An Equivalent Simplification Method for Photovoltaic Inverters Based on Virtual Synchronous Machines

Abstract

The present invention discloses an equivalent simplification method for a photovoltaic inverter based on a virtual synchronous machine, which relates to the technical field of inverters and includes the following steps: Step 1, according to the coherence discrimination method of synchronous generators in an AC power grid, a coherence criterion for an inverter based on virtual synchronous generator control is proposed; Step 2, grouping operations are performed on the photovoltaic inverters; Step 3, according to the coherence equivalence theory in a large power grid, a parameter aggregation method for photovoltaic inverters is proposed, and the photovoltaic inverters in the same group are aggregated into an equivalent inverter. In the process of drawing on the coherence equivalence theory of a large power grid for the photovoltaic cluster inverter based on the virtual synchronous generator control strategy, the present invention clarifies the discrimination criterion for the coherence of photovoltaic inverters, and further studies the equivalent modeling method of inverters in the electromagnetic transient analysis process. It solves the problems that there are no corresponding solutions in the selection of grouping indicators, system parameter aggregation, and equivalent model establishment in the existing photovoltaic system modeling, and the equivalent accuracy is insufficient.

Description

Technical Field

[0001] The present invention relates to the technical field of inverters, and in particular to a photovoltaic inverter equivalent simplification method based on a virtual synchronous machine. Background Art

[0002] In traditional AC systems, there have been a lot of research results on the equivalence of synchronous generator sets, among which the equivalence theory based on coherence has been widely used. In the photovoltaic cluster grid-connected power generation system, when the system operating state changes, the photovoltaic cluster grid-connected power generation system will have complex interactions with the power grid. Therefore, in order to meet the needs of large-scale application of photovoltaic power generation systems, the use of equivalent modeling methods to study the detailed model of photovoltaic clusters using equivalent models has very important research value and significance for efficiently analyzing the interaction between photovoltaic clusters and power systems and the safe and stable operation of power systems. In related fields, existing research is mostly based on grid-following control strategies. In addition, there is still room for optimization in the selection of clustering indicators, system parameter aggregation, and the establishment of equivalent models, and the accuracy of equivalents needs to be improved.

[0003] The paper "Dynamic Clustering Modeling of Regional Centralized Photovoltaic Power Generation System Based on Improved Fuzzy C-means Clustering Algorithm" takes photovoltaic inverters as the core and proposes a dynamic clustering modeling method for regional centralized photovoltaic power generation systems. This method uses the photovoltaic inverter control parameter vector and its sensitivity coefficient to the output trajectory as clustering indicators, and has made certain improvements and innovations in the clustering method; the paper "Research and Application of Equivalent Modeling of Typical Grid-connected Photovoltaic Power Stations" calculates the sensitivity of inverter parameters, obtains the characteristic distance between each inverter and uses this as the equivalence criterion. The equivalent model obtained by this method is difficult to accurately reflect the detailed model; the paper "Coherence Equivalence Method of Grid-connected Inverters and Its Dynamic Analysis Application" proposes a coherence judgment standard based on the generalized Hamiltonian action for grid-connected inverters with multiple state variables, and conducts analysis and experiments on inverters adopting grid-following control strategies.

[0004] In summary, existing research is mostly based on network-following control strategies. In addition, there is still room for optimization in issues such as grouping indicator selection, system parameter aggregation, and equivalent model establishment, and the equivalent accuracy needs to be improved. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a photovoltaic inverter equivalent simplification method based on a virtual synchronous machine. For the photovoltaic inverter based on the virtual synchronous generator control strategy, the judgment criteria for the synchronization of the photovoltaic inverter are clarified by referring to the synchronization equivalent theory of the large power grid, and then the equivalent modeling method of the inverter in the electromagnetic transient analysis process is studied.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A photovoltaic inverter equivalent simplification method based on a virtual synchronous machine includes the following steps:

[0008] Step 1. According to the synchronization judgment method of synchronous generators in AC power grid, an inverter synchronization judgment criterion based on virtual synchronous generator control is proposed;

[0009] Step 2: Perform grouping operations on photovoltaic cluster inverters;

[0010] Step 3. Based on the coherent equivalence theory in large power grids, a parameter aggregation method for photovoltaic inverters is proposed to aggregate photovoltaic inverters in the same group into an equivalent inverter.

