CN-121997068-A - Intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle patrols and examines

CN121997068ACN 121997068 ACN121997068 ACN 121997068ACN-121997068-A

Abstract

The invention discloses an intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection, which relates to the technical field of photovoltaic power station operation and maintenance detection and fault diagnosis, and is characterized in that a multi-frame sequence is constructed on the surface area of the same module, pixel-level time sequence dynamics is calculated, time-varying light shadows are distinguished from candidate defects, a dynamic noise mask and a stable feature mask are generated, the follow-up diagnosis is preferentially based on stable area information, and the probability that the dynamic light shadows are misjudged as defects is reduced; the infrared temperature rise characteristic and the visible light appearance morphological characteristic are jointly extracted in the stable characteristic mask, and the thermal abnormality evidence and the appearance abnormality evidence are put into the same diagnosis chain for mutual evidence, so that the direct triggering of an error conclusion caused by the evidence distortion of a single mode under the strong interference condition is avoided.

Inventors

Yan Jiongjiong
WEI ZHAOCHENG
LIU JUN
LIU DONGDONG
ZHANG MENGYANG
SONG CHANGCHENG
CHEN SHICONG
LI LONG
YANG FAN

Assignees

淮南市国家电投新能源有限公司

Dates

Publication Date: 20260508
Application Date: 20260120

Claims (10)

1. Intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle patrols and examines, its characterized in that includes: the unmanned plane is provided with a thermal infrared imager and a visible light camera; A communication unit; an analysis server receiving infrared image data and visible light image data by means of the communication unit; the analysis server is configured to register a plurality of frames of infrared images and a plurality of frames of visible images obtained at continuous time points of the same photovoltaic module surface area and construct an image sequence; calculating a pixel level time sequence dynamic index based on the image sequence and generating a dynamic noise mask and a stable characteristic mask; Extracting temperature rise characteristics and appearance morphological characteristics from a stable characteristic mask, operating an electrothermal coupling model based on electric branch working electric parameters and environment parameters corresponding to the surface area to obtain theoretical thermal distribution, carrying out consistency scoring on the theoretical thermal distribution, the temperature rise characteristics and the appearance morphological characteristics, and outputting fault diagnosis states of the surface area according to the consistency scoring.
2. The intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection according to claim 1, wherein the image sequence at least comprises 3 frames, and the acquisition time interval of two adjacent frames is smaller than a preset time window, so that dynamic light and shadow interference such as reflection and reflection show observable time-varying property in the sequence.
3. The unmanned aerial vehicle inspection-based intelligent station photovoltaic module fault diagnosis system according to claim 1, wherein the pixel-level time sequence dynamic index comprises at least one of a mean value, a variance or a weighted combination of pixel intensity differences corresponding to adjacent frames after registration.
4. The intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection according to claim 1, wherein the stable feature mask is obtained by comparing the pixel-level time sequence dynamic index with a threshold value, and the threshold value is adaptively determined according to the statistical distribution of the time sequence dynamic index in the surface area of the photovoltaic module.
5. The unmanned aerial vehicle inspection-based intelligent station photovoltaic module fault diagnosis system according to claim 1 or 4, wherein the registration comprises calculating an inter-frame geometric transformation based on feature point matching or optical flow estimation, and performing cross-modal alignment on an infrared image and a visible light image.
6. The intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection according to claim 1, wherein the temperature rise characteristic comprises at least one of a temperature rise amplitude, a temperature rise connected domain area or a temperature gradient relative to a surrounding background; The appearance morphology features include at least one of a bright spot morphology, an edge/crack candidate texture, or a stain blocking candidate texture.
7. The unmanned aerial vehicle inspection-based intelligent station photovoltaic module fault diagnosis system of claim 1, wherein the operating electrical parameters comprise at least one of current and voltage, the environmental parameters comprise at least one of ambient temperature, wind speed or irradiance, and are provided by a station monitoring system or a sensor system.
8. The intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection according to claim 1, wherein the consistency score is obtained by fusing a first check factor and a second check factor, wherein the first check factor represents the matching degree of the temperature rise characteristic and the theoretical heat distribution, and the second check factor represents the matching degree of the appearance morphological characteristic and a pre-stored defect morphological characteristic library.
9. The intelligent station photovoltaic module failure diagnosis system based on unmanned aerial vehicle inspection according to claim 1, wherein the analysis server is further configured to issue a repeat shooting instruction to the unmanned aerial vehicle when the consistency score is in a suspected section, so that the unmanned aerial vehicle changes at least one of shooting angle, exposure or focal length to the suspected section to acquire a supplementary image.
10. The unmanned aerial vehicle inspection-based intelligent station photovoltaic module fault diagnosis system according to claim 9, wherein the repeated shooting instruction further comprises a step of changing a polarization filtering condition or changing an incident view angle range, so as to reduce highlight interference caused by specular reflection.

