CN-122017847-A - Cross-modal fusion monitoring method for millimeter wave radar and microclimate information of urban power station

CN122017847ACN 122017847 ACN122017847 ACN 122017847ACN-122017847-A

Abstract

The invention discloses a cross-modal fusion monitoring method for millimeter wave radar and microclimate information of an urban power station, which comprises the steps of S1 collecting echo amplitude and phase information of the millimeter wave radar, inverting tiny deformation of a structure, S2 synchronously collecting microclimate information and constructing an environment state vector, S3 preprocessing multi-source data such as time synchronization and outlier rejection, S4 extracting structure response characteristics and environment characteristics to construct a cross-modal characteristic set, S5 establishing association relation between a structure and the environment through a fusion mapping function, and S6 constructing a safety state index based on a fusion result to realize anomaly identification and risk output. According to the invention, through a cross-mode fusion technology, the non-contact perception advantage of the radar and the environmental characterization effect of microclimate are comprehensively utilized, the problems of single traditional monitoring means, weak anti-interference capability, difficult early abnormality identification and the like are solved, and the continuous and accurate monitoring of the structure and the environment of the electric power station can be realized in a complex urban scene.

Inventors

ZHENG XIANGTIAN
JI CHEN
HUANG WEI
CAI WEI
ZHAO QUN
CHEN XING

Assignees

南京工程学院

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (7)

1. A cross-mode fusion monitoring method for millimeter wave radar and microclimate information of an urban power station is characterized by comprising the following steps: s1, millimeter wave radar monitoring data acquisition, namely, distributing millimeter wave radars in urban power stations, continuously observing distribution transformers, box-type substations and auxiliary structures thereof, and acquiring radar echo amplitude information And phase information Obtaining a phase difference between adjacent time points based on the phase information And pass through the formula Calculating micro deformation of target along radar sight direction Wherein The working wavelength of the millimeter wave radar; s2, microclimate information acquisition, namely synchronously acquiring microclimate information related to the running environment of the power station area to form an environment state vector changing along with time The microclimate information is used for representing the environmental conditions and the change conditions of the power station area; S3, preprocessing multisource monitoring data, namely respectively performing time synchronization, outlier rejection, noise suppression and data standardization processing on the millimeter wave radar monitoring data and microclimate information, wherein the data standardization processing is performed through a formula Implementation in which And Respectively representing the mean value and standard deviation of the corresponding features; S4, cross-mode feature construction, namely extracting structural response feature vectors from the preprocessed millimeter wave radar monitoring data Wherein Representing the rate of change of deformation over time; extracting environmental feature vectors from the preprocessed microclimate information Uniformly expressing the structural response feature vector and the environment feature vector to construct a cross-modal feature set ; S5, cross-modal fusion analysis, namely based on the cross-modal feature set, mapping functions are fused through cross-modal fusion Performing joint analysis on millimeter wave radar features and microclimate features, and establishing an association relationship between structural state changes and environmental conditions to obtain a fusion result ; S6, safety state evaluation and monitoring output, wherein the safety state evaluation and monitoring output is based on the fusion result By a state evaluation function Table construction safety state index And outputting corresponding monitoring results or risk assessment information when the safety state indexes meet preset abnormal criteria.
2. The method for cross-modal fusion monitoring of millimeter wave radar and microclimate information in an urban power station according to claim 1, wherein the microclimate information in step S2 includes ambient temperature Relative humidity of Rainfall amount Wind speed Wind direction Air pressure At least one of the environmental state vectors 。
3. The method for cross-modal fusion monitoring of millimeter wave radar and microclimate information of an urban power station according to claim 1, wherein the millimeter wave radar in step S1 adopts a non-contact observation mode to realize continuous and stable monitoring under different illumination and meteorological conditions.
4. The cross-modal fusion monitoring method for millimeter wave radar and microclimate information of an urban power station according to claim 1, wherein in step S3, abnormal value rejection adopts an abnormal recognition mechanism based on a 3 sigma criterion or a box graph method, and noise suppression adopts an adaptive filtering or wavelet denoising algorithm.
5. The method for cross-modal fusion monitoring of millimeter wave radar and microclimate information of urban power station according to claim 1, wherein the environmental feature vector in step S4 The method comprises the step of extracting statistical features of microclimate information, wherein the statistical features comprise at least one of mean value, variance, peak value, valley value and change rate.
6. The method for cross-modal fusion monitoring of millimeter wave radar and microclimate information of urban power station according to claim 1, wherein the cross-modal fusion mapping function in step S5 The method is realized by adopting a machine learning model, a statistical modeling method or a deep learning network, wherein the machine learning model comprises a support vector machine, a random forest or a gradient lifting tree, and the deep learning network comprises a convolutional neural network, a cyclic neural network or a concentration mechanism network.
7. The method for cross-modal fusion monitoring of millimeter wave radar and microclimate information of an urban power station according to claim 1, wherein the preset abnormal criteria in step S6 are obtained through training historical normal operation data, and the safety state indexes are obtained by training historical normal operation data The monitoring result or risk assessment information comprises an abnormality type, an abnormality position, a risk level and an intervention suggestion.

