Frequency coordination support optimization method for heterogeneous new energy station based on genetic algorithm

Abstract

The application discloses a heterogeneous new energy station frequency coordination support optimization method based on a genetic algorithm, which establishes a frequency response model of a heterogeneous new energy station considering the frequency space-time distribution characteristics, and analyzes the frequency dynamic characteristics of the station. On this basis, considering the frequency dynamic index and frequency modulation cost constraints of the station, a new energy station frequency coordination support optimization method based on a genetic algorithm is proposed. The proposed method comprehensively considers the coordinated optimization of grid-forming network ratio, key control parameters and energy storage capacity configuration. Compared with a typical scheme, the proposed optimization scheme can improve the frequency support effect of the station while reducing the frequency modulation cost, achieving the optimization goal of coordinated improvement of frequency safety and economic benefits.

Description

Technical Field

[0001] This invention relates to the field of power system frequency stability control and new energy grid connection optimization technology, and in particular to a frequency collaborative support optimization method for heterogeneous new energy power stations based on genetic algorithms. Background Technology

[0002] Guided by the "dual carbon" goal, the proportion of new energy sources such as wind power and photovoltaics in the national energy structure continues to rise, and the power system is undergoing a profound transformation. With the integration of a high proportion of power electronic interface power sources, the rotational inertia and primary frequency regulation capability of the power system, dominated by traditional synchronous generators, are declining, posing a severe challenge to system frequency stability. To address this, control improvements can enhance the active frequency support capability of wind and solar power units. Control improvements typically fall into two categories: grid-following control and grid-building control. Grid-following control enhances the frequency support capability of new energy sources by adding additional inertia simulation components. Grid-building control actively constructs grid voltage and frequency, providing virtual inertia and frequency support for the system.

[0003] However, with the adoption of grid-based control, renewable energy units no longer possess maximum power point tracking (MPPT) capability, and an excessively high proportion of grid-based converters will inevitably reduce the economic efficiency of the power plant. Similarly, optimizing the control parameters of the power plant can improve the frequency regulation effect of renewable energy, but it will also affect the economic efficiency of the power plant. It is worth noting that energy storage has significant advantages in inertia and primary frequency regulation response, effectively enhancing system frequency stability. However, configuring energy storage increases operation and maintenance costs as well as energy storage charging and discharging losses. In particular, conventional energy storage capacity setting methods do not consider the global optimization of grid / network ratio and control parameters, and are usually designed conservatively, resulting in an over-configuration of energy storage capacity and a significant increase in the cost of renewable energy power plants. In summary, the contradiction between the active frequency support effect and economic efficiency of renewable energy power plants urgently needs to be resolved.

[0004] Current research on active frequency support for renewable energy power plants can be divided into three categories: 1) Single-parameter tuning, which proposes methods for estimating minimum synchronous inertia and frequency containment reserves, focusing only on inertia parameter optimization without addressing converter topology matching; 2) Local configuration optimization, such as constructing a grid-type energy storage capacity optimization model and determining energy storage installation locations, but neither of these is coordinated with the control parameters of GFL / GFM converters for optimization; 3) Control method improvement, such as proposing centralized frequency correction pre-control and designing load damping factor controllers, which improve frequency regulation capabilities but do not consider the economic benefits of the power plant. Although existing research has made progress at a single level, it generally lacks research on coordinated optimization support methods from a global perspective of "grid matching - control parameters - energy storage capacity," which is the fundamental reason why it is difficult to balance the frequency support capability and economic benefits of power plants.

[0005] The main challenges in research on active frequency support for renewable energy power plants are: 1) how to analyze the frequency dynamic characteristics of different power sources such as synchronous machines, grid-connected and grid-connected, and construct a frequency response model that considers the spatiotemporal distribution characteristics of frequency; 2) how to reveal the influence of key parameters on frequency security indicators; and 3) how to achieve global optimization of grid-connected / grid-connected ratio, control parameters and energy storage capacity configuration to solve the contradiction between overall frequency support and economic benefits of the power plant.

[0006] To address this, this invention first establishes a frequency response model for heterogeneous renewable energy power stations considering the spatiotemporal distribution characteristics of frequencies, and analyzes the impact of relevant configurations and parameters on the frequency dynamics of renewable energy power stations. Furthermore, considering the frequency dynamic indicators and frequency regulation cost indicators of the power stations, it innovatively proposes a frequency support method for renewable energy power stations that considers the coordinated optimization of "grid matching / structure ratio - control parameters - energy storage capacity," which is the key technical point of this invention. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of existing research by providing a genetic algorithm-based frequency coordination support optimization method for heterogeneous new energy power stations, in order to resolve the contradiction between frequency regulation effectiveness and economic benefits of new energy power stations.

