A vibration and deformation-based photovoltaic array fault monitoring method and system

Abstract

This invention proposes a photovoltaic array fault monitoring method and system based on vibration and deformation, relating to the photovoltaic field. Based on time-series data of horizontal displacement, vertical displacement, and soil moisture content at different depths, the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil are calculated. A dynamic safety sub-index is calculated based on the natural frequency; a static safety sub-index is calculated based on the cumulative deformation and deformation rate; based on the dynamic and static safety sub-indices, a coupled safety assessment model of the support soil is constructed to calculate the comprehensive structural health index under the current state; the natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement, and soil moisture content are input into a trained deep belief neural network to identify potential fault types of the photovoltaic array. This invention takes into account both photovoltaic support vibration and slope deformation, achieving coupled safety assessment and accurate fault identification.

Description

Technical Field

[0001] This invention relates to the field of photovoltaics, and more specifically, to a method and system for monitoring photovoltaic array faults based on vibration and deformation. Background Technology

[0002] As one of the core forms of clean and renewable energy, photovoltaic (PV) power generation has seen exponential growth in installed capacity. The long-term stable operation of PV arrays is directly related to power generation efficiency, operational safety, and economic benefits. PV arrays are exposed to complex outdoor environments year-round, subject to the coupled effects of wind loads, snow loads, temperature stress, ultraviolet radiation, extreme rainfall, and other factors, making them prone to various failures. Among these, damage to the PV support structure and instability of the surrounding slope soil are the two core hidden dangers threatening the safe operation of PV arrays, and these two factors have a significant coupled influence.

[0003] As a key load-bearing component of photovoltaic modules, the structural stiffness degradation of photovoltaic support structures is a typical form of damage. This is mainly caused by factors such as component corrosion, loose bolts, and weld aging. Stiffness degradation directly leads to changes in the natural vibration characteristics of the support structure. If not monitored and identified in time, it may cause serious accidents such as buckling and overturning of the support structure, resulting in damage to the photovoltaic modules and power generation interruption. At the same time, photovoltaic arrays are often built in sloping areas. Changes in the moisture content and displacement deformation of the slope soil directly affect the stability of the support foundation. Under extreme weather conditions such as heavy rainfall, the increased moisture content of the soil will reduce the shear strength, induce slope sliding, and then pull on the support structure to produce additional deformation, accelerating the damage to the support structure and forming a vicious cycle of "support damage-slope instability".

[0004] Currently, existing photovoltaic array fault monitoring methods suffer from insufficient targeting, limited monitoring dimensions, and poor coupling and coordination, making it difficult to meet the comprehensive monitoring needs under complex operating conditions. On the one hand, monitoring of photovoltaic supports often focuses on their deformation mechanical parameters, neglecting the correlation between slope structural characteristics and stiffness attenuation, thus failing to achieve early and accurate identification of support damage. On the other hand, monitoring of slope soil is often independent of support monitoring, failing to consider the coupling effect between the two, making it difficult to reflect the overall safety status of the photovoltaic array. Traditional algorithms lack the ability to fuse multi-dimensional monitoring data, resulting in low fault identification accuracy and an inability to effectively distinguish between support damage, slope instability, and various potential faults caused by their coupling. Summary of the Invention

[0005] The purpose of this invention is to provide a photovoltaic array fault monitoring method and system based on vibration and deformation. This method can take into account both the vibration of the photovoltaic support and the deformation of the slope, and realize a photovoltaic array fault monitoring method that couples safety assessment and accurate fault identification. It solves the problems of insufficient ability of traditional algorithms to integrate multi-dimensional monitoring data, low fault identification accuracy, and inability to effectively distinguish between support damage, slope instability and various potential faults caused by the coupling of the two, thus ensuring the long-term stable operation of the photovoltaic array.

[0006] The embodiments of the present invention are implemented as follows:

[0007] This application provides a photovoltaic array fault monitoring method based on vibration and deformation, including the following steps:

[0008] Based on the collected data on the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support, analyze whether the stiffness of the photovoltaic support has decreased.

