[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN112257331A - Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence - Google Patents

Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence Download PDF

Info

Publication number
CN112257331A
CN112257331A CN202010997929.1A CN202010997929A CN112257331A CN 112257331 A CN112257331 A CN 112257331A CN 202010997929 A CN202010997929 A CN 202010997929A CN 112257331 A CN112257331 A CN 112257331A
Authority
CN
China
Prior art keywords
snow
icing
thickness
panel
battery
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202010997929.1A
Other languages
Chinese (zh)
Inventor
徐尔灵
钟竞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhengzhou Maitou Information Technology Co ltd
Original Assignee
Zhengzhou Maitou Information Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhengzhou Maitou Information Technology Co ltd filed Critical Zhengzhou Maitou Information Technology Co ltd
Priority to CN202010997929.1A priority Critical patent/CN112257331A/en
Publication of CN112257331A publication Critical patent/CN112257331A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B15/00—Measuring arrangements characterised by the use of electromagnetic waves or particle radiation, e.g. by the use of microwaves, X-rays, gamma rays or electrons
    • G01B15/02—Measuring arrangements characterised by the use of electromagnetic waves or particle radiation, e.g. by the use of microwaves, X-rays, gamma rays or electrons for measuring thickness
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N25/00—Investigating or analyzing materials by the use of thermal means
    • G01N25/14—Investigating or analyzing materials by the use of thermal means by using distillation, extraction, sublimation, condensation, freezing, or crystallisation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
    • G01S15/02—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems using reflection of acoustic waves
    • G01S15/06—Systems determining the position data of a target
    • G01S15/08—Systems for measuring distance only

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Evolutionary Computation (AREA)
  • Theoretical Computer Science (AREA)
  • Electromagnetism (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Acoustics & Sound (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Geometry (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a method and a system for cleaning frozen photovoltaic cell panels based on artificial intelligence. The method comprises the following steps: acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data; acquiring the thickness of the accumulated snow and the initial icing grade of the initial data through a DNN network; establishing a mapping model by using the thickness of the accumulated snow, the environmental factors and the initial icing grade; inputting the thickness of the accumulated snow of each battery plate into a mapping model, and predicting the time t for the accumulated snow on the surface of each battery plate to start to freeze0And an icing grade K after a period of time; adjusting the cleaning mode according to the icing grade KIncluding a snow clearing mode and an ice clearing mode. The icing grade after a period of time is accurately predicted by using the mapping model, so that the cleaning mode is adaptively adjusted according to the icing grade, and the reduction of the service life of the battery panel caused by incomplete cleaning is prevented.

