CN-122020065-A - Online evaluation method and system for equivalent inertia of power distribution network driven by random modal switching

CN122020065ACN 122020065 ACN122020065 ACN 122020065ACN-122020065-A

Abstract

The invention relates to the technical field of power system inertia evaluation, in particular to a method and a system for evaluating equivalent inertia of a power distribution network driven by random modal switching on line, wherein the method comprises the steps of defining a group of hidden operation modes and constructing an unobservable Markov chain based on the equivalent inertia of the power distribution network; the method comprises the steps of constructing a double-layer hidden Markov jump system model, constructing a lower layer of an unobservable Markov chain, constructing an upper layer of a continuous dynamic behavior model describing a power distribution network, deducing the probability of a dominant operation mode and each hidden operation mode of the maximum probability of the model inference system based on PMU high-resolution time sequence disturbance data, and carrying out multi-mode probability self-adaptive weighting on each equivalent inertia estimated value based on the probability of the dominant operation mode and each hidden operation mode so as to obtain the equivalent inertia estimated value of the power distribution network on line. The method and the system effectively solve the problem that the equivalent inertia of the power distribution network is difficult to evaluate rapidly, accurately and dynamically in a self-adaptive mode under the conditions of high penetration of new energy and frequent switching of working conditions.

Inventors

WANG CHONG
CAO YANG
JIANG LING
JIN YUQING
JIANG TINGYU

Assignees

河海大学

Dates

Publication Date: 20260512
Application Date: 20260408

Claims (10)

