Distributed resource frequency modulation active support optimization method and system considering time delay characteristics
By constructing an attention scoring function and a multi-objective optimization model, the power flow step size is dynamically adjusted, and the droop coefficient and reactive power compensation of distributed resources are optimized. This solves the frequency regulation problem caused by the time delay characteristics in traditional power systems, and achieves efficient frequency stability and economical operation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-24
AI Technical Summary
With a high proportion of renewable energy integration, traditional power systems cannot meet the requirements of real-time coordinated control through frequency regulation methods. Time delay differences lead to a drop in the minimum frequency point and large overshoot. Furthermore, traditional fixed-step simulation has a contradiction between accuracy and speed, making it difficult to meet the frequency stability and computational efficiency requirements of the power system.
By constructing a physical-guided attention scoring function that correlates frequency change rate with voltage deviation, dynamically adjusting the power flow calculation step size, and combining it with a multi-objective optimization model, the droop coefficient and reactive power compensation command of each heterogeneous resource are optimized to achieve adaptive variable step size mesh partitioning, thus solving the control problem caused by time delay characteristics.
It achieves efficient decoupling of frequency response solution and power flow calculation in distribution networks, resolves the contradiction between long simulation time in the full time domain and the risk of missing transient states in large-step simulation, and improves frequency stability and operational economy.
Smart Images
Figure CN122456533A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system frequency regulation control technology, and in particular to a distributed resource frequency regulation active support optimization method and system that considers time delay characteristics. Background Technology
[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.
[0003] As the penetration rate of new energy sources such as photovoltaics continues to increase, the traditional power system will gradually evolve into a new type of power system dominated by new energy sources. The large-scale integration of new energy sources has led to insufficient grid regulation capabilities and reserve capacity. The strong randomness of new energy output has also made frequency issues increasingly severe. Traditional frequency regulation methods can no longer meet the frequency regulation requirements of the new power system.
[0004] In recent years, with the high proportion of distributed resources such as photovoltaics, energy storage, and electric vehicles being connected to the distribution network, fully tapping and mobilizing the regulation potential of user resources on the distribution network side has become an important development direction for the primary frequency regulation of the future power system.
[0005] However, communication and physical execution in distributed resources involve time delays, and these delays vary significantly among different distributed resources. Given the same droop coefficient, these delays cause a significant drop in the frequency minimum. The greater the delay, the greater the overshoot during frequency recovery, which may induce system instability in severe cases. Furthermore, traditional fixed-step simulations present a trade-off between accuracy and speed. When jointly solving the dynamic frequency response and spatial power flow of the distribution network across the entire time domain, they face severe "curse of dimensionality" and computational timeouts, making it difficult to meet the requirements of real-time collaborative control. Summary of the Invention
[0006] To address the aforementioned issues, this invention proposes a distributed resource frequency regulation active support optimization method and system that considers time delay characteristics. By constructing a physical guidance attention scoring function based on the frequency change rate and voltage deviation, the time step of the power flow calculation in the distribution network is dynamically adjusted. Under the variable step-size solution framework, the optimal droop coefficient and reactive power compensation command for each differentiated resource are obtained through a multi-objective function that integrates minimizing network loss, minimizing overshoot, and maximizing frequency safety margin, thereby obtaining the final frequency regulation control strategy.
[0007] In some implementations, the following technical solutions are adopted: A distributed resource frequency modulation active support optimization method considering latency characteristics includes: Obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources; The minimum frequency point of the system is determined by pre-solving the frequency response model of the system by maximizing the droop control coefficient. Attention scoring is triggered at the lowest frequency point, and the power flow solution step size for the next moment is dynamically calculated based on the attention score; before the lowest frequency point, a preset benchmark power flow solution step size is used. A multi-objective collaborative optimization model considering time-series complementarity is constructed. Under the aforementioned power flow solution step size, the system frequency response model and the distribution network spatial power flow algebraic equation are jointly solved to obtain the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
[0008] As a further solution, a system frequency response model considering the delay characteristics of heterogeneous resources is constructed, specifically including: Establish the system equivalent power balance equations that include distributed heterogeneous resources of the distribution network; A time delay element is introduced into the converter interface device to construct a power response transfer function with time delay.
[0009] As a further approach, attention scoring is triggered at the point of lowest frequency, specifically as follows: ; in, To score attention, σ (·) is the activation function. Let be the rate of change of frequency at time t. To estimate the voltage value, V ref This is the voltage reference value. As a safety threshold, , These are the adjustable gain coefficients.
