A photovoltaic power station inspection management system and method based on an internet of things

Abstract

This invention discloses an IoT-based photovoltaic power plant inspection and management system and method, relating to the field of photovoltaic system monitoring. The system includes: a distributed sensing module that collects real-time operating data of each photovoltaic unit through multiple types of sensors, including current values ​​I. i (t), voltage value U i (t), temperature T i (t) and ambient light intensity L(t); dynamic decision-making module, used to calculate the comprehensive health index H of photovoltaic unit. i (t), and generate dynamic operation and maintenance priority P. i (t); UAV scheduling module, based on P i (t) Generate three-dimensional path planning parameters. This invention achieves full-chain optimization of photovoltaic inspection management through mechanism fusion and algorithm innovation. The comprehensive health index model breaks through the limitations of a single threshold criterion and realizes early fault prediction through dynamic weights and differential operations. Secondly, the synergistic mechanism of adaptive light threshold and nonlinear historical correction significantly improves the data reliability in complex environments.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic system monitoring, and more specifically, relates to an Internet of Things (IoT)-based photovoltaic power plant inspection and management system. Furthermore, this invention also relates to an IoT-based photovoltaic power plant inspection and management method. Background Technology

[0002] Efficient inspection of photovoltaic power plants is a core element in ensuring power generation efficiency. For example, Chinese patent CN117519291A discloses a photovoltaic panel inspection system based on multi-UAV path planning. This system divides regular and irregular areas through panoramic UAV image segmentation and allocates multi-UAV tasks based on power optimization. It adopts a coverage shooting model and an S-shaped inspection path to allocate flight tasks. However, this technology has significant limitations: First, the path planning relies solely on geometric topology division, ignoring the dynamic evolution characteristics of the photovoltaic module's health status. Second, the power management uses a static allocation model, failing to consider abnormal power consumption fluctuations caused by increased impedance in aging modules. In multi-UAV collaborative scenarios, this can easily lead to inaccurate power estimation, resulting in UAV task interruptions. Third, temperature correction relies on fixed environmental thresholds and does not establish a coupling model between light intensity and temperature rise rate. Under low light conditions, normal thermal inertial fluctuations are easily misjudged as faults, leading to a high redundant alarm rate.

[0003] Specifically, the following technical bottlenecks exist:

[0004] 1) The health assessment has a single dimension; Chinese patent CN117519291A discloses a photovoltaic panel inspection system based on multi-UAV path planning, which identifies foreign object obstruction through image analysis, but lacks real-time monitoring of current and voltage parameters and dynamic compensation for aging factors, and cannot quantify the degree of component degradation; for example, when the photovoltaic panel has abnormal output current due to micro-cracks, the traditional solution cannot detect it in time through grayscale image changes.

[0005] 2) The path planning is disconnected from the component status; the S-shaped route and reverse check method used only use battery power as a constraint and do not include the spatiotemporal gradient parameters of the health index in the path weight calculation, resulting in insufficient coverage of high-risk areas.

[0006] 3) Environmental parameter solidification correction; This scheme adjusts the shooting area by predefined tilt angle, but does not introduce light adaptive threshold and time-related aging correction factor. During dawn and dusk or in cloudy weather, the static model has difficulty suppressing temperature sensor noise, and the misjudgment rate increases.

[0007] Based on the above, we propose an IoT-based photovoltaic power plant inspection management system and method to address the problems existing in the current technology. Summary of the Invention

[0008] The purpose of this invention is to address the shortcomings of existing technologies by proposing an Internet of Things-based photovoltaic power plant inspection management system and method. Through mechanism integration and algorithm innovation, the entire chain of photovoltaic inspection management is optimized.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] An IoT-based photovoltaic power plant inspection and management system includes:

[0011] The distributed sensing module collects real-time operating data from each photovoltaic unit through multiple types of sensors, including current value I. i (t), voltage value U i (t), temperature T i (t) and ambient light intensity L(t);

[0012] The dynamic decision-making module is used to calculate the comprehensive health index H of the photovoltaic unit. i (t), and generate dynamic operation and maintenance priorities, and calculate the comprehensive health index H of the photovoltaic unit. i The formula for calculating (t) is:

[0013] ;

