A method and system for evaluating the frequency regulation capability of distribution networks considering the dynamic characteristics of power systems

Abstract

This invention discloses a method and system for evaluating the frequency regulation capability of a distribution network considering the dynamic characteristics of a power system. It acquires operational data from the historical database of the distribution network in the power system, covering the primary frequency regulation and inertia support processes. The operational data is then subjected to state-space mapping, and the mapping results are trained using least-squares to obtain a linear function model of the state-space mapping. A linear function model of the state-space mapping between distribution network parameters is constructed, and the feasible region of the power system droop coefficient and inertia support coefficient is evaluated online in a collaborative iterative manner. This enables a feasible assessment of the final state-of-charge limit and active power limit of distributed photovoltaic-storage systems during the primary frequency regulation process of the distribution network in the power system. Compared with existing technologies, the collaborative iterative process employed in this invention constitutes a collaborative iterative evaluation framework between the distribution network and the power system operation side, achieving accurate analysis of the feasible region boundary of the power system's distribution network frequency regulation capability.

Description

Technical Field

[0001] This invention relates to the field of power system operation and control technology, and in particular to a distribution network state-space mapping evaluation method and system that takes into account the dynamic characteristics of power systems. Background Technology

[0002] The distribution network should assess the feasible region of the maximum droop-inertia support coefficient under different operating conditions and report it to the system operator before participating in primary frequency regulation. Based on the operating state constraints of the distribution network, the maximum feasible region of droop that it can provide can be calculated, where the droop coefficient should be dynamically adjusted according to the reserve power. In model-based assessment methods, droop and inertia support coefficients are generally represented as analytical functions of the distributed photovoltaic-storage (PV-S) state of charge and illumination, respectively, and researchers have assessed the boundary of each feasible region under different conditions. Researchers have used a cubic Kalman filter to assess the boundary of the inertia feasible region based on the rotor speed of the PV-Storage system. However, the above methods only construct the feasible region of each coefficient independently. In reality, droop and inertia control in the distribution network are both achieved by regulating active power output, and therefore they interact. If the interaction mechanism between droop and inertia support coefficients is ignored, it will be difficult for the distribution network to construct an accurate feasible region and achieve optimal primary frequency regulation support.

[0003] In power systems with high renewable energy penetration, frequency dynamics change when numerous distribution networks provide support based on assessment results. This leads to different frequency variations under the same load fluctuations, consequently affecting the capacity of each distribution network. In this context, considering the coupling relationship of primary frequency response capabilities between distribution networks is crucial to obtaining the maximum coefficient boundary. Theoretically, the droop-inertia feasible region of the distribution network can be analytically solved at the system operator's end using physical models and centralized optimization methods. MV Kazemi et al. calculated the feasible region of the droop-inertia support coefficients of the photovoltaic-storage system using the parameter space method based on frequency stability, illumination, and the state of distributed photovoltaic-storage systems. Other researchers have constructed the droop-inertia feasible region for each distribution network using the minimum frequency value and the rate of frequency change, and solved the optimal coefficient scheme based on the rotor stability of distributed photovoltaic-storage systems. However, the centralized assessment in the above methods requires the system operator to obtain state measurement data of all distributed photovoltaic-storage systems in the grid, which leads to a heavy communication burden, and real-time distributed photovoltaic-storage system state data is difficult to upload from the distribution network.

[0004] In contrast, evaluations conducted on the distribution network side exhibit significant advantages such as rapid response, high reliability, and low communication burden. Priyatosh Mahish et al. used aggregated illumination and power reserves, along with the frequency of each distribution network at the point of common coupling, as allocation indicators to iteratively evaluate the droop and inertia feasible regions of each distribution network based on system dynamics. However, some literature sets the power system droop-inertia support coefficient as a constant, which fails to reflect changes in the system's frequency regulation capability. In reality, the power system droop-inertia support coefficient should be aggregated from multiple distribution networks in real time, rather than being set as a constant value. Therefore, it is crucial to propose an iterative evaluation scheme that coordinates between the distribution network and system operators, taking into account changes in the primary frequency regulation response capability of the distribution networks within the system.

[0005] Although the aforementioned coordinated iterative methods do not require centralized modeling and evaluation, they still necessitate obtaining static state parameters of the distribution network. Thus, if the distribution network model is inaccurate or incomplete, the evaluated feasible region will certainly be unreliable. Furthermore, the long solution times and optimization based on large-scale physical models make these methods unsuitable for online applications and tracking illumination changes. To address these issues, data-driven methods have attracted widespread attention because they can utilize historical data to construct equivalent primary frequency response models of the distribution network. Researchers have proposed ANN models to fit the nonlinear dynamics of the primary frequency response of the distribution network and evaluate the active power of the distribution network under different operating conditions. Other researchers have proposed a deep reinforcement learning method to construct a primary frequency response model of the distribution network and evaluate the droop-inertia feasible region. Some researchers have also proposed an improved fuzzy method to evaluate the feasible region of the droop coefficient of the distribution network based on illumination changes. However, these methods require large amounts of sampling data and complex training processes, and lack physical interpretability. Due to the lack of historical samples for extreme scenarios, the accuracy of the maximum boundary of the feasible region of the distribution network cannot be guaranteed.

