing it to collapse. Based on this critical inertia value and the corresponding power imbalance... Using dynamic response relationship model The required inverse derivation of this disturbance The threshold is determined by iterating through all credible perturbations and selecting the most stringent (i.e., smallest) one. Threshold, here is and set one The safety margin is used to determine the safety threshold of the reference frequency change rate. The second step is to calibrate the weighting coefficients of the dynamic compensation index. (corresponding to UFLS) and (Corresponding to the generator frequency controller), in the simulation platform, the predetermined... As a baseline, a scenario where defensive resources are damaged is simulated, and the following settings are made: UFLS capacity offline and The generator frequency controller exits, i.e., UFLS completeness. PFR completeness If the system still uses calculate It will become unstable when encountering disturbances. The calibration goal is to find a set of... and ( ), making And based on this After tightening Calculated This allows the damaged system to recover stability when encountering the same disturbance, while minimizing the increase in total scheduling cost. from to Gradient optimization tests were conducted to ultimately determine... and These are the optimal calibration coefficients.

[0039] In parallel, the calibration procedure calibrates the closed-loop verification logic of the constraint parameters in step b. The first step is to calibrate the calibration coefficients in step b-4. In the simulation platform, a synchronous generator set is selected, and its inertia contribution is set to a known and precise reference value. (Based on its nameplate parameters) Definite Subsequently, the active injection procedure of step b-2 is executed in the simulation system to inject into the system. The power perturbation signal is synchronously used to extract the power response amplitude of the unit from the simulated WAMS data. With global frequency response amplitude Calculate its response ratio Calibration coefficient That is, it was determined to be The second step involves calibrating the credibility threshold used in step b-1 to determine data credibility. On the simulation platform, a long-term dynamic simulation covering 720 consecutive hours, including typical daily load curves and new energy fluctuations, is run. During this process, the WAMS simulator is used to calculate the disturbance richness (global frequency signal in...) within a 1-minute time window. (signal energy within the frequency band), thereby obtaining A batch of perturbation richness samples were used to perform statistical analysis on these samples and generate their cumulative distribution function (CDF). The confidence threshold was determined as the CDF's [value]. The 10th percentile (i.e., the 10th percentile) is considered to be below this threshold, indicating that the data identified passively is not reliable enough and active injection should be triggered.

[0040] Finally, the calibration procedure addresses the economic rationality threshold in step g. Calibration was performed on a simulation platform, constructing 50 typical power grid operating conditions covering high risk and high cost throughout the year. For each condition, constraint optimization step d was executed, and the shadow price corresponding to the minimum system inertia baseline mandatory constraint condition was extracted from the solution results. , these 50 The values ​​are sorted, and the economic rationality threshold is determined. Set as this dataset Quantity (95th percentile), its value is This threshold represents the minimum threshold required to ensure power grid safety. The highest marginal economic cost incurred by the system in a typical high-risk scenario; anything exceeding this threshold... All values ​​are considered extremely low-probability events, triggering manual review; by executing the above calibration procedures, , , , Credibility threshold and All values ​​were determined to have clear engineering and statistical basis, and these values ​​were subsequently written into the configuration file of the new energy power system optimization dispatch system as the benchmark for executing steps a, b, and g during its online operation.

[0041] Example 5: During the on-site commissioning phase before the new energy power system optimization dispatching system is first put into operation in a real large-scale power grid, which includes multiple 750 kV AC transmission lines, in order to verify and correct the benchmark parameters on which the constraint parameter closed-loop verification unit relies, this unit executes the function of step b. The system executes a preliminary on-site calibration procedure. When this procedure is started, the system is placed in a passive monitoring and analysis mode, and its constraint output is not transmitted to the economic dispatch optimization model. Subsequently, the system operators, through the rapid adjustment resources of the power system, specifically a high-voltage direct current (HVDC) transmission system, inject a known power amplitude into the actual power grid according to the procedure in step b-2. , frequency is The power perturbation signal; during signal injection, the constraint parameter closed-loop verification unit synchronously utilizes WAMS real-time data to execute the identification logic in steps b-3 and b-4. For all generator units in the power grid that are already in operation and have inertia contribution capabilities, this type of unit includes synchronous machines and new energy power stations with virtual inertia, and calculates their power perturbation signal one by one. and And utilizing the offline calibration already performed in Example 3 The coefficient is used to calculate the coefficient for each unit. The measured value is used in conjunction with the reference inertia contribution constant derived from the nameplate parameters stored in the system. In comparison, for any given unit, especially a 750 kV synchronous generator unit with a long service life, its The measured values ​​were systematically lower than their values ​​in multiple repeated trials. The value, when it falls below a set threshold, is [value here]. Value At that time, the system will automatically switch to the unit. The value was adjusted down and updated to this field-verified value. The implementation of this procedure ensures that the reference inertia contribution constant of each unit stored in the system is calibrated based on the actual grid response.

