Photovoltaic Power Station Partitioned Distributed Frequency Control Method Based on Consensus Algorithm

Abstract

The present invention discloses a photovoltaic power station partitioned distributed frequency control method based on a consensus algorithm, including: First, this control strategy adopts a partitioned distributed control method to solve problems such as a large number of devices, inconsistent parameters, and communication delays in large-scale photovoltaic power stations. Second, a frequency control strategy is designed based on the consensus algorithm, and photovoltaic inverters actively participate in power grid frequency control, which can effectively suppress the wide-frequency oscillation problem of the power system caused by photovoltaic power generation. This method improves the dynamic stability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of distributed power generation, and more particularly to a photovoltaic power station partitioned distributed frequency control method based on a consensus algorithm. Background Art

[0002] Photovoltaic power generation has the characteristics of randomness, intermittency, and volatility. The high-permeability photovoltaic access to the power grid forms a high proportion of grid-connected interfaces of power electronic devices, resulting in changes in the dynamic behavior of the power system and making it extremely easy to cause oscillation problems. The safe and stable operation of the power grid and the balance of power and electricity are facing tests. Traditional low-frequency oscillation refers to the continuous relative swing between generator rotors due to insufficient damping, and the oscillation frequency is generally between 0.1 and 2.5 Hz, also known as electromechanical oscillation.

[0003] Photovoltaic inverters are grid-connected by the coordinated cooperation of multiple control loops, and their external characteristics involve a wide frequency band due to the control bandwidth. At the same time, there is a frequency band superposition in the multi-loop control structure of photovoltaic inverters, and there is an inherent phase lag in the overlapping frequency band, making the impedance characteristics of the system show negative damping characteristics in different frequency bands. When a large-scale photovoltaic power station is connected to the grid at a high proportion, there are problems such as large differences in equipment parameters and communication delays, resulting in insufficient stability margins of the system in multiple frequency bands and causing wide-frequency oscillations. On the other hand, in a power system with a high proportion of photovoltaic grid connection, in addition to the oscillation phenomena caused by negative damping and weak damping, there will also be phenomena such as external forced oscillation, internal resonance, nonlinear oscillation, and difference-frequency oscillation between them. The frequency oscillation dominated by the frequency regulation process of the power system indicates that the reduction in the stock of adjustable resources in the high-proportion photovoltaic power generation grid-connected system causes insufficient stability in the frequency regulation process, and it is urgent to conduct systematic research on the frequency control process of large-scale photovoltaic power stations.

[0004] The frequency regulation process of the power system includes primary frequency regulation (PFR) and Automatic Generation Control (AGC), which is a dynamic system with closed-loop control. At present, responding to AGC auxiliary services is the main way for photovoltaic power stations to participate in grid frequency regulation. However, this method requires a high-quality communication system, and its reliability and flexibility are relatively low. Existing engineering operation data shows that this method also has serious hysteresis problems. 5G communication technology can significantly improve the response speed of photovoltaic power stations, but in large-scale photovoltaic power stations, this method faces problems such as communication data redundancy, and the cost is too high. On the other hand, a large number of photovoltaic inverters cause an excessive node scale, seriously increasing the burden on the communication network and reducing the safety and stability of the system.

[0005] Photovoltaic inverters need to have the ability of state perception and frequency control, and actively participate in power grid frequency regulation. At present, there are mainly two structures of photovoltaic power stations with frequency control ability: (1) Configure energy storage devices, that is, form a photovoltaic-storage combined power generation system; (2) The photovoltaic inverter unloads and reserves. When the inverter unloads and operates, its operating point deviates from the maximum power point, forming a frequency regulation capacity. Compared with installing energy storage devices, unloading operation is more flexible and avoids the investment and operation and maintenance costs of energy storage devices.