[0006] Due to the advantages of global linearization, Koopman operator theory (KOT) has been applied to fitting nonlinear dynamic models. Based on KOT, nonlinear dynamical algebraic models in low-dimensional state spaces can be transformed into globally linear forms in infinite-dimensional state spaces without loss of accuracy, and linear Koopman operator matrices can be estimated through data-driven training. In practical applications, increasing the dimensionality to an appropriate number is sufficient to transform nonlinear models into high-precision linear mathematical models. The global linearity property enables the establishment of accurate equivalent first-order frequency modulation evaluation models over a wide range without requiring historical samples to cover extreme scenarios. Summary of the Invention

[0007] This invention aims to propose a method and system for evaluating the frequency regulation capability of distribution networks that considers the dynamic characteristics of power systems. It utilizes the state-space mapping of the feasible domain of distribution network droop and inertia support coefficient that considers the dynamic characteristics of power systems, and evaluates the frequency regulation capability of distribution networks based on the obtained feasible boundaries of each distribution network droop coefficient and inertia support coefficient.

[0008] To achieve the above-mentioned objectives, the present invention proposes the following technical solution:

[0009] In a first aspect, the present invention provides a method for evaluating the frequency regulation capability of a distribution network considering the dynamic characteristics of a power system, comprising:

[0010] The system acquires operational data from historical databases of various distribution networks in the power system that participate in primary frequency regulation and inertia support processes, performs state-space mapping on the operational data, and trains the mapping results to obtain a linear function relationship model of state-space mapping.

[0011] The state-space mapping linear function relationship model is solved iteratively. The solution process includes:

[0012] Execute in one iteration:

[0013] (i) Obtain the real-time illumination intensity of the distribution network, the real-time power system droop coefficient and the real-time power system inertial support coefficient calculated and sent to the distribution network by the power system operator.

[0014] (ii) Substitute the real-time light intensity of the distribution network, the real-time droop coefficient of the power system, and the real-time inertial support coefficient into the state-space mapping linear function relationship model between the parameters of the distribution network, and solve the second droop coefficient and the second inertial support coefficient of the distribution network that satisfy the convergence condition in a second iteration, and report to the power system operator side for updating.

[0015] (iii) Repeat (ii) until the total droop coefficient and total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, and output the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient.

[0016] Based on the feasible boundaries of the power system droop coefficient and the inertia support coefficient, a feasible assessment is made of the final state-of-charge limit and the active power limit of distributed photovoltaic energy storage during the primary frequency regulation process of the power system distribution network.

[0017] In some implementations, steps (i), (ii), and (iii) are repeated until the following convergence condition is met:

[0018]

[0019] Among them, Kf,sys (s+1) represents the power system droop coefficient for the next iteration, K f,sys (s) represents the power system droop coefficient in the current iteration, K in (s+1) represents the power system inertia support coefficient for the next iteration.

[0020] K in (s) represents the power system inertia support coefficient for the current iteration, and δ1 and δ2 represent the set threshold values.

[0021] In some implementations, the second droop coefficient and the second inertia support coefficient, which satisfy the convergence condition, are solved iteratively in a second step, further including:

[0022] (a) Set the limits of the final state of charge of distributed photovoltaic energy storage and the limit of active power of distributed photovoltaic energy storage during the primary frequency regulation process of the distribution network, and set the initial values ​​of the second droop coefficient and the second inertia support coefficient.

[0023] (b) Solve the active power limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second inertia support coefficient feasible boundary; solve the final charge state limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second droop coefficient feasible boundary.

[0024] (c) Output the second droop coefficient and the second inertia support coefficient that satisfy the convergence condition.

[0025] In some implementations, steps (a), (b), and (c) are repeated until the following convergence condition is met:

[0026]

[0027] Among them, K in,i (t+1) represents the i-th second inertia support coefficient in the next two iterations, K in,i (t) represents the i-th second inertia support coefficient in the current iteration, K f,i (t+1) represents the i-th second droop coefficient in the next iteration, K f,i (t) represents the i-th second droop coefficient of the current second iteration, and δ1 and δ2 represent the set thresholds.

[0028] In some implementations, the state-space mapping linear function relationship model is shown below;

[0029]

[0030] Among them, M i Represents the least squares data-driven matrix, v m,i Represents light, This represents the power output during the initial frequency modulation startup phase of distributed optical storage in the process of primary frequency modulation and inertia support. Represents the final state of charge of distributed photovoltaic-storage systems, where i represents the i-th distribution network number, and K represents the final state of charge. f,i Represents the droop coefficient of the distribution network, K in,i K represents the inertia support coefficient. f,sys K represents the droop factor of the power system. in Represents a linear model between the inertia support coefficients of the power system, ψ i E represents a scalar function that maps multidimensional nonlinear observation dimensions, and T represents the transpose.