[0011] The above-mentioned Step 1 inverter synchronization criterion is extracted from the inverter power angle dynamic characteristic curve; after the PV inverter grid-connected system disturbance occurs, the characteristic point is extracted As the clustering indicator, for and difference, is the disturbance occurrence time, It is the moment when the power angle dynamic characteristic curve reaches the first swing peak, indicating the time when the power angle dynamic characteristic curve reaches the first swing peak after the disturbance occurs. for and difference, is the power angle of the photovoltaic grid-connected inverter when the system is running stably before the disturbance occurs, It is the power angle of the first swing peak value on the power angle dynamic characteristic curve, indicating the first swing peak value of the power angle dynamic characteristic curve after the disturbance occurs.

[0012] The above Step 2 clustering adopts the coherent inverter clustering method based on the K-means algorithm, and its specific steps are as follows:

[0013] Step 2.1, extract the clustering index vector of all inverters in the detailed model; the detailed model is the actual photovoltaic inverter grid-connected system before the equivalent, and the photovoltaic inverter grid-connected system obtained after the equivalent is the equivalent model;

[0014] Step 2.2, select the initial cluster center and perform preliminary grouping;

[0015] Step 2.3, calculate the objective function value; calculate the Euclidean distance between each clustering index and the initial clustering center and sum them as the objective function;

[0016] Step 2.4, update the cluster center and divide the clusters again;

[0017] Step 2.5, repeat Step 2.3 and Step 2.4 until the change value of the K-means objective function between the two iterations is less than the allowed value, and the iteration ends.

[0018] In the above Step 2, the K-means algorithm is used to cluster the inverters in the photovoltaic cluster, and the following is obtained: K The clustering results of the cluster centers and the inverters of each photovoltaic power generation unit in the photovoltaic cluster are K is an integer.

[0019] In the above Step 3, the photovoltaic cluster structure is obtained from the clustering results obtained in Step 2, which includes c Equivalent photovoltaic system, assuming An equivalent photovoltaic system is The original photovoltaic systems are combined to obtain, 1≤ x ≤ c , the calculation method of equivalent parameters is as follows, so that all c Parameters of an equivalent photovoltaic system:

[0020] Step 3.1. Before calculating equivalent parameters, make constraints and assumptions;

[0021] Step 3.2, divide the inverter parameter aggregation into two links: control parameter aggregation and circuit parameter aggregation;

[0022] Step 3.3, calculate the virtual rotor motion equation;

[0023] Step 3.4, according to the theory of coherent equivalence, the synchronous generator is equivalenced, and the parameter aggregation in Step 3.2 is used for superposition to obtain the rotor motion equation when the synchronous generator speed is equal;

[0024] Step 3.5, according to the equal value constraint conditions, perform superposition of control parameter aggregation;

[0025] Step 3.6. n The synchronous inverters are connected to the same bus, and an equivalent model is established based on the aggregation of inverter LC filter and output line parameters.

[0026] The constraints and assumptions in Step 3.1 above include:

[0027] Step3.1.1, the voltage of the equivalent model and the detailed model at the PCC grid connection point is the same;

[0028] Step 3.1.2, the rated capacity of the equivalent model is the same as that of the detailed model, that is:

[0029] (2);

[0030] In the formula is the rated capacity of the equivalent model, is the rated capacity of each photovoltaic power generation unit in the detailed model, n is the number of synchronous inverters;

[0031] Step 3.1.3, the active power output of the equivalent model and the detailed model at the PCC grid connection point is equal, that is:

[0032] (3);

[0033] In the formula is the rated active power of the equivalent model, is the rated active power of each photovoltaic power generation unit in the detailed model;

[0034] Step 3.1.4, the reactive power output of the equivalent model and the detailed model at the PCC grid connection point is equal, that is:

[0035] (4);

[0036] In the formula is the rated reactive power of the equivalent model, is the rated reactive power of each photovoltaic power generation unit in the detailed model;

[0037] Step 3.1.5, the control parameters and circuit parameters in the detailed model can be aggregated separately.

[0038] The specific steps of Step 3.3 above are:

[0039] The virtual rotor motion equation of the inverter based on the virtual synchronous generator control strategy is:

[0040] (5);

[0041] For the photovoltaic cluster The virtual rotor motion equation of an inverter can be expressed as:

[0042] (6) ;

[0043] Where: represents the mechanical torque, Indicates The mechanical torque of the inverter, represents the electromagnetic torque, Indicates The electromagnetic torque of the inverter, Indicates The moment of inertia of a virtual synchronous generator, is the generator mechanical angular velocity, is the damping coefficient, Indicates The damping coefficient of the inverter is Indicates The synchronous angular velocity of the generator of each inverter, is the synchronous angular velocity of the generator.

