Frequency response characteristic modeling parameter identification method and system for network construction type energy storage

Abstract

The invention belongs to the technical field of power system operation and control, and discloses a method and a system for identifying frequency response characteristic modeling parameters of network construction type energy storage, and the method comprises the steps: constructing a frequency response characteristic model which comprises an inertia response link, a primary frequency modulation link and a secondary frequency modulation link and considers the amplitude limiting constraint of the primary frequency modulation; based on the frequency response characteristic model and the actually measured operation data of the networking type energy storage in the black-start process, a parameter identification optimization problem taking the minimum deviation between the output of the frequency response characteristic model and the actually measured data as a target is established, and the optimization problem is a quadratic programming problem considering the amplitude limiting constraint; and solving a quadratic programming problem to obtain an inertia response time constant of an inertia response link and an identification value of a primary frequency modulation coefficient of a primary frequency modulation link. According to the method, the problem of insufficient precision caused by neglecting multi-time scale coupling and amplitude limiting nonlinearity of a traditional model is solved, and the simulation credibility and the operation safety of the networking type energy storage black-start system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system operation and control, in particular to a frequency response characteristic modeling parameter identification method and system of grid-forming energy storage. BACKGROUND

[0002] Black start of power system refers to the process of gradually restoring power supply by self-starting power sources without relying on external power grid after complete outage. Traditional black start mainly relies on synchronous generators such as hydroelectric generators and gas turbines. However, these methods have many constraints such as long start-up time, geographical location limitation, and the need for external excitation.

[0003] The emergence of grid-forming energy storage system provides a new technical path for black start. It has the innate advantages of fast self-starting, controllable power and energy, and establishing voltage and frequency without external power grid support, and is hailed as the "ideal mobile black start power source". Using grid-forming energy storage as a black start source can greatly speed up the recovery of the power grid and improve the flexibility and reliability of the recovery process.

[0004] In the initial stage of black start, a fragile island grid supported by grid-forming energy storage is formed. The network has very low inertia, and the switching of loads (especially impact loads such as induction motors) will cause severe frequency fluctuations. Frequency stability is the primary key to the success of black start. The frequency response characteristics of grid-forming energy storage directly determine whether the island system can withstand the impact during the load recovery process.

[0005] In order to accurately simulate the dynamic behavior in the black start process and develop a safe load recovery sequence, a high-fidelity grid-forming energy storage frequency response model must be established. However, existing models and methods have serious defects in this specific scenario:

[0006] (1) Ignoring the multi-time scale coupling of frequency response, not fine-tuning for extreme conditions of black start:

[0007] The black start process is a planned and step-by-step power recovery process, and the grid-forming energy storage needs both inertia response and fast frequency regulation to quickly adjust the frequency to avoid too fast or too large frequency variation, and also needs to adjust the output power according to the secondary frequency regulation strategy to accommodate new loads. The "quasi-steady power ramping" process and the fast "inertia response" and "primary frequency regulation" dynamic process are interwoven and interact with each other. Existing models mostly follow the analysis idea of traditional synchronous machines, focus on the influence of the fast response virtual inertia response (response time about tens of milliseconds) and primary frequency regulation (response time about hundreds of milliseconds), and ignore the secondary frequency regulation link which is traditionally considered to be in the time scale of minutes. In fact, due to the fast action characteristics of the energy storage, the time scale of the secondary frequency regulation of the grid-forming energy storage during the black start process can reach seconds, and ignoring the multi-time scale coupling influence of the frequency response will not accurately describe the complete transition process from the load switching impact to the stable power output, resulting in a disconnection between the model and the actual control logic of the black start.

[0008] (2) Model parameter identification is difficult, and it is difficult to accurately evaluate the power regulation contribution of different frequency response processes:

[0009] The control parameters (such as virtual inertia constant, damping coefficient, secondary frequency regulation time constant and amplitude limiting value) of the grid-forming energy storage system have a decisive influence on its frequency response characteristics. However, due to reasons such as manufacturer secrecy and equipment individual differences, these parameters are difficult to obtain directly. Even if the theoretical control diagram is known, due to the existence of software implementation, filters and other links, the equivalent dynamic response parameters are also difficult to obtain directly. Therefore, a method is needed to "invert" its internal equivalent model and key parameters based on measurable external port data (such as frequency deviation and active power output). However, when facing nonlinear systems containing amplitude limiting links, traditional identification methods based on linear systems (such as least squares method and transfer function fitting) cannot be directly applied, and the precision and robustness decrease. It is difficult to accurately distinguish the power response changes caused by the linear link dynamics and the amplitude limiting link trigger, respectively, resulting in non-unique or distorted parameter identification results.

[0010] Therefore, how to provide a grid-forming energy storage frequency response characteristic modeling parameter identification method and system is a problem to be solved at present. SUMMARY

[0011] Embodiments of the present application provide a grid-forming energy storage frequency response characteristic modeling parameter identification method and system to solve the above technical problems in the prior art.

[0012] According to a first aspect of an embodiment of the present application, a grid-forming energy storage frequency response characteristic modeling parameter identification method is provided.