Description

Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a system for cleaning frozen photovoltaic cell panels based on artificial intelligence.
Background
After snowy days, the surface of the panels of the photovoltaic power station can be covered with snow. Under the influence of various environmental factors, when the accumulated snow reaches a certain thickness, the accumulated snow can be slowly melted and then frozen again. Icing on the surface of the solar panel of the photovoltaic power station can not only reduce the output power of the solar panel, but also reduce the service life of the solar panel.
Patent document CN111112270A discloses a photovoltaic intelligent cleaning control system and method based on snow sensing, in which a visual sensor and a snow sensor are used to predict the snow amount of the environment where a photovoltaic module is located, and a meteorological database is combined to timely and accurately judge whether the photovoltaic module starts to snowing, and a cleaning robot automatically sweeps the snow and cleans the photovoltaic module in time, so as to prevent the photovoltaic module from performing photoelectric conversion normally due to the loss of generated energy caused by the snow.
In practice, the inventors have found that the above method does not allow to detect accurately the thickness of the snow on each panel and does not take into account the icing of the surface of the panel, and therefore does not allow to clean each panel thoroughly accordingly.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a photovoltaic cell panel icing cleaning method and system based on artificial intelligence, and the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides an artificial intelligence based method for cleaning an iced photovoltaic cell panel, including the following steps:
acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
acquiring the initial data by a DNN network to obtain the thickness of the accumulated snow and the initial icing grade;
establishing a mapping model by using the snow thickness and the environmental factors and the initial icing grade;
inputting the thickness of the accumulated snow of each battery plate into the mapping model, and predicting the time t for the accumulated snow on the surface of each battery plate to start to freeze0And an icing grade K after a period of time;
adjusting a cleaning mode according to the icing level K, wherein the cleaning mode comprises a snow cleaning mode and an ice cleaning mode;
the mapping model is as follows:
Figure BDA0002693255770000011
Figure BDA0002693255770000012
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are constants; h represents the snow thickness of the surface of the battery plate, T represents the ambient temperature and M represents the ambient humidity;
when t is<t0When K is equal to 0, adjusting to be in a snow clearing mode;
when t is>t0And adjusting the corresponding ice-clearing mode according to the value of K.
The DNN network comprises a panel semantic segmentation network and an accumulated snow thickness detection network, and the step of acquiring the accumulated snow thickness comprises the following steps:
obtaining a semantic area of each battery panel by utilizing the battery panel semantic segmentation network;
combining the semantic area and the ultrasonic data to obtain an ultrasonic signal corresponding to each battery plate;
and inputting the semantic area of the battery plate, the battery plate surface image data and the corresponding ultrasonic signal into the snow thickness detection network to obtain the snow thickness of each battery plate.
The DNN network further comprises an icing level assessment network, the step of obtaining the incipient icing level comprising:
and inputting the semantic area, the panel surface image data and the infrared image data of each panel into the icing grade evaluation network to obtain the initial icing grade of each panel.
The icing factor TcThe calculation formula of (2):
Figure BDA0002693255770000021
wherein HnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hntThe accumulated snow thickness of the solar panel n at the t-th moment is represented;
Figure BDA0002693255770000022
is represented by [ a, t]Average temperature over time;
Figure BDA0002693255770000023
is represented by [ a, t]Average humidity over time; θ is a balance factor and is a constant.
The snow melting factor TbThe calculation formula of (2):
Figure BDA0002693255770000024
wherein Hn1The thickness of the snow cover of the panel n at the 1 st moment is shown; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown;
Figure BDA0002693255770000025
is represented by [1, a ]]Average temperature over time;
Figure BDA0002693255770000026
is represented by [1, a ]]Average humidity over time; θ is a balance factor and is a constant.
In a second aspect, another embodiment of the present invention provides an artificial intelligence based photovoltaic panel icing cleaning system, comprising:
the acquisition unit is used for acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
the DNN network unit is used for acquiring the snow thickness and the initial icing grade of the initial data through a DNN network;
the modeling unit is used for establishing a mapping model by utilizing the snow thickness and the environmental factors and the initial icing level;