[0009] Based on the attenuation of the stiffness of the photovoltaic support, time-series data of horizontal displacement, vertical displacement and soil moisture content of the slope soil below the photovoltaic support at different depths were collected.

[0010] Based on the time-series data of the horizontal displacement, the vertical displacement, and the soil moisture content at different depths, the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil are calculated.

[0011] Calculate the dynamic safety sub-index based on the natural frequency; calculate the static safety sub-index based on the cumulative deformation and the deformation rate; construct a support-soil coupled safety assessment model based on the dynamic safety sub-index and the static safety sub-index, and calculate the comprehensive structural health index under the current state;

[0012] The natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement, and soil moisture content are input into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

[0013] The collection of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support includes: setting up an array of vibration sensors at the connection nodes of the photovoltaic support and at the intersection of the support column and the crossbeam to collect the natural frequency, mode shape, damping ratio, and vibration energy entropy characteristics of the photovoltaic support.

[0014] The step of analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support specifically includes: inputting the characteristic changes of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support into a pre-trained neural network model to determine whether the stiffness of the photovoltaic support has decreased.

[0015] The calculation of the time-series data of the horizontal and vertical displacements of the slope soil to obtain the cumulative deformation and deformation rate of the slope soil specifically includes: calculating the cumulative deformation based on the volume change of the slope soil and the vertical displacement; and calculating the deformation rate based on the rate of change of the cumulative deformation.

[0016] The potential sliding surface location of the slope soil is obtained by calculating the time-series data of the horizontal and vertical displacements of the slope soil. Specifically, this includes: calculating the displacement amount by the horizontal and vertical displacements at different depths; plotting the displacement distribution curve based on the displacement amount at different depths; and determining the potential sliding surface location based on the abrupt change in the slope of the displacement distribution curve.

[0017] The potential failure types of the photovoltaic array specifically include: loose support bolts, fatigue cracking of components, uneven settlement of the foundation, shallow landslides on the slope, and extreme wind-induced vibrations.

[0018] The method for monitoring photovoltaic array faults based on vibration and deformation further includes triggering an early warning when the comprehensive structural health index is lower than a set threshold and the potential fault type is identified.

[0019] A photovoltaic array fault monitoring system based on vibration and deformation includes:

[0020] Photovoltaic support analysis module: Based on the collected natural frequencies, mode shapes, damping ratios, and vibration energy entropy of the photovoltaic support, analyze whether the stiffness of the photovoltaic support has decreased;

[0021] Slope data acquisition module: Based on the attenuation of the stiffness of the photovoltaic support, it simultaneously acquires time-series data of the horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths;

[0022] Slope deformation analysis module: Based on the time series data of the horizontal displacement, the vertical displacement and the soil moisture content at different depths, calculate the cumulative deformation, deformation rate and potential sliding surface location of the slope soil.

[0023] Comprehensive safety assessment module: Calculates the dynamic safety sub-index based on the natural frequency; calculates the static safety sub-index based on the cumulative deformation and the deformation rate; constructs a support-soil coupled safety assessment model based on the dynamic safety sub-index and the static safety sub-index, and calculates the comprehensive structural health index under the current state;

[0024] Potential fault identification module: The natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement and soil moisture content are input into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

[0025] An electronic device includes: a memory for storing one or more programs; a processor; and, when the one or more programs are executed by the processor, implementing any one of the methods for monitoring photovoltaic array faults based on vibration and deformation.

[0026] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the methods for monitoring photovoltaic array faults based on vibration and deformation.

[0027] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0028] A method for monitoring photovoltaic array faults based on vibration and deformation includes the following steps: analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy; based on the decrease in the stiffness of the photovoltaic support, simultaneously collecting time-series data of horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths; calculating the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil based on the time-series data of the horizontal displacement, vertical displacement, and soil moisture content at different depths; calculating a dynamic safety sub-index based on the natural frequency; calculating a static safety sub-index based on the cumulative deformation and deformation rate; constructing a coupled safety assessment model of the support and soil based on the dynamic and static safety sub-indices, and calculating a comprehensive structural health index under the current state; inputting the natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement, and soil moisture content into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

[0029] 1. This invention simultaneously collects the vibration dynamic characteristics of the photovoltaic support and the deformation static characteristics of the slope soil, breaking the limitation of the independent monitoring of the support and the slope in the prior art. It fully considers the coupling influence between the two, and can comprehensively reflect the overall safety status of the photovoltaic array, avoiding the failure to detect faults caused by single-dimensional monitoring.

