CN-121637103-B - Power network topology identification method and system based on intelligent fusion terminal

CN121637103BCN 121637103 BCN121637103 BCN 121637103BCN-121637103-B

Abstract

The invention provides a power network topology identification method and system based on an intelligent fusion terminal, and the technical scheme adopted by the invention is that an orthogonal spread spectrum code sequence is constructed and distributed to each end node, and the power frequency zero crossing point of a power line is used as a physical synchronization reference of the whole network to control all the end nodes to simultaneously carry out micro-power parallel physical modulation; and carrying out differential preprocessing on the acquired superimposed synthesized waveform data at the fusion terminal side, carrying out iterative despreading by utilizing a serial interference cancellation algorithm based on in-situ waveform template learning, analyzing the signal intensity of each node, and finally determining the attribution of the physical topology by combining multiport competition judgment logic. The invention can solve the technical problems of low topology identification efficiency and low identification accuracy under the environment of strong background noise and signal crosstalk.

Inventors

XIONG FEI
WANG GUOLONG
CHEN YONGSHI
WANG LEI
ZHAO FUSHENG

Assignees

航中天启(重庆)微电子股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260205

Claims (10)

1. An intelligent fusion terminal-based power network topology identification method, wherein the method operates in a power network system comprising a central edge computing node and a plurality of end nodes, and comprises the following steps: The central edge computing node constructs a group of mutually orthogonal spread spectrum code sequences, and distributes the spread spectrum code sequences to the end nodes in the power network system respectively; In response to the end node receiving the spreading code sequence, locking a modulation start time according to a uniform time reference; Each end node executes modulation action on the electric physical quantity of the connection point according to the spread spectrum code sequence logic of the end node, and superposition synthesized waveform data formed by a plurality of end node modulation signals and power network background signals is generated on a power line; The central edge computing node collects the overlapped and synthesized waveform data on the power line at a preset sampling rate; preprocessing the overlapped synthesized waveform data, and extracting a characteristic sequence containing modulation characteristics; the central edge computing node carries out iterative processing on the characteristic sequences in an edge computing environment and analyzes the signal component intensities corresponding to the spread spectrum code sequences; and judging the physical topology attribution relation of each end node and the central edge computing node monitoring port according to the signal component intensity.
2. The method of claim 1, wherein determining the uniform time reference according to the steps of: Each end node takes the zero crossing point of the power line voltage waveform as a reference point of a common time reference; and the end node starts a modulation action at the Nth reference point after receiving the spread spectrum code sequence.
3. The method of claim 1, wherein the spreading code sequence logic specifically comprises: Configuring a duration of each symbol in the spreading code sequence to be an integer multiple of a power frequency cycle of a power line or an integer multiple of a half cycle; the spread spectrum code sequence is configured to maintain the number of end nodes performing modulation action and the number of end nodes keeping silence within a preset balance proportion range at any modulation time.
4. The method according to claim 1, characterized in that the modulating action is performed on the electro-physical quantity of the connection point, in particular comprising: the end node controls the impedance element, the load element and/or the power element which are connected in the end node to switch between different states according to the logic level in the spread spectrum code sequence; the modulation action is configured in a micropower modulation mode such that the signal fluctuation amplitude produced by a single end node is lower than the base amplitude of the power line background noise.
5. The method of claim 1, wherein preprocessing the superimposed synthetic waveform data to extract a signature sequence comprising modulation signatures comprises: Removing the power frequency component in the overlapped synthesized waveform data according to the power frequency reference of the power line; And performing normalized mapping on the superimposed synthesized waveform data with the power frequency components removed, and generating a feature sequence logically matched with the spread spectrum code sequence.
6. The method according to claim 1, wherein the step of iteratively processing the signature sequences to resolve the signal component strengths corresponding to the respective spreading code sequences comprises: Carrying out parallel correlation operation on the current characteristic sequence and all unidentified spread spectrum code sequences, locking a target spread spectrum code with the largest correlation peak value, and taking the absolute value of the correlation peak value as the signal component strength of the target spread spectrum code; cutting the current characteristic sequence into a plurality of code element fragments according to the code element time sequence of the target spread spectrum code; all code element fragments corresponding to logic 1 are subjected to superposition average to generate a forward characteristic template; carrying out superposition average on all code element fragments corresponding to logic-1 or 0 to generate a negative characteristic template; Splicing the positive characteristic template and the negative characteristic template according to the sequential logic of the target spread spectrum code to generate a reconstruction waveform of a target terminal node; subtracting the reconstructed waveform from the current characteristic sequence to obtain residual waveform data; In response to the residual waveform data not meeting a preset termination condition, updating the residual waveform data into a current characteristic sequence, and repeatedly executing the parallel correlation operation and subsequent steps; Responding to the residual waveform data to meet a preset termination condition, terminating iteration and outputting all resolved target spread spectrum codes and corresponding signal component intensities thereof; the preset termination condition comprises: calculating the root mean square value of the residual waveform data as a noise base of the current iteration round; Calculating a ratio of the signal component strength recorded in a current iteration round to the noise floor; Responding to the ratio being smaller than a preset minimum signal-to-noise ratio threshold or the current iteration number reaching a preset maximum iteration number, and judging that the termination condition is met; And responding to the ratio being greater than or equal to a preset minimum signal-to-noise ratio threshold or the current iteration number reaching a preset maximum iteration number, and judging that the termination condition is not met.
