A photovoltaic power station cleaning intelligent management system

Abstract

The application discloses a kind of photovoltaic power station cleaning intelligent management systems, it is related to photovoltaic power station cleaning technical field, the present application is periodically collected to the image data of all photovoltaic panel board surface in photovoltaic power station by setting inspection collection module, foreign matter identification module identifies and obtains foreign matter type and foreign matter mark image data in the inspection collection period, illumination prediction module obtains the illumination condition in future predicted cleaning duration;Dispatch allocation module allocates the most suitable photovoltaic cleaning robot quantity, makes full use of the illumination condition of photovoltaic power station to supplement the electric energy of photovoltaic cleaning robot in cleaning process, avoid the problem that photovoltaic cleaning robot endurance mileage is short, path generation optimization unit selects the optimal cleaning strategy based on path deviation duration and path foreign matter equilibrium value, avoid the situation that traditional shortest fixed rule based on cleaning duration does not reasonably cooperate photovoltaic cleaning robot photoelectric conversion capacity and self cleaning power consumption.

Description

Technical Field

[0001] This invention relates to the field of clean energy technology for photovoltaic power plants, and specifically to a clean and intelligent management system for photovoltaic power plants. Background Technology

[0002] In recent years, photovoltaic (PV) energy has gradually become a key player in the energy transition, and the PV industry has shown strong growth momentum. However, while a PV power plant can be built in just six months, its operational lifespan is as long as 25 years. During this operational lifespan, the efficiency of PV power generation largely depends on the cleanliness of the solar panels. Cleaning PV panels can increase power generation efficiency by 20%-30%, but due to natural factors, air pollution, and human factors, outdoor PV modules gradually become covered with "pollutants," reducing their power generation efficiency. To improve the power generation level throughout the entire lifespan of a PV power plant, outdoor PV modules urgently need periodic "major cleaning."

[0003] Currently, for cleaning photovoltaic (PV) panels in photovoltaic (PV) power plants, PV cleaning robots are used. These robots have pre-stored special calculation algorithms that can calculate the shortest and most efficient routes, ensuring that all PV panels within the power plant can be cleaned within their range. For cleaning large numbers of PV panels, multiple PV cleaning robots are used simultaneously. The current method of selecting the number of PV cleaning robots, considering the number of PV cleaning robots in a PV power plant, also allocates as many robots as possible based on the shortest and most efficient routes. However, this allocation is unreasonable. It does not take into account the superior sunlight conditions of the PV power plant and the photoelectric conversion capabilities of the PV cleaning robots themselves. This leads to the unreasonable utilization of cleaning resources in the PV power plant. Furthermore, this route selection does not reasonably coordinate the photoelectric conversion capabilities of the PV cleaning robots with their cleaning time. Therefore, calculating the cleaning path based on the range of the PV cleaning robots without considering their photoelectric conversion capabilities is unreasonable, resulting in low utilization efficiency of the PV cleaning robots and consequently affecting the cleaning efficiency of the PV panels.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] The purpose of this invention is to provide a smart management system for cleaning photovoltaic power plants. This system addresses the problem in existing technologies where the allocation of photovoltaic cleaning robots for cleaning photovoltaic power plants is unreasonable. It fails to consider the superior sunlight conditions of the photovoltaic power plant and the photoelectric conversion capabilities of the cleaning robots themselves. Furthermore, based on this path selection, the photoelectric conversion capabilities of the cleaning robots and their cleaning time are not properly coordinated. Consequently, calculating the cleaning path based on the range of the photovoltaic cleaning robots without taking into account their photoelectric conversion capabilities is unreasonable, resulting in low utilization efficiency of the cleaning robots and consequently affecting the cleaning efficiency of the photovoltaic panels.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A clean and intelligent management system for photovoltaic power plants includes:

[0008] The inspection and data acquisition module collects image data of all photovoltaic panels in the photovoltaic power station every inspection and data acquisition cycle and generates panel image data for that inspection and data acquisition cycle. The image data of the photovoltaic panel contains the serial number of the photovoltaic panel, which is a string of data or letter combination that uniquely identifies the photovoltaic panel.