[0042] To establish the rule base necessary for dynamically determining the current credible disturbance list in step a-2 during the constraint generation and implantation steps (i.e., step c), the system performs a knowledge base construction based on power grid security analysis procedures during the offline phase. This construction process utilizes a power system simulation platform containing a complete power grid topology and dynamic component model. The safety assurance department's operation procedure personnel set a series of benchmark topologies covering typical operating modes of large-scale power grids. These benchmark topologies are set as a series of typical operating conditions, such as high summer load, maintenance of 50 kV line A, or low winter power, large-scale renewable energy generation, and single-line operation of 750 kV line B. For each benchmark topology, the simulation platform automatically traverses and calculates all standard N-1 disturbances, covering generator tripping and line tripping, and the corresponding power imbalance. In addition, based on the settings of the security rule base, the platform also calculates specific Nk disturbances, specifically N-2 disturbances caused by 750 kV double-circuit faults on the same tower or circuit breaker failures, and calculates their... This procedure will integrate all these topologies, disturbances, and The correspondence between the three is stored in the internal database of the constraint generation and implantation unit in step c, in the form of a structured lookup table. Thus, when the method runs online, the real-time topology obtained in step a-1 can be used to instantly match the baseline topology in this database, enabling step a-2 to query and generate a list of current credible disturbances matching the current operating condition, containing N-1 and critical Nk risks, and their corresponding... value.

[0043] Example 6: In an offline simulation phase for calibrating the switching threshold of multiple action setting tables for the UFLS (Underlying Frequency Load Shedding) low-frequency load shedding defense device, this procedure aims to provide a reproducible inertia level benchmark for the automatic selection logic in step h. The UFLS device has three preset action setting tables: A, for high inertia conditions with conservative action setting values; B, for medium inertia conditions; and C, for low inertia conditions with aggressive action setting values. This calibration procedure first uses a defined bottom line for safe grid operation as a constraint target, which is set here to be the threshold when encountering the most severe credible disturbance. The system's lowest frequency is 1100MW. The frequency must not be lower than 49.0Hz to determine the switching threshold between setting tables A and B. In the simulation platform, the system is configured to use only the constant table A. Subsequently, the system employs a bisection method to iteratively adjust the total inertia of the system in the simulation. This continues until a critical inertia value is found, at which the UFLS operation in setpoint table A can just achieve the desired result. Stabilizing at 49.0 Hz, the critical moment of inertia was determined to be... This means that when the system inertia is higher than At that time, only setting table A is needed to meet the safety baseline. In parallel, the switching threshold between setting tables B and C is determined. The system sets UFLS to use only the constant table C and uses the same bisection method to iteratively adjust it again. Find a value such that the constant table C exactly makes The critical moment of inertia that stabilizes at 49.0 Hz is determined to be... ,final, and It is embedded into the control logic of the UFLS device, serving as the quantitative basis for determining the current inertia level and activating the corresponding setpoint table in step h.

[0044] In an offline calibration procedure for the partition thresholds required by the dynamic inertia partitioning algorithm in the calibration step, the procedure aims to provide a reproducible electrical distance criterion based on the physical characteristics of the power grid for the hierarchical clustering algorithm. The system first uses a power system simulation platform to perform a modal analysis on a benchmark power grid topology containing a 750 kV main grid to identify all low-frequency oscillation modes below 1.0 Hz. For the most critical interval oscillation mode with the highest oscillation energy, the system identifies two sets of generator groups oscillating against each other under this mode. These two sets of generator groups are defined as the benchmark dynamic inertia partitions A and B. Subsequently, the procedure calculates the electrical distance between all generator pairs within partition A. The average value was calculated, and the electrical distance between generator pairs in all zones A and B was calculated. The average value, to obtain a criterion with discriminative margin, is the partition threshold. Set as a weighted median of the two averages, its calculation is determined as follows: ,this The value is then embedded into the constraint generation and implantation unit, serving as the quantization threshold for dividing dynamic inertia partitions when the hierarchical clustering algorithm is executed in step ax.

[0045] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing the dispatch of a new energy power system, characterized in that, This includes an internal processing step for generating and verifying power system security constraints before the execution of the economic dispatch optimization model. This internal processing step includes: Step a: In real time, gather the status information of multiple defense resources in the power system, including low-frequency load shedding devices and generator frequency controllers; calculate a dynamic compensation index based on the status information to characterize the overall integrity of the power defense system; and dynamically adjust a frequency change rate safety threshold used to constrain the generation based on the dynamic compensation index, the rule of which is that the frequency change rate safety threshold is automatically tightened when the overall integrity decreases. Step b: Using real-time data from the wide-area measurement system, identify and verify the actual inertia contribution of the powered-on units online; compare the actual inertia contribution with a reference inertia contribution constant determined based on the nameplate parameters or control mode of the powered-on units, and select the smaller value between the actual inertia contribution and the reference inertia contribution constant as the inertia contribution constant for constraint implantation in step c. Step c: Based on the frequency change rate safety threshold obtained after adaptive correction in step a, the minimum system inertia baseline is derived in reverse; and the constraint equation is constructed based on the inertia contribution constant selected after verification in step b; the constraint equation is then incorporated into the economic scheduling optimization model as a mandatory constraint condition. Step d: Execute the economic scheduling optimization model to generate scheduling instructions under the constraints of mandatory conditions.

2. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, The step of reverse derivation of the minimum system inertia baseline in step c further includes: step a-1 obtaining the current power system topology in real time; step a-2 dynamically determining a list of one or more current credible disturbances based on the real-time topology; step a-3 calculating the required disturbance-specific inertia demand for each current credible disturbance in the list; and step a-4 determining the maximum value among the multiple disturbance-specific inertia demands as the minimum system inertia baseline.

3. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, Steps a and c further include: dividing the power system into at least two dynamic inertia partitions based on its dynamic physical characteristics; calculating and determining the corresponding minimum local inertia baseline for each of the at least two dynamic inertia partitions; and in step c, embedding each minimum local inertia baseline as a mandatory constraint into the economic dispatch optimization model to ensure that the sum of inertia contributions within each dynamic inertia partition meets its respective baseline.

4. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, The closed-loop verification step b of the constraint parameters further includes: step b-1 monitoring the disturbance richness in the real-time data of the wide-area measurement system to determine the data credibility of the true inertia contribution; step b-2 when the data credibility is determined to be lower than a credibility threshold set according to the historical statistical characteristics of the real-time data of the wide-area measurement system, actively injecting a standardized power perturbation signal of a specific frequency into the power system through a power system fast adjustment resource; step b-3 synchronously extracting the global frequency response amplitude corresponding to the specific frequency of the power perturbation signal from the real-time data of the wide-area measurement system. and the power response amplitude of each unit Step b-4 is based on the global frequency response amplitude. With the power response amplitude of each unit The physical relationship between them was re-identified and verified to determine the true contribution of inertia. Its identification rules follow: ,in, It is a calibration coefficient characterizing the physical properties of the power system; and by using the re-identified and verified real inertia contribution, the comparison and selection rules defined in step b of the closed-loop verification of the constraint parameters are executed.

5. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, This method also includes: reusing the inertia contribution constant selected after verification in step b, the inertia contribution constant representing the total actual inertia of the operating units in the power system; sending the inertia contribution constant as a power system security status parameter to the low-frequency load shedding defense device in the power system; the low-frequency load shedding defense device automatically selects and activates a set of action setting values ​​that matches the current inertia level from multiple preset action setting value tables based on the received power system security status parameter.

6. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, In step a, the step of calculating the dynamic compensation index includes: using a weighted summation logic based on power system expert rules to aggregate the discrete state information of multiple defense resources into a single normalized value as the dynamic compensation index.

7. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, The step of reverse derivation of the minimum system inertia baseline in step c includes: step c1 establishing a dynamic response relationship model between the total system inertia and the frequency change rate when the power system encounters a preset disturbance; step c2 based on the frequency change rate safety threshold obtained after adaptive correction in step a, the minimum system inertia baseline is derived in reverse through the dynamic response relationship model.

8. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, The constraint equation in step c is in the form of a sum of the product of the generator set start-up state decision variable and the inertia contribution constant selected after verification in step b, and the sum of the product is limited to be greater than or equal to the minimum system inertia baseline derived in reverse in step c.

9. The optimized dispatching method for a new energy power system according to claim 1, characterized in that, The economic dispatch optimization model is an optimization model with the objective function of minimizing total operating cost. Its key feature is that the method further includes: extracting shadow prices corresponding to the minimum system inertia baseline mandatory constraint from the economic dispatch optimization model during or after the constraint optimization process in step d; using the shadow prices as an economic cost indicator for power system security constraints for post-verification; and triggering an alarm or prompting for manual review of the constraint generation in steps a and c when the economic cost indicator exceeds an economic rationality threshold set by the power system dispatch rules.

10. A new energy power system optimization dispatching system, used to implement the new energy power system optimization dispatching method of claim 1, characterized in that, include: The constraint benchmark adaptive correction unit is configured to: collect the status information of multiple defense resources in the power system in real time, including low-frequency load shedding devices and generator frequency controllers. A dynamic compensation index characterizing the overall integrity of the power defense system is calculated based on state information; and a frequency change rate safety threshold used to constrain the generation is dynamically adjusted based on the dynamic compensation index, with the rule that the frequency change rate safety threshold is automatically tightened when the overall integrity decreases. The constraint parameter closed-loop verification unit is configured to: use real-time data from the wide-area measurement system to identify and verify the actual inertia contribution of the powered units online. The actual inertia contribution is compared with a reference inertia contribution constant determined based on the unit's nameplate parameters or control mode, and the smaller value between the actual inertia contribution and the reference inertia contribution constant is selected as the inertia contribution constant output required for subsequent constraint implantation. The constraint generation and implantation unit is configured to: derive the minimum system inertia baseline by reverse-engineering the frequency change rate safety threshold output by the constraint benchmark adaptive correction unit; construct constraint equations based on the inertia contribution constant output by the constraint parameter closed-loop verification unit; and implant the constraint equations as mandatory constraints into an economic scheduling optimization model. The constrained optimization unit is configured to execute an economic scheduling optimization model to generate scheduling instructions under the constraints of mandatory constraints.

Images