CN-121984443-A - Component state detection and early warning method, device, equipment, storage medium and program product

CN121984443ACN 121984443 ACN121984443 ACN 121984443ACN-121984443-A

Abstract

The embodiment of the disclosure provides a component state detection and early warning method, device, equipment, storage medium and program product, the method comprises the steps of collecting motion state data of a target photovoltaic component, extracting structural state characteristic information from the motion state data, inputting a preset component characteristic identification model to extract information characteristics, further eliminating interference information, obtaining an elimination result, judging an initial early warning condition of the target photovoltaic component according to the elimination result, acquiring and adjusting the initial early warning condition according to a photovoltaic field structure diagram, obtaining an adjusted result, carrying out state early warning of the target photovoltaic component based on the adjusted result, determining a related influence component of the target photovoltaic component based on the photovoltaic field structure diagram and photovoltaic field real-time environment data, carrying out extension detection on the related influence component, and analyzing the extension detection result to carry out related early warning. The all-weather automatic monitoring of the wide-area dispersion photovoltaic module is realized, the early fault discovery period is greatly shortened, and the omission ratio is reduced.

Inventors

WEI HAILIANG
JIN ZHENGXIANG
ZHOU TONG
ZHU LIXIN
TENG WEIBO
WANG PENGFEI
YIN QINGLONG

Assignees

三峡新能源红寺堡发电有限公司
中国三峡新能源(集团)股份有限公司宁夏分公司
中国三峡新能源(集团)股份有限公司

Dates

Publication Date: 20260505
Application Date: 20251217

Claims (10)

1. The component state detection and early warning method is characterized by comprising the following steps: Collecting motion state data of a target photovoltaic module, extracting structural state feature information from the motion state data, inputting the structural state feature information into a preset module feature identification model to extract information features, and eliminating interference information in the structural state feature information based on the information features to obtain an elimination result; Judging the initial early warning condition of the target photovoltaic module according to the elimination result, acquiring and adjusting the initial early warning condition according to a photovoltaic field structure diagram to obtain an adjusted result, and carrying out state early warning on the target photovoltaic module based on the adjusted result; and determining a continuous influence component of the target photovoltaic component based on the photovoltaic field structure diagram and the photovoltaic field real-time environment data, carrying out extension detection on the continuous influence component, and analyzing an extension detection result to carry out continuous early warning.
2. The method of claim 1, wherein collecting motion state data of a target photovoltaic module, extracting structural state feature information from the motion state data, comprises: Transmitting a wave beam in a fixed period by utilizing a millimeter wave radar, receiving an echo reflected by the surface of a target photovoltaic module, and analyzing the accurate distance and relative speed information of each monitoring point of the target photovoltaic module according to a preset signal processing algorithm; Calculating the motion state data of the target photovoltaic module according to the spatial relation of a plurality of monitoring points, the accurate distance and the relative speed information, wherein the motion state data comprises three-dimensional coordinates, inclination angles and vibration frequency data; Matching the three-dimensional coordinates with the installation position reference coordinates, the rectangular outline boundary and the normal direction of the target photovoltaic module pre-stored in a database to obtain first information; Screening the vibration frequency data by utilizing the pre-stored vibration frequency in the database to obtain second information; performing time sequence analysis on the first information and the inclination angle, identifying and extracting the inclination angle with stable variation trend in a plurality of continuous sampling periods, and obtaining third information; And obtaining structural state characteristic information according to the first information, the second information and the third information.
3. The method of claim 1, wherein inputting the structural state feature information into a preset component feature recognition model to extract information features, removing interference information in the structural state feature information based on the information features, and obtaining a removal result comprises: Inputting a multidimensional feature vector in the structural state feature information into a preset component feature recognition model, wherein a reference feature range of a normal photovoltaic component is built in the preset component feature recognition model, and comparing the reference feature range with the multidimensional feature vector; and determining information features of the structural state feature information based on the comparison result, synchronously comparing the information features with an interference feature library in the preset component feature recognition model, and marking and eliminating the interference information in the structural state feature information according to the synchronous comparison result to obtain an elimination result.
4. The method of claim 1, wherein determining the initial pre-warning condition of the target photovoltaic module according to the exclusion result, obtaining and adjusting the initial pre-warning condition according to a photovoltaic field structure diagram, obtaining an adjusted result, and performing state pre-warning of the target photovoltaic module based on the adjusted result, comprises: Comparing the removal result with a preset three-level fault threshold library, judging the state of a target photovoltaic module and determining the initial early warning condition of the target photovoltaic module; acquiring a photovoltaic field structure diagram, determining a connecting component of the target photovoltaic component based on the photovoltaic field structure diagram, acquiring motion state data of the connecting component, and performing key analysis on the motion state data of the connecting component to obtain a key analysis result; And adjusting the initial early warning condition by utilizing the key analysis result to obtain an adjusted result, and carrying out early warning according to the adjusted result.
5. The method of claim 4, wherein adjusting the initial pre-warning condition using the key analysis results to obtain adjusted results and pre-warning based on the adjusted results comprises: If the fact that the connection assembly is abnormal in coordination is determined according to the key analysis result, adjusting the early warning property in the initial early warning condition to obtain a property adjustment result; If the connection assembly is determined to be abnormal according to the key analysis result, carrying out auxiliary description supplement on the initial early warning condition to obtain a supplement result; And obtaining an adjusted result according to the property adjustment result and the supplementary result, and carrying out early warning according to the adjusted result.
6. The method of claim 1, wherein determining the joint influence component of the target photovoltaic component based on the photovoltaic field structure map and photovoltaic field real-time environmental data, performing extension detection on the joint influence component, analyzing extension detection results, and performing joint early warning, comprises: analyzing the photovoltaic field structure diagram and the photovoltaic field real-time environment data, analyzing the photovoltaic field structure diagram to determine the physical connection relation of each photovoltaic module, and analyzing the photovoltaic field real-time environment data to determine the potential stress propagation path; determining a connection affecting component of the target photovoltaic component according to the physical connection relation and the potential stress propagation path; and carrying out extension detection on the associated influence components, evaluating the component state of each associated influence component according to an extension detection result and the preset three-level fault threshold library, marking the associated influence components as safe if no fault exists, and carrying out associated early warning on the associated influence components if abnormal states associated with the adjusted result of the target photovoltaic component exist in the associated influence components.
7. A device for component status detection and early warning, the device comprising: the acquisition module is used for acquiring the motion state data of the target photovoltaic module, extracting structural state feature information from the motion state data, inputting the structural state feature information into a preset module feature recognition model to extract information features, and eliminating interference information in the structural state feature information based on the information features to obtain an elimination result; the first early warning module is used for judging the initial early warning condition of the target photovoltaic module according to the elimination result, acquiring and adjusting the initial early warning condition according to a photovoltaic field structure diagram to obtain an adjusted result, and carrying out state early warning on the target photovoltaic module based on the adjusted result; and the second early warning module is used for determining the continuous influence component of the target photovoltaic component based on the photovoltaic field structure diagram and the photovoltaic field real-time environment data, carrying out extension detection on the continuous influence component, and analyzing the extension detection result to carry out continuous early warning.
8. An electronic device, comprising: A memory; processor, and A computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the component status detection and early warning method of any one of claims 1-6.
9. A computer readable storage medium, having stored thereon a computer program/instruction which, when executed by a processor, implements the steps of the method of any of claims 1-6.
10. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1-6.