[0031] In some implementations, the power system droop factor K f,sys (s+1) and inertia support coefficient K in (s+1), formula

[0032] as follows:

[0033]

[0034] K f,i (1) = 0

[0035] K in (1) = 0

[0036] Among them, K f,SG,j K represents the droop coefficient provided by the j-th thermal power unit, h represents the number of thermal power units, and K f,i (s) represents the i-th distribution network droop coefficient in the current iteration step s, K in,i (s) represents the inertia support coefficient of the i-th distribution network in the current iteration step s, l represents the number of distribution networks, and K f,i (s) represents the i-th distribution network droop coefficient in the current iteration step s, K in,i (s) represents the inertia support coefficient of the i-th distribution network power system under the current iteration step s, and l represents the number of distribution networks.

[0037] Secondly, the present invention provides a distribution network frequency regulation capability assessment system considering the dynamic characteristics of a power system. The system includes modules for executing the steps of a distribution network frequency regulation capability assessment method considering the dynamic characteristics of a power system, including a model building module, a model solving module, and an assessment module, wherein:

[0038] The model building module is used to obtain the operation data of the primary frequency regulation and inertia support process in the historical database of each distribution network in the power system, perform state space mapping on the operation data, and perform least squares training on the mapping results to obtain the state space mapping linear function relationship model.

[0039] The model solving module is used to iteratively solve the state-space mapping linear function relationship model. The solution process includes:

[0040] Execute in one iteration:

[0041] (i) Obtain the real-time illumination intensity of the distribution network, the real-time power system droop coefficient and the real-time power system inertial support coefficient calculated and sent to the distribution network by the power system operator.

[0042] (ii) Substitute the real-time light intensity of the distribution network, the real-time droop coefficient of the power system, and the real-time inertial support coefficient into the state-space mapping linear function relationship model between the parameters of the distribution network, and solve the second droop coefficient and the second inertial support coefficient of the distribution network that satisfy the convergence condition in a second iteration, and report to the power system operator side for updating.

[0043] (iii) Repeat (ii) until the total droop coefficient and total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, and output the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient.

[0044] The evaluation module is used to evaluate the feasibility of the final state-of-charge limit and active power limit of distributed photovoltaic energy storage during the primary frequency regulation process of the power system distribution network, based on the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient.

[0045] In some implementations, steps (i), (ii), and (iii) are repeated until the following convergence condition is met:

[0046]

[0047] Among them, K f,sys (s+1) represents the power system droop coefficient for the next iteration, K f,sys (s) represents the power system droop coefficient in the current iteration, K in (s+1) represents the power system inertia support coefficient for the next iteration.

[0048] K in (s) represents the power system inertia support coefficient for the current iteration, and δ1 and δ2 represent the set threshold values.

[0049] In some implementations, the second droop coefficient and the second inertia support coefficient, which satisfy the convergence condition, are solved iteratively in a second step, further including:

[0050] (a) Set the limits of the final state of charge of distributed photovoltaic energy storage and the limit of active power of distributed photovoltaic energy storage during the primary frequency regulation process of the distribution network, and set the initial values ​​of the second droop coefficient and the second inertia support coefficient.

[0051] (b) Solve the active power limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second inertia support coefficient feasible boundary; solve the final charge state limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second droop coefficient feasible boundary.

[0052] (c) Output the second droop coefficient and the second inertia support coefficient that satisfy the convergence condition.

[0053] In some implementations, steps (a), (b), and (c) are repeated until the following convergence condition is met:

[0054]

[0055] Among them, K in,i (t+1) represents the i-th second inertia support coefficient in the next two iterations, K in,i (t) represents the i-th second inertia support coefficient in the current iteration, K f,i (t+1) represents the i-th second droop coefficient in the next iteration, K f,i (t) represents the i-th second droop coefficient of the current second iteration, and δ1 and δ2 represent the set thresholds.

[0056] In some implementations, the state-space mapping linear function relationship model is shown below;

[0057]

[0058] Among them, M i Represents the least squares data-driven matrix, v m,i Represents light, This represents the power output during the initial frequency modulation startup phase of distributed optical storage in the process of primary frequency modulation and inertia support. Represents the final state of charge of distributed photovoltaic-storage systems, where i represents the i-th distribution network number, and K represents the final state of charge. f,i Represents the droop coefficient of the distribution network, K in,i K represents the inertia support coefficient. f,sys K represents the droop factor of the power system. in Represents a linear model between the inertia support coefficients of the power system, ψ i E represents a scalar function that maps multidimensional nonlinear observation dimensions, and T represents the transpose.