[0009] The foreign object identification module is used to identify foreign objects in the photovoltaic panel image data. The foreign object identification module stores a photovoltaic panel foreign object identification model. After receiving the transmitted panel image data for the current inspection and collection cycle, the foreign object identification module inputs it into the photovoltaic panel foreign object identification model to identify foreign objects and obtain panel identification data for the current inspection and collection cycle. The panel identification data for the current inspection and collection cycle includes the serial number of all photovoltaic panels with foreign objects on their surfaces in the photovoltaic power station during the current inspection and collection cycle, the foreign object labeling image data of all photovoltaic panels with foreign objects on their surfaces, and their corresponding foreign object types. Foreign objects on the photovoltaic panel surface are marked with red circles in the photovoltaic panel foreign object labeling image data.

[0010] The foreign object identification module stores the preset basic cleaning time corresponding to all foreign object types. The foreign object identification module calculates the predicted cleaning time A1 of the inspection and collection cycle based on the preset basic cleaning time corresponding to all foreign object types stored in it and the number of red circles corresponding to each foreign object type in the foreign object labeling image data of each photovoltaic panel carried in the panel identification data of the inspection and collection cycle.

[0011] The solar illumination prediction module is used to predict the solar illumination intensity and duration of a photovoltaic power station. The solar illumination prediction module stores a solar illumination prediction model. The solar illumination prediction module collects meteorological data of the location of the photovoltaic power station in real time and inputs it into the solar illumination prediction model to obtain the solar illumination prediction data of the photovoltaic power station within the future A1 time period. The solar illumination prediction data of the photovoltaic power station within the future A1 time period includes the effective solar illumination time and average solar illumination intensity within this A1 time period.

[0012] The cleaning management module is used to schedule and manage the cleaning of photovoltaic panels in a photovoltaic power station. The cleaning management module includes a scheduling and allocation unit and a path generation and selection unit.

[0013] The scheduling and allocation unit calculates and obtains the optimal number of photovoltaic cleaning robots E1 for the current inspection and collection cycle based on the panel identification data of the inspection and collection cycle and the photovoltaic power station's illumination prediction data within the next A1 time period, according to certain calculation and allocation rules.

[0014] The path generation and selection unit generates all cleaning strategies based on the optimal allocation of the number of photovoltaic cleaning robots E1 in the current inspection and collection cycle, the panel recognition data, and the photovoltaic power station's illumination prediction data within the next A1 time period, based on the principle of avoiding collisions and repeated cleaning during collaborative work. A cleaning strategy based on the current inspection and collection cycle includes cleaning paths for E1 photovoltaic cleaning robots. A cleaning path for a photovoltaic cleaning robot includes several cleaning nodes, and each cleaning node corresponds to a photovoltaic panel with foreign objects on its surface.

[0015] The path generation unit selects the optimal cleaning strategy for the inspection and collection cycle based on all cleaning strategies according to certain screening and evaluation rules.

[0016] The path generation and selection unit, based on the optimal cleaning strategy of the inspection and collection cycle, for each photovoltaic cleaning robot cleaning path corresponding to the photovoltaic panel with foreign objects on its surface, inputs the serial number of all photovoltaic panels with foreign objects on their surface, the foreign object labeling image data of all photovoltaic panels with foreign objects on their surface, and their corresponding foreign object types into the corresponding photovoltaic cleaning robot in the panel identification data of the inspection and collection cycle. The corresponding photovoltaic cleaning robot then starts cleaning the surface of the corresponding photovoltaic panel in the photovoltaic power station according to its corresponding cleaning path.

[0017] Furthermore, the serial number of the photovoltaic panel consists of 14 digits.

[0018] Furthermore, the interval between one inspection and collection cycles is P1, where P1 is a preset time threshold.

[0019] Furthermore, the photovoltaic panel foreign object identification model is used to identify and label foreign objects on the surface of the photovoltaic panel.

[0020] Furthermore, the sunshine prediction model is used to predict the sunshine duration and intensity of the photovoltaic power station in the future based on meteorological data of the photovoltaic power station's location.