[0014] In the formula, α and β are weighting coefficients, satisfying α 2 +β 2 =1, and α=ρ·cos(θ), where ρ is the photovoltaic material degradation factor and θ is the installation tilt angle; γ=0.05·ln(t) is the time-related aging correction factor; ΔT i (t) represents the cumulative historical temperature difference;

[0015] The generation of dynamic operation and maintenance priority P i (t), the calculation formula is:

[0016] Where dt is the differential increment of time, and dH i (t) is the derivative of the health index; when P i (t) > P thre Time-marked fault dynamic fault determination threshold, P thre The threshold for dynamic fault determination;

[0017] The drone scheduling module, based on P i (t) Generate 3D path planning parameters, expressed as:

[0018] Among them, ∇ x P i (t) and ∇ y P i(t) represents the fault propagation gradient in the x-direction and the environmental coupling gradient in the y-direction, respectively.

[0019] Preferably, the calculation of the historical cumulative temperature difference of the comprehensive health index in the dynamic decision-making module adopts a nonlinear decay correction, and the expression is:

[0020] Where τ is the time step of historical data, the tanh function is used to suppress temperature fluctuation interference under low light conditions, and T avg L is the dynamic temperature reference value. thre The threshold is set to adapt to ambient light.

[0021] The dynamic temperature reference value T avg The calculation formula is:

[0022] ;

[0023] In the formula, ( ) represents the time decay weight, τ represents the time interval, and N represents the statistical period. Unlike the traditional static temperature threshold, the dynamic benchmark value can reflect the historical inertia and environmental adaptability of the photovoltaic panel's thermal behavior.

[0024] Preferably, the illumination adaptive threshold L thre The critical light intensity used to distinguish between normal operation and abnormal lighting conditions is expressed as follows:

[0025] ;

[0026] In the formula, μL(t) is the historical average sunshine duration for the current season, σL(t) is the standard deviation of sunshine intensity, S(t) is the cloud cover rate over the past 3 days, and S seasonal The baseline seasonal cloud cover rate.

[0027] Preferably, when the UAV scheduling module performs three-dimensional path planning, it prioritizes the fault propagation gradient ∇ in the x-direction. x P i (t) is normalized and weighted with the ambient light intensity attenuation rate, and when the dynamic fault determination threshold P thre The calibration mode is automatically executed when the number of triggers exceeds the preset period;

[0028] The value of the photovoltaic material attenuation factor ρ is determined in real time by a spectral analysis device to measure the light transmittance of the glass cover, and an exponential correlation mapping table is constructed in conjunction with the installation tilt angle θ for dynamic loading.

[0029] Preferably, the dynamic temperature reference value T avgThe value of the statistical period N is adjusted in segments according to the installation tilt angle θ of the photovoltaic module. When the tilt angle θ > 30°, N dynamically shortens the statistical window with the seasonal light distribution.

[0030] The illumination adaptive threshold L thre The baseline seasonal cloud cover S seasonal The moving average is iteratively updated using a historical meteorological database, and measured radiation values ​​are introduced to provide feedback compensation for μL(t) during cloudy weather.

[0031] A photovoltaic power plant inspection and management method based on the Internet of Things (IoT), the method being implemented using the aforementioned IoT-based photovoltaic power plant inspection and management system, includes the following steps:

[0032] S1. Distributed data acquisition and dynamic benchmark calibration: This involves periodically collecting operating parameters of each unit through sensor nodes deployed in the photovoltaic matrix, including:

[0033] Current value I i (t), voltage value U i (t), temperature T i (t) and ambient light intensity L(t);

[0034] S2, Dynamic Health Assessment and Operation and Maintenance Decision Generation: First, a comprehensive health index is generated, then a dynamic priority is generated through differential calculation, and finally, it is updated according to the current seasonal cloud coverage.

[0035] S3. Perform three-dimensional path planning based on operation and maintenance decisions, and at the same time, control the drone for adaptive inspection through the drone scheduling module.

[0036] Preferably, in step S1, the following process is performed:

[0037] S11, regarding the temperature data T i (t) is corrected for nonlinear decay, and the expression is:

[0038] ;

[0039] In the formula, τ is the time step of historical data, and T avg L is the dynamic temperature reference value. thre The illumination adaptive threshold is used; the tanh function is used to suppress light intensity L(t) < L thre Ineffective temperature fluctuations at that time;

[0040] S12. Update the photovoltaic material attenuation factor ρ based on the real-time measured light transmittance of the glass cover, and correct the weighting coefficient α according to ρ·cosθ.