[0008] The objective of this invention is achieved through the following technical solution: a frequency collaborative support optimization method for heterogeneous new energy power stations based on genetic algorithms, comprising: With the objectives of minimizing dynamic frequency indicators and minimizing frequency regulation costs for new energy power plants, a frequency collaborative support optimization model for heterogeneous new energy power plants is constructed to solve for the equivalent frequency regulation coefficient under disturbance conditions. K PE Equivalent virtual inertia H PE Artificial delay of grid-connected converter T DA Grid-type converter capacity P GFM and energy storage configuration capacity P ESS Optimization scheme; The construction of the objective function includes: Objective function 1: Minimize the dynamic frequency index ; In the formula, 1. 2. 3 represents the frequency dynamic weighting coefficient. N The number of load nodes that may experience load fluctuations. M Power nodes and load nodes that are important for frequency dynamics are called frequency-sensitive nodes. o max,i,j Δf max,i,j Δ f ss,i,j The first i When the load fluctuates at the load node, the first j The maximum rate of frequency change, maximum frequency difference, and steady-state frequency deviation of each frequency-sensitive node; Objective function 2: Minimize frequency modulation cost ; In the formula, 4. 5. 6. 7 represents the cost weighting coefficient; The objective function is: min J = J 1 + J 2.

[0009] Furthermore, the analytic hierarchy process (AHP) is used to determine the seven weight coefficients in the objective function, specifically: a. Establish a hierarchical structure model The decision-making objectives are layered, with the objective layer consisting of frequency dynamics and frequency modulation costs; the sub-objective layer is divided into two categories, corresponding to the three indicators of frequency dynamics and the four indicators of frequency modulation costs, respectively. b. Construct the judgment matrix At the same level, based on engineering experience, use integers. p This indicates the importance of each indicator; Define the judgment matrix as follows: ; In the formula: p ij = p i / p j , p i 、p j They represent the first i The and the first j The importance of each indicator; Construct the judgment matrix of the target layer P 1 and the frequency dynamic judgment matrix of the sub-target layer P 2. Frequency modulation cost judgment matrix P 3; c. Hierarchical sorting: Calculate the eigenvector of each judgment matrix; this eigenvector represents the local weight of each indicator at that level; thus, obtain the local weight of the target layer. V 1=[ v11 , v 12 ] T Frequency dynamic local weights of sub-target layers V 2=[ v 21 , v 22 , v 23 ] T Frequency modulation cost local weight V 3=[ v 31 , v 32 , v 33 , v 34 ] T ; d. Calculate the global weights: The local weights of each layer are combined to obtain the frequency-dynamic global weights. Ω 1= v 11 N 1 ·V 2=[ 1, 2, 3] T Global weight of frequency modulation cost Ω 2= v 12 N 2 ·V 3=[ 4, 5, 6, 7] T ;in, N 1=[ n 11 , n 12 , n 13 ] T , N 2=[ n 21 , n 22 , n 23 , n 24 ] T These are the normalized matrices for frequency dynamics and frequency modulation costs, respectively, with the superscript T indicating the transpose operation.

[0010] Furthermore, it also includes adjusting the local weights of the target layer according to the requirements of the solution. V 1=[ v 11 , v 12 ] T Specifically: Option A: Frequency performance priority option, in this case... v 11 > v 12 ; Option B: Prioritizing economic indicators, in this case... v 11 < v 12 ; Option C: Balanced and coordinated optimization scheme. v 11 = v 12 .

[0011] Furthermore, a genetic algorithm is used to solve the frequency coordination support optimization model for heterogeneous new energy power stations, specifically including the following steps: (1) Initialization: Set GA parameters, including population size, number of iterations, and crossover / mutation probabilities; randomly generate an initial population within the boundary constraints of the decision variables; the decision variable vector is... X =[ K PE , H PE , T DA , P GFM , P ESS ]; (2) Scenario calculation: Traverse all scenarios where a load surge of 0.1 pu occurs at all load nodes, and calculate the maximum frequency change rate, maximum frequency deviation, and steady-state frequency deviation of frequency-sensitive nodes under each scenario based on the node frequency response model. (3) Genetic operations: Calculate the overall fitness value of the current individual; check whether all constraints are met, and impose penalties on individuals that violate the constraints; based on the fitness value, generate a new generation of population through selection, crossover, and mutation operations; (4) Termination judgment: If the maximum number of iterations is reached or the fitness converges, output the current best individual as the optimal solution to the problem; otherwise, return to step (2) to continue iterating.

[0012] Furthermore, the construction of the node frequency response model includes: A node frequency response model of the system is constructed, and the frequency modulation dynamics on the network side are characterized by the node phase angle-active response relationship described by the following formula: ; In the formula: Δ P Gen and Δ P Net These represent the power changes at voltage nodes and network nodes, respectively; Δ θ Gen and Δ θ Net These are the corresponding phase angle responses; block Laplace matrix L 11 , L 12 , L 21 , L 22 It is obtained by linearizing the phase angle-active power flow equations of the system; After organizing the above, we get: ; In the formula: L s = L 11 - L 12 L -1 22 L 21 ;Δ P Eq =- L 12 L -1 22Δ P D,Net +Δ P D,Gen This is the disturbance equivalent to the power node; The frequency regulation dynamics of new energy generating units and synchronous generators can be described by the following formula: In the formula: Δ ω g The angular frequency of the power grid; diagonal matrix G The diagonal element of (s) is the frequency-active power transfer function of the device; let the first element be the frequency-active power transfer function of the device. i The transfer function for the unit capacity of the equipment is: g i (s), the per-unit capacity is f i Then its corresponding transfer function G i (s) is the product of the two.