[0030] 2. Based on the stiffness decay state of the photovoltaic support, dynamic and automatic sampling can be performed according to the actual damage of the support, ensuring the complete collection of key damage data and slope deformation data, thus improving monitoring efficiency and data validity.

[0031] 3. By quantitatively calculating the dynamic safety sub-index, static safety sub-index, and comprehensive index, the safety status of the photovoltaic array can be accurately assessed, providing a clear quantitative basis for fault early warning.

[0032] 4. The deep belief neural network is used to fuse and identify multi-dimensional monitoring data, which has strong feature extraction and fault classification capabilities. It can accurately identify various potential faults caused by photovoltaic support stiffness decay, slope soil instability and their coupling, and solve the problems of low fault identification accuracy and inability to distinguish fault types in traditional methods.

[0033] 5. The monitoring method of the present invention does not require complex equipment modification, can be adapted to slope installation scenarios of different types of photovoltaic arrays, and is easy to promote and apply. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart of the photovoltaic array fault monitoring method based on vibration and deformation in Embodiment 1 of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0037] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the various embodiments and features described below can be combined with each other.

[0038] Example 1

[0039] Please see Figure 1 , Figure 1The illustrated embodiment of this application provides a photovoltaic array fault monitoring method based on vibration and deformation, comprising the following steps: analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy; based on the decrease in the stiffness of the photovoltaic support, simultaneously collecting time-series data of the horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths; calculating the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil based on the time-series data of the horizontal displacement, vertical displacement, and soil moisture content at different depths; calculating the dynamic safety sub-index based on the natural frequency; calculating the static safety sub-index based on the cumulative deformation and deformation rate; constructing a coupled safety assessment model of the support and soil based on the dynamic safety sub-index and calculating the comprehensive structural health index under the current state; inputting the natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement, and soil moisture content into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

[0040] The collection of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support includes: setting up an array of vibration sensors at the connection nodes of the photovoltaic support and at the intersection of the support column and the crossbeam to collect the natural frequency, mode shape, damping ratio, and vibration energy entropy characteristics of the photovoltaic support.

[0041] The step of analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support specifically includes: inputting the characteristic changes of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support into a pre-trained neural network model to determine whether the stiffness of the photovoltaic support has decreased.

[0042] The calculation of the time-series data of the horizontal and vertical displacements of the slope soil to obtain the cumulative deformation and deformation rate of the slope soil specifically includes: calculating the cumulative deformation based on the volume change of the slope soil and the vertical displacement; and calculating the deformation rate based on the rate of change of the cumulative deformation.

[0043] The potential sliding surface location of the slope soil is obtained by calculating the time-series data of the horizontal and vertical displacements of the slope soil. Specifically, this includes: calculating the displacement amount by the horizontal and vertical displacements at different depths; plotting the displacement distribution curve based on the displacement amount at different depths; and determining the potential sliding surface location based on the abrupt change in the slope of the displacement distribution curve.

[0044] The potential failure types of the photovoltaic array specifically include: loose support bolts, fatigue cracking of components, uneven settlement of the foundation, shallow landslides on the slope, and extreme wind-induced vibrations.

[0045] The method for monitoring photovoltaic array faults based on vibration and deformation further includes triggering an early warning when the comprehensive structural health index is lower than a set threshold and the potential fault type is identified.

[0046] First, a vibration sensor array is deployed at key locations on the photovoltaic support structure. The sensors are calibrated and tested to ensure the accuracy of the collected data. Specifically, a triaxial acceleration vibration sensor (model: ADXL355, measurement range: ±2g, sensitivity: 0.001g / LSB, sampling frequency: initially set to 100Hz) is deployed at the connection nodes of the photovoltaic support structure (one connection node is selected every two support units, with priority given to the middle and end connection nodes) and at the intersections of the support columns and crossbeams (one deployment point corresponds to each support column). This forms a vibration sensor array covering the critical stress-bearing parts of the entire photovoltaic array.

