Optimization method and device for multi-element heterogeneous frequency modulation resources of power system

Abstract

The invention provides a multi-element heterogeneous frequency modulation resource optimization method and device for a power system. The method comprises the following steps: firstly, determining a space-time coupling relationship of multiple heterogeneous frequency modulation resources; secondly, on the basis of a preset control strategy, key indexes participating in power grid frequency regulation and control are determined; then, constructing a comprehensive evaluation model of the control efficiency; thirdly, based on the coupling relation and the evaluation model, establishing a frequency comprehensive optimization regulation and control model; and finally, multi-element heterogeneous frequency modulation resources are optimized according to the model. According to the method, time and space characteristics and key indexes of resources are considered, the frequency modulation capability of the system is effectively improved, and the limitation of single hydroelectric and thermal power frequency modulation is avoided.

Description

Technical Field

[0001] The present application relates to the field of power grid regulation and control, and in particular to a method and device for optimizing multi-dimensional heterogeneous frequency modulation resources of a power system. Background Art

[0002] The connection of a high proportion of renewable energy to the grid poses severe challenges to the safety, stability and operational optimization of the power grid.

[0003] In order to ensure the frequency stability of the system in the future when new energy accounts for a high proportion, it is urgent to fully explore and integrate various frequency regulation resources. This includes the "source-grid-load-storage" multi-side frequency regulation resources in the new power system, and their synergy is becoming increasingly important for system frequency regulation.

[0004] Although traditional synchronous generators such as thermal power and hydropower were once the main support for system frequency regulation, with the widespread grid connection of renewable energy and the gradual retirement of traditional thermal power units, the system frequency regulation capacity is facing a decline. Summary of the invention

[0005] The purpose of the present application is to overcome the above-mentioned prior art and provide a method and device for optimizing multiple heterogeneous frequency modulation resources of an electric power system.

[0006] The present application provides a method for optimizing multi-dimensional heterogeneous frequency modulation resources of a power system, comprising:

[0007] Determining the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources;

[0008] According to the power grid frequency regulation energy efficiency of the multi-dimensional heterogeneous frequency regulation resources based on the preset control strategy, determining the key indicators characterizing the participation of the multi-dimensional heterogeneous regulation resources in power grid frequency regulation;

[0009] Constructing a comprehensive evaluation model for the control efficiency of multivariate heterogeneous resources based on the key indicators;

[0010] According to the coupling relationship and the comprehensive evaluation model, a frequency comprehensive optimization and control model of multi-dimensional heterogeneous frequency modulation resources is constructed;

[0011] According to the frequency comprehensive optimization and control model, the multi-dimensional heterogeneous frequency regulation resources of the power system are optimized.

[0012] Optionally, determining the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources includes:

[0013] Acquire a first distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources on a time scale;

[0014] Acquire a second spatial distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources;

[0015] According to the correlation between the first distribution characteristic and the second distribution characteristic, a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics is determined.

[0016] Optionally, according to the power grid frequency regulation energy efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy, a key indicator characterizing the participation of the multiple heterogeneous regulation resources in power grid frequency regulation is determined, including:

[0017] Determine the evaluation objectives and criteria for the control strategies of each part of the power system, including the source side, grid side, load side, and storage side;

[0018] Obtain control effect data of the control strategies of the source side, grid side, load side and storage side in different scenarios through simulation or actual operation data;

[0019] Quantitatively evaluate the control effectiveness of each control strategy based on the control effect data;

[0020] Based on the control effectiveness, key indicators that affect the control effectiveness of each control strategy are identified.

[0021] Optionally, according to the coupling relationship and the comprehensive evaluation model, a frequency comprehensive optimization and control model of multi-element heterogeneous frequency modulation resources is constructed, including:

[0022] According to the coupling relationship and the comprehensive evaluation model, relevant elements and parameters of the system frequency security and stability situation are obtained;

[0023] According to the above factors and parameters, a convolutional neural network is used to construct a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources.

[0024] Optionally, the multiple heterogeneous frequency modulation resources include:

[0025] Generator sets, such as solar units, wind units, energy storage systems and electricity demand side.