[0041] Preferably, in step S2, the dynamic decision-making module is invoked to perform the following process:

[0042] S21. Calculate the comprehensive health index:

[0043] ;

[0044] In the formula, U rated For the rated voltage, T max γ represents the upper limit of the component's temperature resistance; γ is the time aging factor.

[0045] S22. Generating dynamic priority through differential operations:

[0046] ;

[0047] When P i (t) > P thre The drone scheduling is triggered in a timely manner, and the number of triggers is recorded for calibration mode determination;

[0048] S23. Update based on the current seasonal cloud cover rate S(t):

[0049] ;

[0050] Where, ΔS = |S(t) - S seasonal |,S seasonal Updated using a moving average from the meteorological database.

[0051] Preferably, in step S3, the UAV scheduling module performs the following process:

[0052] S31. Constructing gradient field parameters:

[0053] ;

[0054] In the formula, The ambient light intensity attenuation rate, The rate of change of light intensity in the x-direction;

[0055] S32. When the installation tilt angle θ > 30°, shorten the dynamic temperature reference T. avg Statistical period:

[0056] ;

[0057] S32. Trigger calibration mode according to preset cycle: If P within Δt thre If the number of triggers is greater than 3 times per cycle, the mapping table will be recalibrated.

[0058] Technical effects and advantages of the present invention: The photovoltaic power station inspection management system and method based on the Internet of Things provided by the present invention have the following effects compared with the prior art:

[0059] Dynamic multi-dimensional health assessment improves the accuracy of operation and maintenance decisions. The dynamic decision-making module of this application constructs a comprehensive health index through four-dimensional parameters of current, voltage, temperature and light intensity, and introduces time-related aging correction factors and material degradation factors to achieve fine quantification of the photovoltaic unit status. In traditional methods, health assessment relies only on a single temperature threshold. However, this application incorporates the photovoltaic panel installation tilt angle and material aging rate into the calculation model by dynamically adjusting the nonlinear weight coefficients α and β. When the module causes local dust accumulation due to excessive tilt angle, the system automatically reduces the temperature weight β and increases the current weight α to avoid misjudgment caused by light intensity fluctuations. Through differential operation and dynamic priority, the instantaneous change trend of the health index is captured to achieve early fault warning.

[0060] Secondly, by constructing a nonlinear correction equation, normal temperature rise and abnormal fluctuations can be effectively distinguished. Traditional linear weighting methods are difficult to suppress temperature noise under low light intensity. However, this application utilizes the saturation characteristics of the tanh function to exponentially decay small temperature fluctuations. In cloudy and rainy weather, short-term temperature rises caused by sudden changes in cloud coverage will be judged as invalid data to avoid triggering error alarms. At the same time, the dynamic temperature reference value is calculated through time decay weights to reflect the thermal inertia characteristics of the photovoltaic panel.

[0061] Finally, the drone scheduling module generates dynamic tracks by constructing a fault propagation gradient field and combining it with the ambient light intensity attenuation rate. Traditional path planning only considers the shortest geometric distance, while this application achieves priority coverage of high-risk areas by normalizing the weighted light intensity and fault gradient parameters. When a photovoltaic unit experiences a sharp drop in its comprehensive health index due to the hot spot effect, the gradient field will enhance the drone inspection density in that area. At the same time, the design of the dynamic statistical cycle enables large-tilt modules to obtain a higher inspection frequency during the high-temperature season. Attached Figure Description

[0062] Figure 1 This is an architecture diagram of the photovoltaic power plant inspection and management system based on the Internet of Things of this invention;

[0063] Figure 2 This is a flowchart of the photovoltaic power plant inspection and management method based on the Internet of Things of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0065] This invention provides an IoT-based photovoltaic power plant inspection management system and method. Through mechanism integration and algorithm innovation, it achieves full-chain optimization of photovoltaic inspection management. First, the comprehensive health index model breaks through the limitations of a single threshold criterion, realizing early fault prediction through dynamic weights and differential operations. Second, the synergistic mechanism of adaptive light threshold and nonlinear historical correction significantly improves data reliability in complex environments. Finally, the gradient-driven path planning technology deeply binds UAV resource allocation with dynamic changes in the light field, solving the resource waste problem of traditional static paths.