[0006] The photovoltaic inverter uses droop control to simulate the frequency response characteristics of traditional units and automatically adjusts its output according to the grid frequency. However, there are problems with control accuracy and power sharing effect in droop control, and secondary control is required to compensate for voltage amplitude and frequency deviation. The existing secondary control mainly includes three types: centralized control, distributed control, and decentralized control. Among them, centralized control requires a high-quality communication system but has low reliability; decentralized control cannot eliminate steady-state errors, and differences in inverter parameters lead to different operating states, which may cause system instability in severe cases. Summary of the Invention

[0007] The embodiment of the present invention provides a method for partitioned distributed frequency control of a photovoltaic power station based on a consensus algorithm, including:

[0008] Obtain the partitions of the photovoltaic inverters and select the typical photovoltaic inverters within the partitions;

[0009] Based on the consensus algorithm, obtain the frequency disturbance of the typical photovoltaic inverters in the current partition;

[0010] According to the frequency disturbance of the typical photovoltaic inverter, the observed frequency, the rated frequency and the rated power of the photovoltaic inverter, calculate the total active power output of the photovoltaic inverters in the partition;

[0011] When the observed frequency in the partition is greater than the given frequency regulation threshold, based on the consensus algorithm, obtain the voltage disturbance in the current partition;

[0012] According to the voltage disturbance and the voltage value corresponding to the maximum power of the photovoltaic inverter, calculate the DC voltage of each photovoltaic inverter;

[0013] According to the DC voltage, the maximum DC voltage, the load shedding rate and the maximum power of each photovoltaic inverter, calculate the active power value of each photovoltaic inverter in the partition;

[0014] Based on the consensus algorithm, adjust the active power distribution and reactive power distribution of each photovoltaic inverter in the partition based on the total output power of the photovoltaic inverters;

[0015] According to the active power and reactive power of each photovoltaic inverter, complete the frequency control of the photovoltaic inverter.

[0016] Preferably, obtaining the photovoltaic inverter partition includes:

[0017] Partitioning the photovoltaic inverters according to the principle that the electrical distances of the photovoltaic inverters are close and the natural environments are the same;

[0018] Each sub-region is a generalized node GN, forming a generalized grid topology graph GGT;

[0019] For the generalized grid topology graph GGT, use the fuzzy C-means clustering algorithm to partition the photovoltaic power station.

[0020] Preferably, the consensus algorithm includes:

[0021] Taking the voltage, current, active power, and reactive power status information of the local and adjacent nodes obtained by the node as the consensus variables for constructing the status information of the consensus protocol, and obtaining the state equation of the i-th node;

[0022]

[0023] Among them, x i represents the state variable of node i, and this node only communicates with its adjacent nodes. When the state variable values of all nodes are the same, the system reaches consensus convergence. u i (t) is the control protocol of node i, which is determined by the information fed back by the adjacent nodes and is expressed as Among them, c is the coupling weight;

[0024] Each node determines the output quantity through the current status information of the adjacent nodes received and its own status information, and finally achieves the network consensus control goal.

[0025] Preferably, obtaining the frequency perturbation of the current partition based on the consensus algorithm includes:

[0026] Introducing a frequency perturbation observer in the typical photovoltaic inverter in the frequency response control layer;

[0027] Calculating the cumulative observed frequency deviation according to the observed values of the observer in the i-th region and the adjacent regions. The cumulative observed frequency deviation formula is expressed as:

[0028]

[0029] In the formula, f inext represents the output frequency at the next moment, f ci represents the current measured frequency, f i and f j respectively represent the observed frequencies of the typical photovoltaic inverters i and j;

[0030] According to the consensus algorithm, obtain the frequency perturbation Δf upper .

[0031] Preferably, calculate the total active power output of the PV inverters in the partition according to the frequency disturbance of the typical PV inverter, the observed frequency, the rated frequency, and the rated power of the PV inverter, including:

[0032] Obtain the total rated power P Ni and the rated frequency f N and the maximum operating frequency f maxi ;

[0033] Calculate the frequency response coefficient

[0034] According to the frequency response coefficient and the frequency disturbance Δf upper Calculate the frequency response characteristics of the PV inverters in the partition:

[0035]

[0036] where P PVi represents the total output power of the PV inverters in the partition.