[0010] As a further approach, the power flow solution step size for the next time step is dynamically calculated based on the attention score, specifically as follows: ; ; Where, Δ T net,t+1 For the next moment's trend step size, Δ T base To preset the baseline step size, β This is an adaptive step size scaling factor. To score attention, The maximum allowable step size for integral solving is preset. This is the preset minimum allowable solution step size.
[0011] As a further solution, the multi-objective collaborative optimization model considering temporal complementarity specifically includes: Consider the first objective function of minimizing network loss and overshoot, and the second objective function of maximizing the system frequency safety margin.
[0012] As a further solution, the first objective function is specifically: ; in, u 1 is the first objective function; and Δ Q , where are decision variables, representing the optimized active power droop coefficient and reactive power compensation power of each heterogeneous resource, respectively; T This is the total control period; N This represents the total number of system nodes. δ ( i ) for nodes i A set of connected nodes; I ij branch road ij The square of the current amplitude on the current is the intermediate variable after second-order cone relaxation; r ij branch road ij The resistance; f The actual frequency of the system. C 1 and C 2 represents the weighting coefficients for minimizing network loss and minimizing frequency overshoot, respectively.
[0013] As a further solution, the second objective function is specifically: ; in, Let this be the first objective function; and Δ Q , where are decision variables, representing the optimized active power droop coefficient and reactive power compensation power of each heterogeneous resource, respectively; The reference frequency is 50Hz. This represents the lowest point of frequency drop.
[0014] In other embodiments, the following technical solutions are adopted: A distributed resource frequency modulation active support optimization system considering latency characteristics includes: The system frequency response model construction module is configured to obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources. The model pre-solution module is configured to pre-solve the system frequency response model by maximizing the droop control coefficient to determine the system's minimum frequency point; An adaptive step size construction module is configured to trigger attention scoring at the lowest frequency point, dynamically solve the power flow solution step size for the next moment based on the attention score, and use a preset benchmark power flow solution step size before the lowest frequency point. The spatiotemporal joint solution module is configured to construct a multi-objective collaborative optimization model that considers temporal complementarity. Under the aforementioned power flow solution step size, it jointly solves the system frequency response model and the distribution network spatial power flow algebraic equation to obtain the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
[0015] In other embodiments, the following technical solutions are adopted: A terminal device includes a processor and a memory, the processor being used to implement instructions; the memory being used to store multiple instructions, the instructions being adapted to be loaded and executed by the processor to perform the aforementioned distributed resource frequency modulation active support optimization method considering latency characteristics.
[0016] In other embodiments, the following technical solutions are adopted: A computer-readable storage medium storing a plurality of instructions adapted for loading and execution by a processor of a terminal device of the above-described distributed resource frequency modulation active support optimization method considering latency characteristics.
[0017] Compared with the prior art, the beneficial effects of the present invention are: This invention efficiently decouples the complex frequency response solution process from power flow calculation in distribution networks. By maximizing the droop control coefficient, it rapidly pre-solves the system frequency response model considering the time delay characteristics of heterogeneous resources under no power flow constraints, obtaining the lowest system frequency point. At the lowest frequency point, an attention mechanism is triggered, and the power flow solution step size for the next moment is dynamically solved based on the attention score, realizing adaptive variable step size mesh partitioning. This effectively resolves the contradiction between the extremely long time consumption of full-time domain fine simulation and the tendency of large step size simulation to miss transient crises, while also overcoming the control difficulties caused by the response delay of different devices.
[0018] This invention constructs a multi-objective function that minimizes overall network loss, overshoot, and frequency safety margin. Under an adaptive variable-step-size grid, it jointly solves the system frequency response model and the spatial power flow algebraic equations of the distribution network to formulate the final frequency regulation control strategy. This achieves efficient and accurate solving of the multi-timescale frequency-power flow coupling model of the distribution network and optimal coordination of heterogeneous time-delay resources.
[0019] Other features and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description
[0020] Figure 1This is a flowchart of the distributed resource frequency modulation active support optimization method considering latency characteristics in an embodiment of the present invention; Figure 2 This is the adaptive variable-step long-time mesh generation result in an embodiment of the present invention; Figure 3 This is a schematic diagram of the frequency modulation control results of heterogeneous resources in an embodiment of the present invention. Detailed Implementation
[0021] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0022] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0023] Example 1 In one or more embodiments, a distributed resource frequency modulation active support optimization method considering latency characteristics is disclosed, combined with... Figure 1 Specifically, it includes the following process: S101: Obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources.