[0047] After the vibration sensor array acquires the vibration signal of the photovoltaic support in real time, four core vibration characteristic parameters are extracted: natural frequency, modal modes, damping ratio, and vibration energy entropy. The natural frequency, which is the inherent frequency of the photovoltaic support under no external force, reflects the structural stiffness of the support and is obtained through spectral analysis of the vibration signal using Fast Fourier Transform (FFT), with a frequency range of 0.1-10Hz. The modal modes are extracted using a modal analysis algorithm (emphasizing polynomial fitting modal identification) to obtain the mode shape distribution of the support at different orders. The damping ratio is calculated using the half-power bandwidth method, with the formula: ζ=(f2-f1) / (2f0), where f0 is the resonance frequency, and f1 and f2 are the frequencies corresponding to the half-power points on either side of the resonance peak. The vibration energy entropy is extracted through wavelet packet decomposition, decomposing the vibration signal into eight layers of wavelet packets and calculating the energy proportion of each frequency band, reflecting the irregularity of the support vibration and used to determine whether the support is damaged.

[0048] The extracted natural frequencies, modal modes, damping ratios, and vibration energy entropy characteristics are input into a pre-trained neural network model to determine whether the stiffness of the photovoltaic support has decreased. A backpropagation (BP) neural network can be used, with 4 neurons in the input layer, 12 neurons in the hidden layer, and 1 neuron in the output layer. The output result is either stiffness decreases or stiffness remains normal. The training process of this BP neural network involves collecting vibration characteristic parameters (natural frequencies, modal modes, damping ratios, and vibration energy entropy) of the photovoltaic support under different stiffness states (normal, slightly decreased, moderately decreased, and severely decreased) as training samples.

[0049] While the neural network model outputs stiffness decay, time-series data on the horizontal displacement, vertical displacement, and soil moisture content of the slope soil surrounding the photovoltaic support are collected using slope data acquisition equipment. Specifically, displacement sensors (such as wire-type displacement sensors with a measurement range of 0-500mm and an accuracy of ±0.1mm) and soil moisture sensors (such as TDR-300 sensors with a measurement range of 0-100% and an accuracy of ±2%) are deployed at different depths (e.g., 0.5m, 1.0m, and 1.5m) on the slope below the photovoltaic support. The horizontal displacement sensors collect the horizontal displacement of the slope soil, the vertical displacement sensors collect the vertical displacement, and the soil moisture sensors collect the soil moisture content.

[0050] Cumulative deformation is the sum of the volumetric changes in horizontal and vertical displacements of the slope soil over a period of time, used to measure the degree of deformation of the slope soil. The volumetric change is estimated based on the collected horizontal and vertical displacements of the slope soil. The formula for calculating cumulative deformation is: S = ∑(Δx_i × Δy_i × h), where Δx_i is the change in horizontal displacement at time i, Δy_i is the change in vertical displacement at time i, and h is the thickness / height of the slope soil (which can be taken as the maximum depth of the sensor deployment). The unit of cumulative deformation is m³. Alternatively, the volume can be automatically calculated by simulating the space based on displacement changes collected at different depths.

[0051] The deformation rate is calculated based on the rate of change of the cumulative deformation. The formula is: v = ΔS / Δt, where ΔS is the difference in cumulative deformation between two adjacent data acquisitions, Δt is the time interval between two adjacent data acquisitions, and the unit of deformation rate v is m³ / h.

[0052] The potential sliding surface is a weak surface in the slope soil that may slide. Identifying it by the location of abrupt changes in the slope of the displacement distribution curve is a core indicator for judging slope instability. The total displacement can be calculated using the formula for the hypotenuse of a triangle: c = √(a² + b²), where c is the displacement, a is the horizontal displacement, and b is the vertical displacement. A displacement distribution curve is plotted with the slope depth on the x-axis and the displacement on the y-axis. The location of the abrupt change in the slope of the displacement distribution curve is determined as the location of the potential sliding surface. The depth of this potential sliding surface and the corresponding horizontal and vertical displacements are recorded. The surface formed by the collected potential sliding surface locations at different depths is confirmed as the potential sliding surface.

