CN-122004196-A - Automatic bird preventing device and method for distributed marine photovoltaic panel

CN122004196ACN 122004196 ACN122004196 ACN 122004196ACN-122004196-A

Abstract

The invention discloses an automatic bird preventing device and method for a distributed marine photovoltaic panel, the method comprises the following steps of S1 networking and sensing collection, S2 identification and threshold adjustment, S3 linkage and driving-off, S4 feedback and data recording, S5 optimization and operation and maintenance calibration, aiming at solving the problem of marine special environment adaptation, integrating multi-module networking and adopting wavelet noise reduction treatment through deploying a salt spray resistant weather-resistant edge node, effectively resisting environmental interference such as high salt spray, light intensity fluctuation and the like, guaranteeing stable operation and data acquisition purity of equipment, filling the blank of insufficient marine environment adaptation of a conventional bird preventing scheme, overcoming inaccurate pain points of bird identification and intention judgment, relying on deep learning, behavior model and track time sequence analysis, combining an environment parameter dynamic adjustment threshold, accurately distinguishing passing birds from residences resident bird, accurately judging landing and excreting intention, greatly reducing false triggering and leaking triggering probability, and improving the reliability of identification decision.

Inventors

Bai Senke
HAN LEI
YU TONG
ZHANG JIANYI
XING XIANGJIE
Jia Ganyuan
LI JINFENG

Assignees

国网山东省电力公司平度市供电公司

Dates

Publication Date: 20260512
Application Date: 20260205

Claims (10)