[0066] like Figure 1 As shown, the architecture of the IoT-based photovoltaic power plant inspection and management system includes:

[0067] The distributed sensing module collects real-time operating data from each photovoltaic unit through multiple types of sensors, including current value I. i (t), voltage value U i (t), temperature T i (t) and ambient light intensity L(t);

[0068] The dynamic decision-making module is used to calculate the comprehensive health index H of the photovoltaic unit. i (t), and generate dynamic operation and maintenance priorities, and calculate the comprehensive health index H of the photovoltaic unit. i The formula for calculating (t) is:

[0069] ;

[0070] In the formula, α and β are weighting coefficients, satisfying α 2 +β 2 =1, and α=ρ·cos(θ), where ρ is the photovoltaic material degradation factor and θ is the installation tilt angle; γ=0.05·ln(t) is the time-related aging correction factor; ΔT i (t) represents the cumulative historical temperature difference;

[0071] Generate dynamic operation and maintenance priority P i (t), the calculation formula is:

[0072] Where dt is the differential increment of time, and dH i (t) is the derivative of the health index; when P i (t) > P thre Time-marked fault dynamic fault determination threshold, P thre The threshold for dynamic fault determination;

[0073] Furthermore, the calculation of the historical cumulative temperature difference of the comprehensive health index in the dynamic decision-making module adopts a non-linear decay correction, the expression of which is:

[0074] Where τ is the time step of historical data, the tanh function is used to suppress temperature fluctuation interference under low light conditions, and T avg L is the dynamic temperature reference value. thre The illumination adaptive threshold; Illumination adaptive threshold L thre The critical light intensity used to distinguish between normal operation and abnormal lighting conditions is expressed as follows:

[0075] ;

[0076] In the formula, μL(t) is the historical average sunshine duration for the current season, σL(t) is the standard deviation of sunshine intensity, S(t) is the cloud cover rate over the past 3 days, and S seasonal Based on seasonal cloud cover;

[0077] Dynamic temperature reference value T avg The calculation formula is:

[0078] ;

[0079] In the formula, ( () represents the time decay weight, τ represents the time interval, and N represents the statistical period. Unlike traditional static temperature thresholds, dynamic benchmark values ​​can reflect the historical inertia and environmental adaptability of photovoltaic panel thermal behavior. The dynamic temperature benchmark value T... avg The value of the statistical period N is adjusted in segments according to the installation tilt angle θ of the photovoltaic module. When the tilt angle θ > 30°, N dynamically shortens the statistical window with the seasonal light distribution.

[0080] Illumination adaptive threshold L thre The baseline seasonal cloud cover S seasonal The moving average is iteratively updated using a historical meteorological database, and measured radiation values ​​are introduced to provide feedback compensation for μL(t) during cloudy weather.

[0081] The drone scheduling module, based on P i (t) Generate 3D path planning parameters, expressed as:

[0082] Among them, ∇ x P i (t) and ∇ y P i (t) represents the fault propagation gradient in the x-direction and the environmental coupling gradient in the y-direction, respectively; when the UAV scheduling module performs 3D path planning, it prioritizes the fault propagation gradient in the x-direction ∇. x P i (t) is normalized and weighted with the ambient light intensity attenuation rate, and when the dynamic fault determination threshold P is... threThe calibration mode is automatically executed when the number of triggers exceeds the preset period;

[0083] The value of the photovoltaic material attenuation factor ρ is determined in real time by measuring the light transmittance of the glass cover using a spectral analysis device, and dynamically loaded by constructing an exponential correlation mapping table in conjunction with the installation tilt angle θ.

[0084] This embodiment also proposes an IoT-based photovoltaic power plant inspection and management method, which is implemented using the aforementioned IoT-based photovoltaic power plant inspection and management system, such as... Figure 2 As shown, it includes the following steps:

[0085] S1. Distributed data acquisition and dynamic benchmark calibration: This involves periodically collecting operating parameters of each unit through sensor nodes deployed in the photovoltaic matrix, including:

[0086] Current value I i (t), voltage value U i (t), temperature T i (t) and ambient light intensity L(t);

[0087] In step S1, the following process is performed:

[0088] S11, Regarding temperature data T i (t) is corrected for nonlinear decay, and the expression is:

[0089] ;

[0090] In the formula, τ is the time step of historical data, and T avg L is the dynamic temperature reference value. thre The illumination adaptive threshold is used; the tanh function is used to suppress light intensity L(t) < L thre Ineffective temperature fluctuations at that time;

[0091] S12. Update the photovoltaic material attenuation factor ρ based on the real-time measured light transmittance of the glass cover, and correct the weighting coefficient α according to ρ·cosθ.

