Simulation method of photovoltaic power station
By clustering processing and equivalent model construction of the inverter of the photovoltaic power station, combining steady-state and transient data, the problem of simulation speed and accuracy in large-scale photovoltaic power station simulation is solved, and efficient and accurate simulation modeling is achieved.
Patent Information
- Application Number
- CN202411140702.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2025-01-14
AI Technical Summary
When simulation modeling large-scale photovoltaic power plants, it is difficult for the existing technology to improve the simulation calculation speed while ensuring simulation accuracy, resulting in too long simulation process time or introduction of simulation errors.
By acquiring the line impedance and output data of all inverters in the photovoltaic power station, collecting steady-state and transient data, performing clustering processing, building an equivalent model, and adjusting the internal parameters of the clustering algorithm to improve simulation accuracy and efficiency.
It realizes the improvement of simulation speed on the basis of ensuring simulation accuracy, reduces simulation errors, improves the fitting accuracy of steady-state and transient processes, simplifies the simulation model, and improves the simulation efficiency.
Smart Images
Figure CN119312657A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of photovoltaic power station design, and in particular to a simulation method for a photovoltaic power station. Background Art
[0002] At present, with the development of photovoltaic power stations, they contain an increasing number of power generation units. Each power generation unit contains a photovoltaic array and an inverter for converting the light energy of the photovoltaic array into electrical energy. The output of the inverter in each power generation unit is connected to the power grid at the collection point. When simulating a photovoltaic power station with a large number of photovoltaic arrays and inverters, if a detailed model is established for each power generation unit, although high-precision simulation can be achieved, the detailed model included in the simulation process is large in scale, high in complexity, and the simulation time is too long. If a single power generation unit is used to represent all power generation units to establish a stand-alone model, although the complexity of the stand-alone model is reduced and the simulation speed is increased, it will inevitably introduce simulation errors and reduce simulation accuracy. Therefore, how to improve the simulation calculation speed while ensuring the simulation accuracy is a technical problem that needs to be solved urgently when simulating and modeling large-scale photovoltaic power stations. Summary of the invention
[0003] The present application provides a simulation method for a photovoltaic power station, which is used to improve the simulation speed while ensuring the simulation accuracy.
[0004] In a first aspect, the present application provides a simulation method for a photovoltaic power station, including: obtaining the line impedance, i.e., RL value, from all inverters in the photovoltaic power station to the grid-connected collection point. Collecting the first output data of all inverters in the first time period and the second output data of the second time period. Specifically, the first time period and the second time period are continuous time periods, and the first time period and the second time period can constitute a continuous time window for collecting data. Adjusting the output data of at least one inverter among all inverters at a time point between the first time period and the second time period, that is, applying a disturbance parameter to the photovoltaic power station at a time point between the first time period and the second time period, so that the second output data changes relative to the first output data. It can be considered that the first output data collected before the disturbance is applied is the output data of all inverters in a steady-state process, and the second output data collected after the disturbance is applied is the output data of all inverters in a transient process. According to the first output data, the second output data and the line impedance of all inverters, all inverters are clustered to obtain multiple clusters. Equivalent models are constructed for multiple clusters respectively. Adjusting the internal parameters of the algorithm for clustering processing so that the error of the equivalent model obtained by the adjusted clustering processing is less than the set value to achieve the target accuracy and efficiency.
[0005] The simulation method provided in the embodiment of the present application performs clustering processing on all inverters to construct an equivalent model, which can improve the simulation speed compared to the detailed model and improve the simulation accuracy compared to the single-machine model. When clustering all inverters, all data in the steady state and transient process are clustered, so that the overall accuracy of the equivalent model finally simulated for the steady state and transient state fitting will be improved.
[0006] In some embodiments of the present application, the first output data of all inverters in the first time period and the second output data of the second time period are collected, and specifically one or more of the output active power P waveform, the output reactive power Q waveform, the output voltage V waveform, and the output current I waveform of all inverters in the first time period and the second time period are collected.
