CN-122024452-A - Water conservancy pump station high-low voltage cabinet early warning method and system based on image detection

CN122024452ACN 122024452 ACN122024452 ACN 122024452ACN-122024452-A

Abstract

The invention discloses a water conservancy pump station high-low voltage cabinet early warning method and system based on image detection, and relates to the technical field of image detection, wherein the method comprises the steps of obtaining physical characteristic information of components in a cabinet, current running state information, structural geometric information and first thermal imaging image data of the high-low voltage cabinet; the method comprises the steps of determining a target space position of an abnormal signal in a high-low voltage cabinet according to structural geometric information, correcting first thermal imaging image data according to the target space position and physical characteristic information, and carrying out abnormality identification on the high-low voltage cabinet according to the physical characteristic information, current running state information and corrected first thermal imaging image data to generate abnormality early warning information. The method can be used for carrying out anomaly identification by combining the structural geometric information, the physical characteristic information and the thermal imaging image data so as to realize anomaly early warning and improve the early warning accuracy and the safety of the hydraulic pump station.

Inventors

LIU JINBAO
PEI JIE
ZHANG JIAN
GONG TIANTIAN
CHEN LI
ZHU ZHEYU
Lv you
LIU XIAOGUANG

Assignees

中苏科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260401

Claims (10)