[0044] The rotor motion equation in the above Step 3.4 is:

[0045] (7);

[0046] Compared with the virtual rotor motion equation of the inverter based on the virtual synchronous generator control strategy, according to the coherence equivalence theory in the traditional AC system, the control parameter aggregation method in the inverter control system is shown in Equation (8) and Equation (9):

[0047] (8);

[0048] (9) ;

[0049] Where: represents the rotor moment of inertia of the equivalent virtual synchronous generator, Indicates The moment of inertia of a virtual synchronous generator, represents the damping coefficient of the equivalent virtual synchronous generator, Indicates The damping coefficient of a virtual synchronous generator.

[0050] The specific process of Step 3.5 above is:

[0051] According to the equivalent constraint requirements, the rated capacity, active power and reactive power of the equivalent model and the detailed model are the same, which is reflected in the control system as the aggregation of active power command value and reactive power command value, as shown in Equation (10) and Equation (11):

[0052] (10);

[0053] (11);

[0054] Where: represents the active power command of the equivalent virtual synchronous generator, Indicates The active power command of the virtual synchronous generator, represents the reactive power command of the equivalent virtual synchronous generator, Indicates Reactive power instructions for virtual synchronous generators.

[0055] In the above Step 3.5, the control parameters also include reactive power voltage droop coefficient and reactive virtual inertia coefficient K , the aggregation process is:

[0056] Reactive voltage droop coefficient and reactive virtual inertia coefficient K Aggregation is also performed in an overlapping manner, as shown in (12) and (13):

[0057] (12);

[0058] (13);

[0059] Where: represents the reactive virtual inertia coefficient of the equivalent virtual synchronous generator, Indicates The reactive virtual inertia coefficient of a virtual synchronous generator, represents the reactive voltage droop coefficient of the equivalent virtual synchronous generator, Indicates Reactive power and voltage droop coefficient of a virtual synchronous generator.

[0060] The specific steps of Step 3.6 above are:

[0061] n When the synchronous inverters are connected to the same bus, the aggregation of the inverter LC filter and output line parameters is equivalent to the parallel connection of each component in the circuit, which is represented by the symbol / / , as shown in equations (14) to (16):

[0062] (14);

[0063] (15);

[0064] (16);

[0065] is the equivalent resistance of the inverter filter, , ··· is the branch resistance of the inverter filter, is the equivalent inductance of the inverter filter, , ··· is the branch inductance of the inverter filter, is the equivalent capacitance of the inverter filter, , ··· is the branch capacitance of the inverter filter.

[0066] The present invention provides a photovoltaic inverter equivalent simplification method based on a virtual synchronous machine. Through the equivalent modeling method, the detailed model of the photovoltaic cluster is studied using an equivalent model. It has very important research value and significance for efficiently analyzing the interaction between the photovoltaic cluster and the power system and the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0068] Figure 1 This is a model diagram of a bipolar photovoltaic power generation unit;

[0069] Figure 2 This is a structural diagram of a grid-connected inverter system based on virtual synchronous generator control;

[0070] Figure 3 It is a schematic diagram of the coherence equivalence process of the AC system;

[0071] Figure 4 This is a typical schematic diagram of the inverter power angle dynamic characteristic curve;

[0072] Figure 5 is a schematic diagram of the photovoltaic cluster structure after clustering;

[0073] Figure 6 yes n Schematic diagram of the circuit structure in which two synchronous inverters are connected to the same bus;

[0074] Figure 7 It is a schematic diagram of the dynamic response curve of active power to fault disturbance;

[0075] Figure 8 It is a schematic diagram of the dynamic response curve of reactive power to fault disturbance;

[0076] Fig. 9 It is a schematic diagram of the dynamic response curve of the output voltage to fault disturbance;

[0077] Fig.10 It is a schematic diagram of the dynamic response curve of the output current to fault disturbance. DETAILED DESCRIPTION

[0078] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments.