[0007] In some embodiments of the present application, the output data of at least one of all the inverters is adjusted, that is, the disturbance parameters are applied to the photovoltaic power station, and one or more of the output active power, output reactive power, output voltage, and output current of one or more of all the inverters can be specifically adjusted, and the output parameters of some inverters can be simulated, for example, when some photovoltaic components are blocked during the actual use of the photovoltaic power station, resulting in changes in the output parameters of some inverters. When simulating the actual use of the photovoltaic power station, for example, when a short circuit to the ground occurs in a certain inverter, a ground branch can be added at the output end of the inverter, which will eventually cause the output voltage parameters of the inverter to change. Alternatively, in the above step S2, the output data of at least one of all the inverters is adjusted, that is, the disturbance parameters are applied to the photovoltaic power station, and one or more of the output active power, output reactive power, output voltage, and output current of all the inverters can be specifically adjusted. When the grid connected to the photovoltaic power station during the actual use of the simulation has a low voltage ride-through fault, a ground branch can be added at the grid connection point, or the grid voltage can be directly adjusted, which will eventually cause the output voltage parameters of all the inverters to change.
[0008] In some embodiments of the present application, the duration of the first time period for collecting the first output data and the second time period for collecting the second output data can be adjusted as needed, and the proportional relationship between the two can also be adjusted. For example, the duration of the first time period for the first output data and the second time period for the second output data can be set to be the same, that is, the amount of data of the acquired steady-state data and transient data is relatively consistent, so that in the subsequent clustering process, the weights of the steady-state data and the transient data are relatively consistent to ensure that the final simulated equivalent model has a relatively consistent overall accuracy for the fitting of the steady state and the transient state. When the transient process is more complicated, the duration of the second time period can also be increased to increase the weight of the transient data. Alternatively, the duration of the first time period can also be increased to increase the weight of the steady-state data.
[0009] In some embodiments of the present application, all inverters are clustered, and specifically, a k-means clustering algorithm, an affinity propagation clustering algorithm, a spectral clustering algorithm, a hierarchical clustering algorithm, or other algorithms may be used to cluster all inverters.
[0010] In some embodiments of the present application, equivalent models are constructed for multiple clusters respectively, which can be implemented in the following way: first, the central line impedance value (RL value) of each cluster in the multiple clusters can be determined, and then, according to the central line impedance value, equivalent models are constructed for the multiple clusters respectively.
[0011] In some embodiments of the present application, equivalent models are constructed for each of the multiple clusters. Specifically, an algorithm such as a capacity weighted method may be used to construct equivalent models for each of the multiple clusters.
[0012] In some embodiments of the present application, the internal parameters of the algorithm for clustering processing are adjusted, which can be specifically implemented in the following manner: first, simulation is performed using an equivalent model and a detailed model constructed for each inverter, respectively, to obtain simulation results; then, the simulation results output by the detailed model are used as a reference, and when the difference between the simulation results of the equivalent model and the simulation results of the detailed model (the difference can be considered as an error) is greater than a set value, such as 5% or 10%, the internal parameters of the algorithm for clustering processing are adjusted to achieve the target accuracy and efficiency.
[0013] In some embodiments of the present application, the simulation results obtained by using the equivalent model and the detailed model constructed for each inverter can be implemented as follows: in the process of using the equivalent model and the detailed model for simulation, the output data is adjusted to obtain the simulation results of the continuous time period including the adjustment time point, that is, the simulation results will include all the output data in the steady state and transient processes.