[0092] S2, Dynamic Health Assessment and Operation and Maintenance Decision Generation: First, a comprehensive health index is generated, then a dynamic priority is generated through differential calculation, and finally, it is updated according to the current seasonal cloud coverage.

[0093] Specifically, in step S2, the dynamic decision-making module is invoked to execute the following process:

[0094] S21. Calculate the overall health index:

[0095] ;

[0096] In the formula, U rated For the rated voltage, T maxγ represents the upper limit of the component's temperature resistance; γ is the time aging factor.

[0097] S22. Generating dynamic priority through differential operations:

[0098] ;

[0099] When P i (t) > P thre The drone scheduling is triggered in a timely manner, and the number of triggers is recorded for calibration mode determination;

[0100] S23. Update based on the current seasonal cloud cover rate S(t):

[0101] ;

[0102] Where, ΔS = |S(t) - S seasonal |,S seasonal Updated using a moving average from the meteorological database.

[0103] S3. Perform three-dimensional path planning based on operation and maintenance decisions, and at the same time, control the drone for adaptive inspection through the drone scheduling module;

[0104] It should be noted that in step S3, the UAV scheduling module performs the following process:

[0105] S31. Constructing gradient field parameters:

[0106] ;

[0107] In the formula, The ambient light intensity attenuation rate, The rate of change of light intensity in the x-direction;

[0108] S32. When the installation tilt angle θ > 30°, shorten the dynamic temperature reference T. avg Statistical period:

[0109] ;

[0110] S32. Trigger calibration mode according to preset cycle: If P within Δt thre If the number of triggers is greater than 3 times per cycle, the mapping table will be recalibrated.

[0111] In summary, compared with the prior art, the present invention has the following advantages:

[0112] Dynamic multi-dimensional health assessment improves the accuracy of operation and maintenance decisions. The dynamic decision-making module of this application constructs a comprehensive health index through four-dimensional parameters of current, voltage, temperature and light intensity, and introduces time-related aging correction factors and material degradation factors to achieve fine quantification of the photovoltaic unit status. In traditional methods, health assessment relies only on a single temperature threshold. However, this application incorporates the photovoltaic panel installation tilt angle and material aging rate into the calculation model by dynamically adjusting the nonlinear weight coefficients α and β. When the module causes local dust accumulation due to excessive tilt angle, the system automatically reduces the temperature weight β and increases the current weight α to avoid misjudgment caused by light intensity fluctuations. Through differential operation and dynamic priority, the instantaneous change trend of the health index is captured to achieve early fault warning.

[0113] Secondly, by constructing a nonlinear correction equation, normal temperature rise and abnormal fluctuations can be effectively distinguished. Traditional linear weighting methods are difficult to suppress temperature noise under low light intensity. However, this application utilizes the saturation characteristics of the tanh function to exponentially decay small temperature fluctuations. In cloudy and rainy weather, short-term temperature rises caused by sudden changes in cloud coverage will be judged as invalid data to avoid triggering error alarms. At the same time, the dynamic temperature reference value is calculated through time decay weights to reflect the thermal inertia characteristics of the photovoltaic panel.

[0114] Finally, the drone scheduling module generates dynamic tracks by constructing a fault propagation gradient field and combining it with the ambient light intensity attenuation rate. Traditional path planning only considers the shortest geometric distance, while this application achieves priority coverage of high-risk areas by normalizing the weighted light intensity and fault gradient parameters. When a photovoltaic unit experiences a sharp drop in its comprehensive health index due to the hot spot effect, the gradient field will enhance the drone inspection density in that area. At the same time, the design of the dynamic statistical cycle enables large-tilt modules to obtain a higher inspection frequency during the high-temperature season.