7. The method of claim 6, wherein performing the superposition average comprises: Extracting K code element fragments belonging to the same logic state from the characteristic sequence; performing arithmetic average on all the code element fragments to obtain an initial reference vector; Calculating a correlation coefficient between each symbol segment and the initial reference vector as a weight of the current symbol segment: Wherein, the Representing the weights of the current symbol fragment, Representing the current symbol fragment, Representing a current symbol fragment The average value of all the elements in (a), The initial reference vector is represented as such, Representing the average of all the initial reference vectors, An L2 norm representing the vector; And carrying out weighted average on the code element fragments by utilizing the weight of the current code element fragments to generate a characteristic template: Wherein, the The feature templates are represented by a set of feature templates, Representing the weights of the current symbol fragment, Representing the current symbol fragment.
8. The method according to claim 1, wherein determining the physical topology affiliation of each end node with the central edge computing node monitoring port based on the signal component strengths, comprises: reading the intensity of all signal components in the current identification period; for each spread spectrum code sequence, calculating the ratio of the corresponding signal component intensity to the original total energy of the characteristic sequence to obtain an energy duty ratio coefficient; In response to the absolute value of the signal component strength of the spread spectrum code sequence being greater than a preset hard decision threshold, and the energy duty cycle coefficient thereof being greater than a preset correlation coefficient threshold, determining that an end node assigned the spread spectrum code sequence is physically connected to the current central edge computing node monitoring port; In response to the signal component strength or the energy duty cycle coefficient not meeting a threshold condition, it is determined that the end node is not connected to a current monitoring port.
9. The method of claim 1, wherein when the central edge computing node includes a plurality of monitoring ports, determining a physical topology home relationship of each end node to a central edge computing node monitoring port, further comprising a multi-port contention determination, the steps comprising: The central edge computing node simultaneously collects the characteristic sequences of the plurality of monitoring ports, and independently executes the iterative processing respectively to construct intensity characteristic vectors aiming at the same end node under different monitoring ports; identifying a maximum value and a next-largest value in the intensity feature vector, and calculating a difference ratio of the maximum value and the next-largest value; responding to the fact that the maximum value exceeds a preset effective signal threshold, and the difference ratio exceeds a preset crosstalk suppression tolerance, and judging that the end node is uniquely attributed to a monitoring port corresponding to the maximum value; And in response to the difference ratio being lower than the crosstalk suppression tolerance, determining that the end node is in a signal crosstalk area, and generating an abnormal alarm without attribution determination.
10. An intelligent fusion terminal-based power network topology identification system is characterized by comprising: The intelligent fusion terminal is used as a central edge computing node and comprises an acquisition module, a communication unit and a processor, wherein the processor provides an edge computing environment; the system comprises a plurality of end nodes, a synchronous trigger unit, a zero crossing detection circuit and a power modulation circuit, wherein the end nodes are distributed on a branch circuit or a user side of a power network; The processor is used for constructing a group of mutually orthogonal spread spectrum code sequences and respectively distributing the spread spectrum code sequences to each end node in the power network topology identification system through the communication unit; the end node is used for responding to the synchronous triggering unit to receive the spread spectrum code sequence, capturing a voltage zero crossing point by utilizing the zero crossing detection circuit to lock a unified time reference, controlling the power modulation circuit to synchronously execute modulation action on the electric physical quantity of the connecting point according to the logic of the spread spectrum code sequence, and generating a mixed characteristic signal on a power line; The acquisition module is used for acquiring superposition synthesized waveform data on a power line at a preset sampling rate, wherein the superposition synthesized waveform data comprises a power network background signal and the mixed characteristic signal; The processor is also used for executing a computer program in the edge computing environment to preprocess the overlapped synthesized waveform data, extract a characteristic sequence containing modulation characteristics, iterate the characteristic sequence, analyze signal component intensities corresponding to the spread spectrum code sequences, and judge the physical topology attribution relation of the end nodes and the central edge computing node monitoring port according to the signal component intensities.