[0014] In some embodiments of the present application, the simulation results may specifically include: output active power waveform, output reactive power waveform, output voltage waveform and output frequency waveform. The above data can be compared respectively to obtain corresponding error values so as to adjust different clustering parameters according to the error values. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 This is a schematic diagram of the structure of a photovoltaic power station;
[0016] Figure 2 A schematic diagram of the architecture of an equivalent model of a photovoltaic power station obtained by the simulation method provided in an embodiment of the present application;
[0017] Figure 3 A schematic diagram of a flow chart of a simulation method for a photovoltaic power station provided in an embodiment of the present application;
[0018] Figure 4 This is a schematic diagram of the architecture of a photovoltaic power station to be simulated in Example 1;
[0019] Figure 5a It is a simulation comparison diagram of the active power output by each model in Example 1;
[0020] Figure 5b It is a simulation comparison diagram of reactive power output by each model in Example 1;
[0021] Figure 5c It is a simulation comparison diagram of the voltage output by each model in Example 1;
[0022] Figure 5d It is a simulation comparison diagram of the frequency output by each model in Example 1;
[0023] Figure 6a It is a simulation comparison diagram of the active power output by each model in Example 2;
[0024] Figure 6b It is a simulation comparison diagram of reactive power output by each model in Example 2;
[0025] Figure 6c It is a simulation comparison diagram of the voltage output by each model in Example 2;
[0026] Figure 6d It is a simulation comparison diagram of the frequency output by each model in Example 2;
[0027] Figure 7a It is a simulation comparison diagram of the active power output by each model in Example 3;
[0028] Figure 7b It is a simulation comparison diagram of reactive power output by each model in Example 3;
[0029] Figure 7c It is a simulation comparison diagram of the voltage output by each model in Example 3;
[0030] Figure 7d This is a simulation comparison diagram of the frequencies output by each model in Example 3. DETAILED DESCRIPTION
[0031] In order to enable ordinary persons in the art to better understand the technical solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.
[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.
[0033] At present, the commonly used simulation methods in photovoltaic power stations are single-machine equivalence method and multi-machine equivalence method. The single-machine equivalence method is to directly average the output parameters of all power generation units in the photovoltaic power station, and use one power generation unit to be equivalent to a photovoltaic sub-array or even a photovoltaic power station in a one-to-one equivalence method. The multi-machine equivalence method is to cluster all power generation units in the photovoltaic power station, and each cluster uses a central value to be equivalent to all inverters in the cluster. In this way, a photovoltaic power station uses several power generation units for equivalence, that is, an N-to-1 equivalence method is adopted, where N is an integer greater than 1.
[0034] If the parameters of all power generation units in a photovoltaic power station are relatively dispersed, the use of the single-machine equivalent method for simulation will result in a large simulation error. When the multi-machine equivalent method is used for simulation, the operating data of the power generation unit at a certain moment is usually used as an indicator for clustering. The equivalent model constructed using this type of data is not suitable for transient processes. If the operating data of the transient process is used as an indicator for clustering, the equivalent model constructed using this type of data is not suitable for steady-state processes.
[0035] In view of this, an embodiment of the present application provides a simulation method for a photovoltaic power station, which is suitable for establishing equivalent modeling of a photovoltaic power station using large-scale power generation units. The simulation method can significantly reduce the simulation scale of the photovoltaic power station and improve the simulation calculation speed, while ensuring the modeling accuracy of steady-state and transient processes, that is, it adopts a high-efficiency and high-precision equivalent modeling method to achieve efficient and accurate simulation of the photovoltaic power station, and provide an accurate reference basis for the stability analysis of the photovoltaic power station and fault experiments using the equivalent model obtained by simulation.
[0036] The simulation method provided in this application can be applied to the modeling process of photovoltaic power stations with a large number of power generation units. Figure 1 In a photovoltaic power station, a large number of photovoltaic arrays PV1, PV2...PV16 are connected to the power grid PCC through connected inverters DC / AC and transformers; refer to Figure 2 ,The simulation method of the present application can simplify the detailed model established for each power generation unit into a smaller number of equivalent models, such as Figure 2The three equivalent models shown in .
[0037] Reference Figure 3 The photovoltaic power station simulation method provided in the embodiment of the present application specifically includes the following steps:
[0038] S1. Obtain the line impedance, i.e., RL value, from all inverters in the photovoltaic power station to the grid-connected collection point. Specifically, the line impedance of each inverter in the photovoltaic power station changes according to the length of the cable connecting the inverter to the collection point. Therefore, the line impedance of each inverter will have certain differences.