[0115] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An Internet of Things-based photovoltaic power station inspection management system, characterized in that, The application relates to a UAV scheduling module for generating three-dimensional path planning parameters. A distributed sensing module collects real-time operation data of each photovoltaic unit through multiple types of sensors, including current value I i (t), voltage value U i (t), temperature T i (t), and ambient light intensity L(t). The history temperature difference accumulation amount calculation of the comprehensive health index in the dynamic decision module adopts nonlinear attenuation correction, and the expression is as follows: A dynamic decision module is configured to calculate a comprehensive health index H of the photovoltaic unit i (t), and generate a dynamic operation and maintenance priority i The calculation formula of the comprehensive health index H (t) is: ； In the formula, a, β are weight coefficients, satisfying a 2 + β 2 = 1, and a = p cos (θ), p is a decay factor of photovoltaic material, and θ is an installation inclination angle; γ = 0.05 ln (t) is a time-dependent aging correction factor; ΔT i (t) is a historical temperature difference cumulative amount; Generating dynamic operation and maintenance priority P i (t), the calculation formula is: where dt is a differential increment of time, dH i (t) is a differential of the health index; when P i (t) > P thre a dynamic failure determination threshold, P thre is a dynamic failure determination threshold; The value of the photovoltaic material attenuation factor p is determined by real-time measurement of the glass cover plate light transmittance through a spectrum analysis device, and an index type correlation mapping table is constructed by combining the installation inclination angle theta to realize dynamic loading. ; where τ is the time step of historical data, the tanh function is used to suppress the temperature fluctuation interference at low light, T avg is the dynamic temperature reference value, L thre is the light adaptive threshold value; the light adaptive threshold value L thre , the critical light intensity for distinguishing normal operation from light abnormal state, the value expression is: ； In the formula, μL(t) is the current seasonal historical light mean value, σL(t) is the light intensity standard deviation, S(t) is the cloud cover in the past 3 days, S seasonal is the reference seasonal cloud cover. Dynamic temperature reference value T avg The calculation formula is: ； wherein is a time decay weight, τ is a time interval, and N is a statistical period. Unlike the conventional static temperature threshold, the dynamic reference value can reflect the historical inertia and environmental adaptability of the thermal behavior of the photovoltaic panel. The dynamic temperature reference value T avg The value of the statistical period N is adjusted in segments according to the installation inclination θ of the photovoltaic module, wherein when the inclination θ > 30°, the statistical window is dynamically shortened with the seasonal light distribution.​ The UAV dispatching module generates the three-dimensional path planning parameters according to P i (t) generating three-dimensional path planning parameters, expressed as: ; wherein, ∇ x P i (t) and ∇ y P i (t) represent the x-direction fault propagation gradient and the y-direction environmental coupling gradient, respectively; when the UAV scheduling module performs three-dimensional path planning, the x-direction fault propagation gradient ∇ x P i (t) is normalized and weighted with the environmental light intensity decay rate, and when the triggering number of the dynamic fault determination threshold P thre exceeds the preset period, the calibration mode is automatically executed. 2.The photovoltaic power station inspection management system based on the Internet of Things according to claim 1, characterized in that, The method is realized by using the photovoltaic power station inspection management system based on the Internet of Things in any one of claims 1-3, and comprises the following steps: 3.The photovoltaic power station inspection management system based on the Internet of Things according to claim 2, characterized in that, The light adaptation threshold L thre The reference seasonal cloud cover S seasonal The sliding average is iteratively updated by a historical meteorological database, and the measured radiation value is introduced to feedback compensate μL(t) when it is cloudy.

4. A photovoltaic power station inspection management method based on the Internet of Things, characterized in that, S1, distributed data acquisition and dynamic reference calibration, the sensing nodes arranged in the photovoltaic matrix periodically collect the operation parameters of each unit, including: S2, dynamic health evaluation and operation and maintenance decision generation, first, the comprehensive health index, then the differential operation generates a dynamic priority, and finally, the current season cloud coverage is updated; a current value I i a voltage value U(t) i a temperature T(t) i a temperature T(t) and an ambient light intensity L(t) S3. According to the operation and maintenance decision, three-dimensional path planning is carried out, and at the same time, the UAV scheduling module is used to control the UAV to adaptively inspect. In step S1, the following process is executed:

5. The photovoltaic power station inspection management method based on the Internet of Things according to claim 4, characterized in that, S11, nonlinear attenuation correction is performed on the temperature data Ti(t); S12, the photovoltaic material attenuation factor p is updated according to the real-time measured glass cover plate light transmittance, and the weight coefficient alpha is corrected according to p.cos theta. ​

Images