[0013] Furthermore, based on the node frequency response model, the maximum frequency change rate, maximum frequency deviation, and steady-state frequency deviation of frequency-sensitive nodes in each scenario are calculated, including: Based on the system frequency response model and applying the final value theorem, the steady-state frequency deviation is obtained as follows: In the formula, a This represents the proportion of the total capacity of all thermal power plants to the total capacity of the system. b This indicates the proportion of grid-connected power generation capacity in the power grid. c The proportion of grid-connected converter capacity for photovoltaic and wind power in the power grid. d The proportion of grid-connected energy storage capacity in the power grid; Δ P L For power disturbance; D sg This is the damping coefficient of the synchronizing machine; K sg This represents the product of the turbine's mechanical power gain coefficient and droop coefficient. K vsg The frequency modulation coefficient of GFM; K p This refers to the frequency regulation coefficient of the grid-type converter; The maximum rate of change of frequency occurs at the instant of the power step disturbance. Find t→0. + The slope is sufficient; according to the initial value theorem, the maximum rate of change of frequency is calculated by the s→∞ limit of the transfer function. The expression for the rate of change of frequency: ; in, o Gen The rate of change of the power node frequency; The maximum frequency change rate of the system under a step power disturbance is: ; because G The highest-order term of (s) is 2. H Gen Therefore, the maximum frequency change rate of the power node is: ; Based on the relationship between the frequency change rate of network nodes and power supply nodes, the frequency change rate of network nodes can be obtained: ; The inverse of the closed-loop transfer function matrix of the system is calculated, and an equivalent power step disturbance Δ is applied. P EqThe voltage phase angle time-domain response curves of each network node are obtained by using numerical inverse transformation or time-domain simulation. Then, the voltage phase angle time-domain response curves are differentiated with respect to time to obtain the instantaneous frequency deviation trajectory of each node. Finally, the instantaneous frequency deviation trajectory is traversed, and the maximum absolute value is extracted, which is determined as the maximum frequency difference of the system under this operating condition.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the above-described method for frequency collaborative support optimization of heterogeneous new energy power stations based on a genetic algorithm.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for frequency collaborative support optimization of heterogeneous new energy power stations based on genetic algorithms.

[0016] The beneficial effects of this invention are as follows: A frequency response model for heterogeneous new energy power stations considering the spatiotemporal distribution characteristics of frequency is established, and the frequency dynamic characteristics of the stations are analyzed. Further considering the frequency dynamic indicators and frequency regulation cost constraints of the stations, a frequency collaborative support optimization method for new energy power stations based on a genetic algorithm is proposed. Depending on different decision objectives, frequency dynamics are improved or frequency regulation costs are reduced. The balanced and coordinated optimization scheme improves the frequency support effect of the power stations while reducing frequency regulation costs. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the heterogeneous new energy power station structure corresponding to the present invention; Figure 2 This is a diagram of the improved IEEE 9-node system topology used in this invention. Figure 3 This is a flowchart of the frequency coordination support optimization method for heterogeneous new energy power stations according to the present invention; Figure 4 The frequency response of different schemes under power disturbance conditions; Figure 5 This represents the frequency change rate response of different schemes under power disturbance conditions. Detailed Implementation

[0019] To describe the present invention in more detail, the following explanation is provided in conjunction with the accompanying drawings and embodiments.

[0020] The present invention provides a frequency collaborative support optimization method for heterogeneous renewable energy power stations based on genetic algorithms, comprising the following steps: (1) Establish a node frequency response model for heterogeneous new energy power stations, specifically including: (1.1) The heterogeneous new energy power station topology studied in this invention is as follows: Figure 1 As shown. Taking the improved IEEE 9-node system as an example, the method proposed in this invention is applied, as follows: Figure 2 As shown in the figure; key parameters are shown in Table 1. The power station contains wind power, photovoltaic, and energy storage units using grid-connected and grid-connected control methods, and is divided into different power generation modules according to type. Each module is collected by a collector line, stepped up by a transformer within the power station, and then connected to the AC power grid via a main transformer. To study the system-level frequency dynamics, the grid side is equivalent to a traditional synchronous power grid containing thermal power generating units.