[0039] S2. Collect the first output data of all inverters in the first time period and the second output data of the second time period. Specifically, the first time period and the second time period are continuous time periods, and the first time period and the second time period can constitute a continuous time window for collecting data. Adjust the output data of at least one inverter among all inverters at the time point between the first time period and the second time period, that is, apply a disturbance parameter to the photovoltaic power station at the time point between the first time period and the second time period, so that the second output data changes relative to the first output data. It can be considered that the first output data collected before the disturbance is applied is the output data of all inverters in the steady state process, and the second output data collected after the disturbance is applied is the output data of all inverters in the transient process.
[0040] S3. Cluster all inverters according to the first output data, second output data and line impedance of all inverters to obtain multiple clusters. Clustering all data in the steady state and transient process increases the amount of information compared to the single moment data, but the clustering is more accurate, so that the overall accuracy of the equivalent model finally simulated for the steady state and transient will be improved.
[0041] S4. Equivalent models are constructed for multiple clusters. Since multiple clusters are obtained based on all data in the steady-state and transient processes, the overall accuracy of the equivalent model finally simulated for the steady-state and transient processes will be improved.
[0042] S5. Adjust the internal parameters of the clustering algorithm so that the error of the equivalent model obtained by the adjusted clustering process is less than the set value to achieve the target accuracy and efficiency.
[0043] The above simulation method provided in the embodiment of the present application performs clustering processing on all inverters to construct an equivalent model, which can improve the simulation speed compared to the detailed model and improve the simulation accuracy compared to the single-machine model. When clustering all inverters, all data in the steady state and transient process are clustered, so that the overall accuracy of the equivalent model finally simulated for the steady state and transient fitting will be improved.
[0044] In some embodiments of the present application, in the above step S2, the first output data of all inverters in the first time period and the second output data of the second time period are collected. Specifically, one or more of the output active power P waveform, the output reactive power Q waveform, the output voltage V waveform, and the output current I waveform of all inverters in the first time period and the second time period can be collected. The following is explained by taking the collection of the output power waveforms PQ of all inverters in steady state and transient state as an example.
[0045] In some embodiments of the present application, in the above step S2, the output data of at least one of all the inverters is adjusted, that is, the disturbance parameters are applied to the photovoltaic power station, and one or more of the output active power, output reactive power, output voltage, and output current of one or more of all the inverters can be specifically adjusted, and the output parameters of some inverters can be changed due to, for example, some photovoltaic components are blocked during the actual use of the photovoltaic power station. When simulating the actual use of the photovoltaic power station, for example, a short circuit to the ground occurs in a certain inverter, a ground branch can be added at the output end of the inverter, which will eventually cause the output voltage parameters of the inverter to change. Alternatively, in the above step S2, the output data of at least one of all the inverters is adjusted, that is, the disturbance parameters are applied to the photovoltaic power station, and one or more of the output active power, output reactive power, output voltage, and output current of all the inverters can be specifically adjusted. When the grid connected to the photovoltaic power station during the actual use of the simulation has a low voltage ride-through fault, a ground branch can be added at the grid connection point, or the grid voltage can be directly adjusted, which will eventually cause the output voltage parameters of all the inverters to change.
[0046] In some embodiments of the present application, the duration of the first time period of the first output data and the second time period of the second output data collected in the above step S2 can be adjusted as needed, and the proportional relationship between the two can also be adjusted. For example, the duration of the first time period of the first output data and the second time period of the second output data can be set to be the same, that is, the amount of data of the acquired steady-state data and transient data is relatively consistent, so that in the subsequent clustering process, the weights of the steady-state data and the transient data are relatively consistent, so as to ensure that the equivalent model finally simulated has a relatively consistent overall accuracy for the steady-state and transient fitting. When the transient process is more complicated, the duration of the second time period can also be increased to increase the weight of the transient data. Alternatively, the duration of the first time period can also be increased to increase the weight of the steady-state data.