[0059] In some implementations, the power system droop factor K f,sys (s+1) and inertia support coefficient Kin (s+1), formula

[0060] as follows:

[0061]

[0062]

[0063] K f,i (1) = 0

[0064] K in (1) = 0

[0065] Among them, K f,SG,j K represents the droop coefficient provided by the j-th thermal power unit, h represents the number of thermal power units, and K f,i (s) represents the i-th distribution network droop coefficient in the current iteration step s, K in,i (s) represents the inertia support coefficient of the i-th distribution network in the current iteration step s, l represents the number of distribution networks, and K f,i (s) represents the i-th distribution network droop coefficient in the current iteration step s, K in,i (s) represents the inertia support coefficient of the i-th distribution network power system under the current iteration step s, and l represents the number of distribution networks.

[0066] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the distribution network frequency regulation capability assessment method considering the dynamic characteristics of the power system provided in any of the above embodiments.

[0067] Fourthly, embodiments of the present invention also provide a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the distribution network frequency regulation capability assessment method considering the dynamic characteristics of the power system provided in any of the above embodiments.

[0068] Compared with the prior art, the present invention can achieve the following beneficial technical effects:

[0069] 1) On the one hand, when the droop coefficient and inertia support coefficient on the power system operator side meet the convergence condition, the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient are output in an iterative manner to evaluate the feasible region of the power system droop coefficient and the inertia support coefficient; on the other hand, the distribution network droop coefficient and inertia support coefficient that meet the convergence condition are solved in a second iteration. The online collaborative iteration process of the two aspects constitutes the evaluation framework of collaborative iteration between the distribution network and the power system operator side.

[0070] 2) This evaluation framework makes full use of the frequency regulation capability of the distribution network, avoids the heavy computational burden of centralized solution on the system side and the communication burden of traditional feasible domain transmission, and further achieves the beneficial effect of fast solution speed.

[0071] 3) This evaluation framework provides an alternating iterative method to realize the search for feasible regions for building the distribution network and achieves accurate analytical solution of the feasible region boundary. Attached Figure Description

[0072] Figure 1 This is a flowchart of a method for evaluating the frequency regulation capability of a distribution network that considers the dynamic characteristics of a power system, according to the present invention.

[0073] Figure 2 This is a detailed flowchart of step 1;

[0074] Figure 3 This is a detailed flowchart of step 2;

[0075] Figure 4 This is a detailed flowchart of steps 2-4;

[0076] Figure 5 This is a block diagram of a distribution network frequency regulation capability assessment system that considers the dynamic characteristics of the power system according to the present invention. Detailed Implementation

[0077] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0078] like Figure 1 As shown, the present invention provides a method for evaluating the frequency regulation capability of a distribution network considering the dynamic characteristics of a power system. The method includes the following steps:

[0079] Step 1: Obtain operational data from the historical databases of each distribution network in the power system that participated in the primary frequency regulation and inertia support process. Taking the i-th distribution network as an example, the above data includes the system frequency distributed photovoltaic-storage primary frequency regulation start-up power of each distribution network in the past participation in the primary frequency regulation and inertia support process. Distributed photovoltaic energy storage final state of charge Distribution network droop coefficient K f,i Inertia support coefficient K in,i Light v m,i Power system droop coefficient K f,sys and the power system inertia support coefficient K in The above parameters are obtained from the distribution network operation database. Subsequently, state-space mapping is performed on the operating data, and the mapping results are trained using least squares to obtain a linear function relationship model of the state-space mapping between distribution network parameters, such as... Figure 2 As shown, this step specifically includes the following processes:

[0080] Step 1-1: Obtain G historical data samples of relevant parameters for each distribution network in the power system. G shall be no less than 1000 sets of data samples. Taking the g-th sample of the i-th distribution network as an example, where g∈G, the parameters of the data sample include the power during the primary frequency regulation startup phase of distributed photovoltaic energy storage. Distributed photovoltaic energy storage final state of charge Distribution network droop coefficient K f,i,g Inertia support coefficient K in,i,g Light v m,i,g Power system droop coefficient K f,sys,g and the power system inertia support coefficient K in,g ;

[0081] Steps 1-2: Perform state-space mapping based on the running data to construct input and output samples, including:

[0082] The input sample is defined as:

[0083] x i,g =[v m,i,g ;K f,i,g ;K in,i,g ;K f,sys,g ;K in,g ] T (1)

[0084] Among them, v m,i,g K represents light. f,i,g Represents the droop coefficient of the distribution network, K f,sys,g Represents the power system droop coefficient, K in,g represents the power system inertia support coefficient, T represents transpose, and represents the vertical concatenation of vectors;

[0085] The output sample is defined as:

[0086]

[0087] in, Represents the final state of charge of distributed photovoltaic energy storage. This represents the power output during the initial frequency modulation startup phase of distributed optical energy storage.