1. An automatic bird prevention method for a distributed marine photovoltaic panel is characterized by comprising the following steps of: The method comprises the steps of S1, networking and sensing collection, namely, deploying salt mist resistant and weather resistant edge nodes at preset points of a photovoltaic panel array, integrating sensing, driving away, communication and energy storage modules, and connecting a remote control platform through remote communication, wherein each node is networking to form a monitoring network, loading a bird database, a verification rule and a sensitive acousto-optic frequency band database, collecting environment and bird data in real time, and preserving effective characteristic data after noise reduction and anti-interference treatment; S2, identifying and threshold adjustment, namely identifying bird types and quantity through deep learning and a basic database, distinguishing passing birds from residents resident bird by means of a behavior model, judging landing and excretion intentions by combining a preset gesture threshold value through track fitting and time sequence gesture analysis; S3, linkage and separation, namely triggering a separation instruction by a node when the existence of the residence resident bird and the related intention is judged by S2, and synchronizing to peripheral nodes by a communication unit to form cooperative protection; s4, feeding back and recording data, namely continuously monitoring bird residues and track changes after driving, collecting response data, energy consumption parameters and equipment states, and uploading the classified data to a network sharing node and a remote management and control platform; And S5, optimizing and operating and maintaining calibration, namely, the remote control platform is used for analyzing S4 feedback data through deep learning, identifying habit characteristics of birds, updating a driving-off parameter library and synchronously issuing the habit characteristics to each node optimizing strategy, and constructing an error compensation model according to environmental parameters to calibrate sensor precision, so that fault accurate positioning, intelligent alarm and remote operating and maintaining scheduling are realized through remote communication.
2. The method for automatically preventing birds of the distributed marine photovoltaic panel according to claim 1, wherein the noise-reducing and anti-interference treatment in S1 specifically comprises the following substeps: s11, collecting original signals output by each perception module of an edge node, including an environment parameter original signal and a bird monitoring original signal, and setting signal sampling frequency and duration; s12, carrying out wavelet decomposition on an original signal, determining the number of decomposition layers, extracting wavelet coefficients of each layer, and separating effective components and noise components of the signal; and S13, carrying out threshold denoising treatment on the decomposed wavelet coefficients, retaining the wavelet coefficients corresponding to the effective components, inhibiting the wavelet coefficients corresponding to the noise components, and carrying out inverse wavelet transform reconstruction to obtain denoised signals.
3. The method for automatically preventing birds of a distributed marine photovoltaic panel according to claim 2, wherein S13 is implemented by adopting a wavelet soft threshold noise reduction function, and the specific logic is that the processed wavelet coefficients are as follows The value rule of (2) is that when the original coefficient after wavelet decomposition Is greater than the soft threshold In the time-course of which the first and second contact surfaces, Is that Is a sign function value of (2) Is subtracted from the absolute value of (2) When the product of the differences of Is less than or equal to the absolute value of In the time-course of which the first and second contact surfaces, Set to 0, wherein the soft threshold The value of (2) is determined by the standard deviation of noise And the number of signal sampling points By passing through The corresponding operation relation is obtained, the noise standard deviation Calculated by adopting a median estimation method, in particular to The values of which are decomposed by wavelet Dividing the median of the absolute value of (2) by 0.6745; as a sign function, output 1 when the input is positive and output-1 when it is negative.
4. The method for automatically preventing birds on a distributed marine photovoltaic panel according to claim 1, wherein the step of dynamically adjusting the detection and intention judgment threshold in S2 comprises the following steps: S21, extracting the light intensity and salt fog concentration environment parameters acquired and denoised in the step S1, and taking the light intensity and salt fog concentration environment parameters as input variables for threshold adjustment; S22, calculating threshold compensation coefficients corresponding to all environment parameters based on a preset environment influence weight model; S23, combining the basic threshold value and the compensation coefficient, generating a real-time detection threshold value and an intention judgment threshold value, and synchronously updating the real-time detection threshold value and the intention judgment threshold value into an edge node recognition algorithm.
5. The method for automatically preventing birds in a distributed marine photovoltaic panel according to claim 4, wherein the dynamic threshold adjustment in S23 is implemented by a linear weighting model, and the specific logic is real-time threshold Is set by a preset basic threshold value Multiplying by an environment correction factor, which is obtained by the light intensity compensation factor Multiplying by its weight coefficient Salt spray concentration compensation coefficient Multiplying by its weight coefficient Post-summation and re-superposition correction factor Is composed of 、 Respectively the weight coefficient of the light intensity and the salt fog concentration, and The value ranges are all 0-1; The light intensity compensation coefficient is determined by the ratio of the actual light intensity to the standard light intensity; the salt spray concentration compensation coefficient is determined by the ratio of the actual salt spray concentration to the standard salt spray concentration; the value of the correction coefficient is 0.9-1.1, and the correction coefficient is used for counteracting the model system error.
6. The method for automatically preventing birds by using the distributed marine photovoltaic panel according to claim 1, wherein the step of performing acousto-optic randomization driving-off in the step S3 comprises the following steps: s31, according to the bird species identified in the S2, calling the corresponding basic driving frequency and intensity from the sensitive acousto-optic frequency band database; s32, generating random disturbance quantity based on a time sequence, and ensuring that the disturbance quantity fluctuates outside a non-target biological sensitive frequency band; And S33, superposing the basic driving parameter and the random disturbance quantity, outputting the final acousto-optic driving parameter and triggering the execution module.
7. The method for automatically preventing birds on a distributed marine photovoltaic panel according to claim 6, wherein the randomized driving-out frequency generation logic in S33 is: time of day drive-off frequency Is driven from the base drive-off frequency by the value of (2) Superposing a sine disturbance quantity which changes along with time to obtain; the maximum amplitude of the sinusoidal disturbance is Angular frequency For controlling the period, initial phase, of frequency variation The value range is 0- Is generated by a random function, so that different frequency change tracks of each driving-off are realized.
8. The method for automatically preventing birds of the distributed marine photovoltaic panel according to claim 1, wherein the step of constructing the error compensation model in S5 comprises the following steps: s51, acquiring sensor calibration data including an actual measurement value and a standard reference value every day, and calculating a corresponding error value; S52, fitting a sensor drift curve by taking temperature and operation time as independent variables and error values as dependent variables to determine a drift rule; And S53, constructing an error compensation model based on the drift curve, substituting the error compensation model into the real-time environment parameters and the operation time length, and outputting the sensor calibration value.
9. The method for automatically preventing birds on a distributed marine photovoltaic panel according to claim 8, wherein the specific logic of the sensor drift error compensation model in S53 is the output value after sensor calibration From raw measurements of sensors Subtracting a total drift error term, wherein the total drift error term is formed by linearly superposing a temperature drift component and a time aging component, and the temperature drift component is a temperature drift coefficient Multiplying the difference between the actual temperature and the standard calibration temperature The temporal aging component is a temporal aging factor Multiplying the sensor continuous run time 。
10. The automatic bird prevention device of a distributed marine photovoltaic panel according to claim 1, comprising a salt spray resistant weather resistant edge node and a remote management and control platform, wherein the edge node integrates sensing, driving away, communication and energy storage modules, each edge node is self-networked to form a monitoring network and establishes connection with the remote management and control platform through remote communication, and the device further comprises the following functional modules: The distributed sensing and networking module is carried on each edge node and is used for deploying the edge nodes at preset points of the photovoltaic panel array, loading a bird database, a verification rule and a sensitive acousto-optic frequency band database, collecting environment and bird data in real time, carrying out noise reduction and anti-interference treatment on the collected data and reserving effective characteristic data, and realizing the self-networking of each node and the remote communication connection with a remote control platform; the intelligent recognition and decision module is used for recognizing the types and the quantity of birds through deep learning in combination with a basic database, distinguishing passing birds from residents resident bird by means of a behavior model, judging landing and excretion intentions of the birds through track fitting and time sequence gesture analysis in combination with a preset gesture threshold value, dynamically adjusting a detection and judgment threshold value in combination with environmental parameters, and realizing multistage verification by cooperating with voiceprint comparison; The cooperative driving execution module is carried on each edge node and is used for triggering driving instructions and synchronizing the driving instructions to peripheral nodes through a communication unit to form cooperative protection when the intelligent recognition and decision module judges that the resident resident bird and the related intention exist, avoiding non-target biological interference according to a sensitive frequency band database, calling specific parameters in combination with bird types and executing acousto-optic randomization driving operation; The effect feedback monitoring module is carried on each edge node and is used for continuously monitoring bird residues and track changes after the driving operation, collecting driving response data, equipment energy consumption parameters and equipment running states, classifying the data and uploading the classified data to the network sharing node and the remote control platform; The self-adaptive operation and maintenance optimization module is carried on the remote control platform and is used for identifying the habitual characteristics of birds and updating the driving-off parameter library according to feedback data uploaded by the deep learning analysis effect feedback monitoring module, synchronously issuing the bird habitual characteristics to each edge node to optimize the driving-off strategy, constructing an error compensation model according to environmental parameters, calibrating sensor precision, and realizing accurate positioning, intelligent alarm and remote operation and maintenance scheduling of equipment faults through remote communication.