[0021] Table 1 Key parameters of the improved IEEE 9-node system in, P sg1 、P sg2 For the capacity of each thermal power plant, P GFL To control the capacity of the grid-connected converter, P GFM For grid-connected control converters used in photovoltaic and wind power applications, P ESS For energy storage installed capacity, Δ P L For power disturbance, H sg The inertial time constant of the synchronous machine D sg This is the damping coefficient of the synchronizing machine. T The time constant of the speed controller, α This refers to the proportion of the steam turbine's output power to its total steam turbine output power. K sg This represents the product of the turbine's mechanical power gain coefficient and droop coefficient. K D To match the virtual inertia coefficient of the grid converter, K p To match the frequency regulation coefficient of the grid converter, T PE To account for the inherent delay of mesh power supplies, T DA The delay was artificially added to the mesh power supply. H vsg The virtual inertial time constant. D vsgThe damping coefficient of GFM is... K vsg For the frequency modulation coefficient of GFM, T vsg Due to the inherent delay of GFM, ω 0 represents the system's nominal electrical angular frequency.

[0022] The node frequency response model of the system is constructed, and the frequency modulation dynamics on the network side are characterized by the node phase angle-active response relationship described by equation (1): (1) In the formula: Δ P Gen and Δ P Net These represent the power changes at voltage nodes and network nodes, respectively; Δ θ Gen and Δ θ Net These are the corresponding phase angle responses; block Laplace matrix L 11 , L 12 , L 21 , L 22 It is obtained by linearizing the phase angle-active power flow equation of the system.

[0023] Equation (1) is rearranged to obtain Equation (2): (2) In the formula: L s = L 11 - L 12 L -1 22 L 21 ;Δ P Eq =- L 12 L -1 22Δ P D,Net +Δ P D,Gen For the disturbance equivalent to the power node, Δ P D,Gen and Δ P D,Net These are the power disturbances of the power nodes and network nodes, respectively.

[0024] The frequency regulation dynamics of new energy generating units and synchronous generators can be described by equation (3): (3) In the formula: Δ ω Gen Here is the electrical angular frequency of the power node; diagonal matrix G The diagonal element of (s) is the frequency-active power transfer function of the device. Let the first element be... i The transfer function for the unit capacity of the equipment is: g i (s), the per-unit capacity is f i Then its corresponding transfer function G i (s) is the product of the two.

[0025] (1.2) Solving for the steady-state frequency deviation of the power station under a step power disturbance, based on the system frequency response model and applying the final value theorem, the steady-state frequency deviation is obtained as follows: (4) In the formula, a This represents the proportion of the total capacity of all thermal power plants to the total capacity of the system. b This indicates the proportion of grid-connected power generation capacity in the power grid. c The proportion of grid-connected converter capacity for photovoltaic and wind power in the power grid. d This represents the proportion of grid-type energy storage capacity in the power grid.

[0026] The maximum rate of frequency change occurs at the instant of the power step disturbance, therefore, we need to find t→0. + The slope is sufficient. According to the initial value theorem, the maximum rate of change of frequency can be calculated using the s→∞ limit of the transfer function.

[0027] From equation (5), the expression for the rate of change of frequency can be derived: (5) in, o Gen The frequency change rate of the power node.

[0028] Furthermore, the maximum frequency change rate of the system under a step power disturbance is: (6) because G The highest-order term of (s) is 2. H Gen Therefore, the maximum frequency change rate of the power node is: (7) Where, Δ P Eq,i for i Equivalent power disturbance of power nodes. HGen,i for i Inertia of the power node.

[0029] Based on the relationship between the frequency change rate of network nodes and power supply nodes, the frequency change rate of network nodes can be obtained: (8) The maximum frequency difference is influenced by many factors, including inertia, damping, frequency regulation coefficient, time delay, grid-to-base ratio, and energy storage capacity. The frequency response model of this system is complex, making it difficult to directly derive an analytical expression for the maximum frequency difference. Numerical calculation methods can be used to calculate the maximum frequency difference.

[0030] First, based on the Laplace matrix in the power network topology construction formula (1), and combined with the dynamic equations of synchronous generators and virtual synchronous machines, a diagonal transfer function matrix is ​​constructed, thereby establishing the system closed-loop transfer function matrix (2) describing the relationship between system power disturbance and voltage phase angle; second, by calculating the inverse matrix of the system closed-loop transfer function matrix, and applying an equivalent power step disturbance Δ... P Eq The voltage phase angle time-domain response curves of each network node are obtained by using numerical inverse transformation or time-domain simulation. Then, the voltage phase angle time-domain response curves are differentiated with respect to time to obtain the instantaneous frequency deviation trajectory of each node. Finally, the instantaneous frequency deviation trajectory is traversed, and the maximum absolute value is extracted, which is determined as the maximum frequency difference of the system under this operating condition.

[0031] (2) With the goal of minimizing the dynamic frequency index and the frequency regulation cost of new energy power stations, a frequency collaborative support optimization model for heterogeneous new energy power stations is constructed to solve the equivalent frequency regulation coefficient under large disturbance conditions. K PE Equivalent virtual inertia H PE Artificial delay of grid-connected converter T DA Grid-type converter capacity P GFM and energy storage configuration capacity P ESS The optimization scheme.