[0026] The present application also provides an optimization device for multiple heterogeneous frequency modulation resources of a power system, comprising:

[0027] A time-space module, used to determine the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time-space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources;

[0028] An indicator module, used to determine key indicators characterizing the participation of the multiple heterogeneous frequency regulation resources in power grid frequency regulation according to the power grid frequency regulation energy efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy;

[0029] A model module is used to construct a comprehensive evaluation model of the control efficiency of multiple heterogeneous resources according to the key indicators; and to construct a frequency comprehensive optimization and control model of multiple heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model;

[0030] The optimization module is used to optimize the multi-dimensional heterogeneous frequency regulation resources of the power system according to the frequency comprehensive optimization and control model.

[0031] Optionally, the time-space module determines the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time-space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources, including:

[0032] Acquire a first distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources on a time scale;

[0033] Acquire a second spatial distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources;

[0034] According to the correlation between the first distribution characteristic and the second distribution characteristic, a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics is determined.

[0035] Optionally, the indicator module determines key indicators characterizing the participation of the multiple heterogeneous frequency regulation resources in power grid frequency regulation according to the power grid frequency regulation energy efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy, including:

[0036] Determine the evaluation objectives and criteria for the control strategies of each part of the power system, including the source side, grid side, load side, and storage side;

[0037] Obtain control effect data of the control strategies of the source side, grid side, load side and storage side in different scenarios through simulation or actual operation data;

[0038] Quantitatively evaluate the control effectiveness of each control strategy based on the control effect data;

[0039] Based on the control effectiveness, key indicators that affect the control effectiveness of each control strategy are identified.

[0040] Optionally, the model module constructs a frequency comprehensive optimization and control model of multi-element heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model, including:

[0041] According to the coupling relationship and the comprehensive evaluation model, relevant elements and parameters of the system frequency security and stability situation are obtained;

[0042] According to the above factors and parameters, a convolutional neural network is used to construct a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources.

[0043] Optionally, the multiple heterogeneous frequency modulation resources include:

[0044] Generator sets, such as solar units, wind units, energy storage systems and electricity demand side.

[0045] The beneficial effects of this application are:

[0046] The present application provides an optimization method for multivariate heterogeneous frequency modulation resources of an electric power system, including: determining the coupling relationship of the multivariate heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to the time characteristics and space characteristics of the multivariate heterogeneous frequency modulation resources; determining the key indicators characterizing the participation of the multivariate heterogeneous regulation resources in grid frequency regulation according to the grid frequency regulation energy efficiency of the multivariate heterogeneous frequency modulation resources based on a preset control strategy; constructing a comprehensive evaluation model for the control efficiency of multivariate heterogeneous resources according to the key indicators; constructing a frequency comprehensive optimization regulation model for multivariate heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model; optimizing the multivariate heterogeneous frequency modulation resources of the electric power system according to the frequency comprehensive optimization regulation model. The present application optimizes multivariate heterogeneous frequency modulation resources based on the time characteristics, space characteristics and key indicators of multivariate heterogeneous frequency modulation resources, avoids the frequency regulation method dominated by hydropower and thermal power, and improves the frequency regulation capability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of the optimization process of the multi-heterogeneous frequency regulation resources of the power system in this application;

[0048] Figure 2 This is a schematic diagram of the framework of the multi-heterogeneous frequency modulation resource optimization method in this application;

[0049] Figure 3 This is a schematic diagram of a typical frequency response process of a power system in this application;

[0050] Figure 4 This is a schematic diagram of the frequency support of the flexible DC interconnection system in this application. DETAILED DESCRIPTION

[0051] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0052] Please refer to Figure 1 As shown, the present application provides a method for optimizing multi-element heterogeneous frequency modulation resources of a power system, and the specific steps of the method include:

[0053] S101. Determine a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources.

[0054] Specifically, the steps are as follows:

[0055] S201, obtaining a first distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources on a time scale;

[0056] S202, obtaining a second spatial distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources;

[0057] S203: Determine a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to a correlation relationship between the first distribution characteristic and the second distribution characteristic.

[0058] like Figure 4 As shown, in the above steps S201 to S203, a simulation model suitable for the frequency-space characteristics analysis of large-scale new energy grid connection is established based on the power grid planning grid, covering five major 10 million kilowatt energy bases, including 10 million kilowatt offshore wind power bases, 10 million kilowatt coastal nuclear power bases, 10 million kilowatt smart energy bases, 10 million kilowatt new energy bases, and 10 million kilowatt new energy storage bases.

[0059] Since the power system contains dynamics across a wide range of time scales, the concept of time scale in traditional power systems and power electronic power systems must be clarified before reduced-order modeling.