[0047] In some embodiments of the present application, in the above step S3, all inverters are clustered, and specifically, all inverters are clustered using algorithms such as k-means clustering (kmeans) algorithm, affinity propagation clustering algorithm, spectral clustering algorithm or hierarchical clustering algorithm.
[0048] In some embodiments of the present application, in the above step S4, equivalent models are constructed for multiple clusters respectively, which can be implemented in the following way: first, the central line impedance value (RL value) of each cluster in the multiple clusters can be determined, and then, according to the central line impedance value, equivalent models are constructed for the multiple clusters respectively.
[0049] In some embodiments of the present application, in the above step S4, equivalent models are respectively constructed for multiple clusters. Specifically, an algorithm such as a capacity weighted method can be used to construct equivalent models for multiple clusters.
[0050] In some embodiments of the present application, adjusting the internal parameters of the algorithm for clustering processing in the above step S5 can be specifically implemented in the following manner: first, using the equivalent model and the detailed model constructed for each inverter to perform simulation respectively to obtain simulation results; then, using the simulation results output by the detailed model as a reference, when the difference between the simulation results of the equivalent model and the simulation results of the detailed model (the difference can be considered as an error) is greater than a set value, such as 5% or 10%, adjust the internal parameters of the algorithm for clustering processing to achieve the target accuracy and efficiency.
[0051] In some embodiments of the present application, the simulation results obtained by using the equivalent model and the detailed model constructed for each inverter can be implemented as follows: in the process of using the equivalent model and the detailed model for simulation, the output data is adjusted to obtain the simulation results of the continuous time period including the adjustment time point, that is, the simulation results will include all the output data in the steady state and transient processes.
[0052] In some embodiments of the present application, the simulation results may specifically include: output active power waveform, output reactive power waveform, output voltage waveform and output frequency waveform. The above data can be compared respectively to obtain corresponding error values so as to adjust different clustering parameters according to the error values.
[0053] The following uses several specific embodiments to illustrate that the simulation method provided by the present application can improve simulation accuracy and simulation speed.
[0054] Embodiment 1:
[0055] Reference Figure 4 , Taking a photovoltaic power station containing 30 photovoltaic arrays and 30 inverters as an example, at t = 2.1s, the grid voltage is reduced to 0.8pu, so that each inverter in the photovoltaic power station enters the low voltage ride-through state.
[0056] The above simulation method provided in the embodiment of the present application is used to establish an equivalent model, which specifically includes the following steps:
[0057] 1. Record the active power P waveform and reactive power Q waveform of all inverter output power in the steady state before the grid voltage is changed, and in the transient process of transitioning to the low voltage ride-through state after the grid voltage is changed.
[0058] 2. Obtain the line impedance RL value between 30 inverters and the transformer, i.e., the grid-connected collection point.
[0059] 3. According to the P waveform, Q waveform and RL value of the 30 inverters, the 30 inverters are clustered to obtain several clusters.
[0060] 4. For each cluster, the capacity weighted method is used to construct an equivalent model.
[0061] The simulation results of the equivalent model output established by the above simulation method provided in the embodiment of the present application are compared with the simulation results of the single-machine model output established by the single-machine equivalent method, taking the simulation results of the detailed model output as a reference. Figure 5a It can be seen that the active power P value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application has a smaller error with the active power P value obtained by the detailed model simulation, while the active power P value obtained by the single machine model simulation has a larger error with the active power P value obtained by the detailed model simulation. Figure 5b It can be seen that the reactive power Q value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application has a smaller error than the reactive power Q value obtained by the detailed model simulation, while the reactive power Q value obtained by the single machine model simulation has a larger error than the reactive power Q value obtained by the detailed model simulation. Figure 5c It can be seen that the output voltage V value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application, the output voltage V value obtained by the single machine model simulation and the output voltage V value obtained by the detailed model simulation have small errors. Figure 5d It can be seen that the output frequency f value obtained by the equivalent model simulation established by the above simulation method provided in the embodiment of the present application has a smaller error than the output frequency f value obtained by the detailed model simulation, while the output frequency f value obtained by the single-machine model simulation has a larger error than the output frequency f value obtained by the detailed model simulation.