[0037] Preferably, the steps to obtain the load shedding rate include:

[0038] Obtain the P-U characteristic curve of the typical inverter in the i-th area;

[0039] Select the load shedding operation mode where the operating point voltage of the inverter operates on the right side of the maximum power point, and obtain the spare capacity of the PV inverter:

[0040] ΔP Bi = P upi - P MPPi

[0041] where P upi is the output power, and P MPPi is the maximum power point corresponding to the maximum DC voltage U MPPi ; approximately linearize the PV output P-U curve to obtain the load shedding rate:

[0042]

[0043] where U i represents the DC voltage, and β i represents the load shedding rate.

[0044] Preferably, calculate the active power value of each PV inverter in the partition according to the DC voltage, the maximum DC voltage, the load shedding rate, and the maximum power of each PV inverter, including:

[0045] Obtain the maximum DC voltage U MPPk of the k-th PV inverter in the partition and its corresponding maximum power point P MPPk ;

[0046] According to the maximum DC voltage U MPPk and its corresponding maximum power P MPPk and the load shedding rate, calculate the output active power of each photovoltaic inverter in the area:

[0047]

[0048] Among them, represents the DC voltage of the kth photovoltaic inverter in the ith area, reference value of the output active power;

[0049] Obtain the voltage value corresponding to the maximum power

[0050] When the photovoltaic inverter operates at the output active power , obtain the voltage disturbance ΔU of the current area based on the consensus algorithm i ;

[0051] Calculate the DC voltage of each photovoltaic inverter:

[0052]

[0053] Preferably, based on the consensus algorithm, adjust the active power distribution of each inverter in the area according to the sum of the output powers of the photovoltaic inverters, including:

[0054] Each photovoltaic inverter in the area distributes active power according to the principle of equal adjustment micro-increment, and the expression of active power distribution includes:

[0055]

[0056] Among them, PV dcim is the adjustment micro-increment of the mth typical photovoltaic inverter in the ith area, N PV is the set of photovoltaic inverters, and the other several parameters respectively represent: α im represents the adjustment micro-increment slope of the mth typical photovoltaic inverter in the ith area, β im represents the adjustment micro-increment deviation of the mth typical photovoltaic inverter in the ith area, and each photovoltaic inverter in the area distributes reactive power according to the principle of equal proportion:

[0057]

[0058] Among them and Q im are the reactive power / voltage droop coefficient and reactive power of the mth photovoltaic inverter in the ith area respectively, and the of the photovoltaic inverter are equal.

[0059] The embodiments of the present invention provide a photovoltaic power station zoning distributed frequency control method based on a consensus algorithm. Compared with the prior art, the beneficial effects are as follows:

[0060] The photovoltaic power station zoning distributed frequency modulation control based on the consensus algorithm proposed by the present invention, which includes three layers of frequency response control, load shedding control, and voltage control, enables the photovoltaic inverters to actively participate in the power grid frequency modulation, solves the problem of non-uniform equipment parameters of large-scale photovoltaic power stations, and improves the dynamic stability during the frequency modulation process of the power system.

[0061] By enabling the photovoltaic inverters to actively participate in frequency regulation, the present invention alleviates many problems such as the decline in the stock of adjustable resources in a high-proportion new energy grid-connected system, the frequent occurrence of wide-band oscillations in the power system, and the insufficient frequency regulation ability of the power system. Description of the Drawings

[0062] Figure 1 It is the photovoltaic inverter control architecture of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;

[0063] Figure 2 It is the photovoltaic output curve of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;

[0064] Figure 3 It is the simulation model circuit of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;

[0065] Figure 4 It is the active power-frequency droop characteristic curve of the photovoltaic inverter of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;.