[0021] Furthermore, the types of foreign objects are classified based on their sources. The types of foreign objects present on the photovoltaic panel surface are classified as dust and particulate matter, plant seedlings, oil stains or tape residue, bird droppings or insects.

[0022] Furthermore, the scheduling and allocation unit calculates the optimal allocation rule for the number of photovoltaic cleaning robots E1 in the current inspection and collection cycle as follows:

[0023] S11: Obtain the number of photovoltaic panels with foreign objects on their surfaces in the panel identification data of this inspection collection cycle, and mark it as B1;

[0024] Obtain the effective illumination time and average illumination intensity within the future A1 time period, and label them as C1 and C2 respectively;

[0025] S12: Calculate the optimal number of photovoltaic cleaning robots E1 for the current inspection and collection cycle using the equation A1×α1+B1×α2=E1*(100-α3+C1×C2×η1×S1). Here, α1 is the preset baseline cleaning power, α2 is the preset energy consumption of a single baseline path, S1 is the surface area of ​​the photovoltaic panels placed on the photovoltaic cleaning robot, η1 is the photoelectric conversion efficiency of the photovoltaic cleaning robot, and α3 is the preset amount of adjustable energy reserved for idle power of the photovoltaic cleaning robot.

[0026] Furthermore, the specific screening and evaluation rules for the path generation and selection unit to obtain the optimal cleaning strategy for the inspection and collection cycle based on all cleaning strategies of the inspection and collection cycle are as follows:

[0027] S21: Sequentially label all cleaning strategies based on this inspection collection cycle as F1, F2, ..., Ff, where f≥1;

[0028] S22: Calculate and obtain the path deviation duration M1 corresponding to the cleaning strategy F1 of the inspection collection cycle according to certain calculation rules;

[0029] S23: Calculate and obtain the path foreign object balance value U1 corresponding to cleaning strategy F1 according to certain calculation rules;

[0030] S24: Calculate the path rating score V1 corresponding to cleaning strategy F1 using the formula V1=M1 / U1;

[0031] S25: Calculate and obtain the path evaluation scores V1, V2, ..., Vf corresponding to cleaning strategies F1, F2, ..., Ff in sequence according to S21 to S25, and obtain the cleaning strategy corresponding to the maximum value Vmax, and recalibrate it as the optimal cleaning strategy for this inspection and collection cycle.

[0032] Furthermore, in step S22, the specific calculation rules for obtaining the path deviation duration M1 corresponding to the cleaning strategy F1 of the inspection collection cycle are as follows:

[0033] S221: Obtain the path travel time G1 of a photovoltaic cleaning robot cleaning path in cleaning strategy F1, and obtain the type of foreign object and the number of red circles corresponding to each type of foreign object in the foreign object annotation image data of each photovoltaic panel containing foreign objects in the cleaning path of the photovoltaic cleaning robot.

[0034] S222: Obtain the type of foreign object in the image data of the photovoltaic panel surface containing foreign objects in the cleaning path of the photovoltaic cleaning robot, and label it as H1, H2, ..., Hh, where h≥1;

[0035] Obtain the number of red circles corresponding to foreign object types H1, H2, ..., Hh in the foreign object annotation image data of the photovoltaic panel, and label them as I1, I2, ..., Ih respectively;

[0036] S223: Utilize formula The path time J1 corresponding to the cleaning path of the photovoltaic cleaning robot is calculated using g = 1, 2, ..., h, where β1, β2, ..., βg are the preset cleaning times corresponding to foreign object types H1, H2, ..., Hh, respectively.

[0037] S224: Calculate and obtain the path consumption times J1, J2, ..., JE1 corresponding to all photovoltaic cleaning robot cleaning paths in cleaning strategy F1 in sequence according to S221 to S223;

[0038] S225: Using the formula Calculate the dispersion K1 of the path consumption times J1, J2, ..., JE1 corresponding to all photovoltaic cleaning robot cleaning paths in cleaning strategy F1, compare K1 and K, where J is the mean of Jk at this time, and K is a preset deviation threshold.