[0079] Example:

[0080] As the basis of photovoltaic clusters, photovoltaic power generation units mainly include single-stage and bipolar types. A bipolar photovoltaic power generation unit model is as follows: Figure 1 shown by Figure 1 It can be seen that the bipolar photovoltaic power generation unit model is mainly composed of photovoltaic array, DC / DC and its control module, DC / AC and its control module, filter and power grid; among them, the photovoltaic array and DC / DC and its control module can be equivalent to a DC source with constant voltage.

[0081] When the inverter and output line impedance are taken into account, the grid-connected inverter system structure based on virtual synchronous generator control is as follows Figure 2 As shown, the DC source is connected to the power grid after passing through the three-phase full-bridge inverter module, the LC filter and the output line, and its working process is carried out under the constraints of the virtual synchronous generator control algorithm and the voltage and current control loop; wherein, the DC source is equivalent to the prime mover of the virtual synchronous generator, which is used to provide energy, and the three-phase full-bridge inverter and the LC filter are equivalent to the virtual synchronous generator, which are used to realize energy conversion.

[0082] With the large-scale development of photovoltaic power generation, the number of photovoltaic arrays and inverters in photovoltaic clusters continues to increase. A large number of power electronic inverters make the power grid show high-order nonlinear characteristics to the outside world. At the same time, for regional distributed photovoltaic clusters, the parameters of each photovoltaic power generation unit are not exactly the same, so there are certain differences in dynamic characteristics. In this case, when the dynamic characteristics of different photovoltaic power generation units are quite different, using a single-machine equivalent model to equate the detailed model to a photovoltaic power generation system will produce a large error, and then the detailed model cannot be accurately characterized. However, if each photovoltaic power generation unit in the system is modeled, it will cause the problem of a very large model scale. Therefore, it is necessary to take into account the accuracy and simplification of the photovoltaic cluster equivalent model, and establish an equivalent model with sufficient simplification and standard accuracy.

[0083] The coherence equivalence process of the traditional AC system includes:

[0084] (1) Divide the generator side area and the grid side area, retain the grid side area in the equalization process, and simplify the generator side area;

[0085] (2) Determine the synchronous generator group in the generator side area;

[0086] (3) Merge the generator buses in the synchronization group;

[0087] (4) Simplify the network;

[0088] (5) Aggregate the homology group parameters to obtain equivalent machine parameters.

[0089] Its principle is as follows Figure 3 shown.

[0090] Based on the synchronous generator characteristics of the photovoltaic inverter using the virtual synchronous generator control strategy, the synchronization identification method of the synchronous generator in the AC system is considered to be used for reference, the power grid containing the photovoltaic inverter is locally equivalent, the network structure is simplified, the system simulation efficiency is improved, and the analysis process of the system is simplified.

[0091] Based on the idea of ​​synchronous equivalence in the dynamic equivalence of power systems, the inverter equivalence method studied takes the inverters with similar dynamic characteristics as the grouping principle, and uses the K-means clustering algorithm to group the inverters so that the inverters in the same group have similar dynamic characteristics. The inverters in the same group are merged into an equivalent inverter to form an equivalent model of regional distributed photovoltaic inverters. The specific process is basically the same as the synchronous equivalence steps of the AC system.

[0092] Although the photovoltaic inverter based on the virtual synchronous generator control strategy has the characteristics of a generator, and the inverter also has an observable power angle after the external system is disturbed, there are still certain differences between the power electronic inverter and the synchronous generator. Therefore, certain corrections need to be made in the equivalent process. The specific problems that need to be solved include:

[0093] ① Find the clustering index that can represent the dynamic characteristics of the inverter based on the synchronous machine control strategy;

[0094] ②Choose a suitable clustering algorithm;

[0095] ③ Based on clustering indicators, use clustering algorithms to group inverters;

[0096] ④ Parameter calculation of equivalent inverter.

[0097] The existing practical criteria for synchronous generator groups believe that in a given time period The two generators have initial power angles of and , if you use and The maximum absolute value of the difference between the two generators reflects the degree of synchronization between the two generators. When the calculated result is less than the given value When , it satisfies:

[0098] (1) ;

[0099] These generators are said to be coherent.

[0100] Generally take , .

[0101] However, there are many factors affecting the dynamic characteristics of photovoltaic inverters, including inverter circuit parameters, inverter control loop parameters, and disturbance types. Therefore, it is difficult to construct an equivalent criterion that can represent the dynamic characteristics from the inherent parameters of photovoltaic inverters.