a prediction unit for inputting the snow thickness of each panel into the mapping model and predicting the snow icing starting time t of each panel surface0And an icing grade K after a period of time; and
the adjusting unit is used for adjusting a cleaning mode according to the icing grade K, and the cleaning mode comprises a snow cleaning mode and an ice cleaning mode;
the mapping model is as follows:
Figure BDA0002693255770000031
Figure BDA0002693255770000032
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are constants; h represents the snow thickness of the surface of the battery plate, T represents the ambient temperature and M represents the ambient humidity;
when t is<t0When K is equal to 0, adjusting to be in a snow clearing mode;
when t is>t0And adjusting the corresponding ice-clearing mode according to the value of K.
The DNN network unit further includes:
the battery board semantic segmentation network unit is used for acquiring the semantic area of each battery board;
the snow thickness detection network unit is used for acquiring the snow thickness of each battery plate; and
and the icing grade evaluation network unit is used for acquiring the initial icing grade of each battery panel.
The modeling unit further includes:
an icing factor unit for acquiring the icing factor TcThe icing factor unit is as follows:
Figure BDA0002693255770000033
wherein HnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hntThe accumulated snow thickness of the solar panel n at the t-th moment is represented;
Figure BDA0002693255770000034
is represented by [ a, t]Average temperature over time;
Figure BDA0002693255770000035
is represented by [ a, t]Average humidity over time; θ is a balance factor and is a constant.
The modeling unit further includes:
an ice-melting factor unit for obtaining the ice-melting factor TbThe ice-melting factor unit is as follows:
Figure BDA0002693255770000036
wherein Hn1The thickness of the snow cover of the panel n at the 1 st moment is shown; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown;
Figure BDA0002693255770000037
is represented by [1, a ]]Average temperature over time;
Figure BDA0002693255770000038
is represented by [1, a ]]Average humidity over time; θ is a balance factor and is a constant.
The prediction unit further includes:
a time prediction unit for predicting the icing start time t of the accumulated snow on the surface of each panel0；
A grade prediction unit for predicting the icing grade K of the accumulated snow on the surface of each panel.
The embodiment of the invention at least has the following beneficial effects:
(1) the collected data and the DNN network can obtain the accurate snow thickness and initial icing grade of the photovoltaic cell panel.
(2) And establishing a strict mapping model according to the snow thickness, the environmental factors and the initial icing grade.
(3) The icing grade after a period of time is accurately predicted by using the mapping model, so that the cleaning mode is adaptively adjusted according to the icing grade, and the reduction of the service life of the battery panel caused by incomplete cleaning is prevented.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a method for cleaning an iced photovoltaic panel based on artificial intelligence according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating steps of a method for cleaning an iced photovoltaic panel based on artificial intelligence according to an embodiment of the present invention;
FIG. 3 is a block diagram of a cleaning system for artificial intelligence based photovoltaic panel icing according to another embodiment of the present invention;
FIG. 4 is a block diagram of a DNN network element provided by one embodiment of the present invention;
FIG. 5 is a block diagram of a modeling unit according to an embodiment of the present invention;
FIG. 6 is a block diagram of a prediction unit according to an embodiment of the present invention;
fig. 7 is a block diagram of an adjusting unit according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the method and system for cleaning the frozen photovoltaic cell panel based on artificial intelligence according to the present invention with reference to the accompanying drawings and preferred embodiments shows the following detailed descriptions. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of a photovoltaic cell panel icing cleaning method and system based on artificial intelligence in detail with reference to the accompanying drawings.
Referring to fig. 1 and 2, an embodiment of the present invention provides a method for cleaning an iced photovoltaic cell panel based on artificial intelligence, which includes the following steps:
step S001, acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
s002, acquiring the thickness of the accumulated snow and the initial icing grade of the initial data through a DNN network;
step S003, a mapping model is established by utilizing the thickness of the accumulated snow, the environmental factors and the initial icing grade;
step S004, inputting the thickness of the accumulated snow of each battery board into a mapping model, and predicting the time t for the accumulated snow on the surface of each battery board to start to freeze0And an icing grade K after a period of time;
and step S005, adjusting a cleaning mode according to the icing grade K, wherein the cleaning mode comprises a snow cleaning mode and an ice cleaning mode.