Description

Power network topology identification method and system based on intelligent fusion terminal Technical Field The invention relates to the technical field of low-voltage power distribution network monitoring, in particular to an electric power network topology identification method and system based on an intelligent fusion terminal. Background Along with the construction of a novel power system, a low-voltage distribution transformer area is connected with massive distributed photovoltaic, electric automobile charging piles and intelligent household loads. In order to realize accurate regulation and lean management of the transformer area source network load storage, it becomes important to define the physical topology attribution relationship between each end node and the side of the distribution transformer (transformer area fusion terminal). The existing power network topology identification technology mainly comprises a method based on power line carrier communication characteristic analysis and a method based on characteristic current or voltage pulse injection. In the prior art based on characteristic signal injection, a time division multiplexing or polling mechanism is generally adopted, that is, a central node controls each end node to sequentially send specific current pulses or manufacture voltage drops, and the central node judges attribution by detecting whether signals exist or not. However, there are significant limitations to this existing serial polling mechanism. Firstly, when the number of nodes in a platform area is large, for example, hundreds of nodes are reached, a very long time is required to traverse all the nodes, the recognition efficiency is low, and the requirement of dynamic topology real-time perception cannot be met. Secondly, the low-voltage distribution network has a bad environment, strong power frequency fundamental waves, abundant harmonic noise and load fluctuation (such as shielding of photovoltaic cloud layers and starting and stopping of heavy loads), the existing simple threshold detection method is extremely easy to be submerged by background noise under the environment of low signal to noise ratio to cause missed judgment, or is interfered by impulse noise to cause misjudgment, and the anti-interference capability is weak. Furthermore, under the scenario of multi-phase or multi-loop parallel laying, crosstalk generated by electromagnetic coupling between wires can cause non-connection phases to detect inductive signals, and the prior art lacks an effective competition judgment mechanism to distinguish real connection signals from inductive crosstalk signals, so that phase or branch identification is difficult. To improve the signal-to-noise ratio, the prior art often requires that the end node generate a larger current jump or voltage drop, which can cause flickering pollution to the power quality of the power grid and even trigger the protection action of sensitive equipment. Meanwhile, the traditional recognition algorithm cannot fully utilize the strong edge computing capability of the new-generation intelligent fusion terminal, and is difficult to process complex signal demodulation and waveform analysis tasks. Therefore, how to utilize the edge computing power to realize high-efficiency, parallel and high-reliability topology identification of mass end nodes on the premise of not affecting the power quality is a technical problem to be solved in the field. Disclosure of Invention Aiming at the defects existing in the prior art, the invention provides an electric power network topology identification method and system based on an intelligent fusion terminal. The invention solves the technical problems of low topology identification efficiency and low identification accuracy in a strong background noise and signal crosstalk environment in the prior art. The technical scheme includes that orthogonal spread spectrum code sequences are constructed and distributed to all end nodes, power frequency zero-crossing points of a power line are used as physical synchronization references of a whole network, all the end nodes are controlled to simultaneously carry out micropower parallel physical modulation, differential preprocessing is carried out on collected overlapped synthesized waveform data at a fusion terminal side, iterative despreading is carried out by utilizing a serial interference elimination algorithm based on in-situ waveform template learning, signal intensity of each node is analyzed, and finally physical topology attribution is determined by combining multiport competition judgment logic. In a first possible implementation, the method operates in a power network system comprising a central edge computing node and a plurality of end nodes, comprising the steps of: The central edge computing node constructs a group of mutually orthogonal spread spectrum code sequences, and distributes the spread spectrum code sequences to all end nodes in the power networ