[0039] If K1≥K, then select the corresponding Jk in descending order of |Jk - J| one by one. After each selection, delete it, recalculate the scatter difference K1 of the remaining Jk, and compare the sizes of K1 and K again until K1 < K. At this time, recalibrate the mean value of the remaining Jk as the average path duration corresponding to all the cleaning paths of the photovoltaic cleaning robots in the cleaning strategy F1, and mark it as L1;

[0040] S226: Calculate and obtain the path deviation duration M1 corresponding to the cleaning strategy F1 using the formula M1 = |L1 - C1|.

[0041] Furthermore, for the above S23, the specific calculation rule for calculating and obtaining the path foreign object balance value U1 corresponding to the cleaning strategy F1 is as follows:

[0042] S231: Obtain the foreign object types that exist in all the foreign object annotation image data of the photovoltaic panels included in the cleaning paths of all the photovoltaic cleaning robots in the cleaning strategy F1, and mark them as N1, N2,..., Nn, where n≥1;

[0043] S232: Obtain the other foreign object types except N1, N2,..., Nn in the foreign object annotation image data of the photovoltaic panels included in the cleaning path of one photovoltaic cleaning robot in the cleaning strategy F1, and mark them as O1, O2,..., Oo, where o≥1, and obtain the number of red circles marked in the foreign object annotation image data of the photovoltaic panels included in the cleaning path of this photovoltaic cleaning robot corresponding to them, and mark them as Q1, Q2,..., Qo;

[0044] S233: Obtain the number of red circles marked in the foreign object annotation image data of the photovoltaic panels included in the cleaning path of this photovoltaic cleaning robot corresponding to the foreign object types N1, N2,..., Nn in the foreign object annotation image data of the photovoltaic panels included in the cleaning path of this photovoltaic cleaning robot in the cleaning strategy F1, and mark them as R1, R2,..., Rn;

[0045] S234: Use the formula where t = 1, 2,..., n and q = 1, 2,..., o to calculate and obtain the path foreign object characteristic value T1 of the cleaning path of this photovoltaic cleaning robot in the cleaning strategy F1. It should be noted here that the path foreign object characteristic value is defined artificially to represent the foreign object type quantity characteristics of the photovoltaic panel surface foreign object annotation image data corresponding to each path node in this path. The μ1 and μ2 are respectively the proportion coefficients of different foreign object types and the same foreign object types, and the λ1, λ2,..., λn are respectively the within-group proportion coefficients of different foreign object types, and the δ1, δ2,..., δq are respectively the within-group proportion coefficients of the same foreign object types;

[0046] S235: Calculate the path debris characteristic values ​​T1, T2, ..., TE1 of all photovoltaic cleaning robot cleaning paths in cleaning strategy F1 according to S231 to S234, and calculate their mean using the summation and averaging formula. Mark the mean value as the path debris equilibrium value corresponding to cleaning strategy F1, and label it as U1.

[0047] The beneficial effects of this invention are:

[0048] (1) This invention sets up an inspection and acquisition module to periodically collect image data of all photovoltaic panels in the photovoltaic power station. The foreign object identification module performs foreign object identification on the panel image data during the inspection and acquisition cycle to obtain the type of foreign object and the foreign object label image data during the inspection and acquisition cycle, and obtains the predicted cleaning time of the inspection and acquisition cycle based on it. The light prediction module obtains the average light intensity and effective light time within the predicted cleaning time. The scheduling and allocation module allocates the most suitable number of photovoltaic cleaning robots for the inspection and acquisition cycle based on it. On the one hand, it makes full use of the light conditions of the photovoltaic power station to replenish the photovoltaic cleaning robots during the cleaning process, avoiding the problem of short driving range of photovoltaic cleaning robots, improving the utilization efficiency of photovoltaic cleaning robots, and further improving the efficiency of photovoltaic panel cleaning. On the other hand, it avoids the problem of unreasonable utilization of cleaning resources caused by unreasonable allocation of the number of photovoltaic cleaning robots.