Preferably, in step S001, gather the initial data of every panel through unmanned aerial vehicle, and unmanned aerial vehicle carries RGB camera, infrared camera and ultrasonic detector, and unmanned aerial vehicle low latitude flight, and the camera overlooks the initial data of downwards gathering every panel.
Further, in step S002, the DNN network includes a panel semantic segmentation network, and the specific step of obtaining the semantic area of each panel is as follows:
1) inputting the battery panel surface image data acquired by the RGB camera into an RGB image encoder to obtain an RGB image characteristic diagram, wherein the RGB characteristic diagram comprises a snow cover characteristic diagram, an ice layer characteristic diagram and a battery panel characteristic diagram;
2) the RGB image feature map is subjected to semantic segmentation by an encoder to obtain a battery panel feature map;
3) and (4) obtaining the semantic area of each battery panel by the battery panel feature map through a semantic segmentation decoder.
Further, the DNN network also includes a snow thickness detection network, and the specific steps of acquiring the snow thickness are as follows:
1) inputting the RGB image feature map and the semantic area of each battery panel into an accumulated snow thickness encoder to obtain an accumulated snow thickness feature map;
2) the accumulated snow thickness characteristic diagram is obtained through first full connection;
3) acquiring the position of the semantic area in the image visual field according to the semantic area of each battery plate, transmitting a beam of ultrasonic signal to the position by an ultrasonic detector, and when receiving the reflected ultrasonic signal, acquiring the ultrasonic signal which is the ultrasonic signal corresponding to the battery plate in the semantic area;
4) the ultrasonic signals corresponding to the cell panels are subjected to second full connection to obtain the fusion characteristics of the ultrasonic signals corresponding to each cell panel;
5) and (4) fully connecting the snow thickness characteristic diagram of each battery board with the fusion characteristic of the ultrasonic signal corresponding to each battery board through a third connection to obtain the actual snow thickness of each battery board.
Further, the DNN network further includes an icing level evaluation network, and the specific steps of obtaining the initial icing level are as follows:
1) inputting an infrared image collected by an infrared camera into an infrared image encoder to obtain an infrared image characteristic diagram of the ice layer;
2) inputting the infrared image characteristic diagram and the RGB image characteristic diagram of the ice layer and the semantic area of each battery panel into an icing coder to obtain an icing characteristic diagram of each battery panel;
3) and fully connecting the icing characteristic diagrams to obtain the icing grade of each battery panel.
The larger the initial icing level, the thicker the ice and the larger the area indicating the icing.
Further, in step S003, the specific steps of establishing the mapping model are as follows:
considering that the thickness change of the snow cover is not only related to environmental factors but also related to the heat effect of the battery plate when the snow cover is not frozen, the snow melting factor T of the battery plate is introduced into the inventionbTo reflect this effect, there are:
Figure BDA0002693255770000051
Figure BDA0002693255770000061
Figure BDA0002693255770000062
wherein HnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hn1The thickness of the snow cover of the panel n at the 1 st moment is shown;
Figure BDA0002693255770000063
is represented by [1, a ]]Average temperature over time;
Figure BDA0002693255770000064
is represented by [1, a ]]Average humidity over time; theta is a balance factor and is a undetermined coefficient.
When the accumulated snow begins to freeze, the thickness change of the accumulated snow is not only related to the environmental factors, but also related to the heat effect of the battery plate and the influence of the ice layer, so the invention introduces the freezing factor TcTo reflect this effect, there are:
Figure BDA0002693255770000065
Figure BDA0002693255770000066
Figure BDA0002693255770000067
wherein HntThe accumulated snow thickness of the solar panel n at the t-th moment is represented; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown;
Figure BDA0002693255770000068
is represented by [ a, t]Average temperature over time;
Figure BDA00026932557700000611
is represented by [ a, t]Average humidity over time; theta is a balance factor and is a undetermined coefficient.
It should be noted that: 1) since the panels are power generating elements, they themselves generate heat, a phenomenon known as thermal effects. The heat effect of panel can influence the speed that melts of snow to when the panel freezes, the temperature distribution on panel surface is inhomogeneous, makes the heat effect of panel change, brings the influence to the formation and the speed that melts of ice on the panel.
