Search

CN-122026359-A - Adaptive photovoltaic grid-connected flexible control method based on multi-source data

CN122026359ACN 122026359 ACN122026359 ACN 122026359ACN-122026359-A

Abstract

The invention relates to the technical field of photovoltaic grid-connected control, in particular to a self-adaptive photovoltaic grid-connected flexible control method based on multi-source data, which comprises the steps of collecting meteorological, power grid and photovoltaic equipment data to form an original data set; obtaining a standardized data set through abnormal elimination and normalization, extracting space/time sequence characteristics, performing attention weighted fusion to generate a fusion characteristic vector, calculating voltage deviation according to the space/time sequence characteristics, generating a first control instruction by using an edge PID, integrating the first control instruction into a layered instruction set, monitoring power deviation and performing online optimization. According to the invention, through comprehensive acquisition and accurate preprocessing of multi-source data, attention mechanism feature fusion, edge node PID hierarchical control and on-line optimization of power deviation driving, the self-adaptive adjustment of a photovoltaic grid-connected system is realized, the running stability, control precision and response speed of a power grid are remarkably improved, and the problems of high communication load and slow response caused by the poor adaptability and low control precision of the traditional photovoltaic grid-connected control method in the prior art are solved.

Inventors

  • WU JINXING

Assignees

  • 北京中科格锐科技有限公司

Dates

Publication Date
20260512
Application Date
20260203

Claims (10)

  1. 1. The adaptive photovoltaic grid-connected flexible control method based on the multi-source data is characterized by comprising the following steps of: step S1, collecting meteorological data, power grid operation state data and photovoltaic power generation equipment operation data, and constructing an original multi-source data set; S2, carrying out abnormal data elimination and normalization processing on the original multi-source data set to obtain a standardized data set; Step S3, extracting spatial features and time sequence features in the standardized data set, splicing the spatial features and the time sequence features, and carrying out weighted fusion on the spliced features based on an attention mechanism to generate fusion feature vectors; Step S4, obtaining a voltage deviation value through a preset calculation mode based on the fusion characteristic vector, determining different control instruction generation modes based on the voltage deviation value to generate a control instruction, wherein, The control instruction generation mode comprises the steps of generating a control instruction by adopting a PID controller deployed at an edge computing node and generating the control instruction by adopting a model predictive control MPC optimizer deployed at a cloud; s5, integrating the control instructions to form a layered control instruction set, monitoring the power deviation of the system in real time, and determining to optimize the layered control instruction set on line based on the power deviation; step S6, determining whether the execution effect meets the standard or not based on the performance index after the execution of the optimized hierarchical control instruction set, and returning to step S3 when the execution effect does not meet the standard, wherein, The performance indexes comprise voltage stabilization time, power regulation precision and system response speed.
  2. 2. The adaptive photovoltaic grid-connected flexible control method according to claim 1, wherein in the step S1, the meteorological data includes solar radiation intensity data, ambient temperature data, wind speed data, and wind direction data; the power grid running state data comprise power grid voltage data, power grid current data, power grid frequency data and power grid power data; The photovoltaic power generation equipment operation data comprise photovoltaic module output voltage data, photovoltaic module output current data, photovoltaic inverter operation temperature data and photovoltaic inverter working state data.
  3. 3. The adaptive photovoltaic grid-connected flexible control method according to claim 2, wherein in step S2, the process of obtaining the standardized dataset comprises: Step S21, calculating the mean value and standard deviation of the data; s22, judging the data exceeding the standard deviation range of the preset multiple as abnormal data and eliminating the abnormal data; step S23, mapping the data to a preset numerical range to complete the normalization processing, and obtaining the standardized data set.
  4. 4. The adaptive photovoltaic grid-connected flexible control method according to claim 3, wherein in the step S3, the process of generating the fusion feature vector includes: s31, extracting spatial features in the standardized data set by adopting a convolutional neural network; Step S32, extracting time sequence characteristics in the standardized data set by adopting a cyclic neural network; and step S33, respectively carrying out weighted summation on each spatial feature and time sequence feature based on the importance weight of each feature in the overall feature so as to obtain a fusion feature vector.
  5. 5. The adaptive photovoltaic grid-connected flexible control method according to claim 4, wherein in the step S4, it is determined that the PID controller disposed at the edge computing node is used to generate the control command based on the result that the voltage deviation value is less than or equal to the standard voltage deviation value.
  6. 6. The adaptive photovoltaic grid-connected flexible control method according to claim 5, wherein in the step S4, it is determined that the MPC optimizer is predicted to generate the control command by using the model deployed in the cloud based on the result that the voltage deviation value is greater than the standard voltage deviation value.
  7. 7. The adaptive photovoltaic grid-connected flexible control method according to claim 6, wherein in step S5, the process of monitoring the system power deviation in real time includes: step S51, acquiring the actual output power and the preset reference power of a photovoltaic grid-connected system in real time; Step S52, calculating a difference between the actual output power and the preset reference power to obtain the system power deviation.
  8. 8. The adaptive photovoltaic grid-connected flexible control method according to claim 7, wherein the start-up optimization program is determined to adjust each control instruction parameter in the hierarchical control instruction set based on the result that the system power deviation is greater than the standard system power deviation.
  9. 9. The adaptive photovoltaic grid-connected flexible control method according to claim 8, wherein the optimization program adopts a gradient descent algorithm or a particle swarm optimization algorithm, takes the reduced system power deviation as an objective function, and performs iterative optimization on control instruction parameters in the hierarchical control instruction set until the system power deviation is within a preset allowable range.
  10. 10. The adaptive photovoltaic grid-connected flexible control method according to claim 9, wherein in the step S6, whether the execution effect meets the standard is determined based on the comparison result of the performance index and the standard performance index; if the performance indexes are all larger than or equal to the standard performance indexes, determining to keep the current hierarchical control instruction set; Otherwise, the determination returns to step S3.

