Novel power system self-organization and self-coordination control method based on data driving

Abstract

The invention provides a novel power system self-organization and self-coordination control method based on data driving based on the characteristics of flexibility, controllability and intelligent calculation of a converter, belongs to the field of power electronics and power systems, and is applied to power electronic power systems. According to the method, a complex power system network is analogous to a neural network, each converter (intelligent node) is analogous to a neuron, and the equivalent impedance of the power network is analogous to a connection weight in the neural network. In the neural network, neurons can update the connection weight through signal receiving and transmitting, and then simulation of a complex nonlinear system is achieved. In a power system, each control link of the converter can be equivalent to port impedance, the purpose of adjusting the system control effect is achieved by adjusting the characteristics of the port impedance, and then the self-organization and self-coordination control function is achieved. The method does not need a system model, and is essentially a method for utilizing computing power to drive system optimization.

Description

Technical Field

[0001] This invention relates to a novel data-driven self-organizing and self-coordinating control method for power systems, belonging to the field of power electronics and power systems, and applied to power electronic power systems. Background Technology

[0002] The massive influx of wind and solar renewable energy has driven the transformation of traditional power systems dominated by synchronous machines into power systems dominated by power electronic equipment, forming a dual-high power system composed of "high proportion of renewable energy" and "high proportion of power electronic equipment." Unlike traditional synchronous machines, power electronic equipment possesses high flexibility, intelligence, and controllability, but these potentials have not yet been fully realized. Conversely, current research does not fully understand the operating mechanisms of power electronic power systems, failing to fully utilize the flexible control characteristics of power electronic equipment. This has led to new problems rarely seen in traditional power systems, such as wideband oscillations, and has also given new characteristics to old problems in traditional power systems, such as synchronization stability and harmonics, requiring further research. The current mainstream approach to address these issues is to simulate synchronous machines, imparting "inertia" to the converter from the control loop and power supply side. However, this is not a route that aligns with the inherent characteristics of power electronic equipment; expensive and fragile power electronic devices and energy storage batteries are the main obstacles restricting this approach. Therefore, fully respecting the flexible and controllable nature of power electronic equipment, leveraging its flexibility, intelligence, and controllability, and using each converter as an intelligent node to form a self-organizing and self-coordinating power system network, can facilitate wide-area mutual support of high-proportion renewable energy sources, achieve power balance across multiple frequency domains, and ensure the stable operation of the power system. This is an economically feasible technical route. Summary of the Invention

[0003] The main objective of this invention is to provide a novel data-driven self-organizing and self-coordinating control method for power systems.

[0004] To achieve the above objectives, the present invention employs the following technical solution:

[0005] A novel data-driven self-organizing and self-coordinating control method for power systems is proposed. This method fully utilizes the intelligent and controllable characteristics of converters, treating the converters as neurons and the power system network as a neural network, and uses a stability evaluation function to optimize the operating state.

[0006] Specifically, the method includes: constructing a power system network based on a neural network, treating each converter as an intelligent node analogous to a neuron, treating the equivalent impedance of the power network analogous to the connection weights in the neural network, and treating the power signal analogous to a neurotransmitter; extracting the power fluctuation characteristics of adjacent converters, and optimizing the operating state of the converters based on the power fluctuation characteristics combined with the power system stability operation capability assessment function, thereby achieving coordinated control of the optimized operation of the power system.

[0007] More specifically, optimizing the operating status of the converter includes optimizing the control parameters of the converter, thereby adjusting the port impedance of the converter, and achieving coordinated control for optimized operation of the power system by reshaping the port impedance characteristics.

[0008] The innovation of this invention lies in:

[0009] Based on the flexible, controllable, and intelligent computing characteristics of converters, a novel data-driven self-organizing and self-coordinating control method for power systems is proposed. The basic idea is to analogize complex power system networks to neural networks, each converter (intelligent node) to a neuron, and the equivalent impedance of the power network to connection weights in a neural network. In a neural network, neurons can update connection weights through signal reception and transmission, thereby simulating complex nonlinear systems. In a power system, each control element of the converter can be equivalent to a port impedance. By adjusting the port impedance characteristics, the system control effect can be adjusted, thus achieving self-organizing and self-coordinating control functions. This method does not require a system model; essentially, it utilizes computing power to drive system optimization.

[0010] There is a mapping relationship between the converter and the neuron; the information received by the neuron from its neighboring neurons is... In the neuron's own activation function Under the influence of the neuron, the output is

[0011]

[0012] The converter simulates the neuron's computational process as follows:

[0013] When the converter From adjacent converters The active power received is reactive power is It can adaptively change its own voltage reference value. , and current reference value Voltage and current reference values ​​can be expressed as active and reactive power command functions, reference...