[0102] Since there are still essential differences between power electronic inverters based on virtual synchronous generator control strategies and synchronous generators, directly applying the coherence criterion of synchronous generators will lead to incorrect discrimination of coherent inverters, thus affecting the equivalent accuracy. However, after the torque equation of the synchronous generator is implanted into the power electronic inverter, to a certain extent, the corresponding relationship between the inverter and the synchronous generator can be used for reference when selecting the coherence criterion.

[0103] Considering that power electronic inverters based on virtual synchronous machine control strategies have a shorter transient process and can enter the steady state faster after being disturbed by the system, after the system disturbance occurs, extract the characteristic points as the clustering index, where is minus , representing the time when the power angle dynamic characteristic curve reaches the first swing peak after the disturbance occurs, is minus , representing the first swing peak of the power angle dynamic characteristic curve after the disturbance occurs. A typical power angle dynamic characteristic curve of the inverter is shown in Figure 4 .

[0104] Based on the above main idea of coherent equivalence, coherent equivalence criterion, and K-means algorithm principle, a method for clustering and grouping coherent inverters based on the K-means algorithm is summarized. The specific steps are as follows:

[0105] S1: Extract the clustering index vectors of all inverters in the detailed model;

[0106] S2: Select the initial clustering centers and perform preliminary grouping;

[0107] S3: Calculate the objective function value;

[0108] S4: Update the clustering centers and perform grouping again; By taking the derivative of the input parameters with respect to the objective function, the constraints of the clustering centers and membership function constraints that minimize the objective function can be obtained. Calculate the membership using the initial clustering centers, and then update the clustering centers using the membership, and iterate continuously in this way;

[0109] S5: Repeat S3 and S4 until the change value of the K-means objective function in two consecutive iterations is less than the allowable value, and the iteration ends.

[0110] Using the K-means algorithm to cluster and group the inverters in the photovoltaic cluster, we can obtain The clustering centers and the clustering results of each photovoltaic power generation unit.

[0111] According to the above method, the inverters in the regional decentralized photovoltaic cluster are clustered. The photovoltaic cluster structure obtained after clustering is as Figure 5 shown.

[0112] Theoretically, according to the Figure 5 clustering results for equivalent modeling, the equivalent model obtained is basically the same as the detailed model in terms of the external characteristics shown at the grid PCC node to the power grid, specifically including power characteristics, frequency characteristics, power angle characteristics, voltage characteristics, current characteristics, etc.

[0113] In Figure 5 , taking the th equivalent photovoltaic system as an example, assuming that the th equivalent photovoltaic system is obtained by merging original photovoltaic systems, the calculation method of its equivalent parameters is as follows, so that the parameters of all Figure 5 equivalent photovoltaic systems in can be obtained.

[0114] Before calculating the equivalent parameters, the following constraints and assumptions are made:

[0115] ① The voltage at the PCC connection point of the equivalent model and the detailed model is the same;

[0116] ② The rated capacities of the equivalent model and the detailed model are the same, that is:

[0117] (2);

[0118] ③ The active power output at the PCC connection point of the equivalent model and the detailed model is equal, that is:

[0119] (3);

[0120] ④ The reactive power output at the PCC connection point of the equivalent model and the detailed model is equal, that is:

[0121] (4);

[0122] ⑤ The control parameters and circuit parameters in the detailed model can be aggregated separately.

[0123] Therefore, the aggregation of inverter parameters can be divided into two links: control parameter aggregation and circuit parameter aggregation.

[0124] For the inverter based on the virtual synchronous generator control strategy, its virtual rotor motion equation is:

[0125] (5);

[0126] For the th inverter in the photovoltaic cluster, its virtual rotor motion equation can be expressed as:

[0127] (6);

[0128] Where: represents the mechanical torque, represents the mechanical torque of the th inverter, represents the electromagnetic torque, represents the electromagnetic torque of the th inverter, represents the moment of inertia of the th virtual synchronous generator, is the mechanical angular velocity of the generator, is the damping coefficient, represents the damping coefficient of the th inverter, represents the synchronous angular velocity of the generator of the th inverter, is the synchronous angular velocity of the generator.