[0013] In an embodiment, the frequency response characteristic modeling parameter identification method of the network-constructed energy storage comprises the following steps:

[0014] A frequency response characteristic model is constructed, which comprises an inertia response link, a primary frequency modulation link and a secondary frequency modulation link, and takes into account the amplitude limiting constraint of the primary frequency modulation;

[0015] Based on the frequency response characteristic model and the measured operation data of the network-constructed energy storage in the black start process, a parameter identification optimization problem is established, which takes the minimum deviation of the output of the frequency response characteristic model from the measured data as the target, and the optimization problem is a quadratic programming problem taking into account the amplitude limiting constraint;

[0016] The quadratic programming problem is solved to obtain the identification values of the inertia response time constant of the inertia response link and the primary frequency modulation coefficient of the primary frequency modulation link.

[0017] In an embodiment, the model of the inertia response link comprises:

[0018] Based on the inertia response time constant to be identified, the rated power of the energy storage, and the system rated frequency, the active power increment and the frequency change are calculated to obtain the inertia response active power increment.

[0019] In an embodiment, the model of the primary frequency modulation link comprises:

[0020] Based on the primary frequency modulation coefficient, the rated power of the energy storage, the system rated frequency and the frequency change, the primary frequency modulation active power increment is calculated;

[0021] The primary frequency modulation active power increment is subjected to amplitude limiting constraint according to the product of the amplitude limiting ratio and the rated power of the energy storage.

[0022] In an embodiment, the model of the secondary frequency modulation link comprises:

[0023] Based on the first-order inertia link time constant of the secondary frequency modulation and the active power of the load, the secondary frequency modulation instruction is calculated.

[0024] In an embodiment, the power balance relationship of the frequency response characteristic model comprises:

[0025] The sum of the secondary frequency modulation instruction, the inertia response power increment and the primary frequency modulation power increment is equal to the active power of the load.

[0026] In an embodiment, the establishment of the parameter identification optimization problem taking the minimum deviation of the output of the frequency response characteristic model from the measured data as the target comprises:

[0027] An optimization problem is constructed, which takes the sum of the squares of the deviations of the output power of the frequency response characteristic model from the measured power as the target, and the objective function is:

[0028] Min J(θ)=Σ[(Pord (t) + ΔP iner (t, θ) + ΔP pri (t, θ) - P L (t) - ΔP 2

[0029] where J(θ) is the sum of squares of the difference between the model output power and the measured power, P ord (t) is the secondary frequency regulation command, ΔP iner (t, θ) is the contribution of the inertia response element to the active power increment through the rate of change of the response frequency, ΔP pri (t, θ) is the contribution of the primary frequency regulation element to the active power increment through the amount of change of the response frequency, P L (t) is the active power of the load.

[0030] In one embodiment, the constraints of the optimization problem include:

[0031] a mixed integer linear constraint that the primary frequency regulation power increment is less than or equal to a clipping value;

[0032] when Δf≤-α / K max ×f N :

[0033] α×P N -α×P N ×(1-v)≤ΔP pri ≤α×P N ;

[0034] -K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf+α×P N )×v≤ΔP pri ≤-K×(P N / f N )×Δf;

[0035] when -α / K max ×f N ≤Δf≤α / K max ×f N :

[0036] K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf;

[0037] when Δf≥α / K max ×f N ​Time:

[0038] -α×P N ≤ΔP pri ≤-α×P N +α×P N ×(1-v);

[0039] -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-α×P N )×v;

[0040] wherein v is whether the primary frequency regulation enters a saturation state, K max is a maximum value of a primary frequency regulation coefficient, α is an amplitude limiting ratio, P N is a rated power of energy storage, f N is a rated frequency of the system, Δf is a frequency variation, and ΔP pri is an active power increment of the primary frequency regulation link.

[0041] The optimization problem is a mixed integer quadratic programming problem including mixed integer linear constraints.

[0042] According to a second aspect of the embodiments of the present application, a system for identifying modeling parameters of frequency response characteristics of grid-forming energy storage is provided.

[0043] In one embodiment, the system for identifying modeling parameters of frequency response characteristics of grid-forming energy storage comprises:

[0044] a model construction module configured to construct a frequency response characteristic model including an inertia response link, a primary frequency regulation link, and a secondary frequency regulation link, and taking into account an amplitude limiting constraint of the primary frequency regulation;

[0045] a model parameter identification module configured to, based on the frequency response characteristic model and measured operation data of the grid-forming energy storage in a black start process, establish a parameter identification optimization problem with a minimum deviation between an output of the frequency response characteristic model and the measured data as an objective, and the optimization problem is a quadratic programming problem taking into account the amplitude limiting constraint; and solve the quadratic programming problem to obtain identification values of an inertia response time constant of the inertia response link and a primary frequency regulation coefficient of the primary frequency regulation link.

[0046] In one embodiment, the model of the inertia response link comprises:

[0047] The active power increment of the inertia response is calculated based on the inertia response time constant, the energy storage rated power, and the system rated frequency of the to-be-identified inertia.

[0048] In one embodiment, the model of the primary frequency modulation link comprises:

[0049] The active power increment of the primary frequency modulation is calculated based on the primary frequency modulation coefficient, the energy storage rated power, the system rated frequency, and the frequency variation.