[0060] The traditional power system can be approximately described by a complex high-order differential dynamics equation. As the basic idea of ​​power system analysis, the system can be simplified by reducing the order and presented in the form of differential algebraic equations.

[0061] Problems in different stability categories can be simplified to varying degrees based on the characteristic time scale.

[0062] As the basis for modeling and stability mechanism analysis, the singular perturbation related theorem is introduced to strictly explain the rationality and establishment conditions of the electromechanical transient DAE reduced-order modeling system.

[0063] The coexistence of multi-time scale dynamics is a ubiquitous physical phenomenon, which can be described by a set of autonomous system equations consisting of slow dynamics and fast dynamics, as shown below:

[0064]

[0065] Among them, ε is a very small constant, called the "singular perturbation parameter"; x is the slow dynamic variable, y is the fast dynamic variable, and the corresponding ordinary differential equations are "dominant slow dynamics" and "follower fast dynamics", or simply slow dynamics and fast dynamics.

[0066] In the equation, f describes the dominant dynamic properties of the system, g represents the rapidly changing dynamics on a low time scale, and the corresponding differential equation can quickly converge to the "quasi-steady state" determined by g=0, that is, the constrained manifold.

[0067] The reduced-order DAE system corresponding to the above formula is:

[0068]

[0069] where x s and s They are respectively called "state variables" and "intermediate variables", and the algebraic equation g=0 is called "algebraic constraint".

[0070] In order to observe the behavior on the microscopic time scale, the time scale can be scaled and the new time variable τ is:

[0071]

[0072] Let y f To characterize the magnitude of y's deviation from the quasi-steady state of the constrained manifold:

[0073] y f =yy s

[0074] Then the "boundary layer system" described by the autonomous system equations can be defined as:

[0075]

[0076] After time scale scaling, the “dominant slow dynamic system” is:

[0077]

[0078] When the parameter ε→0, the x in the formula remains basically unchanged on the microscopic time scale and is input into the boundary layer system as a constant parameter.

[0079] The singular perturbation characteristics require that the time scale can converge quickly to y f =0.

[0080] Tikhonov's theorem can be used to explain the conditions for the trajectory to match before and after the reduction. If the autonomous system equations jointly describe:

[0081] 1. Corresponding to x s (0), y f = 0 is the asymptotically stable equilibrium point of the boundary layer system, and for "t = [0, T]y f (0) = y(0) - y s (0) within its stability domain;

[0082] 2. For "t=[0,T], the partial derivative matrix Along the trajectory (x s (t),y s (t)) The real parts of all characteristic roots are less than a negative number σ; then the original high-order ODE system can be equivalent to the reduced-order DAE system:

[0083] x(t)=x s (t)+O(ε)

[0084] y(t)=y s (t)+y f (t)+O(ε)

[0085] And there exists t1 that satisfies T>t1>0, so that "t=[t1,T] has the following relationship:

[0086] y(t)=y s (t)+O(ε)

[0087] That is, after the deviation of the fast dynamic from the quasi-steady state decays rapidly, the trajectories of the two systems coincide.

[0088] As long as the constraint manifold g=0 satisfies the attractivity to the fast dynamic variable y, the integration of the reduced-order DAE system can be continued and accurately reflect the operating trajectory of the original system.

[0089] On the contrary, if this attraction is lost, y will no longer be adsorbed on the constrained manifold and will immediately diverge.

[0090] After DAE reduction, g = 0 can be set for numerical integration, but the original system dynamics g is not always very close to zero, but only appears to be able to quickly return to 0 after being disturbed and deviating from the constraint. In fact, the deviation of g may become a very large value at the moment of disturbance.

[0091] For modeling problems, the most basic step is to distinguish the number of fast and slow dynamics and select the corresponding variables as state variables or simplify them into intermediate variables.

[0092] Although multi-time scale phenomena are widely present in the real world, they cannot necessarily be explicitly expressed in a form containing a singular perturbation parameter ε. In fact, for any physical system:

[0093]

[0094] Even if the time scales of x and y are similar, we can multiply both sides of the equation by a small parameter ε and set:

[0095] g′(x,y)=εg(x,y)

[0096] All of them can get the form of singular perturbation system.

[0097] Therefore, the method of dividing the time scale by whether the analytical expression of the dynamic process contains a small parameter ε is very limited, especially for complex nonlinear systems.