[0024] Specifically, system frequency response models that consider the delay characteristics of heterogeneous resources mainly include: Establish the equivalent power balance equation for a system that includes heterogeneous resources such as thermal power units, photovoltaics, energy storage, controllable loads, and electric vehicles: (1) in, H Let Δ be the system's inertial constant. P m For the active power increment including thermal power, photovoltaic power, energy storage, controllable loads, and electric vehicles, Δ P e For load disturbance, D This is the load damping coefficient. This represents the frequency change value.
[0025] Meanwhile, time delay elements are introduced for inverter interface equipment such as energy storage, photovoltaics, and electric vehicles. e -sτConstruct the power response transfer function with time delay: (2) (3) in, , , These represent the droop coefficients for photovoltaics, energy storage, and electric vehicles, respectively. , , These represent the time constants for photovoltaics, energy storage, and electric vehicles, respectively. , , , These represent the delay parameters for photovoltaics, energy storage, electric vehicles, and controllable loads, respectively. i = , , .
[0026] The above formulas (1), (2), and (3) constitute the system frequency response model considering the delay characteristics of heterogeneous resources.
[0027] S102: Determine the minimum frequency point of the system by pre-solving the frequency response model of the system by maximizing the droop control coefficient; S103: Attention scoring is triggered at the lowest frequency point, and the power flow solution step size for the next moment is dynamically calculated based on the attention score; before the lowest frequency point, a preset benchmark power flow solution step size is used.
[0028] The time delay introduces a linearly increasing phase lag in the frequency domain, making the closed-loop system more prone to instability or oscillation. Before the control takes effect, the frequency will deteriorate in the opposite direction, causing the minimum frequency point to drop significantly. Furthermore, due to the phase lag, even when the frequency begins to recover, it may continue to apply reverse or excessive regulation, causing the frequency to exceed the steady-state value, resulting in a larger overshoot, which may be accompanied by multiple oscillations and decays.
[0029] To address the aforementioned issues, this embodiment maximizes the support capability of aggregated heterogeneous resources, rapidly pre-solves the frequency response differential model without power flow constraints, and accurately locates key vulnerable feature points such as the lowest system frequency. It also utilizes a physically guided attention mechanism for dynamic mesh partitioning, thereby constructing an adaptive variable-step long-time mesh. At the resource allocation mechanism level, it leverages low-latency resources for rapid response, while during the frequency recovery period, the power of high-latency resources is increased, and low-latency resources are adjusted based on the system frequency deviation and the power of high-latency resources.
[0030] Specifically, addressing the issues of massive computational complexity and difficulty in balancing transient accuracy and solution speed in existing fixed-step full-time domain integral simulations, this embodiment achieves rapid pre-solution of the frequency response differential model without power flow constraints by maximizing the droop control coefficient K. This process removes the time-consuming spatial power flow algebraic equation constraints, thereby enabling rapid and accurate location of the system's vulnerable moments, i.e., the lowest frequency points. t nadir .
[0031] Using the lowest frequency point as a reference, adaptive variable-step long-term mesh generation is performed, as follows: At the lowest frequency point t nadir Previously, a preset baseline power flow solution step size was used; at the point of lowest frequency... t nadir Then, the power flow solution step size for the next time step is dynamically calculated based on the attention score; Figure 2 An example of adaptive variable-step long-time mesh generation is given.
[0032] As a specific implementation method, this embodiment constructs a physically guided attention scoring function. α t The intensity of the physical process is mapped to weighting coefficients: (4) in, σ (·) is the activation function that maps physical quantities to weights; denoted as the rate of change of frequency at time t; the faster the change, the higher the attention level. V es,t To estimate the voltage value; V ref This is the voltage reference value; ε A safety threshold is set; no attention is allocated within the safe range. , These are adjustable gain coefficients, whose main function is to act as scaling factors to map and unify the two heterogeneous physical quantities of frequency and voltage onto the same numerical scale, enabling them to be reasonably linearly superimposed before being fed into the activation function.
[0033] Based on the attention score, the step size for solving the power flow at the next moment is dynamically defined as follows: (5) (6) in, For the next moment's trend step length; This is the preset baseline step size; βThis is the adaptive step size scaling factor. The larger the value, the more stringent the local truncation error control in the critical region and the denser the computational grid. A score is given for attention. The maximum allowable step size for integral solving is preset. This is the preset minimum allowable solution step size.