[0088] Steps 1-3: Upgrade the state space of the input samples to obtain the upgraded input samples, defined as:

[0089]

[0090] Where, ψ i (x i,g ) represents the nonlinear enhanced observation dimension mapping vector function for the i-th distribution network and the g-th data sample;

[0091] Here, assuming the number of augmented observation dimensions is E, the multidimensional nonlinear augmented observation dimension mapping vector function ψ is obtained. i E (x i,g ) is defined as:

[0092] ψ i E (x i,g )=[ψ i,1 (x i,g ),...,ψ i,e (x i,g ),...,ψ i,E (x i,g )] T (4)

[0093] The e-th dimension nonlinearly enhances the observation dimension mapping scalar function ψ i,e (x i,g ) is defined as:

[0094]

[0095] Among them, c i,g,e Representative and input sample x i,g Basis vectors of the same dimension, r i,g,e Representative input sample x i,g With basis vector c i,g,e The geometric distance represented, e∈E;

[0096] Steps 1-4: Arrange the output samples and the upgraded input samples in order to obtain the upgraded input sample set and output sample set;

[0097] The input sample set of the i-th distribution network after dimensionality upgrade is defined as follows:

[0098]

[0099] The i-th distribution network output sample set is defined as:

[0100]

[0101] Steps 1-5: Obtain the state-space mapping linear function relationship model between distribution network parameters through least squares data-driven training;

[0102] Obtain the least squares data-driven matrix M i express The formula is as follows:

[0103]

[0104] in, represent matrix transpose, represent The pseudo-inverse of the matrix;

[0105] The state-space mapping linear function relationship model between distribution network parameters is defined as follows:

[0106]

[0107] Among them, v m,i,g Represents light, This represents the power output during the initial frequency modulation startup phase of distributed optical storage in the process of primary frequency modulation and inertia support. Represents the final state of charge of distributed photovoltaic-storage systems, where i represents the i-th distribution network number, and K represents the final state of charge. f,i Represents the droop coefficient of the distribution network, K in,i K represents the inertia support coefficient. f,sys K represents the droop factor of the power system. in Represents a linear model between the inertia support coefficients of the power system, ψ i E denoted as a scalar function representing a multidimensional nonlinear enhancement of the observation dimension mapping, where T represents the transpose;

[0108] Step 2: Based on the state-space mapping linear function relationship model between distribution network parameters shown in formula (9) obtained in Step 1, online collaborative iterative evaluation of the feasible region of distribution network droop and inertia support coefficient is performed, such as... Figure 3 As shown, this step specifically includes the following processes:

[0109] Step 2-1: Measure the real-time illuminance v of each distribution network at that moment. w,i,real Substituting into the state-space mapping linear function relationship model, we obtain the state-space mapping linear function relationship model between distribution network parameters;

[0110] Step 2-2: Set the iteration step s = 0 and perform one iteration;

[0111] Steps 2-3, iteration step s = s + 1, the power system operator side calculates the power system droop coefficient K for the next iteration step according to formulas (10)-(14). f,sys A linear model K between (s+1) and the power system inertia support coefficient in (s+1), and K f,sys (s+1) and K in (s+1) is distributed to each distribution network, as shown in the following formula:

[0112]

[0113] K f,i (1)=0 (13)

[0114] K in (1)=0 (14)

[0115] Among them, K f,SG,j The droop coefficient provided by the j-th thermal power unit is represented by h, where h represents the number of thermal power units, j∈h, K in (1) K represents the total inertia support coefficient of the system at the start of the iteration. f,sys (1) Represents the total droop coefficient of the system at the start of the iteration, K f,SG K represents the total sag coefficient of the thermal power unit. f (s) represents the total droop coefficient of the distribution network in the current iteration step s, K f,i (s) represents the i-th distribution network droop coefficient in the current iteration step s, K in,i (s) represents the linear model between the inertia support coefficients of the i-th distribution network power system under the current iteration step s, where l represents the number of distribution networks, i∈l;

[0116] Steps 2-4: On each distribution network side, set the power system droop coefficient K. f,sys (s+1) and inertia support coefficient K in Substituting (s+1) into the state-space mapping linear function relationship model between distribution network parameters, we can solve for the distribution network droop coefficient K. f,i (s+1) and inertia support coefficient K in,i (s+1), and report to the power system operator; such as Figure 4 As shown, this step specifically includes the following processes.