[0098] Systems with multiple time scale characteristics do not even necessarily satisfy the requirement that the derivative term in g is much larger than f, so fast and slow dynamics cannot be distinguished simply from the numerical value of the derivative term.

[0099] In this application, based on the consideration of engineering applications, the fast and slow dynamics are divided according to the physical meaning of the singular perturbation characteristics in the time domain or frequency domain, and the corresponding degree of order reduction is performed according to the application scenario and accuracy requirements. Tables 1 and 2 summarize the fast and slow dynamic discrimination methods in the frequency domain and time domain, where the frequency domain characteristics must be observed after linearization near the equilibrium point of the original model.

[0100] Table 1 Frequency characteristics of singular perturbation characteristics

[0101]

[0102]

[0103] Table 2 Time domain characteristics of singular perturbation characteristics

[0104]

[0105] The above judgment method is not sufficient to downgrade the power supply of the power electronic interface.

[0106] Although the object system may satisfy the multi-time scale characteristics, it is not necessarily a classical singular perturbation system. For example, some mixed dynamic variables have both slow and fast dynamic components, then the number of slow dynamics must be greater than the number of slow characteristic roots.

[0107] For linear systems, identification and separation can be performed, but for nonlinear systems in analytical form, it is very difficult. Generally, there are two ways to deal with it:

[0108] The first one is to retain all state variables containing slow dynamic components. Although the number of slow dynamics is greater than the number of slow characteristic roots, the calculation accuracy can still be maintained. The cost is that the stability of the algebraic constraints formed by the remaining variables is different from that of the original boundary layer system. This is called a hybrid simulation model.

[0109] The second type is to choose to ignore the redundant mixed dynamics so that the reduced-order form of the classical singular perturbation system is still retained. This is called the classical DAE model. The advantage is that it still accurately reflects the properties of the original boundary layer system. The consequence is that the accuracy of the calculated trajectory is reduced.

[0110] When determining the fast and slow dynamics in a system, first determine the slow dynamic variables and slow dynamics based on the correlation with the slow characteristic roots, and the remaining dynamics are fast dynamics. Then assume that the parameters of the input fast dynamic equation are calculated to select the fast dynamics that meet the constraints. If the time constants of all characteristic roots are very small and coincide with the fast characteristic roots of the original boundary layer system, it is considered that the singular perturbation characteristics of the original system can be reflected, thereby achieving system simplification.

[0111] like Figure 3 As shown in Figure 1, a typical frequency response process of the power system is shown. Its frequency stability problem is the ability of the power grid to maintain or restore the frequency to the normal allowable operating range without frequency collapse after a large disturbance occurs.

[0112] When there is an active power shortage in the system, the electronic control components in the power grid will make the system transition to a new stable state through a series of dynamic responses. The unbalanced power is mainly absorbed by changing the rotor power output of the generator set (increasing or decreasing), controlling the total active load of the system (increasing or decreasing) and adjusting the active power exchange between regional power grids (connecting or disconnecting the interconnection lines).

[0113] In the process of absorbing unbalanced power, the dynamic characteristics of frequency are related to influencing factors such as the location and type of disturbance in the system, system rated capacity, speed regulation system parameters, grid structure and load effect. The dynamic changes of system frequency after disturbance can be divided into four stages.

[0114] The first stage (0-0.5s after the disturbance) is the initial distribution stage of active power shortage, according to different synchronization coefficients dP e / dδ distributes the system's active power shortage to each generator set.

[0115] At this stage, factors such as the location of the disturbance and the size of the active power shortage it produces affect the distribution of active power. The active power distributed to each generator set is:

[0116]

[0117] Where: δ is the rotor angle of the generator, and Pe is the electromagnetic power of the generator.

[0118] The second stage (0.5-2s after the disturbance) is the stage of redistribution of the disturbance power.

[0119] At this stage, factors such as the inertia size of the generator sets in the system and the oscillation form of the disturbance affect the redistribution of its disturbance power.

[0120] At the same time, the second stage is also the stage where the frequency has a large deviation. The frequency change can be adjusted by changing the mechanical power output of the rotor of the generator set. The disturbance power allocated to each generator set is:

[0121]

[0122] Where: M is the rotor inertia of the generator, ΔP is the power shortage.

[0123] The third stage (2-3s after the disturbance) is the main stage in which the generator speed control element participates in frequency regulation. This stage mainly regulates the frequency dynamic process through the response of the speed control system to the frequency.