[0066] Figure 5 It is the simulation result of Condition 1 of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;

[0067] Figure 6 It is the simulation result of Condition 2 of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention;

[0068] Figure 7 It is the simulation result of Condition 3 of the photovoltaic power station zoning distributed frequency control method based on the consensus algorithm provided by the embodiments of the present invention. Detailed Embodiments

[0069] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0070] See Figures 1 to 7 , the embodiment of the present invention provides a photovoltaic power station partitioned distributed frequency control method based on a consensus algorithm. The method includes:

[0071] (1) Photovoltaic inverter clustering and partitioning

[0072] Clustering is to generate a set of objects according to similarity. These objects are "similar" to each other in the same cluster, while "different" from the objects in other clusters. According to the concept of clustering, the photovoltaic inverters are partitioned according to the principle of close electrical distance and basically the same natural environment (irradiance, temperature), and each sub-region is set as a generalized node (GN) to form a generalized grid topology graph (GGT). The fuzzy C-means clustering algorithm is used to partition the photovoltaic power station. Suppose there are n regions of photovoltaic inverters in the photovoltaic power station participating in AGC auxiliary services, and according to the FCM partitioning algorithm in the previous section, the operating states of the photovoltaic inverters in the i-th region (0 < i ≤ n) are the same. First, select the photovoltaic inverter close to the clustering center in the i-th region as the typical inverter in this region.

[0073] (2) Distributed control based on consensus algorithm

[0074] The photovoltaic inverters participating in frequency modulation adopt a duplex communication mode, and its network topology structure can be treated as an undirected graph. For the undirected communication network G(V, E), let V = {v1, ···, v n} represent the n nodes of the network, represent the communication lines in the network, and {v i , v j} ∈ E represents that there is a communication connection between nodes i and j.

[0075] Let the operating frequency of the typical inverter in each sub-region be the node of the network. The communication network G can be described by the coupling matrix A = (a ij )n×n to represent the relationship between nodes and characterize the network topology structure. Among them, the non-diagonal element a ij is the corresponding element of the adjacency matrix of the node connection graph. If there is an information link between node i and node j, then a ij > 0 ((v i , v j ) ∈ E), indicating that there is a communication connection between devices, otherwise a ij = 0, and the diagonal element a ij = 0.

[0076] According to the consistency control principle, each distributed node exchanges information with adjacent nodes and updates the local photovoltaic inverter control strategy, ultimately enabling the state information of all nodes to achieve consistent convergence. In distributed control based on the consistency principle, each distributed node can only obtain the state information such as voltage, current, active power, and reactive power of the local and adjacent nodes. These state information used to construct the consistency protocol are called consistency variables. Thus, the state equation of the i-th node can be expressed as

[0077]

[0078] where x i represents the state variable of node i, and this node only communicates with its adjacent nodes. The system reaches consistent convergence if and only if the state variable values of all nodes are the same. u i (t) is the control protocol of node i, which is determined by the information fed back by adjacent nodes and satisfies the following formula:

[0079]

[0080] where c represents the coupling strength or coupling weight, and its value determines the speed of network consistent convergence. Each node determines the output quantity only through the state information of the currently received adjacent nodes and its own state information, ultimately achieving the network consistency control goal. Define the Laplacian matrix of the network based on the adjacency matrix

[0081] L = [l ij n×n (3)

[0082] where and when i ≠ j, l ij = -a ij ≤ 0. Therefore, the system equation can be sorted out as

[0083]

[0084] If the system has a fixed network topology structure, then L is time-invariant. The stability of the system is determined by the eigenvalues of matrix L. From the definition of matrix elements, the Laplacian matrix is a real symmetric matrix, which can be diagonalized and the eigenvalues are all real numbers greater than 0. From equation (4), it can be seen that the state of each node in the system ultimately converges to the global average value.

[0085] (3) Photovoltaic power station frequency control strategy

[0086] The present invention constructs a distributed frequency modulation control architecture, which is mainly divided into three aspects: control objective, control strategy, and control voltage.

[0087] (a) Control objective

[0088] The active power is distributed in the PV inverter area according to the principle of equal regulation of micro-increments to achieve active power control:

[0089]

[0090] where PV dcim is the regulation micro-increment of the m-th typical PV inverter in the i-th area, and N PV is the set of PV inverters.

[0091] Support the grid frequency to recover to the reference frequency:

[0092]

[0093] In the formula, f im and f ref are respectively the frequency measurement value and the reference value of the m-th PV inverter in the i-th area.