[0050] The active power increment of the primary frequency modulation is limited based on the product of the limiting amplitude ratio and the energy storage rated power.

[0051] In one embodiment, the model of the secondary frequency modulation link comprises:

[0052] The secondary frequency modulation instruction is calculated based on the first-order inertia link time constant of the secondary frequency modulation and the active power of the load.

[0053] In one embodiment, the power balance relationship of the frequency response characteristic model comprises:

[0054] The sum of the secondary frequency modulation instruction, the inertia response power increment, and the primary frequency modulation power increment is equal to the active power of the load.

[0055] In one embodiment, the parameter identification optimization problem established to minimize the deviation between the output of the frequency response characteristic model and the measured data comprises:

[0056] An optimization problem is constructed to minimize the sum of squares of the deviation between the output power of the frequency response characteristic model and the measured power, and the objective function is:

[0057] Min J(θ)=Σ[(P ord (t)+ΔP iner (t,θ)+ΔP pri (t,θ)-P L (t))] 2 ;

[0058] In the formula, J(θ) is the sum of squares of the deviation between the output power of the model and the measured power, P ord (t) is the secondary frequency modulation instruction, ΔP iner (t,θ) is the active power increment contributed by the inertia response link through the response of the frequency variation rate, ΔP pri (t,θ) is the active power increment contributed by the primary frequency modulation link through the response of the frequency variation, P L (t) is the active power of the load.

[0059] In one embodiment, the constraint condition of the optimization problem comprises:

[0060] a mixed integer linear constraint that the power increment of the first frequency modulation is less than or equal to the clipping value;

[0061] when Δf≤-α / K max ×f N :

[0062] α×P N -α×P N ×(1-v)≤ΔP pri ≤α×P N ;

[0063] -K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf+α×P N )×v≤ΔP pri ≤-K×(P N / f N )×Δf;

[0064] when-α / K max ×f N ≤Δf≤α / K max ×f N :

[0065] K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf;

[0066] when Δf≥α / K max ×f N :

[0067] -α×P N ≤ΔP pri ≤-α×P N +α×P N ×(1-v);

[0068] -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-α×P N )×v;

[0069] wherein v is whether the first frequency modulation enters a saturation state, K maxis the maximum value of the primary frequency modulation coefficient, a is the amplitude limiting ratio, P N is the energy storage rated power, f N is the system rated frequency, and Δf is the frequency variation, and ΔP pri is the active power increment of the primary frequency modulation link;

[0070] The optimization problem is a mixed integer quadratic programming problem containing mixed integer linear constraints.

[0071] According to a third aspect of the embodiments of the present application, a computer device is provided.

[0072] In some embodiments, the computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0073] According to a fourth aspect of the embodiments of the present application, a computer readable storage medium is provided.

[0074] In one embodiment, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0075] The technical solutions provided by the embodiments of the present application can have the following beneficial effects:

[0076] 1. High model accuracy: By considering the inertia response, primary frequency modulation and secondary frequency modulation links at the same time, and taking into account the amplitude limiting constraint of the primary frequency modulation, the frequency response behavior of the grid-forming energy storage in the black start process under multiple time scales can be more realistically reflected, especially for the dynamic process simulation of low inertia islanded power grid.

[0077] 2. Fast and accurate parameter identification: The parameter identification problem is converted into a quadratic programming problem considering the amplitude limiting constraint, which can be quickly solved by an optimization solver, effectively dealing with the identification difficulty brought by the nonlinear amplitude limiting link, and accurately "inverting" the key equivalent parameters (Tj, K) of the system based on the measurable external port data (frequency, power), with strong uniqueness and good robustness of the identification results.

[0078] 3. Strong practicability: The method provides a reliable model basis for the simulation analysis, safety evaluation and control strategy formulation of the grid-forming energy storage black start system, helps to improve the safety and reliability of the black start process, and has important engineering application value.

[0079] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0080] The accompanying drawings, which are incorporated herein and constitute part of this specification, illustrate embodiments consistent with the application and, together with the description, further serve to explain the principles of the application.

[0081] Figure 1 is a flow chart of a modeling parameter identification method for frequency response characteristics of grid-forming energy storage according to an example embodiment;

[0082] Figure 2 is a principle block diagram of a modeling parameter identification system for frequency response characteristics of grid-forming energy storage according to an example embodiment;

[0083] Figure 3 is a structural schematic diagram of a computer device according to an example embodiment;

[0084] Figure 4 is a block diagram of a black start frequency response characteristic model of grid-forming energy storage in an embodiment of the application;

[0085] Figure 5 is a black start test block diagram in an embodiment of the application;

[0086] Figure 6 is a parameter identification verification situation schematic diagram of a 10MW load test in an embodiment of the application;

[0087] Figure 7 is a frequency modulation contribution situation analysis schematic diagram of a 10MW load test in an embodiment of the application;

[0088] Figure 8 is a parameter identification verification situation schematic diagram of a 40MW load test in an embodiment of the application;

[0089] Figure 9 is a frequency modulation contribution situation analysis schematic diagram of a 40MW load test in an embodiment of the application. DETAILED DESCRIPTION

[0090] The following description and drawings are illustrative of specific embodiments of the application and are not intended to be limiting of the application. Particular embodiments of the present application are described herein with reference to the accompanying drawings. The use of the same reference numbers in different figures indicates similar or identical components. The present application is not limited to the illustrative embodiments described and shown.