2)Tb、TcIs a numerical value of the temperature dimension, and is related to the intrinsic factors of the solar panel and the geographic environment factors, such as the type of the solar panel, the heat production quantity, the local average illumination intensity, the solar altitude angle and the like. The initial data, the environmental factor data and the snow thickness obtained by the DNN network collected by the multiple groups of unmanned aerial vehicles are utilized to perform data fitting through the formulas (1) and (2), so that T can be obtainedb、TcAnd the magnitude of the balance factor theta. Therefore, the embodiment of the invention looks at Tb、TcIs a constant.
When the accumulated snow on the surface of the battery plate melts, the thickness of the accumulated snow also changes, and the accumulated snow begins to freeze after reaching a certain thickness, the freezing grade, the thickness of the accumulated snow and the freezing factor TcSnow melting factor TbThe temperature T of the environment and the humidity M.
For a battery plate, the thickness of the accumulated snow on the battery plate is set to be H, the icing grade after t time is set to be K, and therefore the mapping model is as follows:
Figure BDA0002693255770000069
Figure BDA00026932557700000610
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are undetermined coefficients; h denotes the snow thickness of the surface of the panel, T denotes the ambient temperature and M denotes the ringAmbient humidity.
Further, undetermined coefficients α, β, γ in the mapping model may be obtained by performing data fitting according to the above equations (3) and (4) by using multiple sets of initial data acquired by the unmanned aerial vehicle, environmental factor data, and initial icing levels obtained by the DNN network, to obtain specific numerical values of α, β, γ, and determine a final mapping model.
The mapping model reflects the relation between the thickness of the snow cover and the environmental factors and the icing grade, and the icing grade K of the snow cover after the time t can be predicted according to the thickness of the snow cover under the condition of known environmental temperature and humidity. For snow with a certain thickness, after the snow is frozen, the lower the ambient temperature is, the higher the humidity is, the larger the freezing degree is, and the thicker the snow is, the larger the freezing degree is. In addition, the thicker the snow, the lower the temperature and the higher the humidity, the shorter the time to start the ice formation of the snow.
Further, the method for obtaining the data fitting in the final mapping model comprises:
1) is provided with N cell panels HnThe thickness of the accumulated snow of the nth panel is shown. KnThe icing grade of the nth battery plate is represented, and the data of the unmanned aerial vehicle during snow melting is collected once every hour: obtaining accumulated snow thickness data H of the nth cell panel at t moments after t hoursn：{Hn1，Hn2，Hn3，......，HntIcing grade data K at t momentsn：{Kn1，Kn2，Kn3，......，KntThe ambient temperature at T moments is T: { T1，T2，T3，......，TtAnd the ambient humidity at t times is M: { M1，M2，M3，......，Mt}. The unmanned aerial vehicle obtains the data lists in t hours, and the data lists are collectively called a group of data.
2) From the snow thickness and the initial icing level obtained by the DNN network, T can be obtained by performing polynomial equation calculation using the above equations (1), (2), (3) and (4)b、TcSpecific values of θ, α, β, γ.
It should be noted that: 1) the invention only considers the data when the snow melts, and does not consider the data when the snow falls.
For one set of data, the thickness of the snow will slowly decrease as it melts, and will begin to freeze at some point. Setting the icing time of accumulated snow as a: then there are:
Kna＞0
Knaindicating the icing level of the nth panel at the moment a; when the icing level is 0, no icing is indicated.
2) One group of data can be obtained only after t hours of data acquisition in one day, and multiple days of data acquisition are needed to obtain multiple groups of data.
Preferably, in the embodiment of the present invention, when t is 6, data of approximately one day is acquired.
Further, in step S005, the specific steps of adjusting the cleaning mode are as follows:
acquiring initial data of each battery panel by using an unmanned aerial vehicle, and acquiring the accumulated snow thickness H of each battery panel through a DNN (deep neural network);
sending the accumulated snow thickness H to a snow cleaning robot;
the snow cleaning robot calculates the icing starting time t of the accumulated snow on the surface of each battery plate through the mathematical mapping model0And an icing grade K after a period of time;
the snow cleaning robot starts to work, and the snow cleaning mode is used for cleaning the non-frozen snow, namely the snow on the surface of the battery panel with the freezing grade K of 0;
along with the lapse of time, after the snow removing robot cleans one battery panel, when the snow removing robot starts to clean the next battery panel, the snow removing robot calculates the icing grade K of the battery panel according to the cleaning time of the previous battery panel;
and starting different ice cleaning modes according to different icing grades K.