[0094] The PV inverters in the area distribute reactive power according to the principle of equal proportion:

[0095]

[0096] where and Q im are respectively the reactive power / voltage droop coefficient and the reactive power of the m-th PV inverter in the i-th area. In this article, the of each PV inverter are equal to achieve equal sharing of reactive power.

[0097] (b) Control strategy

[0098] To achieve the control objective, this article divides the control architecture into three parts: the frequency response layer, the load shedding control layer, and the voltage control layer, as Figure 1 shown. The main function of the frequency response layer is to actively support the grid frequency through information interaction between PV inverters and avoid problems such as unbalanced active power distribution and system instability caused by frequency instability. When the grid frequency is greater than the frequency modulation threshold, the PV inverter automatically switches from the MPPT mode to the frequency modulation mode, and its output is adjusted by the load shedding control layer.

[0099] Frequency response control layer

[0100] A frequency disturbance observer is introduced in the frequency response control layer. The frequency disturbance observer updates the output value of the next moment according to the frequencies of the observer in the i-th area and its adjacent areas.

[0101]

[0102] In the formula, f inext represents the node frequency at the next moment, and f ciRepresents the current measured value, f i and f j respectively represent the calculated frequency values of nodes i and j. At the same time, Δf is formed according to the consensus algorithm upper , defining the frequency response characteristics of the photovoltaic inverter

[0103]

[0104] and

[0105]

[0106] In the formula, f N represents the rated frequency, P N is the total rated power of the photovoltaic inverters in the i-th area, f maxi is the upper limit of the operating frequency.

[0107] Load shedding control layer

[0108] (1) Photovoltaic inverter load shedding reserve

[0109] Figure 2 is the P-U characteristic curve of the typical inverter in the i-th area, U MPPi is the DC operating voltage at the maximum power point, and the corresponding maximum power is P MPPi .

[0110] If the DC voltage changes, the photovoltaic inverter will deviate from the maximum power point operation. Let the DC voltage be

[0111] U i =U MPPi +ΔU i (11)

[0112] In the formula, ΔU i is the change in the DC voltage of the inverter. From Figure 2 it can be seen that if ΔU i >0, then the photovoltaic inverter operates on the right side of the maximum power point, and the output power is P upi ; if ΔU i <0, then the inverter operates on the left side of the maximum power point, and the output power is P lowi . Considering the charge and discharge effect of the DC side capacitor, select the load shedding operation mode with the operating point voltage of the inverter on the right side, that is, ΔU i >0. The spare capacity of the photovoltaic inverter is

[0113] ΔP Bi =P upi -P MPPi (12)

[0114] To improve the response speed of the control system and simplify the control equation, within the adjustable range, the P-U curve of the photovoltaic output is approximately linearized as

[0115]

[0116] (2) Load shedding control

[0117] If the photovoltaic inverters in the same area are "similar" to the typical photovoltaic inverter in this area, and there are m photovoltaic inverters in the i-th area, then for the k-th (0 < k ≤ m) photovoltaic inverter, the relationship between the load shedding rate, the disturbance voltage, and the regulation power is expressed as

[0118]

[0119] Among them, represents the reference value of the active power output of the photovoltaic inverter, and

[0120]

[0121] Then the DC-side voltage of the photovoltaic inverter

[0122]

[0123] Voltage control

[0124] The voltage control layer of the photovoltaic inverter adopts a double-loop control. The outer loop uses a PI controller to track the DC voltage reference value output by the load shedding control layer, and the inner loop uses a PI controller with current feedback.

[0125] (3) Case analysis

[0126] Based on the MATLAB / Simulink platform, establish Figure 3 the simulation model as shown. The photovoltaic inverters are divided into 6 areas, and the model parameters are shown in Table 1.

[0127] Table 1 Simulation model parameters

[0128] Tab.1 Main parameters of the simulation model

[0129]

[0130] Referring to the primary frequency regulation dead zone parameters of conventional hydro and thermal power units, set the primary frequency regulation dead zone of the photovoltaic inverter to ±0.033 Hz, and set the adjustment coefficients k1 and k2 to 0.15. The active power-frequency droop characteristic curve of the photovoltaic inverter is as Figure 4 shown.