[0062] Embodiment 2:
[0063] The difference between the second embodiment and the first embodiment is that the second embodiment lowers the grid strength compared with the first embodiment. The second embodiment lowers the grid strength to SCR<2, and adopts the same steps as the first embodiment to establish an equivalent model.
[0064] The simulation results of the equivalent model output established by the above simulation method provided in the embodiment of the present application are compared with the simulation results of the single-machine model output established by the single-machine equivalent method, taking the simulation results of the detailed model output as a reference. Figure 6a It can be seen that the active power P value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application has a smaller error with the active power P value obtained by the detailed model simulation, while the active power P value obtained by the single machine model simulation has a larger error with the active power P value obtained by the detailed model simulation. Figure 6b It can be seen that the reactive power Q value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application has a smaller error than the reactive power Q value obtained by the detailed model simulation, while the reactive power Q value obtained by the single machine model simulation has a larger error than the reactive power Q value obtained by the detailed model simulation. Figure 6c It can be seen that the output voltage V value obtained by the equivalent model simulation established by the above simulation method provided by the embodiment of the present application has a smaller error than the output voltage V value obtained by the detailed model simulation, while the output voltage V value obtained by the stand-alone model simulation has a larger error than the output voltage V value obtained by the detailed model simulation. Figure 6d , it can be seen that the output frequency f value obtained by the simulation of the equivalent model established by the simulation method provided in the embodiment of the present application has a smaller error than the output frequency f value obtained by the simulation of the detailed model, and the output frequency f value obtained by the simulation of the stand-alone model has a larger error than the output frequency f value obtained by the simulation of the detailed model. By analyzing the simulation data, it can be obtained that the simulation result error of the equivalent model established by the simulation method provided in the embodiment of the present application relative to the detailed model is 1.0%, while the simulation result error of the stand-alone model relative to the detailed model is 12.0%.
[0065] Embodiment three:
[0066] In Example 3, the same clustering algorithm is used, and different collected data are used for clustering, and the clustering results are significantly different. The simulation results of the equivalent model output established by the above simulation method provided in the embodiment of the present application are compared with the simulation results of the model output constructed by transient data and the simulation results of the model output constructed by steady-state data, with reference to the simulation results of the detailed model output. Figure 7a It can be seen that the active power P value obtained by the simulation of the equivalent model established by the above simulation method provided in the embodiment of the present application has a smaller error with the active power P value obtained by the simulation of the detailed model, and the active power P value obtained by the simulation of the model constructed by the steady-state data has a larger error with the active power P value obtained by the simulation of the detailed model. Figure 7bIt can be seen that the reactive power Q value obtained by the simulation of the equivalent model established by the above simulation method provided in the embodiment of the present application has a small error with the reactive power Q value obtained by the simulation of the detailed model, the reactive power Q value obtained by the simulation of the model constructed by transient data has a large error with the reactive power Q value obtained by the simulation of the detailed model, and the reactive power Q value obtained by the model constructed by steady-state data has a large error with the reactive power Q value obtained by the simulation of the detailed model. Figure 7c It can be seen that the output voltage V value obtained by the simulation of the equivalent model established by the above simulation method provided in the embodiment of the present application, the output voltage V value obtained by the simulation of the model constructed by steady-state data, the output voltage V value obtained by the simulation of the model constructed by transient data, and the output voltage V value obtained by the simulation of the detailed model are all small in error. Figure 7d It can be seen that the output frequency f value obtained by the simulation of the equivalent model established by the above simulation method provided by the embodiment of the present application has a smaller error than the output frequency f value obtained by the simulation of the detailed model, and the output frequency f value obtained by the simulation of the model constructed by the steady-state data has a larger error than the output frequency f value obtained by the simulation of the detailed model. In addition, it can be seen from the table below that the simulation time of the model constructed by transient data is relatively long.