In summary, the embodiment of the present invention provides an artificial intelligence based photovoltaic cell panel icing cleaning method, in which collected data of each cell panel surface is used to obtain snow thickness and initial icing level through a DNN network, and snow thickness, environmental factors and initial icing level are usedEstablishing a mapping model, and then predicting the icing starting time t of the accumulated snow on the surface of each battery plate through the mapping model0And the icing grade K after a period of time, and adjusting the cleaning mode according to the icing grade K. The DNN network can be used for more accurately acquiring the thickness of the accumulated snow and the initial icing grade, a more rigorous mapping model can be established, and the result error is reduced; the icing grade K can be accurately acquired by utilizing the mapping model, so that the snow cleaning robot can adjust the cleaning mode in a self-adaptive manner, the aim of thorough cleaning is fulfilled, and the service life of the battery panel is prevented from being reduced.
Based on the same inventive concept as the method, the embodiment also provides a cleaning system for the icing of the photovoltaic cell panel based on artificial intelligence.
Referring to fig. 3, an embodiment of the present invention provides an artificial intelligence based photovoltaic panel icing cleaning system, including: an acquisition unit 10, a DNN network unit 20, a modeling unit 30, a prediction unit 40, and an adjustment unit 50.
The acquisition unit 10 is used for acquiring initial data of each battery panel by using the unmanned aerial vehicle, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
a DNN network unit 20, which is used for acquiring the snow thickness and the initial icing grade from the initial data through a DNN network;
the modeling unit 30 is used for establishing a mapping model by using the snow thickness and the environmental factors and the initial icing level;
a prediction unit 40 for inputting the snow thickness of each panel into the mapping model and predicting the snow icing starting time t of each panel surface0And an icing grade K after a period of time;
the adjusting unit 50 is used for adjusting cleaning modes according to the icing level K, wherein the cleaning modes comprise a snow cleaning mode and an ice cleaning mode;
the mapping model is:
Figure BDA0002693255770000081
Figure BDA0002693255770000082
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are constants; h represents the snow thickness of the surface of the battery plate, T represents the ambient temperature and M represents the ambient humidity;
further, referring to fig. 4, the DNN network unit 20 further includes a panel semantic segmentation network unit 21, a snow thickness detection network unit 22, and an icing level evaluation network unit 23.
The panel semantic segmentation network unit 21 is used for acquiring a semantic area of each panel;
the snow thickness detection network unit 22 is used for acquiring the snow thickness of each battery panel;
the icing level assessment network unit 23 is used to obtain an initial icing level for each panel.
Further, referring to fig. 5, the modeling unit 30 further includes an icing factor unit 31 and an ice melting factor unit 32.
The icing factor unit 31 is:
Figure BDA0002693255770000091
the ice-melting factor unit 32 is:
Figure BDA0002693255770000092
wherein Hn1The thickness of the snow cover of the panel n at the 1 st moment is shown; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hntThe accumulated snow thickness of the solar panel n at the t-th moment is represented;
Figure BDA0002693255770000093
is represented by [ a, t]Average temperature over time;
Figure BDA0002693255770000094
is represented by [ a, t]Average humidity over time;
Figure BDA0002693255770000095
is represented by [1, a ]]Average temperature over time;
Figure BDA0002693255770000096
is represented by [1, a ]]Average humidity over time; θ is a balance factor and is a constant.
Further, referring to fig. 6, the prediction unit 40 further includes a temporal prediction unit 41 and a level prediction unit 42.
The time prediction unit 41 is used for predicting the snow icing starting time t of the surface of each battery board0；
The level prediction unit 42 is used to predict the icing level K of the snow cover on the surface of each panel.
Further, referring to fig. 7, the adjusting unit 50 further includes a snow cleaning unit 51 and an ice cleaning unit 52.
The snow cleaning unit 51 is used for the time t<t0When K is equal to 0, adjusting to be in a snow clearing mode;
the ice-removing unit 52 is used for the time t>t0And adjusting the corresponding ice-clearing mode according to the value of K.
In summary, the present invention provides a cleaning system for icing on photovoltaic panels, which comprises an acquisition unit 10, a DNN network unit 20, a modeling unit 30, a prediction unit 40, and an adjustment unit 50. The DNN network unit 20 is used for acquiring the thickness of the accumulated snow and the initial icing grade, so that the accuracy of data is ensured, the rigor is brought to the establishment of a mapping model, and the error of a result is reduced; the icing grade K after a period of time can be accurately predicted through the prediction unit 40 and the adjusting unit 50, so that the snow cleaning robot can adaptively adjust the cleaning mode conveniently, and the reduction of the service life of the battery panel due to incomplete cleaning is prevented.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (10)