[0117] Step 2-4-1: Set the iteration step t=0 and perform a second iteration;

[0118] Step 2-4-2: Accumulate iteration steps t = t + 1, and set the allowable final state-of-charge limit ω of distributed photovoltaic energy storage during the primary frequency regulation process of the distribution network. limit,i and the active power limit of distributed photovoltaic energy storage Among them, the final state-of-charge boundary ω of distributed photovoltaic energy storage limit,i Defined as:

[0119]

[0120] Distributed photovoltaic energy storage active power limit Defined as:

[0121]

[0122] Among them, M 1,i Least squares data-driven linear matrix M i The first line, M 2,i M represents the least squares data-driven linear matrix iIn the second line, K in,i (t+1) represents the i-th distribution network inertia support coefficient updated in iteration t+1, K f,i (t+1) represents the i-th distribution network droop coefficient updated in the (t+1)-th iteration;

[0123] Step 2-4-3: Set the initial values ​​for the distribution network droop coefficient and inertia support coefficient, where the initial value of the distribution network droop coefficient K is... f,i (0) = 0, initial value of distribution network inertia support coefficient K in,i (0) = 0;

[0124] Step 2-4-4: Solve for the active power limit of distributed photovoltaic storage using the bisection method. The equation yields the feasible boundary K for the i-th distribution network inertia support coefficient, updated iteratively. in,i (t+1);

[0125] Furthermore, the bisection method is used to solve the final state-of-charge boundary ω of distributed photovoltaic energy storage. limit,i The equation yields the feasible boundary K for the i-th distribution network droop coefficient, which is updated iteratively. f,i (t+1);

[0126] Steps 2-4-5: Determine whether the inertia support coefficient and droop coefficient of each distribution network meet the following convergence conditions:

[0127]

[0128] Among them, δ1 and δ2 are set thresholds, which are generally set to within one-thousandth of the initial value of the above coefficients.

[0129] If the condition in formula (17) is met, then output the droop coefficient K for each distribution network. f,i (s+1)=K f,i (t+1) and inertia support coefficient K in,i (s+1)=K in,i (t+1), and report to the power system operator; if the above conditions are not met, let the iteration step t = t+1, and go to step 2-4-3 to continue iterating and perform the site-side evaluation.

[0130] Steps 2-5: Obtain the power system droop coefficient and inertia support coefficient using the distribution network droop coefficient and inertia support coefficient that meet the convergence conditions;

[0131] Determine whether the power system droop coefficient and inertia support coefficient satisfy the power system-side convergence condition in formula (18):

[0132]

[0133] Among them, Kf,sys (s+1) represents the power system droop coefficient for the next iteration, K f,sys (s) represents the power system droop coefficient in the current iteration, K in (s+1) represents the power system inertia support coefficient for the next iteration, K in (s) represents the power system inertia support coefficient in the current iteration, and δ1 and δ2 represent the set threshold K. in (s) represents the power system inertia support coefficient in the current iteration;

[0134] If the condition in formula (18) holds, then the feasible boundary K of the output power system droop coefficient is... f (s+1) and feasible boundary K of inertia support coefficient in (s+1); If the above conditions are not met, let the iteration step s = t+1, and proceed to step 2-3 to continue the collaborative iteration of the distribution network-power system operation side; Based on the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient, realize the feasible assessment of the final state of charge limit of distributed photovoltaic energy storage and the active power limit of distributed photovoltaic energy storage in the primary frequency regulation process of each distribution network in the power system.

[0135] like Figure 5 As shown, the present invention provides a distribution network frequency regulation capability assessment system that considers the dynamic characteristics of the power system, comprising a model building module, a model solving module, and an assessment module.

[0136] Among them: the model building module is used to obtain the operation data of the primary frequency regulation and inertia support process in the historical database of each distribution network in the power system, perform state space mapping on the operation data, and perform least squares training on the mapping results to obtain the state space mapping linear function relationship model.

[0137] The model solving module is used to iteratively solve the state-space mapping linear function relationship model. The solution process includes:

[0138] Execute in one iteration:

[0139] (i) Obtain the real-time illumination intensity of the distribution network, the real-time power system droop coefficient and the real-time power system inertial support coefficient calculated and sent to the distribution network by the power system operator.

[0140] (ii) Substitute the real-time light intensity of the distribution network, the real-time droop coefficient of the power system, and the real-time inertial support coefficient into the state-space mapping linear function relationship model between the parameters of the distribution network, and solve the second droop coefficient and the second inertial support coefficient of the distribution network that satisfy the convergence condition in a second iteration, and report to the power system operator side for updating.

[0141] (iii) Repeat (ii) until the total droop coefficient and total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, and output the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient.

[0142] The evaluation module is used to evaluate the feasibility of the final state-of-charge limit and active power limit of distributed photovoltaic energy storage during the primary frequency regulation process of the power system distribution network, based on the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient.

[0143] In some implementations, steps (i), (ii), and (iii) are repeated until the following convergence condition is met:

[0144]

[0145] Among them, K f,sys (s+1) represents the power system droop coefficient for the next iteration, K f,sys (s) represents the power system droop coefficient in the current iteration, K in (s+1) represents the power system inertia support coefficient for the next iteration.

[0146] K in (s) represents the power system inertia support coefficient for the current iteration, and δ1 and δ2 represent the set threshold values.

[0147] In some implementations, the second droop coefficient and the second inertia support coefficient, which satisfy the convergence condition, are solved iteratively in a second step, further including:

[0148] (a) Set the limits of the final state of charge of distributed photovoltaic energy storage and the limit of active power of distributed photovoltaic energy storage during the primary frequency regulation process of the distribution network, and set the initial values ​​of the second droop coefficient and the second inertia support coefficient.