[0131] To test the response characteristics of a photovoltaic inverter, three different operating conditions were set, and it was assumed that the irradiance and temperature remained constant under these three conditions.

[0132] Condition 1: The photovoltaic inverter operates in the MPPT mode. At the 3s moment, the grid frequency increases and exceeds the photovoltaic frequency modulation threshold, and the photovoltaic inverter switches to the frequency modulation mode. The simulation results are as Figure 5 shown.

[0133] Condition 2: The grid voltage contains odd harmonics, and the harmonic voltage amplitudes are: u5 = 0.01 p.u., u7 = 0.01 p.u. The grid frequency continuously increases from 3s to 3.4s. The simulation results are as Figure 6 shown..

[0134] Condition 3: The grid voltage contains negative sequence and zero sequence components with an amplitude of 0.05 p.u. The grid frequency continuously increases from 3s to 3.4s. The simulation results are as Figure 7 shown.

[0135] It can be seen from Figure 5 that initially, the grid frequency is the rated value, and the photovoltaic inverter operates in the MPPT mode; after the photovoltaic inverter detects that the grid frequency is greater than its frequency modulation threshold, it switches from the MPPT mode to the frequency modulation mode, and the frequency decreases; after the grid frequency deviation is less than the frequency modulation threshold, the photovoltaic inverter increases its output power, the frequency increases and is greater than the frequency modulation threshold again, and the photovoltaic inverter reduces its output again to participate in frequency modulation.

[0136] It can be seen from Figure 6 that the grid frequency continuously increases from 3s to 3.4s, and the photovoltaic inverter continuously reduces its output to perform frequency modulation. The grid background harmonics have no influence on the operating characteristics.

[0137] It can be seen from Figure 7 that the grid frequency continuously increases from 3s to 3.4s, and the photovoltaic inverter continuously reduces its output to perform frequency modulation. The positive sequence and zero sequence components of the grid have no influence on the operating characteristics.

[0138] The above discloses only several specific embodiments of the present invention. Those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the present invention. However, the embodiments of the present invention are not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A photovoltaic power station partitioned distributed frequency control method based on a consensus algorithm, characterized in that Including: Obtain the partitions of photovoltaic inverters and select typical photovoltaic inverters within the partitions; Obtain the frequency disturbances of the typical photovoltaic inverters in the current partition based on the consensus algorithm, including: introducing a frequency disturbance observer in the typical photovoltaic inverters at the frequency response control layer; Calculate the cumulative observed frequency deviation according to the observed values of the observers in the i-th area and its adjacent areas. The formula for the cumulative observed frequency deviation is expressed as: Where, f inext represents the output frequency at the next moment, f ci represents the currently measured frequency, f i and f j respectively represent the observed frequencies of the typical PV inverters i and j; According to the consensus algorithm, a frequency perturbation Δf is obtained based on the cumulative observed frequency deviation upper ; The consensus algorithm includes: Use the voltage, current, active power, and reactive power status information of the local node and its adjacent nodes obtained by the node as the consensus variables for constructing the status information of the consensus protocol, and obtain the state equation of the i-th node; Among them, x i represents the state variable of node i, which only communicates with its adjacent nodes. When the state variable values of all nodes are the same, the system reaches consensus convergence. u i (t) is the control protocol of node i, which is determined by the information fed back by adjacent nodes and is expressed as where c is the coupling weight; Each node determines the output quantity based on the status information of the adjacent nodes received currently and its own status information, and finally achieves the network consensus control goal; Calculate the total active power output of the photovoltaic inverters in the partition according to the frequency disturbances, observed frequencies of the typical photovoltaic inverters, rated frequencies, and rated powers of the photovoltaic inverters; When the observed frequency in the partition is greater than the given frequency modulation threshold, obtain the voltage disturbance in the current partition based on the consensus algorithm; Calculate the DC voltage of each photovoltaic inverter according to the voltage disturbance and the voltage value corresponding to the maximum power of the photovoltaic inverter; Calculate the active power values of each photovoltaic inverter in the partition according to the DC voltages, maximum DC voltages, load shedding rates, and maximum powers of each photovoltaic inverter; Based on the consensus algorithm, adjust the active power distribution and reactive power distribution of each photovoltaic inverter in the partition based on the total output power of the photovoltaic inverters; Complete the frequency control of the photovoltaic inverters according to the active power and reactive power of each photovoltaic inverter.