[0053] The dynamic safety sub-index and static safety sub-index are calculated separately. Then, a coupled safety assessment model of the support and soil is constructed based on the two to calculate the comprehensive structural health index under the current state. The specific calculation process is as follows:

[0054] Calculation of Dynamic Safety Sub-index: Based on the collected natural frequencies of the photovoltaic support, the natural frequencies are converted into a dynamic safety sub-index using a normalization formula. The value of the dynamic safety sub-index D ranges from 0 to 1. The closer D is to 1, the better the dynamic safety of the photovoltaic support; the closer D is to 0, the worse the dynamic safety of the support, and the greater the risk of serious damage. Calculation of Static Safety Sub-index (S): The static safety sub-index S is calculated by summing the weighted values ​​(0-1) obtained from the calculated cumulative deformation and deformation rate. The value of S ranges from 0 to 1. The closer S is to 1, the better the static safety of the slope soil; the closer S is to 0, the higher the risk of slope instability.

[0055] A weighted summation method can be used to construct a safety assessment model for the coupled support and soil to calculate a comprehensive index. The weights of the dynamic safety sub-index and the static safety sub-index can be preset according to the actual operating scenario of the photovoltaic array. For example, they can be dynamically adjusted according to the structural stability importance of different slope soil types and support types. The resulting comprehensive structural health index H ranges from 0 to 1.

[0056] Optionally, a warning threshold for the structural health comprehensive index can be set, divided into three levels: Level 1 warning threshold H≥0.8, Level 2 warning threshold 0.5≤H<0.8, and Level 3 warning threshold H<0.5, used for subsequent fault warning triggering. Optionally, Level 1 warning only displays normal operation on the data monitoring platform and does not trigger audible and visual warnings; Level 2 warning displays the alert status and potential fault type on the monitoring platform, and issues a low-frequency alarm via audible and visual alarms on-site, while simultaneously sending a warning SMS containing the potential fault type, location, and current monitoring data to the mobile phones of maintenance personnel; Level 3 warning displays emergency warning information on the monitoring platform, issues a high-frequency alarm via on-site audible and visual alarms, and simultaneously sends an emergency warning information containing the potential fault type, location, and current monitoring data to the mobile phones of maintenance personnel and the monitoring center, and also activates on-site emergency linkage devices, such as cutting off power supply to the photovoltaic array in the fault area. Maintenance personnel promptly rush to the site to investigate the fault based on the warning information. After the problem is resolved, the system re-collects monitoring data and performs a new round of stiffness analysis, safety assessment, and fault identification until the warning is lifted.

[0057] A deep belief neural network was trained for identifying potential fault types in photovoltaic (PV) arrays. The specific training process is as follows: The input features of the neural network are the natural frequencies, modal modes, damping ratios, and vibration energy entropy of the PV support, as well as the horizontal displacement, vertical displacement, and soil moisture content of the slope soil. Multiple sets of monitoring data for each fault type and multiple sets of monitoring data collected under fault-free conditions were used as training samples. The deep belief neural network structure was set as follows: 7 neurons in the input layer, 3 hidden layers with 16, 12, and 8 neurons respectively, and 6 neurons in the output layer (corresponding to the 5 potential fault types and the fault-free category). A sigmoid activation function was used. During training, adaptive momentum gradient descent was employed to optimize the model parameters. After training until the accuracy on the validation set was ≥96% and the accuracy on the test set was ≥95%, the trained deep belief neural network model was saved. The deep belief neural network is a deep learning model composed of multiple restricted Boltzmann machines and a backpropagation network. It possesses strong feature extraction and fault classification capabilities, and is used for the accurate identification of PV array faults. Five potential fault types were selected as output results: loose support bolts, fatigue cracking of components, uneven settlement of foundation, shallow landslide of slope and extreme wind-induced vibration. At the same time, no fault was set as the sixth output category.