[0149] (b) Solve the active power limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second inertia support coefficient feasible boundary; solve the final charge state limit equation of distributed photovoltaic energy storage using the bisection method to obtain the iteratively updated second droop coefficient feasible boundary.

[0150] (c) Output the second droop coefficient and the second inertia support coefficient that satisfy the convergence condition.

[0151] Repeat steps (a), (b), and (c) until the following convergence condition is met:

[0152]

[0153] Among them, K in,i(t+1) represents the i-th second inertia support coefficient in the next two iterations, K in,i (t) represents the i-th second inertia support coefficient in the current iteration, K f,i (t+1) represents the i-th second droop coefficient in the next iteration, K f,i (t) represents the i-th second droop coefficient of the current second iteration, and δ1 and δ2 represent the set thresholds.

[0154] Furthermore, based on a similar inventive concept, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0155] Furthermore, based on a similar inventive concept, embodiments of the present invention also provide a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described method.

[0156] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0157] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0158] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0159] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0160] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0161] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0162] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered to fall within the scope of protection of the present invention.

Claims

1. A power system dynamic characteristic considering power distribution network frequency modulation capability evaluation method, characterized in that, The method comprises the following steps: obtaining operation data of each distribution network in the power system participating in the primary frequency modulation and inertia support process in the historical database of the distribution network, performing state space mapping on the operation data, and training the mapping result to obtain a state space mapping linear function relationship model; iteratively solving the state space mapping linear function relationship model, and the solving process comprising: performing the following steps once: (i) obtaining real-time light intensity of a real-time distribution network, real-time power system droop coefficient and real-time power system inertia support coefficient calculated by a power system operator and delivered to the distribution network; Power system droop coefficient and inertia support coefficient , as follows: ; ; ; ; ; wherein, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, (ii) substituting the real-time light intensity of the real-time distribution network, the real-time power system droop coefficient and the real-time inertia support coefficient into the state space mapping linear function relationship model of the distribution network parameters, and iteratively solving the second droop coefficient and the second inertia support coefficient of the distribution network that meet the convergence condition, and reporting to the power system operator side for updating; further comprising: (a) setting the final state of charge limit of the distributed light storage and the active power limit of the distributed light storage allowed in the primary frequency modulation process of the distribution network, and setting the initial value of the second droop coefficient and the second inertia support coefficient; (b) solving the active power limit equation of the distributed light storage by using the bisection method to obtain the feasible boundary of the iteratively updated second inertia support coefficient; solving the final state of charge limit equation of the distributed light storage by using the bisection method to obtain the feasible boundary of the iteratively updated second droop coefficient; (c) outputting the second droop coefficient and the second inertia support coefficient that meet the convergence condition; (iii) repeating (ii) until the total droop coefficient and the total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, and outputting the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient; According to the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient, the final state of charge limit of the distributed light storage and the active power limit of the distributed light storage in the primary frequency modulation process of the distribution network in the power system are evaluated. 2.The power system dynamic characteristic considering power distribution network frequency modulation capability evaluation method according to claim 1, wherein, The convergence condition is: ; wherein, a power system droop coefficient representative of a next iteration, a power system droop coefficient representative of a current iteration, a power system inertia support coefficient representative of a next iteration, a power system inertia support coefficient representative of a current iteration, and representing a set threshold value.

3. The method of claim 1, wherein the method further comprises: repeat steps (a), (b) and (c) until the following convergence condition is met: ; wherein, the ith second inertia support coefficient representing the next quadratic iteration, the ith second inertia support coefficient representing the current quadratic iteration, the ith second droop coefficient representing the next quadratic iteration, the ith second droop coefficient representing the current quadratic iteration, and representing a set threshold value.

4. The method of claim 1, wherein the method further comprises: The state space mapping linear function relationship model is as follows: ; wherein, represents the least square data-driven matrix, represents the light, represents the primary frequency modulation and inertia support process in the distributed light storage primary frequency modulation starting stage power, represents the final state of charge of the distributed light storage, i represents the number of the i-th power distribution network, represents the droop coefficient of the power distribution network, represents the inertia support coefficient, represents the droop coefficient of the power system, represents the linear model between the inertia support coefficients of the power system, represents a multi-dimensional nonlinear augmented observation dimension mapping scalar function, and T represents transposition.

