Search

CN-121984100-A - Hierarchical self-adaptive power decoupling control method for network-structured converter and related equipment

CN121984100ACN 121984100 ACN121984100 ACN 121984100ACN-121984100-A

Abstract

The embodiment of the application provides a hierarchical self-adaptive power decoupling control method for a grid-connected converter and related equipment, belonging to the technical field of power electronics and smart grids. According to the method, firstly, a power model of a grid-structured converter is structurally layered through a rotation matrix, and is decomposed into an independent impedance angle coupling layer and an independent power angle coupling layer. And aiming at the impedance angle coupling layer, dynamically updating an impedance angle estimated value on line by adopting an event triggering mechanism, and realizing self-adaptive decoupling of impedance angle coupling. And aiming at the power angle coupling layer, a dynamic updating equation of the power angle is established, a reinforcement learning algorithm based on an Actor-Critic framework is introduced, and the self-adaptive compensation is carried out on the power angle and the voltage amplitude through an intelligent agent, so that the dynamic accurate decoupling of the power angle coupling is realized. The method effectively solves the problems of incomplete decoupling and poor robustness caused by inaccurate line parameters, model mismatch and working point change in the prior art, and remarkably improves the power control precision and dynamic stability of the grid-structured converter under complex working conditions.

Inventors

  • HU HUIJUN
  • PENG YUNJIAN

Assignees

  • 华南理工大学

Dates

Publication Date
20260505
Application Date
20251202

Claims (10)

  1. 1. The hierarchical self-adaptive power decoupling control method for the network-structured converter is characterized by comprising the following steps of: carrying out structural layering on a power model of the grid-structured converter through a rotation matrix to obtain an impedance angle coupling layer and a power angle coupling layer which are independent of each other, wherein the impedance angle coupling layer is only related to a line impedance angle, and the power angle coupling layer is only related to a power angle between the output voltage of the converter and the voltage of a power grid; an impedance angle decoupling module is constructed, and an impedance angle estimated value is dynamically updated through an event triggering mechanism so as to realize self-adaptive dynamic decoupling of the impedance angle coupling layer; And constructing a power angle decoupling module, and performing self-adaptive compensation on the power angle coupling layer by establishing a dynamic update equation of the power angle and introducing a reinforcement learning algorithm based on an Actor-Critic framework so as to realize dynamic decoupling control.
  2. 2. The method of claim 1, wherein the structuring layering of the power model of the grid-formed converter by a rotation matrix comprises: establishing an original power equation simultaneously containing an impedance angle coupling term and a power angle coupling term based on an output equivalent circuit model of the grid-structured converter; The original power equation is transformed into a form in which the rotation transformation matrix is multiplied by a gain matrix by introducing a rotation transformation matrix, wherein elements of the rotation transformation matrix are related to the line impedance angle only for characterizing the impedance angle coupling layer, and elements of the gain matrix are related to the power angle only for characterizing the power angle coupling layer.
  3. 3. The method of claim 2, wherein after the structured layering, the method of controlling the adaptive power decoupling of the grid-formed converter layering further comprises: Carrying out small signal linearization treatment on a network construction control method adopted by the network construction type converter to obtain a small signal model of the controller; And combining the small signal model of the controller with the layered power coupling model to construct a closed-loop small signal model of the grid-structured converter system containing power coupling.
  4. 4. The method of claim 1, wherein the constructing an impedance angle decoupling module comprises: Setting an initial value of the impedance angle estimation value; Designing an event triggering condition, and triggering the updating of the primary impedance angle estimated value when the error between the actual impedance angle and the impedance angle estimated value exceeds a preset threshold value; based on the measured values of the output voltage of the converter and the power grid voltage, a new impedance angle estimated value is calculated, and the current impedance angle estimated value is corrected according to an updating algorithm, so that the on-line self-adaptive identification of the impedance angle parameters is realized.
  5. 5. The method of claim 4, wherein the impedance angle decoupling module projectively transforms the active power and the reactive power through an orthogonal rotation matrix to eliminate impedance angle coupling; and the decoupling module is arranged behind the power control loop, and a decoupling coefficient related to the impedance angle estimated value is introduced, so that the steady-state error of the power control is reduced.
  6. 6. The method of claim 1, wherein the constructing a power angle decoupling module comprises: Establishing a dynamic updating equation of the power angle, and predicting the power angle at the current moment according to the voltage set value at the previous sampling moment; The reinforcement learning algorithm based on the Actor-Critic framework is designed, and an agent of the algorithm outputs compensation actions on the power angle and the output voltage amplitude by observing the system state, and guides the agent to learn by a reward function so as to minimize the power coupling error.
  7. 7. The method of claim 6, wherein in the reinforcement learning algorithm: State space Comprises active power Reactive power Angle of power Amplitude of output voltage of converter ; Action space Includes a power angle compensation amount And output voltage amplitude compensation amount ; Reward function Is configured as a weighted sum of a power factor angular deviation reward, a power amplitude deviation reward, and an output smoothing reward to promote dynamic stability of the system while ensuring power tracking accuracy.
  8. 8. An electronic device comprising a memory storing a computer program and a processor implementing the method of any of claims 1 to 7 when the computer program is executed by the processor.
  9. 9. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method of any one of claims 1 to 7.
  10. 10. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of any one of claims 1 to 7.