[0091] The various modules in the device or system of the present application can be implemented by software, hardware and combinations thereof in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above modules.

[0092] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0093] Figure 1 An embodiment of a frequency response characteristic modeling parameter identification method of a networked energy storage is shown.

[0094] In this optional embodiment, the frequency response characteristic modeling parameter identification method of the networked energy storage comprises:

[0095] S101, a frequency response characteristic model comprising an inertia response link, a primary frequency modulation link and a secondary frequency modulation link is constructed, and a primary frequency modulation limiting constraint is taken into account.

[0096] S102, based on the frequency response characteristic model and the measured operation data of the networked energy storage in the black start process, a parameter identification optimization problem is established with the minimum deviation of the frequency response characteristic model output and the measured data as the target, and the optimization problem is a quadratic programming problem taking into account the limiting constraint.

[0097] S103, the quadratic programming problem is solved to obtain the identification values of the inertia response time constant of the inertia response link and the primary frequency modulation coefficient of the primary frequency modulation link.

[0098] In this optional embodiment, the model of the inertia response link comprises:

[0099] Based on the inertia response time constant to be identified, the energy storage rated power and the system rated frequency, the active power increment and the frequency change are calculated to obtain the inertia response active power increment.

[0100] In this optional embodiment, the model of the primary frequency modulation link comprises:

[0101] Based on the primary frequency modulation coefficient, the energy storage rated power, the system rated frequency and the frequency change, the primary frequency modulation active power increment is calculated; and the primary frequency modulation active power increment is limited by the product of the limiting ratio and the energy storage rated power.

[0102] In this optional embodiment, the model of the secondary frequency modulation link comprises:

[0103] Based on the first-order inertia link time constant of the secondary frequency modulation and the load active power, the secondary frequency modulation instruction is calculated.

[0104] In this alternative embodiment, the power balance relationship of the frequency response characteristic model comprises:

[0105] The sum of the secondary frequency modulation instruction, the inertia response power increment and the primary frequency modulation power increment is equal to the active power of the load.

[0106] In this alternative embodiment, the parameter identification optimization problem is established to minimize the deviation between the output of the frequency response characteristic model and the measured data, which comprises:

[0107] An optimization problem is constructed to minimize the sum of squares of the deviation between the output power of the frequency response characteristic model and the measured power, and the objective function is:

[0108] Min J(θ) = Σ[(P ord (t) + ΔP iner (t, θ) + ΔP pri (t, θ) - P L (t))]2 2 ;

[0109] In the formula, J(θ) is the sum of squares of the deviation between the output power of the model and the measured power, P ord (t) is the secondary frequency modulation instruction, ΔP iner (t, θ) is the active power increment contributed by the inertia response link through the response to the frequency change rate, ΔP pri (t, θ) is the active power increment contributed by the primary frequency modulation link through the response to the frequency change, P L (t) is the active power of the load.

[0110] In this alternative embodiment, the constraint condition of the optimization problem comprises:

[0111] A mixed integer linear constraint that the primary frequency modulation power increment is less than or equal to the amplitude limiting value;

[0112] When Δf ≤ -α / K max ×f N :

[0113] α×P N -α×P N ×(1-v) ≤ ΔP pri ≤ α×P N ;

[0114] -K×(P N / f N )×Δf + (K max ×(P N / f N )×Δf + α×P N )×v ≤ ΔP pri ≤ -K×(P N / f N )×Δf;

[0115] when -a / K max ×f N ≤Δf≤a / K max ×f N :

[0116] K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf;

[0117] when Δf≥a / K max ×f N :

[0118] -a×P N ≤ΔP pri ≤-a×P N +a×P N ×(1-v);

[0119] -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-a×P N )×v;

[0120] where v is whether the primary frequency regulation enters the saturation state, K max is the maximum value of the primary frequency regulation coefficient, a is the amplitude limiting ratio, P N is the energy storage rated power, f N is the system rated frequency, Δf is the frequency change, and ΔP pri is the active power increment of the primary frequency regulation link; the optimization problem is a mixed integer quadratic programming containing mixed integer linear constraints.

[0121] Figure 2 An embodiment of a frequency response characteristic modeling parameter identification system of a network-constructed energy storage is shown.

[0122] In this optional embodiment, the frequency response characteristic modeling parameter identification system comprises:

[0123] a model construction module 201 configured to construct a frequency response characteristic model comprising an inertia response link, a primary frequency regulation link, and a secondary frequency regulation link, and taking into account the primary frequency regulation amplitude limiting constraint;

[0124] The model parameter identification module 202 is configured to: based on the frequency response characteristic model and the measured operation data of the network-constructed energy storage in the black start process, establish a parameter identification optimization problem with the minimum deviation between the output of the frequency response characteristic model and the measured data as the target, and the optimization problem is a quadratic programming problem considering the amplitude limiting constraint; and solve the quadratic programming problem to obtain the identification values of the inertia response time constant of the inertia response link and the primary frequency modulation coefficient of the primary frequency modulation link.