[0058] Example 2

[0059] This application provides a photovoltaic array fault monitoring system based on vibration and deformation, used to implement the photovoltaic array fault monitoring method based on vibration and deformation described in Embodiment 1. The system includes: a photovoltaic support analysis module: analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy; a slope data acquisition module: based on the decrease in the stiffness of the photovoltaic support, simultaneously acquiring time-series data of the horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths; and a slope deformation analysis module: based on the time-series data of the horizontal displacement, vertical displacement, and soil moisture content at different depths. The system calculates the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil; the comprehensive safety assessment module calculates the dynamic safety sub-index based on the natural frequency; calculates the static safety sub-index based on the cumulative deformation and deformation rate; constructs a coupled safety assessment model of the support soil based on the dynamic and static safety sub-indices, and calculates the comprehensive structural health index under the current state; the potential fault identification module inputs the natural frequency, the mode shape, the damping ratio, the vibration energy entropy, the horizontal displacement, the vertical displacement, and the soil moisture content into a trained deep belief neural network to identify the potential fault types of the photovoltaic array.

[0060] The photovoltaic support analysis module is connected to the vibration sensor array, receiving the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support from the sensor array. It also outputs a judgment result indicating whether the support stiffness is attenuating or normal, using a pre-trained BP neural network model as described in Example 1. The photovoltaic support analysis module is connected to the slope data acquisition module to transmit the judgment result, triggering attenuation based on the photovoltaic support stiffness. Simultaneously, it acquires time-series data on the horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths. The slope data acquisition module is connected to the photovoltaic support analysis module, displacement sensors, and soil moisture sensors, transmitting the time-series data to the slope deformation analysis module and the potential fault identification module.

[0061] The slope deformation analysis module is connected to the slope data acquisition module. Through the built-in deformation parameter calculation algorithm, it calculates the time series data of horizontal and vertical displacement of the slope soil, obtains the cumulative deformation, deformation rate and potential sliding surface location of the slope soil, and transmits it to the comprehensive safety assessment module to provide data support for safety assessment.

[0062] The comprehensive safety assessment module is connected to the photovoltaic support analysis module and the slope deformation analysis module. It incorporates calculation logic for dynamic safety sub-indices, static safety sub-indices, and a comprehensive index. The dynamic safety sub-indices are calculated based on natural frequencies, and the static safety sub-indices are calculated based on cumulative deformation and deformation rate. Based on these two factors, a coupled safety assessment model of the support and soil is constructed to calculate the comprehensive structural health index under the current condition. Simultaneously, the comprehensive structural health index is compared with a set threshold to determine the current safety level.

[0063] The potential fault identification module is connected to the photovoltaic support analysis module and the slope data acquisition module. It receives and inputs the natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement and soil moisture content into the built-in trained deep belief neural network to identify the potential fault types of the photovoltaic array.

[0064] In addition to the modules mentioned above, the system may also include a data storage module, which can use solid-state drives to store monitoring data, model parameters, fault records, and other information, supporting data querying and export. It can connect to the comprehensive safety assessment module and the potential fault identification module through the early warning module to receive safety level information and fault type information, thereby triggering corresponding warnings through audible and visual alerts, SMS alerts, or emergency response linkage. Furthermore, a visual monitoring interface displays real-time monitoring data, structural health comprehensive index, fault type, and early warning information, supporting operations and maintenance personnel to manually adjust parameters, view historical data, and cancel warnings.

[0065] The method and system of this application embodiment were applied to a photovoltaic power station on a mountain slope. The photovoltaic array consisted of 1000 sets of supports, the slope gradient was 35°, and the soil was silty clay. A three-month field test was conducted. Five potential faults were simulated (loose support bolts, fatigue cracking of components, uneven foundation settlement, shallow slope landslide, and extreme wind-induced vibration). Each fault was simulated 10 times. The system could accurately identify all faults, achieving an accuracy rate of 96.2%, far exceeding the accuracy of a pre-trained deep belief neural network predicting potential fault types, as well as traditional monitoring methods such as monitoring abnormal vibration data of the slope or photovoltaic supports. The test results show that the method and system of this invention can take into account both photovoltaic support vibration and slope deformation, achieving coupled safety assessment and accurate fault identification. This overcomes the shortcomings of traditional algorithms, ensures the long-term stable operation of the photovoltaic array, and possesses good practicality and promotional value.