[0032] Objective function 1: Minimize the dynamic frequency index (9) In the formula, 1. 2. 3 represents the frequency dynamic weighting coefficient. N The number of load nodes that may experience load fluctuations. MPower supply nodes and load nodes that are critical for frequency dynamics are called frequency-sensitive nodes. max,i,j Δ f max,i,j Δ f ss,i,j The first i When the load fluctuates at the load node, the first j The maximum rate of frequency change, maximum frequency difference, and steady-state frequency deviation of each frequency-sensitive node.

[0033] Objective function 2: Minimize frequency modulation cost (10) In the formula, 4. 5. 6. 7 represents the cost weighting coefficient.

[0034] The objective function is: min J = J 1 + J 2.

[0035] The priority of the seven weight coefficients in the objective function is determined using the analytic hierarchy process (AHP).

[0036] a. Establish a hierarchical structure model The decision-making objectives are stratified, with the objective layer consisting of frequency dynamics and frequency modulation cost; the sub-objective layer is divided into two categories, corresponding to the three indicators of frequency dynamics (maximum rate of change, maximum frequency difference, and steady-state frequency deviation) and the four indicators of frequency modulation cost (equivalent frequency modulation coefficient). K PE Equivalent virtual inertia H PE Grid-type converter capacity P GFM and energy storage configuration capacity P ESS ).

[0037] b. Construct the judgment matrix At the same level, based on engineering experience, use integers from 1 to 9. p This indicates the importance of each indicator.

[0038] Define the judgment matrix as follows: (11) In the formula: p ij = p i / p j , pi 、p j They represent the first i The and the first j The importance of each indicator.

[0039] Construct the judgment matrix of the target layer P 1 and the frequency dynamic judgment matrix of the sub-target layer P 2. Frequency modulation cost judgment matrix P 3.

[0040] c. Hierarchical sorting: Calculate the eigenvector of each judgment matrix; this eigenvector represents the local weight of each indicator at that level.

[0041] Obtain the local weights of the target layer V 1=[ v 11 , v 12 ] T Frequency dynamic local weights of sub-target layers V 2=[ v 21 , v 22 , v 23 ] T Frequency modulation cost local weight V 3=[ v 31 , v 32 , v 33 , v 34 ] T The superscript T indicates the transpose operation. v 11 , v 12 v 21 , v 22 , v 23 、 v 31 , v 32 , v 33 and v 34 These are used to characterize the importance of frequency dynamics, frequency regulation cost, maximum frequency change rate, maximum frequency difference, steady-state frequency deviation, effective frequency regulation coefficient, equivalent virtual inertia, grid-type converter capacity, and energy storage configuration capacity, respectively.

[0042] d. Calculate the global weights: The local weights of each layer are combined to obtain the global weight of the bottom-level indicator relative to the overall goal.

[0043] Obtain the frequency dynamic global weights Ω 1= v 11 N 1 ·V 2=[ 1, 2, 3] T Global weight of frequency modulation cost Ω 2= v 12 N 2 ·V 3=[ 4, 5, 6, 7] T .in, N 1=[ n 11 , n 12 , n 13 ] T , N 2=[ n 21 , n 22 , n 23 , n 24 ] T These are the normalized matrices for frequency dynamics and frequency modulation costs, respectively.

[0044] When obtaining weighting coefficients using AHP, three optimization schemes can be derived by setting the importance of frequency dynamics and frequency modulation costs. In the specific implementation of the invention, the frequency dynamics judgment matrix... P 2 and frequency modulation cost judgment matrix P 3 is: (12) Find the eigenvectors to obtain the local weights of each element in the frequency dynamic hierarchy. V 2 = [0.88, 0.44, 0.18] T .

[0045] (13) Find the eigenvectors to obtain the local weights of each element in the frequency dynamic hierarchy. V3 = [0.06, 0.15, 0.44, 0.88] T .

[0046] Based on the range of the frequency modulation parameters, the normalized matrix is ​​obtained as follows: N =[1, 1 / 0.5, 1 / 0.3, 1 / 6, 1 / 30, 1 / 120, 1 / 120] T .

[0047] Option A: Frequency performance priority option, setting the importance of target layer frequency dynamics to frequency modulation cost at a ratio of 2:1. In this case, V 1 = [0.89, 0.45] T ;Calculation obtained Ω =[0.78, 0.78, 0.53, 0.0193, 0.0023, 0.0017, 0.0033] T .

[0048] Option B: The economic indicator priority option sets the importance of frequency dynamics and frequency modulation cost at the target layer to 1:2. In this case, V 1 = [0.45, 0.89] T ;Calculation obtained Ω =[0.4, 0.4, 0.27, 0.0381, 0.0045, 0.0033, 0.0065] T .

[0049] Option C: Balanced Coordination and Optimization Scheme, setting the importance of frequency dynamics and frequency modulation cost at the target layer to 1:1. In this case, V 1=[0.7,0.7] T Calculations yielded Ω =[0.62, 0.62, 0.42, 0.03, 0.0035, 0.0026, 0.0051] T .