2. The photovoltaic power station zoning distributed frequency control method based on the consensus algorithm according to claim 1, characterized in that The obtaining of the partitions of photovoltaic inverters includes: Partition the photovoltaic inverters according to the principle that the electrical distances of the photovoltaic inverters are close and the natural environments are consistent; Each sub-region is a generalized node GN, forming a generalized power grid topology graph GGT; For the generalized power grid topology graph GGT, use the fuzzy C-means clustering algorithm to partition the photovoltaic power station.

3. The photovoltaic power station zoning distributed frequency control method based on the consensus algorithm according to claim 1, characterized in that The calculating of the total active power output of the photovoltaic inverters in the partition according to the frequency disturbances, observed frequencies of the typical photovoltaic inverters, rated frequencies, and rated powers of the photovoltaic inverters includes: Obtain the total rated power P within the partition Ni , rated frequency f N , maximum operating frequency f maxi ; Calculate the frequency response coefficient According to the frequency response coefficient and the frequency perturbation Δf upper Calculate the frequency response characteristics of the partitioned photovoltaic inverter: Among them, P PVi represents the total output power of the photovoltaic inverters in the partition.

4. The photovoltaic power station zoning distributed frequency control method based on the consensus algorithm according to claim 3, wherein, The steps of obtaining the load shedding rate include: Obtain the P-U characteristic curve of the typical inverter in the i-th area; Select the load shedding operation mode where the operating point voltage of the inverter operates on the right side of the maximum power point, and obtain the reserve capacity of the photovoltaic inverter: ΔP Bi = P upi - P MPPi Among them, P upi is the output power, and P MPPi is the maximum power point corresponding to the maximum DC voltage U MPPi . Approximate the P-U curve of the photovoltaic output as linear to obtain the load shedding rate; Among them, U i represents the DC voltage, and β i represents the load shedding rate.

5. The photovoltaic power station zoning distributed frequency control method based on the consensus algorithm according to claim 4, characterized in that The calculating of the active power values of each photovoltaic inverter in the partition according to the DC voltages, maximum DC voltages, load shedding rates, and maximum powers of each photovoltaic inverter includes: Obtain the maximum DC voltage U of the k-th photovoltaic inverter within the partition MPPk and its corresponding maximum power point P MPPk ; According to the maximum DC voltage U MPPk and its corresponding maximum power P MPPk and the load shedding rate, calculate the active power output of each PV inverter in the zone: Among them, represents the DC voltage of the k-th PV inverter in the i-th area, output active power reference value; Obtain the voltage value corresponding to the maximum power When the photovoltaic inverter operates at the output active power the voltage disturbance ΔU of the current partition is obtained based on the consensus algorithm i ; Calculate the DC voltage of each photovoltaic inverter:

6. The photovoltaic power station zoning distributed frequency control method based on the consensus algorithm according to claim 1, characterized in that The adjusting of the active power distribution of each inverter in the partition based on the total output power of the photovoltaic inverters according to the consensus algorithm includes: Each photovoltaic inverter in the partition distributes the active power according to the principle of equal adjustment micro-increment. The expression for the active power distribution includes: Among them, PVdcim is the adjustment micro-increment of the m-th typical photovoltaic inverter in the i-th area, NPV is the set of photovoltaic inverters, and the other several parameters respectively represent: α im represents the adjustment micro-increment slope of the m-th typical photovoltaic inverter in the i-th area, β im represents the adjustment micro-increment deviation of the m-th typical photovoltaic inverter in the i-th area. Each photovoltaic inverter within the area distributes reactive power according to the principle of equal proportion distribution: wherein and Qim are respectively the reactive power / voltage droop coefficient and the reactive power of the m-th photovoltaic inverter in the i-th region, and the are equal.

Images