[0124] In the third stage, the characteristic parameters of the generator set speed regulator, the spinning reserve capacity and the characteristics of the generator sets (nuclear power, hydropower, thermal power, and new energy) in the system will affect the frequency regulation capability.

[0125] Stage 4: The main stage where frequency modulation components participate in frequency regulation control.

[0126] During this stage, automatic generation control is mainly involved in frequency regulation.

[0127] The fourth stage mainly regulates the system frequency by adjusting the active power on the system interconnection line.

[0128] By analyzing the dynamic changes of frequency after the system disturbance, we can get:

[0129] The geographical location of the disturbance, the type of fault, the size of the system capacity, the difference in the control parameters of the speed regulation system, the characteristics of the grid structure and the load effect affect the entire process of frequency change.

[0130] Therefore, typical instability scenarios include severe AC faults, DC lockout, DC commutation failure, and sudden changes in new energy power.

[0131] Due to the differences in inertia and electrical connection relationships at various locations in the system, when the power system suffers a large disturbance, the kinetic energy increment of the generator rotor cannot spread throughout the entire power grid instantly, resulting in differences in the propagation of the disturbance at various locations in the power grid.

[0132] The measured data of the power grid show that the system frequency has obvious temporal and spatial distribution phenomenon, but the current theoretical basis of frequency regulation is still based on the aggregated average frequency or unified frequency response model, which obviously cannot achieve precise frequency zero-difference control.

[0133] This application first analyzes the frequency dynamic response mechanism and key factors of wide-area complex systems.

[0134] Secondly, the inertia distribution of the generator in the system is studied, and an inertia distribution method for the generator set based on the Floyd algorithm is proposed.

[0135] Then, the influence of unit inertia distribution and grid structure on frequency spatiotemporal dynamics is deeply analyzed, and the spatiotemporal distribution sequence of system frequency is deduced and determined.

[0136] From the time dimension and space dimension, the essence of the dynamic equations of inertia, active power and speed of traditional units is studied. Based on the idea of ​​"splitting, correlating and integrating", the three variables of the dynamic equation are replaced by three variable sets, and the equivalent inertia and frequency response capability of the multivariate wide-area frequency regulation resources of each subsystem are taken into account. The calculation method of the index set is proposed, thus forming a dynamic equation that describes the multi-temporal and spatial characteristics of inertia-power-frequency of the new power system.

[0137] According to different disturbances and propagation processes, the inertia-power-frequency coupling relationship between multiple inertia subsystems of the new power system is studied from the time dimension and the space dimension. An index system that characterizes the inertia-power-frequency coupling of each subsystem is established. The relationship between the inertia, power, frequency of each subsystem and the comprehensive inertia and frequency regulation capability of the system is formulated. An effective analytical method for the system inertia-power-frequency coupling mechanism is proposed, revealing the system inertia-power-frequency coupling mechanism.

[0138] S102. Determine key indicators characterizing the participation of the multiple heterogeneous frequency regulation resources in grid frequency regulation according to the grid frequency regulation efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy.

[0139] Here are the steps:

[0140] S301. Determine the evaluation objectives and criteria for the control strategies of each part of the power system, including the source side, the grid side, the load side, and the storage side.

[0141] Starting from the physical model of the new power system networking equipment, the multi-category frequency regulation resources are divided into four parts: source side, grid side, load side and storage side. The control can be divided into inertia and primary frequency regulation capacity control.

[0142] On the source side, according to the frequency response and control characteristics of the source side equipment, it can be divided into conventional units, rotating equipment + power electronic converter new energy units and stationary equipment + power electronic converter units.

[0143] Conventional units mainly include hydropower, thermal power, pumped storage units, phase regulators, gas turbines and other units that do not require power electronic converters to be connected to the grid. This part of the inertia is represented by the unit's rotational inertia, and its research is relatively mature.

[0144] The units with rotating equipment + power electronic converter connected to the grid mainly refer to direct-drive, doubly-fed and other wind turbines.

[0145] S302. Obtain control effect data of the control strategies of the source side, grid side, load side and storage side in different scenarios through simulation or actual operation data.

[0146] By studying the topological structures and virtual inertia control strategies of different types of wind turbine grid-connected converters, wind turbine kinetic energy-power support strategies, and wind turbine speed recovery strategies, the control and characterization methods of virtual inertia of different types of wind turbines at different time scales are obtained.