[0129] In the traditional AC system, according to the coherency equivalence theory, when equivalenting synchronous generators, the aggregation method of each parameter in the rotor motion equation is achieved by superposition. When the rotational speeds of synchronous generators are equal, it can be expressed as:

[0130] (7);

[0131] Comparing the virtual rotor motion equation of the inverter based on the virtual synchronous generator control strategy and referring to the coherency equivalence theory in the traditional AC system, the aggregation method of control parameters in the inverter control system is shown in Equations (8) and (9):

[0132] (8);

[0133] (9);

[0134] According to the requirements of the equivalent constraint conditions, the rated capacity, active power, and reactive power of the equivalent model and the detailed model are the same. In the control system, it is reflected in the aggregation of the active power command value and the reactive power command value, as shown in Equations (10) and (11) specifically:

[0135] (10);

[0136] (11);

[0137] Other control parameters are also aggregated in a superposition manner, as specifically shown in (12) and (13):

[0138] (12);

[0139] (13);

[0140] n The circuit structure of the same - homology inverters connected to the same bus is as Figure 6 shown; the aggregation of the inverter LC filter and output line parameters can be regarded as the parallel equivalent of each component in the circuit, represented by the symbol / / , as specifically shown in equations (14) to (16).

[0141] (14);

[0142] (15);

[0143] (16);

[0144] is the equivalent resistance of the inverter filter, , ··· are the branch resistances of the inverter filter, is the equivalent inductance of the inverter filter, , ··· are the branch inductances of the inverter filter, is the equivalent capacitance of the inverter filter, , ··· are the branch capacitances of the inverter filter.

[0145] Theoretically, according to the above - mentioned equivalent modeling method, the established equivalent model is basically consistent with the detailed model in terms of the external characteristics presented to the power grid at the PCC node of the power grid, specifically including power characteristics, frequency characteristics, power - angle characteristics, voltage characteristics, current characteristics, etc.

[0146] In order to verify the feasibility and accuracy of the proposed equivalent simplification method for photovoltaic inverters based on virtual synchronous machines, a multi - inverter parallel simulation platform based on the virtual synchronous machine control strategy was built in the Matlab / Simulink simulation environment. The simulation platform includes 5 grid - connected inverters controlled based on virtual synchronous generators with different parameters, which are operated in parallel on the same AC bus; during the modeling process, the inverter - side area is equivalently simplified while keeping the grid - side area unchanged.

[0147] In the process of example analysis, by extracting the clustering indexes in the power angle curves of each inverter, the coherence between each inverter is judged and the clustering groups are divided, and then the parameter aggregation is completed to obtain the multi-machine equivalent model of the inverter; finally, a simulated disturbance is applied to the grid-connected system, and the dynamic response curves of the detailed model and the equivalent model are compared and analyzed, and then the efficiency and accuracy of the equivalent model are calculated.

[0148] During the simulation process, a fault disturbance is set, and the dynamic response curves of the active power, reactive power, output voltage, and output current at the boundary line between the inverter side area and the grid side area before and after equivalence are compared. The results are as Figure 7~Figure 10 shown.

[0149] From Figure 7 and Figure 8 it can be seen that after a single-phase grounding short-circuit fault occurs in the system, after the active power and reactive power drop briefly, they quickly rise to provide frequency and voltage support, and then enter an oscillating state and gradually tend to be stable; the simulation analysis is carried out on the detailed model and the equivalent model of the system respectively. By extracting the data of the dynamic response curves of the active power and reactive power, the simulation results of the two models are compared and calculated: the maximum error in the transient process of the active power is 6.45%; the maximum error in the transient process of the reactive power is 3.85%.

[0150] In Fig. 9 and Fig.10 by extracting the data of the dynamic response curves of the output voltage and output current, the simulation results of the two models are compared and calculated: the maximum error in the transient process of the output voltage is 5.13%; the maximum error in the transient process of the output current is 1.01%.

[0151] Figure 7-10 The corresponding data table is as follows:

[0152] Table 1 Comparison of dynamic response data of active power, reactive power, voltage and current under fault disturbance

[0153]

[0154] It can be seen from the above analysis that whether it is the dynamic response curve of the active power, reactive power, output voltage or output current of the equivalent model, its similarity with the dynamic response curve of the detailed model is very high. Compared with the detailed model, the equivalent error is very small, and this error is within the allowable error range acceptable for power grid analysis.

[0155] As mentioned before, one of the purposes of constructing the inverter equivalent model is to improve the efficiency of system analysis. In this note, the analysis efficiency is reflected by the simulation time required for the detailed model and the equivalent model, as shown in Table 2 specifically.