[0125] In this optional embodiment, the model of the inertia response link comprises:

[0126] Based on the inertia response time constant to be identified, the rated power of the energy storage, and the system rated frequency, the active power increment of the inertia response is calculated.

[0127] In this optional embodiment, the model of the primary frequency modulation link comprises:

[0128] Based on the primary frequency modulation coefficient, the rated power of the energy storage, the system rated frequency, and the frequency variation, the active power increment of the primary frequency modulation is calculated; and the active power increment of the primary frequency modulation is subjected to amplitude limiting constraint according to the product of the amplitude limiting ratio and the rated power of the energy storage.

[0129] In this optional embodiment, the model of the secondary frequency modulation link comprises:

[0130] Based on the first-order inertia link time constant of the secondary frequency modulation and the active power of the load, the secondary frequency modulation instruction is calculated.

[0131] In this optional embodiment, the power balance relationship of the frequency response characteristic model comprises:

[0132] The sum of the secondary frequency modulation instruction, the inertia response power increment, and the primary frequency modulation power increment is equal to the active power of the load.

[0133] In this optional embodiment, the establishment of the parameter identification optimization problem with the minimum deviation between the output of the frequency response characteristic model and the measured data as the target comprises:

[0134] An optimization problem with the minimum sum of the square deviations between the output power of the frequency response characteristic model and the measured power as the target is constructed, and the objective function is:

[0135] Min J(θ)=Σ[(P ord (t)+ΔP iner (t,θ)+ΔP pri (t,θ)-P L (t))] 2 ;

[0136] In the formula, J(θ) is the sum of the square deviations between the model output power and the measured power, P ord (t) is the secondary frequency modulation instruction, and ΔPiner (t, 0) is the contribution of the active power increment by the inertia response element through the rate of change of the response frequency, Af pri (t, 0) is the contribution of the active power increment by the primary frequency regulation element through the amount of change of the response frequency, P L (t) is the active power of the load.

[0137] In this alternative embodiment, the constraints of the optimization problem include:

[0138] a mixed-integer linear constraint that the primary frequency regulation power increment is less than or equal to the limit value;

[0139] when Af < -a / K max xf N :

[0140] a x P N -a x P N x (1 - v) < AP pri < a x P N ;

[0141] - K x (P N / f N ) x Af + (K max x (P N / f N ) x Af + a x P N ) x v < AP pri < - K x (P N / f N ) x Af;

[0142] when -a / K max xf N < a / K max xf N :

[0143] K x (P N / f N ) x Af < AP pri < - K x (P N / f N ) x Af;

[0144] when Af > a / K max xf N :

[0145] - a x P N < AP pri < - a x P N + a x P N x (1 - v);

[0146] - K x (P N / f N) x Delta f <= Delta P pri <= -K x (P N / f N ) x Delta f + (K max x (P N / f N ) x Delta f - alpha x P N ) x v;

[0147] In the formula, v is whether primary frequency modulation enters a saturation state, K max is a maximum value of a primary frequency modulation coefficient, alpha is an amplitude limiting ratio, P N is a storage rated power, f N is a system rated frequency, Delta f is a frequency change amount, and Delta P pri is an active power increment of a primary frequency modulation link.

[0148] In order to facilitate understanding of the above technical solutions of the application, the above technical solutions of the application are further described from the aspects of architecture and principle, as follows:

[0149] The application comprises modeling and parameter identification of frequency response characteristics considering frequency modulation amplitude limiting constraints during network-constructed energy storage black start, so as to solve the problems of inaccurate modeling of island power grid frequency at the initial stage of black start and difficulty in accurately identifying key control parameters. Firstly, a frequency response model comprising inertia response, primary frequency modulation (considering amplitude limiting) and secondary frequency modulation link is constructed, so as to accurately describe the frequency dynamics of a low-inertia island system at the initial stage of black start. In view of the difficulty in identifying key parameters of the model, the problem is converted into a quadratic programming problem considering amplitude limiting constraints, and the inertia response time constant and the primary frequency modulation coefficient are accurately identified based on measured data by using an optimization algorithm. The application solves the problem of insufficient accuracy caused by ignoring multi-time scale coupling and amplitude limiting nonlinearity in the traditional model, and improves the simulation credibility and operation safety of the network-constructed energy storage black start system.

[0150] I. Constructing a frequency response characteristic model considering frequency modulation amplitude limiting constraints

[0151] The model comprises an inertia response link, a primary frequency modulation link and a secondary frequency modulation link, and considers the power output amplitude limiting constraint of the primary frequency modulation link. The inertia response link contributes to an active power increment Delta P iner by responding to a frequency change rate, and the mathematical model thereof is as follows:

[0152] Delta P iner = - (T j x P N / f N ) x (d (Delta f) / dt) ;

[0153] wherein, T jLet P be the inertial response time constant to be identified. N f is the rated power of energy storage. N Here, denoted as the system's rated frequency, Δf is the frequency change, and t is time.