[0147] Stationary equipment + power electronic converter unit mainly refers to photovoltaic grid-connected system.

[0148] The energy storage side also belongs to stationary equipment + power electronic converter, which is similar to the photovoltaic grid-connected system on the source side. It can be classified as the source side or separately as the storage side.

[0149] Since both the source side and the storage side involve multiple levels such as units-stations-clusters, it is necessary to first aggregate the model based on Study 3.1 and then study it. In the new power system, this part of the units generally adopts a current source type virtual inertia control strategy, which directly changes the active output according to the frequency change rate. It is necessary to study the primary frequency regulation and virtual inertia control allocation method.

[0150] On the grid side, traditional DC and flexible DC topologies and control characteristics are studied.

[0151] On this basis, the influence mechanism of virtual inertia and power modulation method on DC voltage and frequency of AC power grids on both sides under different DC working modes is studied, and the analytical expression of virtual inertia and frequency modulation capability of DC system is studied.

[0152] like Figure 3 As shown in the figure, the flexible DC converter station enhances the system equivalent inertia and damping by increasing the outer loop control command. When the PLL bandwidth is large and the integral coefficient of the outer loop proportional integral controller is large, the equivalent inertia and damping of the flexible DC converter station can be approximated as:

[0153] M eq (s) = h / (1 + T F s),D eq (s) = d / (1 + T F s)

[0154] Where: M eq Denotes the equivalent inertia of the AC system, D eq represents the equivalent damping of the AC system, T F Represents the first-order filter time constant.

[0155] On the load side, based on the aggregation model of the wide-area frequency regulation resources on the load side, the relationship between load-side power regulation and frequency is studied, as well as the inertia of the wide-area frequency regulation resources on the load side and the control capability of primary frequency regulation.

[0156] S303: According to the control effectiveness, identify key indicators that affect the control effectiveness of each control strategy.

[0157] Based on the research on frequency-related control strategies on the source side, grid side, load side and storage side, the control effectiveness and key influencing factors of each control strategy are analyzed.

[0158] Taking into account the uneven temporal and spatial distribution, the division and equivalent methods of inertia centers at different spatial positions under new power system disturbances are studied, and the key indicators that characterize the ability of multiple heterogeneous resources to participate in power grid frequency regulation are studied. A multidimensional indicator system for its control efficiency is constructed, including inertia center and coverage area, frequency change rate, frequency extreme point, frequency adjustment time, etc.

[0159] S103. Construct a comprehensive evaluation model for the control efficiency of multi-dimensional heterogeneous resources based on the key indicators.

[0160] From the system level, a comprehensive evaluation model for the control efficiency of multiple heterogeneous resources is established, and the influence of nonlinear links such as virtual inertia of different types of new energy and DC frequency regulation resources on the maximum frequency difference after disturbance, such as emergency support power, dead zone setting, and speed regulator limiting, is studied. An analytical model of the inertia center frequency response under different degrees of disturbance is established with the maximum frequency deviation as an indicator, to achieve analytical analysis and rapid calculation of the dynamic frequency response of the wide-area power grid after disturbance.

[0161] For different inertia centers, the influence mechanism of power transfer between different inertia centers on the maximum frequency deviation is studied.

[0162] Within the safe and stable range, the maximum support power at different time scales between different inertia centers is studied, the decoupling of the maximum frequency difference calculation of multiple inertia centers is achieved, and the rapid calculation method of key frequency response indicators of new power systems under different inertia centers is further studied.

[0163] For disturbances of different degrees, the influence of equivalent inertia of new power system and primary frequency regulation strategies of different time scales on the maximum frequency deviation is studied. With the goal of ensuring that the maximum frequency deviation does not cause safety and stability measures such as low-frequency load reduction and high-frequency machine cutting, the minimum inertia and frequency regulation requirements of the new power system under disturbances of different degrees are studied.

[0164] Therefore, taking into account the multivariate flexible controllable characteristics and the uneven temporal and spatial distribution, the relationship between the control efficiency of different frequency modulation resources and the temporal and spatial characteristics of system frequency is studied, and an evaluation method for the comprehensive control efficiency of system frequency is proposed.

[0165] S104: Construct a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model.

[0166] S401. Obtain relevant factors and parameters of the system frequency security and stability situation according to the coupling relationship and the comprehensive evaluation model.