[0066] Example 3

[0067] This embodiment provides an electronic device for implementing the photovoltaic array fault monitoring method based on vibration and deformation described in Embodiment 1. This electronic device can serve as the control core of the system and be deployed in the photovoltaic power station monitoring center. The electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The memory uses DDR4 RAM and an SSD solid-state drive to store one or more programs (including the execution program for the monitoring method, neural network model parameters, monitoring data, fault records, etc.), and supports real-time data reading and writing as well as long-term storage. It also has a data backup function to ensure that data is not lost.

[0068] Processor: The processor is an Intel Core i7-12700H with a clock speed of ≥2.7GHz and ≥14 cores. It is used to execute programs stored in memory. When one or more programs are executed by the processor, all the monitoring steps described in Example 1 are implemented, including vibration feature acquisition and analysis, adaptive sampling triggering, slope deformation calculation, safety index assessment, fault identification and early warning triggering, etc. The processing latency is ≤100ms, which meets the real-time monitoring requirements.

[0069] Example 4

[0070] This embodiment provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements all the steps of the photovoltaic array fault monitoring method based on vibration and deformation described in Embodiment 1. The computer-readable storage medium can be any medium capable of storing a computer program, such as a USB flash drive, portable hard drive, read-only memory, random access memory, magnetic disk, or optical disk. For example, a USB flash drive can be used to store the computer program. The USB flash drive is inserted into the USB port of an electronic device, and the processor reads the computer program from the USB flash drive and executes the relevant steps of the monitoring method. Alternatively, the computer program can be stored on a server, and the electronic device can connect to the server via a network to download and execute the computer program to achieve the monitoring function. The computer program can also be stored in the built-in memory of the electronic device, automatically starting upon power-on for real-time monitoring.

[0071] The computer program includes a series of instructions, which, when executed by the processor, sequentially perform the following operations: receive vibration signals from the photovoltaic support collected by the vibration sensor array, extract vibration characteristic parameters, and analyze whether the support stiffness has decreased; trigger adaptive sampling based on the stiffness decrease state to collect time-series data of the slope soil; preprocess the displacement time-series data to calculate the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil; calculate the dynamic safety sub-index, static safety sub-index, and structural health comprehensive index; input a deep confidence neural network to identify the fault type; trigger corresponding level early warnings based on the comprehensive index and fault type; and store monitoring data, fault records, and model parameters.

[0072] In summary, the photovoltaic array fault monitoring method and system based on vibration and deformation provided in this application embodiment simultaneously collects the dynamic vibration characteristics of the photovoltaic support and the static deformation characteristics of the slope soil, breaking the limitation of independent support monitoring and slope monitoring in the prior art. It fully considers the coupling influence between the two, comprehensively reflecting the overall safety status of the photovoltaic array and avoiding fault omissions caused by single-dimensional monitoring. Based on the stiffness attenuation state of the photovoltaic support, it can dynamically and automatically sample according to the actual damage situation of the support, ensuring the complete collection of key damage data and slope deformation data, improving monitoring efficiency and data validity. By quantitatively calculating the dynamic safety sub-index, static safety sub-index, and comprehensive index, the safety status of photovoltaic arrays can be accurately assessed, providing clear quantitative basis for fault early warning. A deep belief neural network is used to fuse and identify multi-dimensional monitoring data, possessing strong feature extraction and fault classification capabilities. It can accurately identify various potential faults caused by photovoltaic support stiffness decay, slope soil instability, and their coupling, solving the problems of low fault identification accuracy and inability to distinguish fault types in traditional methods. The monitoring method of this invention requires no complex equipment modification, can be adapted to slope installation scenarios of different types of photovoltaic arrays, and is easy to promote and apply.