[0156] Table 2 Time Required for Model Simulation

[0157]

[0158] As can be seen from Table 1, under the same operating conditions, the detailed model takes 108 s to complete the simulation, while the equivalent model only needs 38 s. Compared with the detailed model, the simulation efficiency of the equivalent model has increased by 184%. The example only calculates the equivalent models of 5 inverters. When the number of inverters increases, the difference in the simulation time required for the detailed model and the equivalent model will be even more significant. At the same time, the significance of constructing the equivalent model of the inverter will become more prominent.

[0159] This description focuses on the equivalent simplification method of photovoltaic inverters based on virtual synchronous machines, establishes an equivalent model and conducts simulations, indicating that the proposed coherence criterion and grouping algorithm suitable for power electronic inverters fully consider the operating conditions of the inverters, making the discrimination of the inverter coherence groups more accurate; the power electronic inverter parameter aggregation method designed in combination with the parameter aggregation process of synchronous generators in traditional large power grids is simpler and more feasible than existing methods and is more suitable for the scenario of equivalent connection of inverters with large-scale virtual synchronous generator control strategies to the grid.

Claims

1. A photovoltaic inverter equivalent simplification method based on a virtual synchronous machine, characterized in that: The following steps are involved: Step 1. According to the synchronization judgment method of synchronous generators in AC power grid, an inverter synchronization judgment criterion based on virtual synchronous generator control is proposed; The inverter synchronization criterion is extracted from the dynamic characteristic curve of the inverter power angle; after the PV inverter grid-connected system disturbance occurs, the characteristic points are extracted As the clustering indicator, for and difference, is the disturbance occurrence time, It is the moment when the power angle dynamic characteristic curve reaches the first swing peak, indicating the time when the power angle dynamic characteristic curve reaches the first swing peak after the disturbance occurs. for and difference, is the power angle of the photovoltaic grid-connected inverter when the system is running stably before the disturbance occurs, is the power angle of the first swing peak value on the power angle dynamic characteristic curve, indicating the first swing peak value of the power angle dynamic characteristic curve after the disturbance occurs; Step 2: Perform grouping operations on photovoltaic cluster inverters; Step 3: Based on the coherent equivalence theory in large power grids, a parameter aggregation method for photovoltaic inverters is proposed to aggregate photovoltaic inverters in the same group into an equivalent inverter; The Step 2 clustering adopts the coherent inverter clustering method based on the K-means algorithm, and its specific steps are as follows: Step 2.1, extract the clustering index vector of all inverters in the detailed model; the detailed model is the actual photovoltaic inverter grid-connected system before the equivalent, and the photovoltaic inverter grid-connected system obtained after the equivalent is the equivalent model; Step 2.2, select the initial cluster center and perform preliminary grouping; Step 2.3, calculate the objective function value; calculate the Euclidean distance between each clustering index and the initial clustering center and sum them as the objective function; Step 2.4, update the cluster center and divide the clusters again; Step 2.5, repeat Step 2.3 and Step 2.4 until the change value of the K-means objective function in the two iterations is less than the allowed value, and the iteration ends; In Step 2, the K-means algorithm is used to cluster the inverters in the photovoltaic cluster, and the following can be obtained: K The clustering results of the cluster centers and the inverters of each photovoltaic power generation unit in the photovoltaic cluster are K is an integer; In the above Step 3, the photovoltaic cluster structure is obtained from the clustering results obtained in Step 2, which includes c Equivalent photovoltaic system, assuming An equivalent photovoltaic system is The original photovoltaic systems are combined to obtain, 1≤ x ≤ c , the calculation method of equivalent parameters is as follows, so that all c Parameters of an equivalent photovoltaic system: Step 3.

1. Before calculating equivalent parameters, make constraints and assumptions; Step 3.2, divide the inverter parameter aggregation into two links: control parameter aggregation and circuit parameter aggregation; Step 3.3, calculate the virtual rotor motion equation; Step 3.4, according to the theory of coherent equivalence, the synchronous generator is equivalenced, and the parameter aggregation in Step 3.2 is used for superposition to obtain the rotor motion equation when the synchronous generator speed is equal; Step 3.5, according to the equal value constraint conditions, perform superposition of control parameter aggregation; Step 3.