Description

Adaptive photovoltaic grid-connected flexible control method based on multi-source data Technical Field The invention relates to the technical field of photovoltaic grid-connected control, in particular to a self-adaptive photovoltaic grid-connected flexible control method based on multi-source data. Background Along with the transformation of global energy structures to clean energy, photovoltaic power generation is widely applied due to the characteristics of cleanliness and reproducibility, and the scale and installed capacity of a photovoltaic grid-connected system are continuously increased. However, photovoltaic power generation is significantly affected by meteorological conditions (such as solar radiation and ambient temperature), the output power has strong fluctuation and intermittence, and meanwhile, the dynamic change of the running state (such as voltage and frequency) of a power grid and the aging or failure of photovoltaic power generation equipment (such as photovoltaic modules and inverters) can cause the reduced matching of a photovoltaic grid-connected system and the power grid. CN116760097a discloses a photovoltaic grid-connected flexible control system, which comprises an electric quantity prediction module, a load calculation module, a flexible calculation module, a strategy configuration module and a strategy distribution module, wherein the flexible analysis can be performed on power supply requirements according to actual load values of photovoltaic grid-connected nodes and flexible response values of users, the utilization of photovoltaic electric energy is maximized, the rationality and the utilization rate of resource utilization are improved, on the other hand, as all control information is analyzed and generated in a data center way, although the calculation force can be concentrated, the communication load is higher, and certain transmission and response time are required, the invention is used for solving the problems that the traditional photovoltaic grid-connected control system is not comprehensive in analysis and not timely in control, but the control method tries to introduce multi-source data, and has the following defects: firstly, the data preprocessing is only simple to filter missing values, and the abnormal data (such as extremum caused by sensor faults) is not accurately removed, so that the subsequent feature extraction error is large, secondly, the feature fusion adopts a simple splicing mode, the importance of spatial features (such as output differences among multiple components) and time sequence features (such as time sequence changes of power) is not distinguished, the feature utilization rate is low, thirdly, the control instruction generation is not used for selecting a controller according to voltage deviation in a grading mode, and the problems of delayed cloud control response or insufficient edge control precision are outstanding. Therefore, a self-adaptive control method capable of integrating multi-source data, accurately processing the data, intelligently fusing the characteristics, generating control instructions in a grading manner and optimizing on line is needed, so that the problems of poor adaptability, low control precision and weak power grid compatibility of a traditional method are solved, and the flexible control requirement of a photovoltaic grid-connected system is met. Disclosure of Invention Therefore, the invention provides a self-adaptive photovoltaic grid-connected flexible control method based on multi-source data, which is used for solving the problems of high communication load and slow response caused by poor adaptability and low control precision of the traditional photovoltaic grid-connected control method in the prior art. In order to achieve the above purpose, the invention provides a multi-source data-based adaptive photovoltaic grid-connected flexible control method, which comprises the following steps: step S1, collecting meteorological data, power grid operation state data and photovoltaic power generation equipment operation data, and constructing an original multi-source data set; S2, carrying out abnormal data elimination and normalization processing on the original multi-source data set to obtain a standardized data set; Step S3, extracting spatial features and time sequence features in the standardized data set, splicing the spatial features and the time sequence features, and carrying out weighted fusion on the spliced features based on an attention mechanism to generate fusion feature vectors; Step S4, obtaining a voltage deviation value through a preset calculation mode based on the fusion characteristic vector, determining different control instruction generation modes based on the voltage deviation value to generate a control instruction, wherein, The control instruction generation mode comprises the steps of generating a control instruction by a PID controller deployed at an edge computing node and generatin