[0014] ;

[0015] in, It is a converter Voltage reference value, It is a converter The current reference value;

[0016] Furthermore, the method of the present invention also includes constructing a power system stability operation capability assessment function, and quantifying the power system stability operation status index based on the assessment function.

[0017] Define the loss function (evaluation function).

[0018]

[0019] Referring to the learning rules of neural sources, the response time Stabilization time , and overshoot A smaller loss function L indicates a faster converter response, faster stabilization, and smaller overshoot, indicating better system performance. The converter command function is updated based on the gradient of the loss function. Internal parameters.

[0020] Furthermore, the evaluation function quantifies the stable operating state of the power system based on the power fluctuation characteristics and in conjunction with the state variables of the converter. Furthermore, the state variables include, but are not limited to, voltage, current, active power, reactive power, and frequency state variables.

[0021] Furthermore, the evaluation function is calculated using the gradient descent method, and the optimization process aims to minimize the value of the evaluation function. When the evaluation function reaches the target minimum value, the power system is considered to have reached the optimization objective, and the optimization stops.

[0022] Furthermore, this invention constructs the network structure features of the power system based on the neural network structure of the power storage pool, updates the connection weights within the power storage pool, and optimizes the operation of the power system.

[0023] The present invention also provides a novel data-driven self-organizing and self-coordinating control system for power systems, the system being used to implement the aforementioned novel self-organizing and self-coordinating control method for power systems.

[0024] Furthermore, the self-organizing and self-coordinating control system of the present invention includes a converter and a controller, wherein the controller is used to execute the above-described novel power system self-organizing and self-coordinating control method.

[0025] Compared with the prior art, the advantages of the present invention are:

[0026] 1) This invention utilizes the self-organizing and self-optimizing characteristics of neural networks to propose a new analogy method for power electronic power systems with neural networks. Unlike the traditional method of configuring a neural network for each converter, this method deeply analogizes the converter to neurons, greatly reducing the energy consumed by each converter in deep calculations.

[0027] 2) This method employs a novel analogical approach, achieving a high degree of similarity between power systems and neural networks. It establishes a multi-layered and comprehensive analogy: neuron-converter, neurotransmitter-power signal, synaptic weight-equivalent impedance, and neural network-power network. Compared to traditional methods, it demonstrates better compatibility between power systems and neural networks.

[0028] 3) Utilizing the self-renewal characteristics of neurons and synapses, a self-renewal method for converter control systems was constructed. Based on the stability capability evaluation function, a self-organizing and self-coordinating method among converters was realized, which is a brand-new control paradigm. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of two converters interconnected by line impedance provided by the present invention;

[0030] Figure 2 This is a schematic diagram of the converter control structure provided by the present invention, which is equivalent to the port impedance.

[0031] Figure 3 This is a schematic diagram of neurons connected by synapses and undergoing weight optimization calculations, provided by the present invention.

[0032] Figure 4 This invention provides a 128-node power interconnection system.

[0033] Figure 5 This is a schematic diagram of the neural network structure based on reservoir calculation provided by the present invention;

[0034] Figure 6 This is a schematic diagram of the self-organizing and self-coordinating optimization control method provided by the present invention. Detailed Implementation

[0035] The technical content of the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0036] The impedance model between converters can be divided into line impedance (such as...) Figure 1 (as shown) and the converter's own port impedance (such as Figure 2 (As shown). During operation, the port impedance is determined by the control characteristics of the converter. Different control structures and control parameters exhibit different port impedance characteristics. Coordinated control for optimized system operation is achieved through port impedance reshaping.

[0037] In this invention, the characteristics of neurons are as follows: Figure 3 As shown, two neurons are connected via synapses and release neurotransmitters to excite or inhibit the connected neurons. The cell membrane potential of an excited neuron changes, while the inhibited neuron remains in its original state. Through continuous reinforcement of external stimuli, the synapses between neurons change, forming gradually converging connection weights. Externally, this manifests as the nervous system developing a "habitual" response to a certain stimulus.

[0038] There is a mapping relationship between converters and neurons. By continuously applying disturbances to the converter system (such as sudden load changes, changes in new energy output, etc.), the control characteristics of the converter itself are continuously adjusted, the port impedance characteristics are reshaped, and thus the control of the system is optimized.

[0039] There is a mapping relationship between the converter and the neuron; the information received by the neuron from its neighboring neurons is... In the neuron's own activation function Under the influence of the neuron, the output is

[0040]

[0041] In an embodiment of the present invention, the converter-type neuron and the converter-simulated neuron calculation process are as follows:

[0042] When the converter From adjacent converters The active power received is reactive power is It can adaptively change its own voltage reference value. , and current reference value Voltage and current reference values ​​can be expressed as active and reactive power command functions, reference...