Description

Component state detection and early warning method, device, equipment, storage medium and program product Technical Field The disclosure relates to the technical field of intelligent operation and maintenance, and in particular relates to a component state detection and early warning method, device, equipment, storage medium and program product. Background The current large-scale photovoltaic power station generally has the characteristics of wide occupied area, large number of components and distributed dispersion, and the configuration of on-site operation and maintenance personnel is usually only tens of people, so that all-weather and full-coverage manual inspection is difficult to realize. According to industry data statistics, in a traditional manual inspection mode, the detection period of early faults such as loosening and slight displacement of component bolts is longer, and the omission rate is higher. When the state of the component of the photovoltaic component changes, namely the bolt of the photovoltaic component loosens, high-frequency vibration can be generated under the action of wind force, so that abrasion and falling of other bolts are accelerated; when the quantity of the bolts falling off is higher, the photovoltaic module can swing greatly, and finally the module is overturned or falls off. The fault can not only cause the damage of a single component, but also damage a plurality of adjacent components, and even more serious, the fault can cause short circuit due to the pulling and breaking of component cables, so that the inverter is in failure and is stopped, the traditional inspection (manual inspection and unmanned aerial vehicle visible light inspection) is low in efficiency and long in period, and millimeter-level micro-displacement cannot be quantized, and the vibration sensor and other contact type monitoring devices are complex in installation and high in cost, and the health state perception of a large-area, high-precision and real-time component structure is difficult to realize. Disclosure of Invention To solve or at least partially solve the above technical problems, the present disclosure provides a method, an apparatus, a device, a storage medium, and a program product for detecting and pre-warning a component status. The embodiment of the disclosure provides a component state detection and early warning method, which comprises the following steps: Collecting motion state data of a target photovoltaic module, extracting structural state feature information from the motion state data, inputting the structural state feature information into a preset module feature identification model to extract information features, and eliminating interference information in the structural state feature information based on the information features to obtain an elimination result; Judging the initial early warning condition of the target photovoltaic module according to the elimination result, acquiring and adjusting the initial early warning condition according to a photovoltaic field structure diagram to obtain an adjusted result, and carrying out state early warning on the target photovoltaic module based on the adjusted result; and determining a continuous influence component of the target photovoltaic component based on the photovoltaic field structure diagram and the photovoltaic field real-time environment data, carrying out extension detection on the continuous influence component, and analyzing an extension detection result to carry out continuous early warning. The method provided by the embodiment of the disclosure acquires the motion state data of the target photovoltaic module, and extracts the structural state characteristic information from the motion state data, and comprises the following steps: Transmitting a wave beam in a fixed period by utilizing a millimeter wave radar, receiving an echo reflected by the surface of a target photovoltaic module, and analyzing the accurate distance and relative speed information of each monitoring point of the target photovoltaic module according to a preset signal processing algorithm; Calculating the motion state data of the target photovoltaic module according to the spatial relation of a plurality of monitoring points, the accurate distance and the relative speed information, wherein the motion state data comprises three-dimensional coordinates, inclination angles and vibration frequency data; Matching the three-dimensional coordinates with the installation position reference coordinates, the rectangular outline boundary and the normal direction of the target photovoltaic module pre-stored in a database to obtain first information; Screening the vibration frequency data by utilizing the pre-stored vibration frequency in the database to obtain second information; performing time sequence analysis on the first information and the inclination angle, identifying and extracting the inclination angle with stable variation trend in a pl