[0154] The primary frequency regulation stage contributes the active power increment ΔP by responding to frequency changes. pri Its mathematical model is:

[0155] ΔP pri =-K×(P N / f N )×Δf, where ΔP pri The primary frequency modulation power increment is K, where K is the primary frequency modulation coefficient; and the primary frequency modulation power increment satisfies the amplitude limiting constraint: |ΔP pri |≤α×P N , where α is the amplitude limiting ratio.

[0156] When Δf≤0:

[0157] ΔP pri =min{-K×(P N / f N )×Δf,α×P N};

[0158] When Δf≥0:

[0159] ΔP pri =max{-K×(P N / f N )×Δf,-α×P N};

[0160] Where K is the primary frequency modulation coefficient to be identified; α is the limiting ratio, and when the frequency deviation is too large, the power output of the primary frequency modulation stage is constrained by the limiting ratio. It should be noted that this constraint condition is non-convex and nonlinear, and cannot be solved quickly using an optimization solver.

[0161] The secondary frequency modulation stage measures the active power change of the load and generates a secondary frequency modulation command P through a first-order inertial element. ord Its mathematical model (frequency domain) is:

[0162] P ord (s)=[1 / (T sec ×s+1)]×P L (s);

[0163] Among them, P ord This is a secondary frequency modulation command, T sec P is the time constant of the first-order inertial element in the second-frequency modulation. L This represents the active power of the load.

[0164] The power balance equation of the model is:

[0165] P ord +ΔP iner +ΔP pri =P L .

[0166] II. Model parameter identification based on a quadratic programming algorithm:

[0167] Define the parameter vector to be identified θ = [T j , K]. Based on the model and the measured operating data of the grid-forming energy storage during the black start process, an optimization problem (objective function of the parameter identification optimization problem) is constructed, which aims to minimize the sum of squares of the deviation between the model output power and the measured power. The optimization problem is a quadratic programming problem considering the amplitude constraint:

[0168] Min J(θ) = Σ[(P ord (t) + ΔP iner (t, θ) + ΔP pri (t, θ) - P L (t))] 2 .

[0169] The constraint condition is a mixed integer linear constraint that the incremental primary frequency power does not exceed the amplitude limit value. The optimization problem is a mixed integer quadratic programming problem with linear constraints, which can be quickly solved using an optimization solver.

[0170] When Δf ≤ -α / K max ×f N (K max is the maximum value of the primary frequency coefficient, for example, the primary frequency coefficient is generally between 10-50, and the value can be set to 100):

[0171] α×P N -α×P N ×(1-v)≤ΔP pri ≤α×P N ;

[0172] -K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf+α×P N )×v≤ΔP pri ≤-K×(P N / f N )×Δf;

[0173] When -α / K max ×f N ≤Δf≤α / K max×f N hour:

[0174] K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf;

[0175] When Δf≥α / K max ×f N hour:

[0176] -α×P N ≤ΔP pri ≤-α×P N +α×P N ×(1-v);

[0177] -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-α×P N )×v;

[0178] Where v indicates whether the primary frequency modulation has entered saturation, 0 for unsaturated and 1 for saturated.

[0179] The parameter identification method proposed in this invention transforms the optimization problem into a mixed-integer quadratic programming problem with linear constraints. An optimization solver is then used to solve the quadratic programming problem to obtain the parameter T to be identified. j The accurate value of K. That is, the identification value of the inertial response time constant of the inertial response element and the primary frequency modulation coefficient of the primary frequency modulation element.

[0180] III. Model Validation:

[0181] By using a hardware-in-the-loop simulation platform or on-site black-start test, active power output and system frequency data of grid-type energy storage are collected. Simulation is then performed using the models established in Process 1 and Process 2 and the identified parameters to verify the model accuracy.

[0182] The following provides a specific application of the frequency response characteristic modeling and parameter identification method for grid-type energy storage during black start-up, taking into account frequency regulation and amplitude limiting constraints. A model block diagram is shown below. Figure 4 As shown, ΔP is...

[0183] First, a black-start test was conducted at a real-world grid-connected energy storage power station, such as... Figure 5As shown, part of the energy storage is set to construct a network operation, and the other part is set to follow the network operation. The change of load during black start is simulated by adjusting the energy storage power of the follow network operation, and the active power P meas (t) of the energy storage port of the constructed network type is collected by the high-frequency recording device

[0184] 1. Calculate the frequency change Δf(t) and the frequency change rate d(Δf) / dt from the collected data.

[0185] 2. According to the known secondary frequency modulation time constant T sec (usually set by the controller, which is 2s in the test at this time) and the measured active power P L (t) (approximately equal to P meas (t) after the load is put into steady state), calculate the secondary frequency modulation instruction P ord (t).

[0186] 3. Construct the quadratic programming problem as described in the present application. The objective function is the sum of squares of the difference between the total power (P ord +ΔP iner +ΔP pri ) calculated by the model and the measured power P meas . The constraint condition is that the primary frequency modulation power increment does not exceed the amplitude limit value of the mixed integer linear constraint.