[0067] Data used Number of clusters Active power error Simulation time All data points in the time window 5 0.8％ 27.0 Using steady-state moment data 5 5.47％ 26.6 Using transient moment data 11 1.39％ 84.4
[0068] The above-mentioned simulation method provided by the embodiment of the present application collects all data of all inverters in steady state and transient process in the data collection stage, and then clusters all inverters according to the data and line impedance to obtain multiple clusters, respectively constructs equivalent models for multiple clusters, and adjusts the internal parameters of the algorithm for clustering processing so that the error of the equivalent model obtained by the adjusted clustering processing is less than the set value. Clustering all inverters to construct equivalent models can improve the simulation speed compared to the detailed model, and can improve the simulation accuracy compared to the stand-alone model. When clustering all inverters, clustering is performed on all data in steady state and transient process, so that the overall accuracy of the equivalent model finally simulated for steady state and transient fitting will be improved.
[0069] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A simulation method for a photovoltaic power station, characterized in that: include: Obtaining line impedances from all inverters in the photovoltaic power station to a grid-connected collection point; Collecting first output data of all the inverters in a first time period and second output data of all the inverters in a second time period, wherein the first time period and the second time period are continuous time periods, and adjusting output data of at least one inverter among all the inverters at a time point between the first time period and the second time period; Clustering all the inverters according to the first output data, the second output data and the line impedance of all the inverters to obtain a plurality of clusters; Constructing equivalent models for each of the multiple clusters; The internal parameters of the clustering algorithm are adjusted so that the error of the equivalent model obtained by the adjusted clustering process is less than a set value.
2. The simulation method according to claim 1, characterized in that: The collecting first output data of all inverters in the first time period and second output data of all inverters in the second time period includes: One or more of the output active power waveforms, output reactive power waveforms, output voltage waveforms, and output current waveforms of all the inverters in the first time period and the second time period are collected.
3. The simulation method according to claim 1 or 2, characterized in that: The adjusting output data of at least one inverter among all the inverters comprises: Adjust one or more of the output active power, output reactive power, output voltage, and output current of one or more inverters among all the inverters; or, Adjust one or more of the output active power, output reactive power, output voltage, and output current of all the inverters.
4. The simulation method according to any one of claims 1 to 3, characterized in that: Also includes: The ratio of the duration of the first time period to the duration of the second time period is adjusted.
5. The simulation method according to any one of claims 1 to 4, characterized in that: The clustering of all the inverters comprises: All the inverters are clustered using a k-means clustering algorithm, an affinity propagation clustering algorithm, a spectral clustering algorithm or a hierarchical clustering algorithm.
6. The simulation method according to any one of claims 1 to 5, characterized in that: The constructing equivalent models respectively for the multiple clusters includes: determining a central line impedance value for each cluster in the plurality of clusters; According to the central line impedance value, equivalent models are respectively constructed for the multiple clusters.
7. The simulation method according to any one of claims 1 to 6, characterized in that: The constructing equivalent models respectively for the multiple clusters includes: Using a capacity weighted method, equivalent models are constructed for each of the multiple clusters.
8. The simulation method according to any one of claims 1 to 7, characterized in that: The adjusting of the internal parameters of the clustering algorithm includes: Using the equivalent model and the detailed model constructed for each inverter to perform simulation respectively, to obtain simulation results; When the difference of the simulation results is greater than the set value, the internal parameters of the clustering algorithm are adjusted.
9. The simulation method according to claim 8, characterized in that: The simulation is performed using the equivalent model and the detailed model constructed for each inverter to obtain simulation results, including: In the process of performing simulation using the equivalent model and the detailed model respectively, the output data is adjusted to obtain simulation results of a continuous time period including the adjusted time point.
10. The simulation method according to claim 8 or 9, characterized in that: The simulation results include: output active power waveform, output reactive power waveform, output voltage waveform and output frequency waveform.