[0034] In this embodiment, the physical-guided attention scoring function integrates the frequency change rate and the predicted voltage deviation; the greater the frequency change rate or the further the predicted voltage deviates from the safety threshold, the higher the attention score; by using an adaptive variable-step long-time grid, a fine time step is used to solve the differential and algebraic equations together at the lowest frequency point; and a coarse-resolution solution step is used during the steady-state period to reduce the amount of computation, thereby significantly eliminating redundant computation.
[0035] When solving simultaneously across multiple time scales, at the point of lowest frequency t nadir Afterwards, the system followed The simulation time is advanced. Within this time step, the dynamic state of each distributed resource and the active / reactive output boundaries are updated by solving the system frequency response model. Then, the updated power injection is used as the boundary condition of the distribution network spatial power flow algebraic equation. This step size is used to trigger the spatial power flow calculation of the distribution network, thereby realizing the variable step size alternating iterative solution of dynamic frequency in the time dimension and static power flow in the spatial dimension.
[0036] This embodiment effectively resolves the contradiction between the extremely long time consumption of full-time domain fine simulation and the tendency of large-step-size simulation to miss transient crises by using adaptive variable-step-size mesh partitioning, thereby improving the frequency stability and overall operational economy of the distribution network under high-proportion heterogeneous resource access.
[0037] S104: Construct a multi-objective collaborative optimization model that considers time-series complementarity. Under the above power flow solution step size, jointly solve the system frequency response model and the distribution network spatial power flow algebraic equation to obtain the frequency regulation control droop coefficient and reactive power compensation command of each heterogeneous resource.
[0038] Specifically, we establish a first objective function that considers minimizing network loss and minimizing overshoot: (7) In the formula, u 1 is the first objective function; and Δ Q , where are decision variables, representing the optimal active power droop coefficient and reactive power compensation power of each resource, respectively; T This is the total control period; N This represents the total number of system nodes. δ ( i ) for nodes iA set of connected nodes; I ij branch road ij The square of the current amplitude on the current is the intermediate variable after second-order cone relaxation; r ij branch road ij The resistance; f The actual system frequency is 50, and the reference value for the rated frequency is Hz. C 1 and C 2 represents the weighting coefficients for minimizing network loss and minimizing frequency overshoot, respectively.
[0039] Simultaneously, a second objective function is established to maximize the system's frequency safety margin: (8) in, f N The reference frequency is 50Hz. f nadir This represents the lowest point of frequency drop.
[0040] Formulas (7) and (8) above constitute a multi-objective collaborative optimization model. Under the adaptive variable step long-time grid divided in step S103, the optimization model jointly solves the system frequency response model and the distribution network spatial power flow algebra equation to obtain the frequency regulation control droop coefficient and reactive power compensation command of each heterogeneous resource.
[0041] In this embodiment, the spatial power flow algebraic equations of the distribution network mainly consist of second-order cone relaxation AC power flow constraints, node voltage over-limit constraints, and inverter apparent capacity constraints, as detailed below: (1) The AC power flow constraint of the second-order cone relaxation is as follows: (9) (10) (11) (12) Where, π( j ) as a node j The set of the first nodes of the endpoint; P ij,t and Q ij,t They are respectively t Time flows through the side road ij The active and reactive power; P ik,t and Q ik,t They are respectively t Time flows through the side road ik The active and reactive power;P j,t and Q j,t They are respectively t Time Node j Net injected active and reactive power; x ij branch road ij The reactance; V i,t and V j,t They are respectively t Time Node i and j The square of the voltage amplitude; For the resistance of power distribution network lines, for t Branches of time ij The current.
[0042] Equation (12) is a standard second-order cone relaxation constraint, used to transform the non-convex power flow equation into a convex optimization problem to ensure solution efficiency.
[0043] (2) The node voltage constraints are as follows: (13) In the formula, V min and V max These are the lower and upper limits of the allowable safe voltage at distribution network nodes, respectively.
[0044] (3) The specific constraints on the capacity and adjustment of distributed resource devices are as follows: (14) (15) in, P inv,i,t and Q inv,i,t These represent the basic active and reactive power outputs of inverter node i, respectively; Δ Q i,t To optimize the amount of reactive power compensation commands; S inv,i,max This is the maximum apparent capacity limit for inverter interface devices; For heterogeneous resources i The droop coefficient, K i,min and K i,max Heterogeneous resources i The upper and lower limits of the droop coefficient can be configured based on its physical characteristics.