1. A cleaning method for icing of a photovoltaic cell panel based on artificial intelligence is characterized by comprising the following steps:
acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
acquiring the initial data by a DNN network to obtain the thickness of the accumulated snow and the initial icing grade;
establishing a mapping model by using the snow thickness and the environmental factors and the initial icing grade;
inputting the thickness of the accumulated snow of each battery plate into the mapping model, and predicting the time t for the accumulated snow on the surface of each battery plate to start to freeze0And an icing grade K after a period of time;
adjusting a cleaning mode according to the icing level K, wherein the cleaning mode comprises a snow cleaning mode and an ice cleaning mode;
the mapping model is as follows:
Figure FDA0002693255760000011
Figure FDA0002693255760000012
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are constants; h represents the snow thickness of the surface of the battery plate, T represents the ambient temperature and M represents the ambient humidity;
when t is<t0When K is equal to 0, adjusting to be in a snow clearing mode;
when t is>t0And adjusting the corresponding ice-clearing mode according to the value of K.
2. The method of claim 1, wherein the DNN network comprises a panel semantic segmentation network and a snow thickness detection network, the step of obtaining the snow thickness comprising:
utilizing the battery board semantic segmentation network to obtain a semantic area of the battery board;
combining the semantic area and the ultrasonic data to obtain an ultrasonic signal corresponding to each battery plate;
and inputting the semantic area of the battery plate, the battery plate surface image data and the corresponding ultrasonic signal into the snow thickness detection network to obtain the snow thickness of each battery plate.
3. The method of claim 2, wherein the DNN network further comprises an icing level assessment network, the step of obtaining the incipient icing level comprising:
and inputting the semantic area, the panel surface image data and the infrared image data of each panel into the icing grade evaluation network to obtain the initial icing grade of each panel.
4. The method of claim 1, wherein the icing factor TcThe calculation formula of (2):
Figure FDA0002693255760000013
wherein HnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hntThe accumulated snow thickness of the solar panel n at the t-th moment is represented;
Figure FDA0002693255760000021
is represented by [ a, t]Average temperature over time;
Figure FDA0002693255760000022
is represented by [ a, t]Average humidity over time; θ is a balance factor and is a constant.
5. Method according to claim 1 or 4, characterized in that said snow melting factor TbThe calculation formula of (2):
Figure FDA0002693255760000023
wherein Hn1The thickness of the snow cover of the panel n at the 1 st moment is shown; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown;
Figure FDA0002693255760000024
is represented by [1, a ]]Average temperature over time;
Figure FDA0002693255760000025
is represented by [1 ],a]Average humidity over time; θ is a balance factor and is a constant.
6. A cleaning system for photovoltaic cell panel icing based on artificial intelligence, the system comprising:
the acquisition unit is used for acquiring initial data of each battery panel, wherein the initial data are battery panel surface image data, infrared image data and ultrasonic data;
the DNN network unit is used for acquiring the snow thickness and the initial icing grade of the initial data through a DNN network;
the modeling unit is used for establishing a mapping model by utilizing the snow thickness and the environmental factors and the initial icing level;
a prediction unit for inputting the snow thickness of each panel into the mapping model and predicting the snow icing starting time t of each panel surface0And an icing grade K after a period of time; and
the adjusting unit is used for adjusting a cleaning mode according to the icing grade K, and the cleaning mode comprises a snow cleaning mode and an ice cleaning mode;
the mapping model is as follows:
Figure FDA0002693255760000026
Figure FDA0002693255760000027
wherein T is more than 0; t is0Is a constant and represents the optimal temperature of the battery plate when the battery plate works under the normal environment; t is t0Represents the time when the accumulated snow on the surface of the battery plate begins to freeze; t represents a time; t iscRepresents an icing factor; t isbRepresents a snow melting factor; alpha, beta and gamma are constants; h represents the snow thickness of the surface of the battery plate, T represents the ambient temperature and M represents the ambient humidity;
when t is<t0When K is equal to 0, the snow is clearedA mode;
when t is>t0And adjusting the corresponding ice-clearing mode according to the value of K.
7. The system of claim 6, wherein the DNN network unit further comprises:
the battery board semantic segmentation network unit is used for acquiring the semantic area of each battery board;
the snow thickness detection network unit is used for acquiring the snow thickness of each battery plate; and
and the icing grade evaluation network unit is used for acquiring the initial icing grade of each battery panel.
8. The system of claim 6, wherein the modeling unit further comprises:
an icing factor unit for acquiring the icing factor TcThe icing factor unit is as follows:
Figure FDA0002693255760000031
wherein HnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown; hntThe accumulated snow thickness of the solar panel n at the t-th moment is represented;
Figure FDA0002693255760000032
is represented by [ a, t]Average temperature over time;
Figure FDA0002693255760000033
is represented by [ a, t]Average humidity over time; θ is a balance factor and is a constant.
9. The system of claim 8, wherein the modeling unit further comprises:
an ice-melting factor unit for obtaining the ice-melting factor TbThe ice-melting factor unit is as follows:
Figure FDA0002693255760000034
wherein Hn1The thickness of the snow cover of the panel n at the 1 st moment is shown; hnaThe thickness of the accumulated snow of the battery plate n at the a-th moment is shown;
Figure FDA0002693255760000035
is represented by [1, a ]]Average temperature over time;
Figure FDA0002693255760000036
is represented by [1, a ]]Average humidity over time; θ is a balance factor and is a constant.
10. The system of claim 6, wherein the prediction unit further comprises:
a time prediction unit for predicting the icing start time t of the accumulated snow on the surface of each panel0；
A grade prediction unit for predicting the icing grade K of the accumulated snow on the surface of each panel.
CN202010997929.1A 2020-09-21 2020-09-21 Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence Withdrawn CN112257331A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010997929.1A CN112257331A (en) 2020-09-21 2020-09-21 Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010997929.1A CN112257331A (en) 2020-09-21 2020-09-21 Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence

Publications (1)

Publication Number Publication Date
CN112257331A true CN112257331A (en) 2021-01-22

Family

ID=74232730

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010997929.1A Withdrawn CN112257331A (en) 2020-09-21 2020-09-21 Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence

Country Status (1)

Country Link
CN (1) CN112257331A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114169229A (en) * 2021-11-24 2022-03-11 华能新能源股份有限公司 Data-driven identification method and device for optimizing cleaning time of photovoltaic array
CN114264259A (en) * 2021-04-02 2022-04-01 湖南国戎科技有限公司 Equivalent ice observation method and system
CN118508867A (en) * 2024-07-18 2024-08-16 江苏国强兴晟能源科技股份有限公司 Snow protection system of photovoltaic cell panel

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114264259A (en) * 2021-04-02 2022-04-01 湖南国戎科技有限公司 Equivalent ice observation method and system
CN114264259B (en) * 2021-04-02 2023-03-10 湖南国戎科技有限公司 Equivalent ice observation method and system
CN114169229A (en) * 2021-11-24 2022-03-11 华能新能源股份有限公司 Data-driven identification method and device for optimizing cleaning time of photovoltaic array
CN118508867A (en) * 2024-07-18 2024-08-16 江苏国强兴晟能源科技股份有限公司 Snow protection system of photovoltaic cell panel

Similar Documents

Publication Publication Date Title
CN112257331A (en) Cleaning method and system for icing of photovoltaic cell panel based on artificial intelligence
CN103453867B (en) Electric transmission line ice coating thickness monitoring method
EP3198311A1 (en) Detection unit and method for identifying and monitoring clouds in an observed area of the sky
CN110097787A (en) A kind of ship collision warning monitoring system and method based on monitoring navigation light
Dierer et al. Wind turbines in icing conditions: performance and prediction
CN100359339C (en) Method of determining the risk of ice deposition due to precipitation and apparatus for exercising the method
CN111042456B (en) Heating snow melting method, system and readable storage medium
CN108508372A (en) A kind of calculating of unmanned electricity and method for early warning and system based on environmental visual fusion
CN109104152A (en) Auto cleaning system and clean method for photovoltaic plant
US20200304058A1 (en) Smart shingles
JP6793508B2 (en) Insolation amount estimation device and solar radiation amount estimation method
CN113700231B (en) Highway photovoltaic ceiling system of intelligence snow removing
Zhu et al. Study on Image Recognition Algorithm for Residual Snow and Ice on Photovoltaic Modules.
CN117033935B (en) Prediction method of rainfall characteristic under statistics and monitoring based on Bayesian fusion
CN115454182B (en) Grain storage method, system, equipment and storage medium
CN117517335A (en) System and method for monitoring pollution of insulator of power transformation equipment
CN112452861B (en) Artificial intelligence-based ice and snow removal adjusting method and device for photovoltaic cleaning robot
JP3220606B2 (en) Transmission line snow damage alarm system
CN115014216A (en) Method and device for detecting icing of power transmission line
Wang et al. Feasibility Study for an Ice-Based Image Monitoring System for Polar Regions Using Improved Visual Enhancement Algorithms
JPH09166666A (en) Estimation method for freezing of road surface
CN114049755B (en) Environment testing system and method
CN114327914B (en) Mountain scenic spot mountain climbing decision-making method and system based on multi-factor edge calculation
CN118770129A (en) Intelligent water spray control wiper system, method and device
CN118469247B (en) Unmanned ship self-organizing replenishing method and system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20210122