[0167] Based on the spatiotemporal distribution characteristics and control efficiency of massive heterogeneous frequency modulation resources, the relevant elements and parameters for the safe and stable frequency of the system are constructed in view of the risks of primary frequency modulation configured with the frequency consistency characteristics of the entire network, the risk of inertia response being unsuitable, the risk of disorderly cross-regional fluctuations in transient frequency-related control response power, and even causing other chain reactions.

[0168] S202: Based on the above mentioned factors and parameters, a frequency comprehensive optimization control model of multi-dimensional heterogeneous frequency modulation resources is constructed using a convolutional neural network.

[0169] Establish as Figure 2 The perception data set shown and its mapping model with the power grid frequency characteristics use a convolutional neural network to generate a multi-information perception model for power grid security and stability, and perform situational awareness of system frequency stability.

[0170] Based on grid situation awareness, the control schemes for heterogeneous resources under different frequency stability conditions of the system are studied, the feasible control strategies for each equipment participating in system frequency and voltage control under different networking modes and operation scenarios are clarified, and a frequency and voltage control parameter and scheme library for new energy DC transmission systems under flexible networking modes is formed.

[0171] Furthermore, the controllable scale of the frequency regulation resources of each inertia center and the frequency fluctuation differences of different inertia centers after large-scale power disturbances are clarified, the influence of the inertia response of the frequency regulation resources of different inertia centers, the primary frequency regulation control parameters, and the control strategy on the key disturbance indicators of the system frequency are studied, and a system frequency active coordination defense optimization system taking into account the security situation of different frequencies and the frequency fluctuation differences of different inertia centers is constructed.

[0172] The equipment response time and response speed under each control strategy are studied, the fault ride-through process is divided finely, the detailed switching behavior of each equipment control strategy in the fault ride-through process is studied in stages, and the technical principles and multi-dimensional coupling constraint relationships of each equipment participating in the system frequency control are proposed.

[0173] Study the effects and influencing paths of the active behavior of diversified equipment on the system frequency, analyze the regulation laws of active power resources at different time scales during the entire fault crossing process, analyze the conversion mechanism and characteristics of active-frequency response, clarify the trade-off relationship of multiple objectives, establish an active collaborative defense optimization model for the system under multi-dimensional coupling constraints, and propose a comprehensive frequency optimization and control method that takes into account the characteristics of diversified frequency regulation resources.

[0174] S105. Optimize the multi-dimensional heterogeneous frequency regulation resources of the power system according to the comprehensive frequency optimization and control model.

[0175] Based on the results of grid frequency situation perception and the designed system coordinated control method, comprehensive consideration is given to new energy control mode, output level, spatiotemporal correlation, stable power supply startup method, output level, transmission capacity of key branches and sections, power modulation of DC transmission system, and full consideration is given to the uncertainty of new energy. The online optimization scheme of the system's comprehensive frequency regulation performance is studied.

[0176] In order to maximize the use of offline simulation computing capabilities and improve the efficiency of online applications, the sample distribution of the operating space is used to transform the frequency control multi-objective optimization model into a sample space mapping problem. The intelligent algorithm is used to extract the characteristic vectors of relevant data under uncertain conditions of new energy output, complete model training, determine the nonlinear characteristics between input variables and parameters to be optimized, and input them into the trained model to derive the online optimization control strategy and parameters. At the same time, the offline database is enriched, thereby forming an online optimization solution for active coordinated defense of system frequency and improving the stability and controllability of system frequency.

Claims

1. A method for optimizing multi-element heterogeneous frequency modulation resources in a power system, characterized in that: include: Determining the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources; According to the power grid frequency regulation energy efficiency of the multi-dimensional heterogeneous frequency regulation resources based on the preset control strategy, determining the key indicators characterizing the participation of the multi-dimensional heterogeneous regulation resources in power grid frequency regulation; Constructing a comprehensive evaluation model for the control efficiency of multivariate heterogeneous resources based on the key indicators; According to the coupling relationship and the comprehensive evaluation model, a frequency comprehensive optimization and control model of multi-dimensional heterogeneous frequency modulation resources is constructed; According to the frequency comprehensive optimization and control model, the multi-dimensional heterogeneous frequency regulation resources of the power system are optimized.

2. The optimization method of multi-element heterogeneous frequency modulation resources in the power system according to claim 1, characterized in that: According to the time characteristics and spatial characteristics of the multiple heterogeneous frequency modulation resources, the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics is determined, including: Acquire a first distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources on a time scale; Acquire a second spatial distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources; According to the correlation between the first distribution characteristic and the second distribution characteristic, a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics is determined.