1. The water conservancy pump station high-low voltage cabinet early warning method based on image detection is characterized by comprising the following steps of: Acquiring physical characteristic information of components in a cabinet, and current operation state information, structure geometric information and first thermal imaging image data of a high-low voltage cabinet, wherein the physical characteristic information comprises target thermal radiation emissivity, the current operation state information comprises current, voltage and power, and the structure geometric information comprises a three-dimensional shape, a three-dimensional space position and an insulating partition board layout; Determining the target space position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information; Correcting the first thermal imaging image data according to the target space position and the physical characteristic information; And carrying out abnormality identification on the high-low voltage cabinet according to the physical characteristic information, the current running state information and the corrected first thermal imaging image data to generate abnormality early warning information.
2. The method of claim 1, wherein the obtaining information of the physical characteristics of the internal components of the cabinet comprises: Collecting the environmental parameters in the cabinet, the operation load of a pump station, the optical characteristics of the components in the cabinet, the thickness of a surface oxide layer, the second thermal imaging image data and the material heat conductivity coefficient, wherein the environmental parameters in the cabinet comprise the temperature in the cabinet, the humidity in the cabinet and the density of corrosive gas, and the optical characteristics comprise the intensity and the polarization degree of reflected light; calculating a first emissivity based on the optical characteristic and the surface oxide thickness; calculating a second thermal emissivity based on the second thermographic image data and the material thermal conductivity; calculating a third heat radiation emissivity by using a preset mapping function according to the environmental parameters in the cabinet; And carrying out weighted average on the first heat radiation emissivity, the second heat radiation emissivity and the third heat radiation emissivity according to the operation load of the pump station to obtain the target heat radiation emissivity.
3. The method of claim 1, wherein determining the target spatial location of the anomaly signal in the high-low voltage cabinet based on the structural geometry information comprises: collecting visible light image data of components in the cabinet; denoising the visible light image data to inhibit reflection light spots; Performing image enhancement processing on the denoised visible light image data to enhance texture information; Extracting characteristic points from visible light image data after image enhancement processing, wherein the characteristic points comprise bolt edges or connecting piece corner points; And determining the target space position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information and the characteristic points.
4. A method according to claim 3, wherein determining the target spatial position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information and the feature point comprises: Monitoring deformation data of key connection points or easily deformable parts in the high-low voltage cabinet through a miniature displacement sensor; According to the deformation data, carrying out local adjustment on the structural geometric information; carrying out three-dimensional reconstruction on the characteristic points according to the structure geometric information after the local adjustment to obtain a three-dimensional reconstruction result; And determining the target space position of the abnormal signal in the high-low voltage cabinet according to the three-dimensional reconstruction result.
5. The method according to claim 4, wherein after performing three-dimensional reconstruction on the feature points according to the locally adjusted structural geometry information, the method further comprises: Performing deviation analysis on the three-dimensional reconstruction result and the structure geometric information subjected to local adjustment to obtain a deviation analysis result, wherein the deviation analysis result comprises spatial distribution and time continuity; identifying a point in time when a deviation occurs based on the spatial distribution and the temporal continuity; Acquiring parameter fluctuation data of the miniature displacement sensor corresponding to the deviation occurrence time point; Identifying deviation inducements according to the spatial distribution, the time continuity and the micro displacement sensor parameter fluctuation data; selecting a target adjustment strategy from a preset adjustment strategy library according to the deviation incentive; and adjusting three-dimensional reconstruction parameters according to the target adjustment strategy.
6. The method according to claim 1, wherein the performing abnormality recognition on the high-low voltage cabinet according to the physical characteristic information, the current operation state information and the corrected first thermal imaging image data to generate abnormality pre-warning information includes: acquiring physical parameters of components in the cabinet, wherein the physical parameters comprise component temperature parameters and vibration level parameters; determining the normal operation parameter range of the internal parts of the cabinet according to the current operation state information and the physical characteristic information; correcting the physical parameters according to the normal operation parameter range; Extracting features of the corrected physical parameters to obtain multidimensional features, wherein the multidimensional features comprise temperature rise rate and local vibration spectrum change; Identifying a fault type based on the multi-dimensional features and an early fault pattern, the early fault pattern including partial discharge, poor contact, and insulation aging; extracting a historical fault evolution trend corresponding to the fault type from a historical fault database; And carrying out abnormality identification on the high-low voltage cabinet according to the fault type, the historical fault evolution trend and the corrected first thermal imaging image data to generate abnormality early warning information.
7. The method of claim 6, wherein the identifying a fault type based on the multi-dimensional feature and an early failure mode comprises: Acquiring operation condition parameters and cabinet external environment parameters of a high-low voltage cabinet, wherein the operation condition parameters comprise load levels, and the cabinet external environment parameters comprise cabinet external temperature and cabinet external humidity; According to the operation condition parameters and the external environment parameters of the cabinet, adjusting a fuzzy membership function between the multi-dimensional features and the early fault mode, wherein the parameters of the fuzzy membership function comprise the shape and the threshold value of the fuzzy membership function, and the fuzzy membership function is used for reflecting the matching relation between different multi-dimensional features and the target fault mode; Calculating fuzzy matching degree between the multi-dimensional features and the early fault modes by using the fuzzy membership function, wherein the fuzzy matching degree is used for reflecting the degree of the multi-dimensional features belonging to each fault mode; and identifying the fault type according to the fuzzy matching degree.
8. The method of claim 6, wherein the feature extraction of the corrected physical parameters to obtain multi-dimensional features comprises: acquiring noise data in a high-low voltage cabinet and material properties of components in the cabinet, wherein the material properties comprise natural frequency and thermal expansion coefficient; According to the noise data and the corrected physical parameters, extracting statistical characteristics, frequency spectrum characteristics and energy distribution characteristics; calculating the local vibration spectrum change according to the natural frequency, the spectrum feature and the statistical feature; And calculating the temperature rise rate according to the thermal expansion coefficient, the energy distribution characteristic and the statistical characteristic.
9. The method of claim 8, wherein the extracting statistical, spectral and energy distribution features from the noise data and corrected physical parameters comprises: According to the noise data, adjusting parameters of a signal filter; filtering the corrected physical parameters by using the adjusted signal filter; performing time domain analysis on the physical parameters after the filtering treatment to obtain the statistical characteristics, wherein the statistical characteristics comprise mean value, variance, peak value and kurtosis; Carrying out frequency domain analysis on the physical parameters after the filtering treatment to obtain the frequency spectrum characteristics, wherein the frequency spectrum characteristics comprise main frequency, harmonic components and bandwidth; and carrying out time-frequency domain analysis on the physical parameters after the filtering treatment to obtain the energy distribution characteristics, wherein the energy distribution characteristics comprise an energy concentration area.
10. Water conservancy pump station high-low pressure cabinet early warning system based on image detection, its characterized in that includes: The information acquisition module is used for acquiring physical characteristic information of components in the cabinet, current operation state information of the high-low voltage cabinet, structural geometric information and first thermal imaging image data, wherein the physical characteristic information comprises target thermal radiation emissivity, the current operation state information comprises current, voltage and power, and the structural geometric information comprises a three-dimensional shape, a three-dimensional space position and an insulating partition board layout; The space position determining module is used for determining the target space position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information; a thermal imaging image data correction module for correcting the first thermal imaging image data according to the target spatial position and the physical characteristic information; And the abnormality early warning information generation module is used for carrying out abnormality recognition on the high-low voltage cabinet according to the physical characteristic information, the current running state information and the corrected first thermal imaging image data to generate abnormality early warning information.