5. A power system dynamic characteristic considering power distribution network frequency modulation capability evaluation system, characterized in that, The method comprises a model construction module, a model solving module and a distribution network frequency modulation capability evaluation module, wherein: The model construction module is used to obtain operation data of each distribution network in the power system participating in the primary frequency modulation and inertia support process in the historical database of the distribution network, perform state space mapping on the operation data, and perform least square training on the mapping result to obtain a state space mapping linear function relationship model; The model solving module is used to iteratively solve the state space mapping linear function relationship model, and the solving process comprising: performing the following steps once: (i) obtaining real-time light intensity of a real-time distribution network, real-time power system droop coefficient and real-time power system inertia support coefficient calculated by a power system operator and delivered to the distribution network; (ii) substituting the real-time light intensity of the real-time distribution network, the real-time power system droop coefficient and the real-time inertia support coefficient into the state space mapping linear function relationship model of the distribution network parameters, and iteratively solving the second droop coefficient and the second inertia support coefficient of the distribution network that meet the convergence condition, and reporting to the power system operator side for updating; further comprising: (a) setting the final state of charge limit of the distributed light storage and the active power limit of the distributed light storage allowed in the primary frequency modulation process of the distribution network, and setting the initial value of the second droop coefficient and the second inertia support coefficient; (b) solving the active power limit equation of the distributed light storage by using the bisection method to obtain the feasible boundary of the iteratively updated second inertia support coefficient; solving the final state of charge limit equation of the distributed light storage by using the bisection method to obtain the feasible boundary of the iteratively updated second droop coefficient; (c) outputting the second droop coefficient and the second inertia support coefficient that meet the convergence condition; (iii) repeating (ii) until the total droop coefficient and the total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, and outputting the feasible boundary of the power system droop coefficient and the feasible boundary of the inertia support coefficient; Power system droop coefficient and inertia support coefficient , as follows: ; ; ; ; ; wherein, represents the number of thermal power units, represents the number of thermal power units, represents the number of thermal power units, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, represents the number of distribution networks, (ii), the real-time power distribution network real-time light intensity, real-time power system droop coefficient and real-time inertia support coefficient are substituted into the state space mapping linear function relationship model of the power distribution network parameter, the second droop coefficient and the second inertia support coefficient of the power distribution network which meet the convergence condition are solved by quadratic iteration, and the power system operator side is reported for updating; further comprising: (a), setting the distributed light storage final state of charge limit and the distributed light storage active power limit allowed in the power distribution network primary frequency modulation process, setting the second droop coefficient and the second inertia support coefficient initial value; (b), the distributed light storage active power limit equation is solved by using dichotomy method, and the iterative updated second inertia support coefficient feasible boundary is obtained; the distributed light storage final state of charge limit equation is solved by using dichotomy method, and the iterative updated second droop coefficient feasible boundary is obtained; (c), output the second droop coefficient and the second inertia support coefficient which meet the convergence condition; (iii), repeat (ii) until the total droop coefficient and the total inertia support coefficient of the power system obtained by using the latest second droop coefficient and the latest second inertia support coefficient meet the convergence condition, output the power system droop coefficient feasible boundary and the inertia support coefficient feasible boundary; The evaluation module is used for evaluating the distributed light storage final state of charge limit and the distributed light storage active power limit in the power distribution network primary frequency modulation process in the power system according to the power system droop coefficient feasible boundary and the inertia support coefficient feasible boundary.

6. The power system dynamics considering power distribution network frequency regulation capability evaluation system according to claim 5, wherein, The convergence condition is: ; wherein, a power system droop coefficient representative of a next iteration, a power system droop coefficient representative of a current iteration, a power system inertia support coefficient representative of a next iteration, a power system inertia support coefficient representative of a current iteration, and representing a set threshold value.

7. The power system dynamics considering power distribution network frequency regulation capability evaluation system according to claim 5, wherein, The model solving module is specifically used for repeating steps (a), (b) and (c) until the following convergence condition is met: ; wherein, the ith second inertia support coefficient representative of the next quadratic iteration, the ith second inertia support coefficient representative of the current quadratic iteration, the ith second droop coefficient representative of the next quadratic iteration, the ith second droop coefficient representative of the current quadratic iteration, and representing a set threshold value.

8. The power system dynamics considering power distribution network frequency regulation capability evaluation system according to claim 5, wherein, The state space mapping linear function relationship model is as follows: ; wherein, represents the least square data-driven matrix, represents the light, represents the primary frequency modulation and inertia support process in the distributed light storage primary frequency modulation starting stage power, represents the final state of charge of the distributed light storage, i represents the number of the i-th power distribution network, represents the droop coefficient of the power distribution network, represents the inertia support coefficient, represents the droop coefficient of the power system, represents the linear model between the inertia support coefficients of the power system, represents the multi-dimensional nonlinear augmented observation dimension mapping scalar function, T represents the transpose.

9. An electronic device, comprising: Including: The memory, the processor and the computer program stored on the memory and executable on the processor, when the processor executes the computer program, the power distribution network frequency modulation capacity evaluation method considering the dynamic characteristics of the power system is realized according to any one of claims 1-4.

10. A non-transitory computer-readable storage medium, comprising: The computer program is stored on the memory, and the computer program is executed by the processor to realize the power distribution network frequency modulation capacity evaluation method considering the dynamic characteristics of the power system according to any one of claims 1-4.

Images