[0187] 4. Call the optimization solver (such as CPLEX, Gurobi, etc.) to solve the mixed integer quadratic programming problem with linear constraints, and obtain the optimal parameter estimation values T j and K.

[0188] Finally, the model is verified. Substitute the identified parameters T j and K into the frequency response model shown in Figure 4 , and perform simulation on the black start test data. Compare the system frequency curve output by the model with the measured frequency curve. When the load is 10MW, T j = 2.54, K = 33.40 can be identified from the measured data. As can be seen from the comparison of the frequency curves of Figure 6 , the simulation results are consistent with the measured data, Figure 7 showing the active contribution of different frequency response links. In order to be general, further identification is performed on the measured data of the load of 40MW. T j = 2.48, K = 31.48 can be identified from the measured data, which is close to the identification result when the load is 10MW. As can be seen from the comparison of the frequency curves of Figure 8 , the simulation results are consistent with the measured data, Figure 9The active contribution of different frequency response links is demonstrated. In summary, the test results verify the effectiveness of the proposed method, and the contribution of different frequency response links of the grid-forming energy storage to the active power can be further analyzed by the proposed method, providing a reference for subsequent grid-forming energy storage black start evaluation work.

[0189] It should be noted that T sec , alpha, P N and other parameters are usually known or settable controller parameters. For different types or control modes of grid-forming energy storage, the frequency response structure may differ slightly, but the core modeling idea and parameter identification method described in the present application are still applicable.

[0190] The present application includes frequency response characteristic modeling and parameter identification of grid-forming energy storage during black start, which has great theoretical significance and engineering application value for improving the simulation credibility, operation safety and control effectiveness of the grid-forming energy storage black start system.

[0191] In one embodiment, a computer device, which can be a server, has an internal structure diagram as shown in Figure 3 The computer device includes a processor, a memory and a network interface connected by a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store static information and dynamic information data. The network interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement the steps in the above method embodiments.

[0192] Those skilled in the art can understand, Figure 3 the structure shown in the figure, only a block diagram of part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0193] In addition, the present application also provides a computer device including a memory and a processor, the memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0194] In addition, the present application also provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by the processor to implement the steps in the above method embodiments.

[0195] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments of the methods. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0196] The present application is not limited to the structures described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present application is only limited by the appended claims.

Claims

1. A method for identifying modeling parameters of frequency response characteristics in grid-type energy storage, characterized in that, include: A frequency response characteristic model is constructed, which includes an inertial response stage, a primary frequency modulation stage, and a secondary frequency modulation stage, and takes into account the primary frequency modulation amplitude limiting constraint. Based on the frequency response characteristic model and the measured operation data of grid-type energy storage during the black start process, a parameter identification optimization problem is established with the goal of minimizing the deviation between the output of the frequency response characteristic model and the measured data. The optimization problem is a quadratic programming problem that takes into account the amplitude constraint. Solving the quadratic programming problem yields the identification values ​​of the inertial response time constant of the inertial response element and the primary frequency regulation coefficient of the primary frequency regulation element.

2. The method for identifying frequency response characteristic modeling parameters of grid-type energy storage according to claim 1, characterized in that, The model for the inertial response element includes: Based on the inertial response time constant to be identified, the rated power of energy storage, and the rated frequency of the system, the active power increment and frequency change are calculated, and the active power increment of the inertial response is obtained.

3. The method for identifying frequency response characteristic modeling parameters of grid-type energy storage according to claim 1, characterized in that, The model of the primary frequency modulation stage includes: Based on the primary frequency regulation coefficient, the rated power of energy storage, the rated frequency of the system, and the frequency change, the active power increment of primary frequency regulation is calculated. The increment of active power for primary frequency regulation is constrained by multiplying the limiting ratio and the rated power of energy storage.

4. The method for identifying modeling parameters of frequency response characteristics of grid-type energy storage according to claim 1, characterized in that, The model of the secondary frequency modulation stage includes: The secondary frequency regulation command is calculated based on the time constant of the first-order inertial element and the active power of the load in the secondary frequency regulation.

5. The method for identifying frequency response characteristic modeling parameters of grid-type energy storage according to claim 1, characterized in that, The power balance relationship of the frequency response characteristic model includes: The sum of the secondary frequency regulation command, the inertia response power increment, and the primary frequency regulation power increment equals the load active power.

6. The method for identifying frequency response characteristic modeling parameters of grid-type energy storage according to claim 1, characterized in that, The parameter identification and optimization problem, which aims to minimize the deviation between the frequency response characteristic model output and the measured data, includes: Construct an optimization problem with the objective of minimizing the sum of squared deviations between the output power of the frequency response characteristic model and the measured power, and the objective function is: Min J(θ)=Σ[(P ord (t)+ΔP iner (t,θ)+ΔP pri (t,θ)-P L (t))] 2 ; In the formula, J(θ) is the sum of squares of the deviations between the model output power and the measured power, and P ord (t) represents the secondary frequency modulation command, ΔP iner (t,θ) represents the active power increment contributed by the inertial response element through the rate of change of the response frequency, ΔP. pri (t,θ) represents the active power increment contributed by the primary frequency regulation element through the response to frequency changes, P L (t) represents the active power of the load.