[0050] Typical solution: H PE and K PE The settings are based on the normalized values ​​of typical synchronous generator sets; no consideration is taken into account. T DA The impact on frequency support is taken as 0.5s; the grid connection ratio is taken as 1:1 to balance maximum power point tracking and grid support capabilities; the energy storage capacity is taken as 20% of the total capacity of the site, based on the industry's common recommended value.

[0051] After renewable energy power plants participate in frequency regulation, the system frequency should meet the dynamic frequency index constraints: (14) In the formula, o MAX Δ is the maximum allowable rate of frequency change of the system. f MAX The frequency deviation value set for triggering low-frequency load shearing in the system; Δ f ss,MAX This represents the maximum allowable steady-state frequency deviation of the system.

[0052] When renewable energy power plants participate in frequency regulation, the frequency support parameters should meet the following constraints: (15) The load power disturbance type is set to a power deficit of 0.1 pu. The system reference frequency is 50 Hz, and the maximum RoCoF value is... o MAX Set to 1Hz / s, low-frequency load shearing operation frequency Δ f MAX Set to 0.5Hz, the system allows a maximum steady-state frequency deviation Δ f ss,MAX Set to 0.3Hz.

[0053] (2.2) The above model is solved using a genetic algorithm, and the decision variable vector is: X =[ K PE , H PE , T DA , P GFM , P ESS See also ]. Figure 3 The solution steps specifically include: 1) Population Initialization and Parameter Settings: Population size is 50, maximum number of iterations is 50, crossover probability is set to 0.8, and mutation probability is set to 0.01. Based on equipment physical limitations, boundary constraints are set for the decision variables: virtual inertia. H PE ∈[5, 6]s, droop coefficient K PE ∈[10, 30], with network latency T DA ∈[0.1, 3.0]s, network capacity P GFM ∈[10,100]MW, energy storage capacity P ESS ∈[10, 20]MW. Within the above constraint space, an initial population of 50 individuals is randomly generated using a uniform distribution.

[0054] 2) Multi-scenario fitness assessment: Substitute the parameters of each individual in the population into the system frequency response model (1)-(3). Traverse the preset fault set, and all load nodes are subjected to a 0.1 pu load step disturbance in sequence. Calculate the maximum frequency change rate of the frequency-sensitive nodes under each scenario. o max Maximum frequency difference Δ f max and steady-state frequency difference Δ f ss .

[0055] 3) Constraint handling and penalty function calculation: Check whether the above frequency indicators meet the system security constraints. o max ≤1.0Hz / s, Δ f max ≤ 0.5Hz, Δ f ss ≤ 0.3Hz. A penalty function method is used to handle constraints: for any individual that violates a constraint, its objective function value is forcibly set to a maximum penalty value, thereby reducing its survival probability in subsequent evolution.

[0056] 4) Genetic Evolution Operations: Based on the calculated individual fitness values, the following genetic operations are performed to generate the offspring population. Selection operation: Individuals with high fitness are selected from the current population as parents; Crossover operation: Arithmetic crossover is performed on the selected parent individuals with a probability of 0.8 to generate offspring with new parameter combinations, in order to retain superior gene segments; Mutation operation: Non-uniform mutation is performed on the gene positions of the offspring individuals with a probability of 0.01 to introduce random perturbation to maintain population diversity and prevent the algorithm from getting trapped in local optima.

[0057] 5) Convergence Criterion and Output: Determine whether the maximum number of iterations has been reached or whether the average fitness of the population has stopped decreasing for several consecutive generations. If the termination condition is met, output the parameter combination corresponding to the individual with the best fitness in the current population as the final optimal solution; otherwise, return to step 2 to continue the next generation of evolution.

[0058] Based on the above algorithm, by adjusting the ratio of the frequency security index weight in objective function (9) to the economic cost weight in objective function (10), schemes A, B, and C are obtained, thus yielding the optimal parameter combinations for the three different optimization schemes shown in Table 2: Table 2 Comparison of optimization results under different schemes Substituting the optimization parameters of the typical scheme and three optimized schemes into the model, a power disturbance of -0.1 pu is introduced at load node 8. Taking load node 8 and power supply node 3 as representatives, their frequency and rate of change response are observed, such as... Figure 4 and Figure 5 As shown. By Figure 4 It can be seen that, with the same power disturbance, for controlling the maximum frequency difference and steady-state frequency difference, scheme A is the most effective, followed by scheme C, then the typical scheme, and scheme B is the worst. Figure 5 It can be seen that, with the same power disturbance, scheme A is the most effective for controlling the maximum rate of frequency change, followed by scheme C, while the typical scheme and scheme B are the worst. This is consistent with the target of the optimization scheme.

[0059] Multiplying the indicators by their respective weighting coefficients and summing them yields a comprehensive evaluation of the system's frequency dynamics or frequency regulation costs. Therefore, the frequency dynamics index and the frequency regulation cost index are defined as follows: (16) (17) in, , , ; FDI is the frequency dynamic index; the smaller it is, the better the system's frequency dynamics. FRCI is the frequency regulation cost index; the smaller it is, the lower the frequency regulation cost of the site.