[0045] Figure 3The droop coefficient allocation results for coordinated frequency regulation of different types of resources are presented. Based on the optimization results, this embodiment adopts a timing complementary mechanism, using low-latency, high-speed response resources to compensate for the power idle period during the initial startup of high-latency resources, and generates the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
[0046] As a further implementation method, a dynamic exit mechanism to prevent overshoot is formulated for the frequency recovery period: when the power of the high-latency resource gradually responds and outputs, the fast-response resource dynamically adjusts its own power exit slope according to the actual power ramp-up of the high-latency resource, so as to avoid the secondary frequency drop or overshoot oscillation caused by the asynchronous actions of different types of resources.
[0047] As a further implementation method, a dynamic exit mechanism to prevent overshoot is formulated for the frequency recovery period. Specifically, the traditional fixed droop coefficient is prone to causing excessive total compensation power in heterogeneous resource collaboration; therefore, this invention achieves smooth and dynamic power exit by dynamically changing the equivalent droop coefficient of low-latency fast-response resources.
[0048] During the recovery period after a system frequency drop, the droop coefficient of low-latency, fast-response resources such as energy storage and photovoltaics is no longer kept constant, but is instead constructed as a dynamic function negatively correlated with the actual output power of high-latency resources such as traditional units or slow-controllable loads: (16) in, To quickly respond to resources t The dynamic droop coefficient at any given time; This refers to the maximum allowable droop coefficient for the resource allocation obtained by solving the above multi-objective optimization model; Ω slow A collection of high-latency resources in the distribution network; Δ P j ( t ) for high latency resources j exist t The actual active power response at any given time; Δ Pmax j ( t ) for high latency resources j The expected maximum active power support for this frequency regulation; γ This is an adjustable exit sensitivity coefficient, typically ranging from (0,1) to fine-tune the exit rate.
[0049] Based on a dynamically updated droop coefficient, the system provides low-latency, fast-response real-time active power compensation commands for resources. Adjusted to: (17) in, f ( t) for system t The actual frequency at any given moment.
[0050] As high-latency resources gradually overcome latency limitations, the actual output power... Its physical gradient gradually increases. Based on the aforementioned decay function, it is forced... Synchronous and smooth scaling down. At this point, the fast response resource automatically and smoothly reduces its size based on its own decreasing droop factor. The output of this mechanism achieves temporal complementarity of resources with different response characteristics, thereby avoiding frequency overshoot caused by the direct superposition of power from different resources.
[0051] This embodiment achieves precise timing compensation for the startup window of high-latency resources by employing multi-objective collaborative optimization and a dynamic exit mechanism to prevent overshoot. Specifically, it leverages the responsiveness of low-latency resources such as energy storage to prioritize providing urgently needed active power support during the initial stage (window period) when high-latency resources such as thermal power / loads have issued commands but not yet started operating, thus curbing deep frequency drops. Subsequently, as the power of slow-response resources gradually increases, fast-response resources dynamically exit at a matched rate. This timing control resolves frequency degradation caused by slow response and avoids secondary overshoot caused by asynchronous actions, improving the frequency stability and overall operational economy of the distribution network under high-proportion heterogeneous resource access.
[0052] Example 2 In one or more embodiments, a distributed resource frequency modulation active support optimization system considering latency characteristics is disclosed, comprising: The system frequency response model construction module is configured to obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources. The model pre-solution module is configured to pre-solve the system frequency response model by maximizing the droop control coefficient to determine the system's minimum frequency point; An adaptive step size construction module is configured to trigger attention scoring at the lowest frequency point, dynamically solve the power flow solution step size for the next moment based on the attention score, and use a preset benchmark power flow solution step size before the lowest frequency point. The spatiotemporal joint solution module is configured to construct a multi-objective collaborative optimization model that considers temporal complementarity. Under the aforementioned power flow solution step size, it jointly solves the system frequency response model and the distribution network spatial power flow algebraic equation to obtain the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
[0053] It should be noted that the specific implementation methods of the above modules are exactly the same as those in Example 1, and will not be described in detail again.
[0054] Example 3 In one or more embodiments, a terminal device is disclosed, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions adapted to be loaded by the processor and executed by the processor to perform the distributed resource frequency modulation active support optimization method considering latency characteristics as described in Embodiment 1.
[0055] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.
[0056] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.
[0057] In the implementation process, each step of the above method can be completed by the integrated logic circuits in the processor hardware or by software instructions.