[0049] (2) This invention sets up a path generation optimization unit to allocate the most suitable number of photovoltaic cleaning robots based on the inspection and collection cycle and the foreign object labeling image data corresponding to the inspection and collection cycle. Based on the principle of avoiding collisions and repeated cleaning during collaborative work, it generates multiple cleaning strategies for the inspection and collection cycle. Based on the path deviation time and path foreign object balance value in the cleaning path of each photovoltaic cleaning robot in the cleaning strategy, it calculates and obtains the path evaluation score corresponding to the cleaning path of each photovoltaic cleaning robot. Based on the path evaluation score, the optimal cleaning strategy is selected, which avoids the situation where the traditional rule of selecting the shortest cleaning time does not reasonably coordinate the photovoltaic cleaning robot's photoelectric conversion capability and its own cleaning power consumption. Attached Figure Description

[0050] The invention will now be further described with reference to the accompanying drawings.

[0051] Figure 1 This is a system block diagram of the present invention;

[0052] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] like Figure 1 As shown in Figure 2, a photovoltaic power station clean intelligent management system includes an inspection and data collection module, a foreign object identification module, a light prediction module, and a cleaning management module.

[0055] The inspection and data acquisition module collects image data of all photovoltaic panels within the photovoltaic power station at regular intervals and generates panel image data for that inspection and data acquisition cycle. The module then transmits this panel image data to the foreign object identification module. It should be noted that the image data of the photovoltaic panels includes the serial number of the photovoltaic panel. The serial number is a string of data or letter combinations that uniquely identifies the photovoltaic panel. In this embodiment, the serial number consists of 14 digits. The interval between one inspection and data acquisition cycle is P1, where P1 is a preset time threshold.

[0056] The foreign object identification module is used to identify foreign objects in the photovoltaic panel image data. The module pre-stores a photovoltaic panel foreign object identification model, which is used to identify and label foreign objects on the photovoltaic panel surface. After receiving the panel image data for the current inspection and collection cycle from the inspection and collection module, the module inputs it into the photovoltaic panel foreign object identification model to obtain panel identification data for that inspection and collection cycle. This data includes the serial numbers of all photovoltaic panels with foreign objects on their surfaces within the photovoltaic power station during that cycle, the labeled image data of the foreign objects on all photovoltaic panels with foreign objects, and the corresponding foreign object types. The foreign object types are categorized based on their source. In this embodiment, the types of foreign objects on the photovoltaic panel surface are classified as dust and particulate matter, plant seedlings, oil stains or tape residue, bird droppings, or insects. Foreign objects on the photovoltaic panel surface are marked with red circles in the labeled image data.

[0057] The foreign object identification module stores the preset basic cleaning time corresponding to all foreign object types. The foreign object identification module calculates the predicted cleaning time of the inspection and collection cycle based on the preset basic cleaning time corresponding to all foreign object types stored in it and the number of red circles corresponding to each foreign object type in the foreign object labeling image data of each photovoltaic panel carried in the panel identification data of the inspection and collection cycle, and marks it as A1.

[0058] The foreign object identification model generates a prediction acquisition instruction based on the predicted cleaning time of the inspection and collection cycle and transmits it to the illumination prediction module.

[0059] The solar illumination prediction module is used to predict the solar intensity and duration of solar power stations. The solar illumination prediction module has a pre-stored solar illumination prediction model, which is used to predict the solar intensity and duration of solar power stations. The solar illumination prediction module collects meteorological data of the location of the solar power station in real time. The meteorological data of the location of the solar power station includes wind speed, wind direction, temperature, humidity, solar radiation intensity, illuminance, sunshine hours, atmospheric pressure and rainfall.

[0060] After receiving the prediction acquisition instruction transmitted by the foreign object identification module, the illumination prediction module inputs the meteorological data of the photovoltaic power station location acquired in real time into the illumination prediction model to obtain the illumination prediction data of the photovoltaic power station within the future A1 time period. The illumination prediction data of the photovoltaic power station within the future A1 time period includes the effective illumination time and average illumination intensity within this A1 time period.

[0061] The illumination prediction module transmits the illumination prediction data of the photovoltaic power plant within the future A1 time period to the clean management module.

[0062] The cleaning management module is used to schedule and manage the cleaning of photovoltaic panels in a photovoltaic power station. The cleaning management module includes a scheduling and allocation unit and a path generation and selection unit.