Description

Automatic bird preventing device and method for distributed marine photovoltaic panel Technical Field The invention relates to the field of bird prevention of ocean photovoltaic panels, in particular to an automatic bird prevention device and method for a distributed ocean photovoltaic panel. Background With the rapid development of the global new energy industry, the ocean photovoltaic is used as a clean and efficient energy utilization form, and has become an important development direction of the photovoltaic industry by virtue of the advantages of no land occupation, stable illumination condition and high power generation efficiency; however, the marine photovoltaic panel array is exposed in a marine environment for a long time and is easily seriously influenced by the activities of seabirds, namely, the seabirds stay on the surface of the photovoltaic panel and excrete to form stains, shelter from sunlight and corrode the surface layer of the photovoltaic panel, so that the photoelectric conversion efficiency is obviously reduced; At present, the existing bird prevention technology mainly comprises two main types of physical bird prevention and electronic driving; the physical bird prevention mode comprises the steps of mounting a bird prevention net and bird thorns, the problems of heavy structure and high mounting and maintenance difficulty, inapplicability to large-scale deployment of a distributed marine photovoltaic panel, possibility of interference to normal migration of marine birds, the electronic driving mode comprises the steps of fixing frequency sound waves and single strong light driving, and has certain flexibility, but has obvious defects that on one hand, the marine environment has the characteristics of high salt fog, strong humidity fluctuation and severe light intensity change, the conventional sensor is easy to be interfered by the environment, so that bird recognition accuracy is low, a fixed detection threshold cannot be adapted to complex environment change, on the other hand, the single-mode driving means are easy to habituate birds, the driving effect is greatly attenuated after long-term use, and in addition, the existing scheme lacks a distributed node cooperative mechanism, single-point driving is difficult to cover a large-scale photovoltaic array, a closed-loop optimization system is not established, and the problems of difficult positioning of equipment faults and high operation and maintenance cost under the marine environment cannot be effectively solved according to bird behavior change and equipment operation state are not established; The prior art fails to fully consider the suitability of the ocean special environment, the bird-preventing long-acting property and the operation and maintenance convenience, and is difficult to meet the actual bird-preventing requirement of the distributed ocean photovoltaic panel, so that an automatic bird-preventing technical scheme with strong pertinence, high efficiency and stability is needed. Disclosure of Invention Aiming at the problems existing in the prior art, the invention aims to provide an automatic bird prevention device and an automatic bird prevention method for a distributed marine photovoltaic panel, which aim to solve the problems that the suitability of a marine special environment is insufficient, the environment with high salt fog, fluctuation of light intensity and severe temperature and humidity changes is easy to cause large detection errors of sensors and poor equipment stability, the conventional bird prevention scheme lacks environment adaptation design and cannot realize precise bird identification and stable operation, solve the problems that the conventional bird identification and intention judgment are inaccurate, the conventional scheme mainly adopts fixed threshold judgment, does not combine with dynamic adjustment of environmental parameters, is difficult to distinguish a passing bird from a resident resident bird, cannot accurately identify the falling and excreting intention of birds, is easy to generate false triggering or missed triggering, solves the problems that the long-lasting effect of the driving effect is insufficient, the single-point driving means in a fixed mode is easy to cause habituation of birds, the single-point driving coverage range is limited, the comprehensive protection cannot be formed, the conventional scheme lacks closed loop optimization, and the conventional scheme does not have feedback iteration mechanism and cannot optimize according to the behavior change of birds and the driving effect and the running data of equipment, and the bird prevention efficiency is lowered after long-term use. In order to solve the problems, the invention adopts the following technical scheme. An automatic bird prevention method for a distributed marine photovoltaic panel comprises the following steps: The method comprises the steps of S1, networking and sensing collection, deploying sal