3. The optimization method of multi-element heterogeneous frequency modulation resources in the power system according to claim 1, characterized in that: According to the power grid frequency regulation energy efficiency of the multi-dimensional heterogeneous frequency regulation resources based on the preset control strategy, key indicators characterizing the participation of the multi-dimensional heterogeneous regulation resources in power grid frequency regulation are determined, including: Determine the evaluation objectives and criteria for the control strategies of each part of the power system, including the source side, grid side, load side, and storage side; Obtain control effect data of the control strategies of the source side, grid side, load side and storage side in different scenarios through simulation or actual operation data; Quantitatively evaluate the control effectiveness of each control strategy based on the control effect data; Based on the control effectiveness, key indicators that affect the control effectiveness of each control strategy are identified.

4. The optimization method of multi-element heterogeneous frequency modulation resources in the power system according to claim 1, characterized in that: According to the coupling relationship and the comprehensive evaluation model, a frequency comprehensive optimization and control model of multi-dimensional heterogeneous frequency modulation resources is constructed, including: According to the coupling relationship and the comprehensive evaluation model, relevant elements and parameters of the system frequency security and stability situation are obtained; According to the above factors and parameters, a convolutional neural network is used to construct a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources.

5. The method for optimizing multi-element heterogeneous frequency modulation resources of a power system according to claim 1, characterized in that: The multiple heterogeneous frequency modulation resources include: Generator sets, such as solar units, wind units, energy storage systems and electricity demand side.

6. An optimization device for multi-element heterogeneous frequency modulation resources in a power system, characterized in that: include: A time-space module, used to determine the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time-space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources; An indicator module, used to determine key indicators characterizing the participation of the multiple heterogeneous frequency regulation resources in power grid frequency regulation according to the power grid frequency regulation energy efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy; A model module is used to construct a comprehensive evaluation model of the control efficiency of multiple heterogeneous resources according to the key indicators; and to construct a frequency comprehensive optimization and control model of multiple heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model; The optimization module is used to optimize the multi-dimensional heterogeneous frequency regulation resources of the power system according to the frequency comprehensive optimization and control model.

7. The device for optimizing multi-element heterogeneous frequency modulation resources of a power system according to claim 6, characterized in that: The time-space module determines the coupling relationship of the multiple heterogeneous frequency modulation resources in terms of time-space distribution characteristics according to the time characteristics and space characteristics of the multiple heterogeneous frequency modulation resources, including: Acquire a first distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources on a time scale; Acquire a second spatial distribution characteristic of the multi-dimensional heterogeneous frequency modulation resources; According to the correlation between the first distribution characteristic and the second distribution characteristic, a coupling relationship between the multiple heterogeneous frequency modulation resources in terms of time and space distribution characteristics is determined.

8. The device for optimizing multi-element heterogeneous frequency modulation resources of a power system according to claim 6, characterized in that: The indicator module determines key indicators characterizing the participation of the multiple heterogeneous frequency regulation resources in power grid frequency regulation according to the power grid frequency regulation energy efficiency of the multiple heterogeneous frequency regulation resources based on a preset control strategy, including: Determine the evaluation objectives and criteria for the control strategies of each part of the power system, including the source side, grid side, load side, and storage side; Obtain control effect data of the control strategies of the source side, grid side, load side and storage side in different scenarios through simulation or actual operation data; Quantitatively evaluate the control effectiveness of each control strategy based on the control effect data; Based on the control effectiveness, key indicators that affect the control effectiveness of each control strategy are identified.

9. The device for optimizing multi-element heterogeneous frequency modulation resources of a power system according to claim 6, characterized in that: The model module constructs a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources according to the coupling relationship and the comprehensive evaluation model, including: According to the coupling relationship and the comprehensive evaluation model, relevant elements and parameters of the system frequency security and stability situation are obtained; According to the above factors and parameters, a convolutional neural network is used to construct a frequency comprehensive optimization and control model for multi-dimensional heterogeneous frequency modulation resources.

10. The device for optimizing multi-element heterogeneous frequency modulation resources of a power system according to claim 6, characterized in that: The multiple heterogeneous frequency modulation resources include: Generator sets, such as solar units, wind units, energy storage systems and electricity demand side.

Images