Description

Water conservancy pump station high-low voltage cabinet early warning method and system based on image detection Technical Field The invention relates to the technical field of image detection, in particular to a water conservancy pump station high-low voltage cabinet early warning method and system based on image detection. Background The water conservancy pump station is used as an important infrastructure, and the stable operation of the high-low voltage power distribution cabinet in the water conservancy pump station is the core for guaranteeing the normal operation of the whole water conservancy system. In the long-term operation of the high-low voltage cabinets, due to the influence of various factors such as environmental temperature and humidity, power load change and the like, equipment components are easy to age, contact failure, partial discharge and other abnormal conditions. Traditional modes of manual inspection and periodic maintenance are often difficult to find the potential risks in real time, equipment is easy to suddenly fail, the operation efficiency and the overall safety of a pump station are further affected, and the fault early warning accuracy is low. In summary, the technical problems in the related art are to be improved. Disclosure of Invention The embodiment of the invention mainly aims to provide a water conservancy pump station high-low voltage cabinet early warning method and system based on image detection, which can be used for carrying out abnormality recognition by combining structural geometric information, physical characteristic information and thermal imaging image data so as to realize abnormality early warning and improve early warning accuracy and safety of the water conservancy pump station. On the one hand, the embodiment of the invention provides a water conservancy pump station high-low voltage cabinet early warning method based on image detection, which comprises the following steps: Acquiring physical characteristic information of components in a cabinet, and current operation state information, structure geometric information and first thermal imaging image data of a high-low voltage cabinet, wherein the physical characteristic information comprises target thermal radiation emissivity, the current operation state information comprises current, voltage and power, and the structure geometric information comprises a three-dimensional shape, a three-dimensional space position and an insulating partition board layout; Determining the target space position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information; Correcting the first thermal imaging image data according to the target space position and the physical characteristic information; And carrying out abnormality identification on the high-low voltage cabinet according to the physical characteristic information, the current running state information and the corrected first thermal imaging image data to generate abnormality early warning information. On the other hand, the embodiment of the invention provides a water conservancy pump station high-low voltage cabinet early warning system based on image detection, which comprises: The information acquisition module is used for acquiring physical characteristic information of components in the cabinet, current operation state information of the high-low voltage cabinet, structural geometric information and first thermal imaging image data, wherein the physical characteristic information comprises target thermal radiation emissivity, the current operation state information comprises current, voltage and power, and the structural geometric information comprises a three-dimensional shape, a three-dimensional space position and an insulating partition board layout; The space position determining module is used for determining the target space position of the abnormal signal in the high-low voltage cabinet according to the structural geometric information; a thermal imaging image data correction module for correcting the first thermal imaging image data according to the target spatial position and the physical characteristic information; And the abnormality early warning information generation module is used for carrying out abnormality recognition on the high-low voltage cabinet according to the physical characteristic information, the current running state information and the corrected first thermal imaging image data to generate abnormality early warning information. The embodiment of the application has the advantages that the physical characteristic information of the components in the cabinet and the current operation state information, the structural geometric information and the first thermal imaging image data of the high-low voltage cabinet are firstly obtained, then the target space position of an abnormal signal in the high-low voltage cabinet is determined according to the structural geometric informatio