1. An online evaluation method for equivalent inertia of a power distribution network driven by random modal switching is characterized by comprising the following steps: Based on equivalent inertia source constitution and operation modes of the power distribution network, a group of discrete mutually exclusive hidden operation modes are defined a priori, and an unobservable Markov chain driven by a transition probability matrix is constructed to represent random switching of the hidden operation modes; Constructing a double-layer hidden Markov jump system model, wherein the bottom layer of the double-layer hidden Markov jump system model is the invisible Markov chain, the upper layer is a continuous dynamic behavior model for describing the frequency dynamic response of the power distribution network, and the state space equation parameters of the upper layer are determined by the hidden operation mode of the bottom layer; performing an improved online sequential expectation maximization algorithm on the double-layer hidden Markov jump system model based on PMU high-resolution time sequence disturbance data so as to infer the probability of the dominant operation mode of the maximum probability of the system and the probability of each hidden operation mode according to mode identification and parameter estimation; Based on the maximum probability dominant operation mode and the probability of each hidden operation mode, carrying out multi-mode probability self-adaptive weighting on the equivalent inertia estimated value corresponding to each hidden operation mode so as to obtain the equivalent inertia estimated value of the power distribution network on line.
2. The method for online evaluation of equivalent inertia of a stochastic mode switch-driven power distribution network according to claim 1, wherein constructing an unobservable markov chain driven by a transition probability matrix comprises: defining the hidden operation modes, wherein the hidden operation modes at least comprise a low equivalent inertia mode of high renewable energy, a high equivalent inertia mode of synchronous machine leading, a virtual equivalent inertia mode of strong energy storage support and a load reduction mode in emergency; representing the implicit operating modality as a set of discrete states of a markov chain; constructing a discrete time Markov chain meeting first-order Markov as the underlying hopping mechanism, and driving modal switching by adopting the transition probability matrix so as to enable elements of the transition probability matrix to be switched Representing the probability of a transition from modality i to modality j in a time step; the residence time of each modality is estimated based on historical data and a jump rate associated with the residence time to generate a modality duration model that characterizes the implicit operating modality residence time.
3. The method for online evaluation of equivalent inertia of a power distribution network driven by random modal switching as set forth in claim 1, wherein constructing a double-layer hidden markov jump system model includes: And constructing a continuous dynamic behavior model of the power distribution network frequency dynamic response as a model of the upper layer, modeling the power distribution network frequency dynamic response after power disturbance into a linear state space equation for each implicit operation mode, so that the linear state space equation comprises a state equation and an observation equation, and respectively introducing process noise and observation noise related to the implicit operation mode.
4. The method for online evaluation of equivalent inertia of a power distribution network driven by random modal switching as recited in claim 3, wherein constructing a double-layer hidden markov jump system model further comprises: The parameters of the model of the upper layer jump along with the hidden operation mode of the bottom layer, and a system matrix, an input matrix and an observation matrix related to the mode are expressed as parameter matrices related to equivalent inertia, damping coefficients and power response time constants under the hidden operation mode.
5. The method for online evaluation of equivalent inertia of a power distribution network driven by random modal switching according to claim 1, wherein the online sequential expectation maximization algorithm is executed based on an empirical maximization iterative algorithm framework; The empirical maximization iterative algorithm framework comprises an E step and an M step, wherein the E step is used for calculating posterior probability distribution of a hidden modal sequence based on current parameter estimation values, and the M step is used for updating parameter estimation values based on the posterior probability distribution so as to maximize expected complete data log likelihood functions.
6. The method for online evaluation of equivalent inertia of a power distribution network driven by random mode switching according to claim 5, wherein the step of deducing the probability of the system maximum probability dominant operation mode and each hidden operation mode comprises the following steps: Calculating a modal probability by a forward-backward algorithm, wherein a forward variable represents a joint probability that the system is in a mode j at a time T given observation data from time 1 to T, and a backward variable represents a conditional probability that an observation sequence from t+1 to T appears in the future given that the system is in a mode i at the time T; and obtaining the posterior probability of the time t in the mode i and the joint probability of the time t in the mode i and the transition from the time t+1 to the mode j based on the forward variable and the backward variable.
7. The method for online evaluation of equivalent inertia of a power distribution network driven by random mode switching according to claim 6, wherein the method for estimating the probability of the highest probability dominant operation mode of the system and each hidden operation mode further comprises: Maximizing a Q function constructed based on the posterior probability and the joint probability to update parameters, wherein a portion of the continuous dynamic parameters is converted into a weighted least squares problem weighted by the posterior probability to update equivalent inertia and damping parameters for each modality, and introducing a forgetting factor to adjust the decay rate of historical information in the current parameter estimation, while updating the transition probability matrix based on the ratio of the number of hops from modality i to modality j to the total number of hops from modality i.
8. The method for online evaluation of equivalent inertia of a power distribution network driven by random modal switching according to claim 1, wherein the power distribution network equivalent inertia evaluation value is an adaptive weighting result of the equivalent inertia evaluation values of all modes with posterior probability of each implicit operation mode as a weight.
9. The utility model provides a distribution network equivalent inertia on-line evaluation system of random mode switching drive which characterized in that, the system includes: The chain construction module is used for defining a group of discrete mutually exclusive hidden operation modes in a priori based on equivalent inertia source constitution and operation modes of the power distribution network, and constructing an unobservable Markov chain driven by a transition probability matrix to represent random switching of the hidden operation modes; The jump system module is used for constructing a double-layer hidden Markov jump system model, wherein the bottom layer of the double-layer hidden Markov jump system model is an invisible Markov chain, the upper layer is a continuous dynamic behavior model for describing the frequency dynamic response of the power distribution network, and the state space equation parameters of the upper layer are determined by the hidden operation mode of the bottom layer; The probability inference module executes an improved online sequential expectation maximization algorithm on the double-layer hidden Markov jump system model based on PMU high-resolution time sequence disturbance data so as to infer the probability of the system maximum probability dominant operation mode and each hidden operation mode according to mode identification and parameter estimation; The weighting evaluation module carries out multi-mode probability self-adaptive weighting on the equivalent inertia estimated value corresponding to each hidden operation mode based on the probability of the maximum probability dominant operation mode and the probability of each hidden operation mode so as to obtain the equivalent inertia estimated value of the power distribution network on line.
10. The random modal switching driven power distribution network equivalent inertia online assessment system of claim 9, wherein the markov chain building module comprises: The system comprises a mode definition unit, a load reduction unit and a load reduction unit, wherein the mode definition unit defines a hidden operation mode, and the hidden operation mode comprises a low equivalent inertia mode of high renewable energy, a high equivalent inertia mode of synchronous machine leading, a virtual equivalent inertia mode of strong energy storage support and a load reduction mode in emergency; A discrete representation unit that represents the implicit operating modality as a set of discrete states of a Markov chain; a chain construction unit for constructing a discrete time Markov chain meeting first-order Markov as a bottom layer jump mechanism and driving mode switching by using a transition probability matrix so as to enable elements of the transition probability matrix Representing the probability of a transition from modality i to modality j in a time step; And a residence characterization unit for estimating residence time of each mode and jump rate related to the residence time based on the historical data to generate a mode duration model for characterizing the residence time of the implicit operation mode.