[0043] ;

[0044] in, It is a converter Voltage reference value, It is a converter The current reference value;

[0045] Unlike the clearly defined layers of artificial intelligence neural networks, self-organizing and self-coordinating power systems do not have distinct input and output layers; all nodes belong to the same layer, such as... Figure 4 As shown, therefore, in terms of network structure characteristics, it is similar to Reservoir Computing, such as... Figure 5 As shown, by utilizing the structure of the power storage pool neural network, the connection weights within the power storage pool are updated to optimize the operation of the power system.

[0046] like Figure 6 As shown, by constructing a power system stability operation capability assessment function, the system stability operation status index is quantified, the power fluctuation characteristics between converters are extracted, and the converter operation status is optimized. The optimization objects include controller parameters (control parameters), suppressing the occurrence of power oscillations and other phenomena, thereby enhancing the system stability operation capability.

[0047] Define the loss function (evaluation function).

[0048]

[0049] Referring to the learning rules of neural sources, the response time Stabilization time , and overshoot A smaller loss function L indicates a faster converter response, faster stabilization, and smaller overshoot, indicating better system performance. The converter command function is updated based on the gradient of the loss function. Internal parameters.

[0050] The optimization process aims to minimize the value of the evaluation function, which evaluates state variables, including but not limited to voltage. Current meritorious No results ,frequency The evaluation function assesses the response time, settling time, and overshoot of these state variables. It's important to note that these state variables are all valued in the frequency domain, which facilitates optimization of various oscillation modes of the system. The evaluation function is calculated using gradient descent; when the evaluation function reaches its target minimum, the system is considered to have reached the optimization objective, and training ceases.

[0051] The embodiments of the present invention also disclose a novel data-driven self-organizing and self-coordinating control system for power systems, the system being used to implement the aforementioned novel self-organizing and self-coordinating control method for power systems.

[0052] Specifically, the self-organizing and self-coordinating control system of the present invention includes a converter and a controller, wherein the controller is used to execute the above-mentioned novel power system self-organizing and self-coordinating control method.

[0053] Those skilled in the art should understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the methods and techniques disclosed above to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

[0054] The self-organizing and self-coordinating optimization control method provided by the present invention has been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A novel data-driven self-organizing and self-coordinating control method for power systems, characterized in that, A power system network is constructed based on a neural network. Each converter is treated as an intelligent node, analogous to a neuron, and the equivalent impedance of the power network is analogous to the connection weights in the neural network. Power fluctuation characteristics of adjacent converters are extracted, and the operating state of the converters is optimized based on these power fluctuation characteristics and the power system stability operation capability assessment function, thereby achieving coordinated control for optimized operation of the power system.

2. The novel self-organizing and self-coordinating control method for power systems according to claim 1, characterized in that, Optimizing the operating status of the converter includes optimizing the converter's control parameters, thereby adjusting the converter's port impedance. By reshaping the port impedance characteristics, coordinated control for optimized operation of the power system can be achieved.

3. The novel self-organizing and self-coordinating control method for power systems according to claim 1, characterized in that, The converter simulates the neuron's computational process as follows: When the converter From adjacent converters The active power received is reactive power is It can adaptively change its own voltage reference value. , and current reference value The reference values ​​for voltage and current can be expressed as functions of active and reactive power: .

4. The novel power system self-organizing and self-coordinating control method according to claim 1, characterized in that, It also includes constructing the evaluation function, which quantifies the stable operation state indicators of the power system based on the power fluctuation characteristics and in combination with the state variables of the converter.

5. The novel self-organizing and self-coordinating control method for power systems according to claim 1, characterized in that, The evaluation function is calculated using the gradient descent method. The optimization process aims to minimize the value of the evaluation function. When the evaluation function reaches the target minimum value, the power system is considered to have reached the optimization target, and the optimization stops.

6. The novel power system self-organizing and self-coordinating control method according to claim 4, characterized in that, The state quantities include, but are not limited to, voltage, current, active power, reactive power, and frequency.

7. The novel self-organizing and self-coordinating control method for power systems according to claim 1, characterized in that, The network structure features of the power system are constructed based on the neural network structure of the power storage pool, and the connection weights within the power storage pool are updated to optimize the operation of the power system.

8. A novel data-driven self-organizing and self-coordinating control system for power systems, characterized in that, The system is used to implement the novel power system self-organizing and self-coordinating control method as described in any one of claims 1-7.

9. The novel data-driven self-organizing and self-coordinating control system for power systems according to claim 8, characterized in that, It includes a converter and a controller, the controller being used to execute the novel power system self-organizing and self-coordinating control method.

Images