6. n The synchronous inverters are connected to the same bus, and the equivalent model is established based on the aggregation of the inverter LC filter and output line parameters; The constraints and assumptions in Step 3.1 include: Step3.1.1, the voltage of the equivalent model and the detailed model at the PCC grid connection point is the same; Step 3.1.2, the rated capacity of the equivalent model is the same as that of the detailed model, that is: (2)； In the formula is the rated capacity of the equivalent model, is the rated capacity of each photovoltaic power generation unit in the detailed model, n is the number of synchronous inverters; Step 3.1.3, the active power output of the equivalent model and the detailed model at the PCC grid connection point is equal, that is: (3)； In the formula is the rated active power of the equivalent model, is the rated active power of each photovoltaic power generation unit in the detailed model; Step 3.1.4, the reactive power output of the equivalent model and the detailed model at the PCC grid connection point is equal, that is: (4)； In the formula is the rated reactive power of the equivalent model, is the rated reactive power of each photovoltaic power generation unit in the detailed model; Step 3.1.5, the control parameters and circuit parameters in the detailed model are aggregated separately.

2. The photovoltaic inverter equivalent simplification method based on virtual synchronous machine according to claim 1, characterized in that: The specific steps of Step 3.3 are: The virtual rotor motion equation of the inverter based on the virtual synchronous generator control strategy is: (5)； For the photovoltaic cluster The virtual rotor motion equation of an inverter is expressed as: (6) ； Where: represents the mechanical torque, Indicates The mechanical torque of the inverter, represents the electromagnetic torque, Indicates The electromagnetic torque of the inverter, Indicates The moment of inertia of a virtual synchronous generator, is the generator mechanical angular velocity, is the damping coefficient, Indicates The damping coefficient of the inverter is Indicates The synchronous angular velocity of the generator of each inverter, is the synchronous angular velocity of the generator.

3. A photovoltaic inverter equivalent simplification method based on a virtual synchronous machine according to claim 2, characterized in that: The rotor motion equation in Step 3.4 is: (7)； Compared with the virtual rotor motion equation of the inverter based on the virtual synchronous generator control strategy, according to the coherence equivalence theory in the traditional AC system, the control parameter aggregation method in the inverter control system is shown in Equation (8) and Equation (9): (8)； (9) ； Where: represents the rotor moment of inertia of the equivalent virtual synchronous generator, Indicates The moment of inertia of a virtual synchronous generator, represents the damping coefficient of the equivalent virtual synchronous generator, Indicates The damping coefficient of a virtual synchronous generator.

4. The photovoltaic inverter equivalent simplification method based on virtual synchronous machine according to claim 3 is characterized in that: The specific process of Step 3.5 is as follows: According to the equivalent constraint requirements, the rated capacity, active power and reactive power of the equivalent model and the detailed model are the same, which is reflected in the control system as the aggregation of active power command value and reactive power command value, as shown in Equation (10) and Equation (11): (10)； (11)； Where: represents the active power command of the equivalent virtual synchronous generator, Indicates The active power command of the virtual synchronous generator, represents the reactive power command of the equivalent virtual synchronous generator, Indicates Reactive power instructions for virtual synchronous generators.

5. A photovoltaic inverter equivalent simplification method based on a virtual synchronous machine according to claim 4, characterized in that: In the above Step 3.5, the control parameters also include reactive voltage droop coefficient and reactive virtual inertia coefficient K , the aggregation process is: Reactive voltage droop coefficient and reactive virtual inertia coefficient K Aggregation is also performed in an overlapping manner, as shown in (12) and (13): (12)； (13)； Where: represents the reactive virtual inertia coefficient of the equivalent virtual synchronous generator, Indicates The reactive virtual inertia coefficient of a virtual synchronous generator, represents the reactive voltage droop coefficient of the equivalent virtual synchronous generator, Indicates Reactive power and voltage droop coefficient of a virtual synchronous generator.

6. A photovoltaic inverter equivalent simplification method based on a virtual synchronous machine according to claim 5, characterized in that: The specific steps of Step 3.6 are: n When the synchronous inverters are connected to the same bus, the aggregation of the inverter LC filter and output line parameters is equivalent to the parallel connection of each component in the circuit, which is represented by the symbol / / , as shown in equations (14) to (16): (14)； (15)； (16)； is the equivalent resistance of the inverter filter, , ··· is the branch resistance of the inverter filter, is the equivalent inductance of the inverter filter, , ··· is the branch inductance of the inverter filter, is the equivalent capacitance of the inverter filter, , ··· is the branch capacitance of the inverter filter.

Images