[0073] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for monitoring photovoltaic array faults based on vibration and deformation, characterized in that, Includes the following steps: Based on the collected data on the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support, analyze whether the stiffness of the photovoltaic support has decreased. Based on the attenuation of the stiffness of the photovoltaic support, time-series data of horizontal displacement, vertical displacement and soil moisture content of the slope soil below the photovoltaic support at different depths were collected. Based on the time-series data of the horizontal displacement, the vertical displacement, and the soil moisture content at different depths, the cumulative deformation, deformation rate, and potential sliding surface location of the slope soil are calculated. Calculate the dynamic safety sub-index based on the natural frequency; calculate the static safety sub-index based on the cumulative deformation and the deformation rate; Based on the dynamic safety sub-index and the static safety sub-index, a support-soil coupled safety assessment model is constructed to calculate the comprehensive structural health index under the current state. The natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement, and soil moisture content are input into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

2. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 1, characterized in that, The collection of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support includes: setting up an array of vibration sensors at the connection nodes of the photovoltaic support and at the intersection of the support column and the crossbeam to collect the natural frequency, mode shape, damping ratio, and vibration energy entropy characteristics of the photovoltaic support.

3. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 2, characterized in that, The step of analyzing whether the stiffness of the photovoltaic support has decreased based on the collected natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support specifically includes: inputting the characteristic changes of the natural frequency, mode shape, damping ratio, and vibration energy entropy of the photovoltaic support into a pre-trained neural network model to determine whether the stiffness of the photovoltaic support has decreased.

4. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 1, characterized in that, The calculation of the time-series data of the horizontal and vertical displacements of the slope soil to obtain the cumulative deformation and deformation rate of the slope soil specifically includes: calculating the cumulative deformation based on the volume change of the slope soil and the vertical displacement; and calculating the deformation rate based on the rate of change of the cumulative deformation.

5. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 1, characterized in that, The potential sliding surface location of the slope soil is obtained by calculating the time-series data of the horizontal and vertical displacements of the slope soil. Specifically, this includes: calculating the displacement amount by the horizontal and vertical displacements at different depths; plotting the displacement distribution curve based on the displacement amount at different depths; and determining the potential sliding surface location based on the abrupt change in the slope of the displacement distribution curve.

6. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 1, characterized in that, The potential failure types of the photovoltaic array specifically include: loose support bolts, fatigue cracking of components, uneven settlement of the foundation, shallow landslides on the slope, and extreme wind-induced vibrations.

7. The photovoltaic array fault monitoring method based on vibration and deformation as described in claim 1, characterized in that, Also includes: An early warning is triggered when the overall structural health index falls below a set threshold and the potential fault type is identified.

8. A photovoltaic array fault monitoring system based on vibration and deformation, characterized in that, include: Photovoltaic support analysis module: Based on the collected natural frequencies, mode shapes, damping ratios, and vibration energy entropy of the photovoltaic support, analyze whether the stiffness of the photovoltaic support has decreased; Slope data acquisition module: Based on the attenuation of the stiffness of the photovoltaic support, it simultaneously acquires time-series data of the horizontal displacement, vertical displacement, and soil moisture content of the slope soil below the photovoltaic support at different depths; Slope deformation analysis module: Based on the time series data of the horizontal displacement, the vertical displacement and the soil moisture content at different depths, calculate the cumulative deformation, deformation rate and potential sliding surface location of the slope soil. The comprehensive safety assessment module calculates the dynamic safety sub-index based on the natural frequency and the static safety sub-index based on the cumulative deformation and the deformation rate. Based on the dynamic safety sub-index and the static safety sub-index, a support-soil coupled safety assessment model is constructed to calculate the comprehensive structural health index under the current state. Potential fault identification module: The natural frequency, mode shape, damping ratio, vibration energy entropy, horizontal displacement, vertical displacement and soil moisture content are input into a trained deep belief neural network to identify potential fault types of the photovoltaic array.

9. An electronic device, characterized in that, include: Memory, used to store one or more programs; processor; When the processor executes the one or more programs, it implements a photovoltaic array fault monitoring method based on vibration and deformation as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a photovoltaic array fault monitoring method based on vibration and deformation as described in any one of claims 1-7.

Images