CN-122000855-A - Port microgrid direct-current bus voltage intelligent control method based on virtual direct-current motor

CN122000855ACN 122000855 ACN122000855 ACN 122000855ACN-122000855-A

Abstract

The invention provides an intelligent control method for port micro-grid direct current bus voltage based on a virtual direct current motor, and relates to the technical field of port micro-grid control. The method comprises the steps of firstly collecting port microgrid direct current bus voltage, establishing and discretizing a virtual direct current motor model, then establishing a compact dynamic linearization data model between virtual mechanical torque and direct current bus voltage, calculating a pseudo partial derivative based on input and output data, designing a model-free self-adaptive controller to realize self-adaptive compensation of the virtual mechanical torque, carrying out self-adaptive adjustment on parameters of the model-free self-adaptive controller by reinforcement learning, and finally simulating external characteristics of the direct current motor by controlling a power converter to realize stable control of the direct current bus voltage. The method can effectively inhibit the voltage fluctuation of the direct current bus in the port microgrid caused by the start and stop of the loading and unloading equipment, load switching and the fluctuation of new energy output, and improves the dynamic response performance and the system operation stability of the voltage regulation of the direct current bus.

Inventors

YAO WENLONG
YANG LEI
YUE YAOBIN
CHENG PENGFEI
DONG SONGLI
LIU YICHEN

Assignees

青岛科技大学

Dates

Publication Date: 20260508
Application Date: 20260210

Claims (6)

1. The port microgrid direct current bus voltage intelligent control method based on the virtual direct current motor is characterized by comprising the following steps of: s1, collecting port microgrid direct current bus voltage, and establishing and discretizing a virtual direct current motor model; s2, constructing a compact dynamic linearization data model between the virtual mechanical torque and the direct current bus voltage; S3, estimating a pseudo partial derivative based on input and output data, and designing a model-free self-adaptive controller to realize self-adaptive compensation of the virtual mechanical torque; s4, adaptively adjusting parameters of the model-free adaptive controller by using reinforcement learning; s5, simulating the external characteristics of the direct current motor by controlling the power converter to realize the stable control of the direct current bus voltage.
2. The method for controlling the direct current bus voltage of the port micro-grid based on the virtual direct current motor according to claim 1, wherein in the step S1, the step of collecting the direct current bus voltage of the port micro-grid, and the step of establishing and discretizing the virtual direct current motor model specifically comprises the following steps: (1) The simulation of the output external characteristics of the direct current generator is realized by establishing a virtual direct current motor model: The armature equation and the mechanical equation of the armature loop of the virtual direct current motor are respectively as follows: Wherein, the ; Representing a torque coefficient; Representing magnetic flux; The actual angular velocity of the direct current motor; the rated angular speed of the direct current motor is set; Is the moment of inertia; is a damping coefficient; Is a mechanical torque; Is electromagnetic torque; Is electromagnetic power; (2) Discretizing an armature equation by combining a mechanical equation of the virtual direct current generator, wherein the expression is as follows: Wherein, the The pole pair number of the generator is; Indicating time of day Direct current bus voltage of (2); The DC bus voltage of the next sampling moment is represented; Representation of Virtual mechanical torque at the moment.
3. The port micro-grid direct current bus voltage control method based on the virtual direct current motor of claim 1, wherein in step S2, the construction of the compact dynamic linearization data model between the virtual mechanical torque and the direct current bus voltage specifically comprises the following steps: (1) Establishing a discrete time nonlinear system: Wherein, the Indicating that the system is at discrete moments Response output of the time system; representing the system at discrete moments Control input of (a); And Respectively representing the historical hysteresis orders corresponding to the output variable and the input variable; An unknown nonlinear function describing the dynamic characteristics of the system; (2) The discrete-time nonlinear system satisfies the following constraint: The systems being respectively related to And (3) with Is present and continuous; the system meets the bounded condition of generalized Lipschitz, namely, at any discrete sampling time t, when the input torque is increased When the output variable quantity is not zero, the output variable quantity of the system meets the following conditions: Wherein , B is a positive constant; (3) Establishing a compact dynamic linearization data model only related to the input virtual mechanical torque and the output direct current bus voltage by a compact dynamic linearization method: Wherein, the Is that The direct current bus voltage at moment; Is that The direct current bus voltage at moment; Is a virtual mechanical torque increment; the pseudo partial derivative of the virtual mechanical torque to the DC bus voltage represents the equivalent sensitivity of the voltage to the torque change, and is estimated on line by a model-free self-adaptive algorithm.
4. The method according to claim 1, wherein in step S3, the estimating of the pseudo-partial derivative based on the input/output data and the designing of the model-free adaptive controller, the adaptive compensation of the virtual machine torque comprises: (1) Calculating a pseudo partial derivative estimation law of the DC bus voltage: Wherein, the Is that Is a function of the estimated value of (2); Is a virtual mechanical torque increment; the voltage increment of the direct current bus is shown; is a step factor; Is a weight factor; (2) And designing a model-free self-adaptive controller of the DC bus voltage: establishing criterion functions To (3) pair The derivation is carried out and the derivation is equal to zero, and the control law of obtaining the virtual mechanical torque is as follows: wherein lambda >0 is a weight factor for limiting the control input variation amplitude; the voltage of the direct current bus is expected; is a step size factor.
5. The method of claim 4, wherein the step (1) comprises the steps of: (11) The set-up criterion function is as follows: (12) For both sides of the criterion function Deriving and making zero to obtain a pseudo Jacobian matrix number estimation law: Wherein, the Representing a pseudo partial derivative estimation step factor for adjusting the speed of pseudo partial derivative estimation update; Regularization parameters are represented to prevent the mother from approaching zero when the virtual machine torque delta is small.
6. The method according to claim 1, wherein in step S4, the adaptively adjusting model-free adaptive controller parameters using reinforcement learning specifically comprises: (1) Constructing a reinforcement learning state vector, taking current parameters of the model-free self-adaptive controller as reinforcement learning state input, wherein the state vector is expressed as: Wherein, the For controlling gain weight parameters; to control regularization parameters; estimating a regular parameter for the pseudo-bias guide; Estimating a step size parameter for the pseudo-partial derivative; (2) Constructing a reinforcement learning action space based on the state vector, and weighting the control gain parameters Controlling regularization parameters Estimating regular parameters by pseudo-bias leads Pseudo-bias estimation step size parameter And performing joint adjustment to update parameters according to the following mode: (3) Constructing a reinforcement learning reward function, wherein the reward function is related to the DC bus voltage tracking error and the control input variation, and is defined as: Wherein, the Is a direct current bus voltage reference value; Is the actual DC bus voltage; Is a virtual mechanical torque increment; is a penalty coefficient; (4) Building a discount return function according to the rewarding function: Wherein, the Indicating the slave time A starting discount return value; Indicating time of day The value of the immediate return function; Represent the first Discount weights reported at a future time; (5) Updating the reinforcement learning strategy based on the discount return function, and enabling the control gain weight parameter to be under the premise of meeting parameter constraint conditions Controlling regularization parameters Estimating regular parameters by pseudo-bias leads Pseudo-bias estimation step size parameter And (3) self-adaptively adjusting, so that the dynamic and steady-state control performance of the direct-current bus voltage is optimized while the stability of the system is ensured.