Description

Intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle patrols and examines Technical Field The invention relates to the technical field of operation and maintenance detection and fault diagnosis of photovoltaic power generation stations, in particular to an intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection. Background At present, an unmanned aerial vehicle is utilized to develop an automatic inspection scheme in the operation and maintenance of a photovoltaic power station, wherein the unmanned aerial vehicle is provided with a radiation type infrared thermal imager and a high-resolution visible light camera to acquire component images, and the thermal anomalies and appearance anomalies are identified and positioned by combining methods such as deep learning, wherein the thermal infrared images are favorable for finding out abnormal isothermal lifting characteristics related to hot spots, substring/bypass diodes, the visible light images are more suitable for recording appearance defects such as glass breakage, delamination, shielding and severe dust accumulation/bird droppings, and the like, and the existing acquisition process generally needs to restrict the flying height, the overlapping degree and the camera view angle to reduce the influence of reflection glare on the surface of the component on imaging in order to ensure the data quality. However, when the technical system is applied to a water surface/floating type photovoltaic field station, because the accessibility of the array is limited, the inspection is more prone to relying on aerial modes such as unmanned aerial vehicles, and meanwhile, the field environment presents stronger operation and maintenance characteristics such as moisture, corrosion and bird droppings, solar flare (sunglint) and sky reflection which changes along with wind waves are easy to occur on the water surface on an imaging level, the highlight area and texture can change rapidly along with time, and moreover, the fluctuation of an observation visual angle is caused by slight fluctuation of a floating body, so that irregular highlight spots, local overexposure or reflection texture are easier to occur in a visible light image, and continuous interference is caused to defect identification based on the texture/brightness characteristics. Aiming at complex illumination interference, strategies such as multi-mode fusion of infrared light and visible light and screening/de-duplication by utilizing multi-frame consistency are proposed to improve robustness, but false detection and omission detection can still occur under the condition that strong reflection and dynamic light shadow are highly coincident with the physical position of a component, so that defect identification confidence coefficient is reduced, and reliability of a diagnosis conclusion is affected. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. The invention provides an intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection, which solves the problem that the defect identification misjudgment confidence coefficient is seriously reduced due to the photovoltaic reflection dynamic noise of a water surface. In order to solve the technical problems, the invention provides the following technical scheme: The embodiment of the invention provides an intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection, which comprises the following steps: the unmanned plane is provided with a thermal infrared imager and a visible light camera; A communication unit; an analysis server receiving infrared image data and visible light image data by means of the communication unit; the analysis server is configured to register a plurality of frames of infrared images and a plurality of frames of visible images obtained at continuous time points of the same photovoltaic module surface area and construct an image sequence; calculating a pixel level time sequence dynamic index based on the image sequence and generating a dynamic noise mask and a stable characteristic mask; Extracting temperature rise characteristics and appearance morphological characteristics from a stable characteristic mask, operating an electrothermal coupling model based on electric branch working electric parameters and environment parameters corresponding to the surface area to obtain theoretical thermal distribution, carrying out consistency scoring on the theoretical thermal distribution, the temperature rise characteristics and the appearance morphological characteristics, and outputting fault diagnosis states of the surface area according to the consistency scoring. The intelligent station photovoltaic module fault diagnosis system based on unmanned aerial vehicle inspection is a preferab