7. The method for identifying frequency response characteristic modeling parameters of grid-type energy storage according to claim 6, characterized in that, The constraints of the optimization problem include: Mixed-integer linear constraint where the primary frequency modulation power increment is less than or equal to the amplitude limit; When Δf≤-α / K max ×f N hour: α×P N -α×P N ×(1-v)≤ΔP pri ≤α×P N ; -K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf+α×P N )×v≤ΔP pri ≤-K×(P N / f N )×Δf; When -α / K max ×f N ≤Δf≤α / K max ×f N hour: K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf; When Δf≥α / K max ×f N hour: -α×P N ≤ΔP pri ≤-α×P N +α×P N ×(1-v); -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-α×P N )×v; In the formula, v represents whether the primary frequency modulation has entered saturation, and K... max The maximum value of the primary frequency modulation coefficient, α is the limiting ratio, and P is the maximum value of the primary frequency modulation coefficient. N f is the rated power of energy storage. N The system's rated frequency is given by Δf, where ΔP is the frequency change. pri This represents the active power increment of the primary frequency regulation stage; The optimization problem is a mixed-integer quadratic programming problem with mixed-integer linear constraints.

8. A frequency response characteristic modeling parameter identification system for grid-type energy storage, characterized in that, include: The model building module is used to build a frequency response characteristic model that includes an inertial response element, a primary frequency modulation element, and a secondary frequency modulation element, and takes into account the primary frequency modulation amplitude limiting constraint. The model parameter identification module is used to establish a parameter identification optimization problem based on the frequency response characteristic model and the measured operating data of the grid-type energy storage during the black start process. The objective is to minimize the deviation between the output of the frequency response characteristic model and the measured data. The optimization problem is a quadratic programming problem that takes into account the amplitude limit constraint. Solving the quadratic programming problem yields the identified values ​​of the inertial response time constant of the inertial response element and the primary frequency regulation coefficient of the primary frequency regulation element.

9. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 8, characterized in that, The model for the inertial response element includes: Based on the inertial response time constant to be identified, the rated power of energy storage, and the rated frequency of the system, the active power increment and frequency change are calculated, and the active power increment of the inertial response is obtained.

10. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 8, characterized in that, The model of the primary frequency modulation stage includes: Based on the primary frequency regulation coefficient, the rated power of energy storage, the rated frequency of the system, and the frequency change, the active power increment of primary frequency regulation is calculated. The increment of active power for primary frequency regulation is constrained by multiplying the limiting ratio and the rated power of energy storage.

11. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 8, characterized in that, The model of the secondary frequency modulation stage includes: The secondary frequency regulation command is calculated based on the time constant of the first-order inertial element and the active power of the load in the secondary frequency regulation.

12. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 8, characterized in that, The power balance relationship of the frequency response characteristic model includes: The sum of the secondary frequency regulation command, the inertia response power increment, and the primary frequency regulation power increment equals the load active power.

13. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 8, characterized in that, The parameter identification and optimization problem, which aims to minimize the deviation between the frequency response characteristic model output and the measured data, includes: Construct an optimization problem with the objective of minimizing the sum of squared deviations between the output power of the frequency response characteristic model and the measured power, and the objective function is: Min J(θ)=Σ[(P ord (t)+ΔP iner (t,θ)+ΔP pri (t,θ)-P L (t))] 2 ; In the formula, J(θ) is the sum of squares of the deviations between the model output power and the measured power, and P ord (t) represents the secondary frequency modulation command, ΔP iner (t,θ) represents the active power increment contributed by the inertial response element through the rate of change of the response frequency, ΔP. pri (t,θ) represents the active power increment contributed by the primary frequency regulation element through the response to frequency changes, P L (t) represents the active power of the load.

14. The frequency response characteristic modeling parameter identification system for grid-type energy storage according to claim 13, characterized in that, The constraints of the optimization problem include: Mixed-integer linear constraint where the primary frequency modulation power increment is less than or equal to the amplitude limit; When Δf≤-α / K max ×f N hour: α×P N -α×P N ×(1-v)≤ΔP pri ≤α×P N ; -K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf+α×P N )×v≤ΔP pri ≤-K×(P N / f N )×Δf; When -α / K max ×f N ≤Δf≤α / K max ×f N hour: K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf; When Δf≥α / K max ×f N hour: -α×P N ≤ΔP pri ≤-α×P N +α×P N ×(1-v); -K×(P N / f N )×Δf≤ΔP pri ≤-K×(P N / f N )×Δf+(K max ×(P N / f N )×Δf-α×P N )×v; In the formula, v represents whether the primary frequency modulation has entered saturation, and K... max The maximum value of the primary frequency modulation coefficient, α is the limiting ratio, and P is the maximum value of the primary frequency modulation coefficient. N f is the rated power of energy storage. N The system's rated frequency is given by Δf, where ΔP is the frequency change. pri This represents the active power increment of the primary frequency regulation stage; The optimization problem is a mixed-integer quadratic programming problem with mixed-integer linear constraints.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images