Description

Port microgrid direct-current bus voltage intelligent control method based on virtual direct-current motor Technical Field The invention belongs to the technical field of port micro-grid control, and particularly relates to an intelligent port micro-grid direct-current bus voltage control method based on a virtual direct-current motor. Background With the continuous promotion of port electrification, intellectualization and low carbonization level, a micro-grid taking new energy and power electronic equipment as cores is increasingly widely applied to port shore power systems, loading and unloading equipment power supply systems and port area comprehensive energy systems. The micro-grid has the advantages of less energy conversion links, high system efficiency, easiness in accessing photovoltaic, energy storage and other distributed power sources, and has become an important development direction of a port green energy system. However, because port load has strong impact, frequent fluctuation and complex operation conditions, the voltage of the direct current bus of the micro-grid is easily influenced by abrupt load change, power unbalance and parameter uncertainty, thereby generating voltage fluctuation and even causing system stability problems. Therefore, the realization of the rapid, stable and robust control of the DC bus voltage is a key technical problem in the operation of the port micro-grid. Aiming at the problems of insufficient inertia of the micro-grid and busbar voltage fluctuation, related researches provide a control method based on a virtual direct current motor. Huang et al propose a virtual DC motor control strategy, by introducing the mechanical equation and the armature equation of the DC generator in the control of the power electronic converter, the rotational inertia and damping characteristics of the DC generator are simulated, so that the DC bus presents dynamic response characteristics similar to those of a synchronous generator when the load disturbance and the new energy output fluctuation, and the stability of the DC bus voltage is improved. The method provides a certain energy buffering capacity for the micro-grid by introducing virtual inertia and damping. However, the above-mentioned virtual dc motor control method generally uses fixed virtual moment of inertia and damping parameters, and the adjustment manner is essentially to indirectly influence the dc bus voltage by changing the dynamic characteristics of the system. When the port micro-grid load is suddenly changed or the power is seriously unbalanced, the port micro-grid load is regulated only by virtue of virtual inertia and damping parameters, certain hysteresis exists in the voltage compensation process, the direct-current bus voltage deviation is difficult to effectively correct in time, and the problems of low voltage recovery speed or larger steady-state deviation are easy to occur. In order to improve the adaptability of the virtual direct current motor control method to the working condition change, the patent discloses a virtual direct current generator control method based on parameter self-adaption, which is used for carrying out on-line adjustment on virtual rotational inertia according to the voltage deviation of a direct current bus by introducing a Proportional Integral (PI) control link in the design of virtual rotational inertia parameters, thereby improving the dynamic response speed of the system under load disturbance. However, the adjustment object of the method is still a virtual rotational inertia parameter, the compensation path still belongs to an indirect adjustment mode, and the proportional integral control introduces an error integral link, so that system oscillation is easy to be caused under the condition of frequent switching of loads or strong uncertainty of system parameters, and the requirement of a port micro-grid on quick and stable adjustment of the voltage of a direct current bus is difficult to be met. On the other hand, in order to reduce the dependence on accurate system models, model-free adaptive control methods are receiving attention. Hou Zhongsheng et al propose a model-free adaptive control theory, and the method realizes the adaptive control without accurate system model by dynamic linearization in a compact format and only utilizing the input and output data of the system to estimate the pseudo partial derivative of the system on line and designing a control law according to the pseudo partial derivative. The model-free adaptive control shows good adaptability in a system with strong nonlinearity and obvious model uncertainty. However, in the existing model-free adaptive control method, direct current bus voltage is mostly used as a direct control object, and is not effectively combined with a physical model of a virtual direct current motor, so that the key energy adjustment amount in the virtual direct current motor is difficult to control in a targeted manner. Meanwh