Description

Hierarchical self-adaptive power decoupling control method for network-structured converter and related equipment Technical Field The application relates to the technical field of power electronic control and intelligent power grids, in particular to a hierarchical self-adaptive power decoupling control method for a grid-connected converter and related equipment. Background With the large-scale access of renewable energy sources (such as photovoltaic and wind power), the grid-structured converter plays an increasingly important role in a modern power system, and the control performance of the grid-structured converter is directly related to the stability and the electric energy quality of a power grid. However, in a practical medium-low voltage distribution network, the line impedance exhibits a mixed resistive-inductive characteristic, rather than pure inductive. This is in contradiction with the assumption that the existing grid-formed inverter control method (droop control, virtual synchronous machine control) is generally established that the transmission line is purely inductive and that the power angle (the phase angle difference between the inverter output voltage and the grid voltage) is small. This contradiction brings about a significant coupling effect between the active power and the reactive power output by the converter. The power coupling not only increases the control difficulty of the grid-formed inverter, but also has double negative effects on the dynamic performance and stability of the grid-formed inverter, and the dynamic response speed of the system is reduced, the oscillation phenomenon is aggravated, and safety accidents such as system instability and the like can be caused when the dynamic response speed of the system is serious. In order to solve the power coupling problem, some decoupling control methods are proposed in the prior art. For example, the virtual impedance approach compensates for the effects of line impedance by introducing a virtual impedance in the control loop. However, such methods typically only consider the problem with line impedance angles, ignoring the larger power angles that may occur when the system is running, resulting in incomplete decoupling and possibly significant reactive power shortage. Another class of small signal model-based decoupling methods, while effective near certain operating points, relies heavily on the accuracy of the model for their decoupling performance. When the system running state changes, deviates from a preset working point, or the model parameter itself has uncertainty or fluctuation, the decoupling performance of the small signal model based on the fixed parameter can be rapidly deteriorated. In summary, the existing decoupling technology has obvious defects in coping with complex power coupling problems caused by non-purely inductive lines and high power angles, namely 1) the decoupling control strategy is incomplete, impedance angle coupling and power angle coupling cannot be properly processed at the same time, 2) the decoupling performance is excessively sensitive to system working points and model parameters, and has poor robustness, and 3) the control precision and dynamic stability are difficult to ensure when the parameters are uncertain or the working conditions are changed. Disclosure of Invention The embodiment of the application mainly aims to provide a hierarchical self-adaptive power decoupling control method, electronic equipment, a storage medium and a program product for a networking type converter based on an Actor-Critic, which solve the problems that the prior art cannot completely decouple, the parameters of a controller depend on an accurate model and the robustness is poor by processing the problems of impedance angle coupling and power angle coupling in a layering way and adopting an event triggering self-adaptive mechanism and reinforcement learning intelligent compensation respectively. In order to achieve the above objective, an aspect of an embodiment of the present application provides a hierarchical adaptive power decoupling control method for a grid-connected converter, where the method includes: carrying out structural layering on a power model of the grid-structured converter through a rotation matrix to obtain an impedance angle coupling layer and a power angle coupling layer which are independent of each other, wherein the impedance angle coupling layer is only related to a line impedance angle, and the power angle coupling layer is only related to a power angle between the output voltage of the converter and the voltage of a power grid; an impedance angle decoupling module is constructed, and an impedance angle estimated value is dynamically updated through an event triggering mechanism so as to realize self-adaptive dynamic decoupling of the impedance angle coupling layer; And constructing a power angle decoupling module, and performing self-adaptive compensation on the power angle coupl