[0058] Example 4 In one or more embodiments, a computer-readable storage medium is disclosed, wherein a plurality of instructions are stored, the instructions being adapted to be loaded by a processor of a terminal device and executed by the distributed resource frequency modulation active support optimization method considering latency characteristics described in Embodiment 1.
[0059] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.
Claims
1. A distributed resource frequency modulation active support optimization method considering delay characteristics, characterized in that, include: Obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources; The minimum frequency point of the system is determined by pre-solving the frequency response model of the system by maximizing the droop control coefficient. Attention scoring is triggered at the lowest frequency point, and the power flow solution step size for the next moment is dynamically calculated based on the attention score; before the lowest frequency point, a preset benchmark power flow solution step size is used. A multi-objective collaborative optimization model considering time-series complementarity is constructed. Under the aforementioned power flow solution step size, the system frequency response model and the distribution network spatial power flow algebraic equation are jointly solved to obtain the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
2. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, Constructing a system frequency response model that considers the latency characteristics of heterogeneous resources, specifically including: Establish the system equivalent power balance equations that include distributed heterogeneous resources of the distribution network; A time delay element is introduced into the converter interface device to construct a power response transfer function with time delay.
3. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, Attention scoring is triggered at the point of lowest frequency, specifically as follows: ； in, To score attention, σ (·) is the activation function. Let be the rate of change of frequency at time t. To estimate the voltage value, V ref This is the voltage reference value. As a safety threshold, , These are the adjustable gain coefficients.
4. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, The step size for power flow calculation at the next time step is dynamically determined based on attention scores, specifically: ； ； Where, Δ T net,t+1 For the next moment's trend step size, Δ T base To preset the baseline step size, β This is an adaptive step size scaling factor. To score attention, The maximum allowable step size for integral solving is preset. This is the preset minimum allowable solution step size.
5. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, The multi-objective collaborative optimization model considering temporal complementarity specifically includes: Consider the first objective function of minimizing network loss and overshoot, and the second objective function of maximizing the system frequency safety margin.
6. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, The first objective function is specifically: ； in, u 1 is the first objective function; and Δ Q , where are decision variables, representing the optimized active power droop coefficient and reactive power compensation power of each heterogeneous resource, respectively; T This is the total control period; N This represents the total number of system nodes. δ ( i ) for nodes i A set of connected nodes; I ij branch road ij The square of the current amplitude on the current is the intermediate variable after second-order cone relaxation; r ij branch road ij The resistance; f The actual frequency of the system. C 1 and C 2 represents the weighting coefficients for minimizing network loss and minimizing frequency overshoot, respectively.
7. The distributed resource frequency modulation active support optimization method considering delay characteristics as described in claim 1, characterized in that, The second objective function is as follows: ； in, Let this be the first objective function; and Δ Q , where are decision variables, representing the optimized active power droop coefficient and reactive power compensation power of each heterogeneous resource, respectively; The reference frequency is 50Hz. This represents the lowest point of frequency drop.
8. A distributed resource frequency modulation active support optimization system considering delay characteristics, characterized in that, include: The system frequency response model construction module is configured to obtain the time delay parameters and physical regulation capabilities of distributed heterogeneous resources in the distribution network, and construct a system frequency response model that considers the time delay characteristics of heterogeneous resources. The model pre-solution module is configured to pre-solve the system frequency response model by maximizing the droop control coefficient to determine the system's minimum frequency point; An adaptive step size construction module is configured to trigger attention scoring at the lowest frequency point, dynamically solve the power flow solution step size for the next moment based on the attention score, and use a preset benchmark power flow solution step size before the lowest frequency point. The spatiotemporal joint solution module is configured to construct a multi-objective collaborative optimization model that considers temporal complementarity. Under the aforementioned power flow solution step size, it jointly solves the system frequency response model and the distribution network spatial power flow algebraic equation to obtain the frequency regulation control droop coefficient and reactive power compensation command for each heterogeneous resource.
9. A terminal device comprising a processor and a memory, the processor for implementing instructions; the memory for storing multiple instructions, characterized in that, The instructions are adapted to be loaded by a processor and executed by the distributed resource frequency modulation active support optimization method considering latency characteristics as described in any one of claims 1-7.
10. A computer-readable storage medium storing a plurality of instructions, characterized in that, The instructions are adapted to be loaded and executed by the processor of the terminal device, and to be the distributed resource frequency modulation active support optimization method considering latency characteristics as described in any one of claims 1-7.