[0060] Substituting the optimization parameters of the typical scheme and the three optimized schemes into the model, we can obtain the FDI and FRCL under this scheme, as shown in Tables 3 and 4.

[0061] Table 3 Comparison of FDI for different schemes Table 4 Comparison of FRCI under different schemes As shown in Tables 3 and 4, Scheme A has a frequency dynamic index that is 12.1% lower than the typical scheme and a frequency modulation cost index that is 9.0% higher than the typical scheme. Scheme A achieves the best frequency dynamics, but its frequency modulation cost is relatively high. Scheme B has a frequency dynamic index that is 8.9% higher than the typical scheme and a frequency modulation cost index that is 21.1% lower than the typical scheme. Scheme B achieves the lowest frequency modulation cost, but its frequency dynamics are relatively poor. Scheme C has a frequency dynamic index that is 1.5% lower than the typical scheme and a frequency modulation cost index that is 10.9% lower than the typical scheme. It is better than the typical scheme in both frequency dynamics and frequency modulation cost.

[0062] In summary, different optimization schemes set different priorities for frequency dynamics and frequency regulation costs, and all achieved their respective objectives. Among them, the balanced and coordinated optimization scheme C outperformed the typical scheme in both frequency dynamics and frequency regulation costs. Therefore, the frequency collaborative support optimization method for heterogeneous new energy power stations proposed in this invention is effective.

[0063] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the above-described method for frequency collaborative support optimization of heterogeneous new energy power stations based on a genetic algorithm.

[0064] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for frequency collaborative support optimization of heterogeneous new energy power stations based on genetic algorithms.

[0065] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied 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.

[0066] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0067] 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.

[0068] 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 1The steps of the function specified in one or more boxes.

[0069] The above embodiments are only used to illustrate the design concept and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes or modifications made based on the principles and design ideas disclosed in the present invention are within the protection scope of the present invention.

Claims

1. A frequency collaborative support optimization method for heterogeneous new energy power stations based on genetic algorithms, characterized in that, include: With the objectives of minimizing dynamic frequency indicators and minimizing frequency regulation costs for new energy power plants, a frequency collaborative support optimization model for heterogeneous new energy power plants is constructed. A genetic algorithm is used to solve for the equivalent frequency regulation coefficient under disturbance conditions. K PE Equivalent virtual inertia H PE Artificial delay of grid-connected converter T DA Grid-type converter capacity P GFM and energy storage configuration capacity P ESS Optimization scheme; The construction of the objective function includes: Objective function 1: Minimize the dynamic frequency index ； In the formula, 1.

2. 3 represents the frequency dynamic weighting coefficient. N The number of load nodes that may experience load fluctuations. M Power nodes and load nodes that are important for frequency dynamics are called frequency-sensitive nodes. o max,i,j Δ f max,i,j Δ f ss,i,j The first i When the load fluctuates at the load node, the first j The maximum rate of frequency change, maximum frequency difference, and steady-state frequency deviation of each frequency-sensitive node; Objective function 2: Minimize frequency modulation cost ； In the formula, 4.

5.

6. 7 represents the cost weighting coefficient; The objective function is: min J = J 1+ J 2.

2. The method according to claim 1, characterized in that, The seven weight coefficients in the objective function were determined using the analytic hierarchy process (AHP), as follows: a. Establish a hierarchical structure model The decision-making objectives are layered, with the objective layer consisting of frequency dynamics and frequency modulation costs; the sub-objective layer is divided into two categories, corresponding to the three indicators of frequency dynamics and the four indicators of frequency modulation costs, respectively. b. Construct the judgment matrix At the same level, based on engineering experience, use integers. p This indicates the importance of each indicator; Define the judgment matrix as follows: ； In the formula: p ij = p i / p j , p i 、p j They represent the first i The and the first j The importance of each indicator; Construct the judgment matrix of the target layer P 1 and the frequency dynamic judgment matrix of the sub-target layer P 2. Frequency modulation cost judgment matrix P 3; c. Hierarchical sorting: Calculate the eigenvector of each judgment matrix; this eigenvector represents the local weight of each indicator at that level; thus, obtain the local weight of the target layer. V 1=[ v 11 , v 12 ] T Frequency dynamic local weights of sub-target layers V 2=[ v 21 , v 22 , v 23 ] T Frequency modulation cost local weight V 3=[ v 31 , v 32 , v 33 , v 34 ] T ; d. Calculate the global weights: The local weights of each layer are combined to obtain the frequency-dynamic global weights. Ω 1= v 11 N 1 ·V 2=[ 1, 2, 3] T Global weight of frequency modulation cost Ω 2= v 12 N 2 ·V 3=[ 4, 5, 6, 7] T ;in, N 1=[ n 11 , n 12 , n 13 ] T , N 2=[ n 21 , n 22 , n 23 , n 24 ] T These are the normalized matrices for frequency dynamics and frequency modulation costs, respectively, with the superscript T indicating the transpose operation.