Description

Online evaluation method and system for equivalent inertia of power distribution network driven by random modal switching Technical Field The invention relates to the technical field of power system inertia evaluation, in particular to a method and a system for evaluating equivalent inertia of a power distribution network driven by random modal switching on line. Background The equivalent inertia of the power system is used as a core index for measuring the frequency stability of the system, and the capability of the power grid for resisting power disturbance is reflected. With the large-scale grid connection of renewable energy sources, new energy sources such as wind power generation, photovoltaic power generation and the like are connected into a power grid through a power electronic inverter interface, so that the duty ratio of the traditional rotary machine is obviously reduced. The transformation makes the power system show two major characteristics of overall reduced equivalent inertia level and high time-varying inertia characteristic, and particularly dynamically changes along with fluctuation of new energy output, load change and operation mode switching. In particular, for a power distribution network or a receiving end power network with a complex topological structure, uncertainty of inertia space-time distribution becomes a great hidden danger for threatening the safe operation of the power network. The current inertia evaluation technology has limitations, firstly, a disturbance dependent method is required to rely on dynamic response data after a large-scale power disturbance event (such as a machine set off-grid), an effective evaluation result is difficult to obtain during steady-state operation of a system, secondly, a steady-state identification method is aimed at carrying out parameter identification based on a slowly-changing quasi-steady-state working condition, instantaneous inertia characteristics under the scenes of rapid start-stop of new energy, abrupt load change and the like can not be captured, and a static probability modeling method adopts a Markov chain to describe a historical operation state, but only can realize post state classification, can not sense the disturbance response characteristic of the system in real time, and also can not realize online identification capability of random switching modes. The technical defects cause the defects of insufficient dynamic adaptability, difficulty in tracking inertia fluctuation caused by high-permeability new energy access, limited real-time performance, incapability of meeting on-line evaluation requirements depending on complete disturbance events or long-period steady-state data, weak working condition adaptability and lack of effective modeling means for the operation characteristics of multi-mode random switching of the power distribution network. The information disclosed in this background section is only for enhancement of understanding of the general background of the disclosure and is not to be taken as an admission or any form of suggestion that this information forms the prior art that is well known to a person skilled in the art. Disclosure of Invention The invention provides a method and a system for evaluating equivalent inertia of a power distribution network driven by random modal switching on line, which can effectively solve the problems in the background technology. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: an online evaluation method for equivalent inertia of a power distribution network driven by random modal switching, comprising the following steps: Based on equivalent inertia source constitution and operation modes of the power distribution network, a group of discrete mutually exclusive hidden operation modes are defined a priori, and an unobservable Markov chain driven by a transition probability matrix is constructed to represent random switching of the hidden operation modes; Constructing a double-layer hidden Markov jump system model, wherein the bottom layer of the double-layer hidden Markov jump system model is the invisible Markov chain, the upper layer is a continuous dynamic behavior model for describing the frequency dynamic response of the power distribution network, and the state space equation parameters of the upper layer are determined by the hidden operation mode of the bottom layer; performing an improved online sequential expectation maximization algorithm on the double-layer hidden Markov jump system model based on PMU high-resolution time sequence disturbance data so as to infer the probability of the dominant operation mode of the maximum probability of the system and the probability of each hidden operation mode according to mode identification and parameter estimation; Based on the maximum probability dominant operation mode and the probability of each hidden operation mode, carrying out multi-mode probability self-adaptiv