Description

Cross-modal fusion monitoring method for millimeter wave radar and microclimate information of urban power station Technical Field The invention belongs to the technical field of urban power distribution network operation safety monitoring and intelligent perception, and particularly relates to a cross-mode fusion monitoring method for millimeter wave radar and microclimate information of an urban power station. Background With the continuous promotion of the urban process, the scale of the urban power distribution network is continuously enlarged, and the urban power distribution area is used as a core basic unit directly facing to end users in the power distribution network, so that the number and the distribution density of the urban power distribution network are increasingly increased. Urban power stations usually cover distribution transformers, box-type substations, cable terminals, supporting structures and auxiliary building structures, are widely distributed in residential areas, commercial areas and along roads, face the practical problems of complex space environment and multiple peripheral interference factors, and the operation safety of the urban power stations is directly related to urban power supply reliability and public safety. Urban electric utility facilities are susceptible to the integration of multiple adverse factors during long-term operation. On one hand, the uneven settlement of the foundation, the ageing of the structure and the degradation of the material performance can cause the slow deformation and the stability reduction of the structure of the power station, and on the other hand, the periodical vibration, the peripheral construction disturbance, the heavy rainfall, the strong wind, the severe temperature change and other microclimate conditions caused by the traffic of vehicles can continuously act on the structural state and the running environment of the facilities of the power station. The factors are often expressed as micro deformation, abnormal vibration response or inclination accumulation of millimeter-level even smaller scale at the early stage, and the urban power supply system has the characteristics of strong concealment and uneasy detection, and if the factors cannot be recognized and intervened in time, the factors are easily gradually evolved into equipment instability, structural damage and even local collapse accidents, so that serious threat is caused to the safe and stable operation of the urban power supply system. Currently, safety monitoring of urban power stations mainly relies on manual inspection, electrical quantity monitoring and a small number of contact sensors. The manual inspection is limited by personnel experience and inspection period, continuous and real-time monitoring is difficult to realize, the identification capability of early anomalies such as micro deformation is limited, the electric quantity monitoring mainly reflects the electric operation state, the influence of structural deformation and environmental effect on facilities cannot be directly perceived, the contact sensor can acquire local physical quantity information, but has the problems of high layout cost and limited coverage range, is easily interfered by the environment or fails in complex urban scenes, and is difficult to meet the actual requirements of large-scale, long-term and fine monitoring. The millimeter wave radar is used as a non-contact sensing means, has the outstanding advantages of all-weather operation, strong illumination interference resistance, high deformation and vibration sensing precision and the like, and is gradually paid attention to in the field of structural health monitoring in recent years. However, in the application scene of the urban power station, the monitoring information acquired by the millimeter wave radar mainly reflects the geometric and motion characteristics of the target, the interference of complex environmental factors on the monitoring result is difficult to be independently represented, the multipath reflection in the urban environment is serious, the types of the target are various, and the single radar monitoring mode has obvious limitations in the aspects of anomaly identification and risk discrimination. In addition, microclimate conditions are important external factors affecting the operating environment of the power grid, and their changes have a significant effect on the facility structural states and monitoring signals. In the prior art, millimeter wave radar monitoring data and microclimate information are generally obtained and used in a scattered mode, an effective combined modeling and collaborative analysis means is lacked, and an internal association relation between environmental factors and structural response is difficult to comprehensively describe, so that accuracy and reliability of a monitoring result are limited. Therefore, there is a need for a monitoring method for urban power stations, which can e