3. The method according to claim 2, characterized in that, This also includes adjusting the local weights of the target layer according to the requirements of the solution. V 1=[ v 11 , v 12 ] T Specifically: Option A: Frequency performance priority option, in this case... v 11 > v 12 ; Option B: Prioritizing economic indicators, in this case... v 11 < v 12 ; Option C: Balanced and coordinated optimization scheme. v 11 = v 12 .

4. The method according to claim 1, characterized in that, The genetic algorithm is used to solve the frequency coordination support optimization model for heterogeneous new energy power stations. The specific steps include: (1) Initialization: Set GA parameters, including population size, number of iterations, and crossover / mutation probabilities; randomly generate an initial population within the boundary constraints of the decision variables; the decision variable vector is... X =[ K PE , H PE , T DA , P GFM , P ESS ]; (2) Scenario calculation: Traverse all scenarios where a load surge of 0.1 pu occurs at all load nodes, and calculate the maximum frequency change rate, maximum frequency deviation, and steady-state frequency deviation of frequency-sensitive nodes under each scenario based on the node frequency response model. (3) Genetic operations: Calculate the overall fitness value of the current individual; check whether all constraints are met, and impose penalties on individuals that violate the constraints; based on the fitness value, generate a new generation of population through selection, crossover, and mutation operations; (4) Termination judgment: If the maximum number of iterations is reached or the fitness converges, output the current best individual as the optimal solution to the problem; otherwise, return to step (2) to continue iterating.

5. The method according to claim 4, characterized in that, The construction of the node frequency response model includes: A node frequency response model of the system is constructed, and the frequency modulation dynamics on the network side are characterized by the node phase angle-active response relationship described by the following formula: ； In the formula: Δ P Gen and Δ P Net These represent the power changes at voltage nodes and network nodes, respectively; Δ θ Gen and Δ θ Net These are the corresponding phase angle responses; block Laplace matrix L 11 , L 12 , L 21 , L 22 It is obtained by linearizing the phase angle-active power flow equations of the system; After organizing the above, we get: ； In the formula: L s = L 11 - L 12 L -1 22 L 21 ;Δ P Eq =- L 12 L -1 22Δ P D,Net +Δ P D,Gen This is the disturbance equivalent to the power node; The frequency regulation dynamics of new energy generating units and synchronous generators can be described by the following formula: In the formula: Δ ω g The angular frequency of the power grid; diagonal matrix G The diagonal element of (s) is the frequency-active power transfer function of the device; let the first element be the frequency-active power transfer function of the device. i The transfer function for the unit capacity of the equipment is: g i (s), the per-unit capacity is f i Then its corresponding transfer function G i (s) is the product of the two.

6. The method according to claim 5, characterized in that, Based on the node frequency response model, the maximum frequency change rate, maximum frequency deviation, and steady-state frequency deviation of frequency-sensitive nodes in each scenario are calculated, including: Based on the system frequency response model and applying the final value theorem, the steady-state frequency deviation is obtained as follows: In the formula, a This represents the proportion of the total capacity of all thermal power plants to the total capacity of the system. b This indicates the proportion of grid-connected power generation capacity in the power grid. c The proportion of grid-connected converter capacity for photovoltaic and wind power in the power grid. d The proportion of grid-connected energy storage capacity in the power grid; Δ P L For power disturbance; D sg This is the damping coefficient of the synchronizing machine; K sg This represents the product of the turbine's mechanical power gain coefficient and droop coefficient. K vsg The frequency modulation coefficient of GFM; K p This refers to the frequency regulation coefficient of the grid-type converter; The maximum rate of change of frequency occurs at the instant of the power step disturbance. Find t→0. + The slope is sufficient; according to the initial value theorem, the maximum rate of change of frequency is calculated by the s→∞ limit of the transfer function. The expression for the rate of change of frequency: ； in, o Gen The rate of change of the power node frequency; The maximum frequency change rate of the system under a step power disturbance is: ； because G The highest-order term of (s) is 2. H Gen Therefore, the maximum frequency change rate of the power node is: ； Based on the relationship between the frequency change rate of network nodes and power supply nodes, the frequency change rate of network nodes can be obtained: ； The inverse of the closed-loop transfer function matrix of the system is calculated, and an equivalent power step disturbance Δ is applied. P Eq The voltage phase angle time-domain response curves of each network node are obtained by using numerical inverse transformation or time-domain simulation. Then, the voltage phase angle time-domain response curves are differentiated with respect to time to obtain the instantaneous frequency deviation trajectory of each node. Finally, the instantaneous frequency deviation trajectory is traversed, and the maximum absolute value is extracted, which is determined as the maximum frequency difference of the system under this operating condition.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a frequency collaborative support optimization method for heterogeneous new energy power stations based on genetic algorithms as described in any one of claims 1-6.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the frequency collaborative support optimization method for heterogeneous new energy power stations based on genetic algorithms as described in any one of claims 1-6.

Images