[0063] The cleaning management module receives the panel identification data of the inspection collection cycle transmitted by the foreign object identification module, and after receiving the photovoltaic power station's illumination prediction data for the future A1 time period transmitted by the illumination prediction module, it transmits them to the scheduling and allocation unit and the path generation and selection unit respectively.

[0064] After receiving the panel identification data for the current inspection and collection cycle and the photovoltaic power station's irradiance prediction data for the next A1 time period from the cleaning management module, the scheduling and allocation unit calculates the optimal number of photovoltaic cleaning robots E1 for the current inspection and collection cycle according to a certain calculation and allocation rule, as follows:

[0065] S11: Obtain the number of photovoltaic panels with foreign objects on their surfaces in the panel identification data of this inspection collection cycle, and mark it as B1;

[0066] Obtain the effective illumination time and average illumination intensity within the future A1 time period, and label them as C1 and C2 respectively;

[0067] S12: Calculate the optimal number of photovoltaic cleaning robots E1 for the current inspection and collection cycle using the formula A1×α1+B1×α2=E1*(100-α3+C1×C2×η1×S1), where α1 is the preset baseline cleaning power, α2 is the preset energy consumption of a single baseline path, S1 is the surface area of ​​the photovoltaic panels placed on the photovoltaic cleaning robot, η1 is the photoelectric conversion efficiency of the photovoltaic cleaning robot, and α3 is the preset amount of adjustable energy reserved for idle power of the photovoltaic cleaning robot.

[0068] The scheduling and allocation unit transmits the optimal number of photovoltaic cleaning robots E1 for the current inspection and collection cycle to the path generation and selection unit. Upon receiving the optimal number of photovoltaic cleaning robots E1 from the scheduling and allocation unit, and after receiving the panel identification data for the current inspection and collection cycle and the photovoltaic power station's illumination prediction data for the next A1 time period from the cleaning management module, the path generation and selection unit generates all cleaning strategies based on the principles of avoiding collisions and redundant cleaning during collaborative work. A cleaning strategy for this inspection and collection cycle contains cleaning paths for E1 photovoltaic cleaning robots. A cleaning path for a photovoltaic cleaning robot contains several cleaning nodes, each corresponding to a photovoltaic panel with foreign objects on its surface.

[0069] The scheduling and allocation unit and the path generation unit select the optimal cleaning strategy for the inspection and collection cycle based on all cleaning strategies according to certain screening and evaluation rules. The specific screening and evaluation rules are as follows:

[0070] S21: Sequentially label all cleaning strategies based on this inspection collection cycle as F1, F2, ..., Ff, where f≥1;

[0071] S22: Calculate the path deviation duration M1 corresponding to the cleaning strategy F1 for this inspection collection cycle according to certain calculation rules. The specific steps are as follows:

[0072] S221: Obtain the path travel time G1 of a photovoltaic cleaning robot cleaning path in cleaning strategy F1, and obtain the type of foreign object and the number of red circles corresponding to each type of foreign object in the foreign object annotation image data of each photovoltaic panel containing foreign objects in the cleaning path of the photovoltaic cleaning robot.

[0073] S222: Obtain the type of foreign object in the image data of the photovoltaic panel surface containing foreign objects in the cleaning path of the photovoltaic cleaning robot, and label it as H1, H2, ..., Hh, where h≥1;

[0074] Obtain the number of red circles corresponding to foreign object types H1, H2, ..., Hh in the foreign object labeled image data on the photovoltaic panel surface, and mark them as I1, I2, ..., Ih respectively;

[0075] S223: Use the formula g = 1, 2, ..., h to calculate and obtain the path consumption time J1 corresponding to the cleaning path of the photovoltaic cleaning robot, where β1, β2, ..., βg are the preset cleaning times corresponding to foreign object types H1, H2, ..., Hh respectively;

[0076] S224: Calculate and obtain the path consumption times J1, J2, ..., JE1 corresponding to the cleaning paths of all photovoltaic cleaning robots in cleaning strategy F1 in sequence according to S221 to S223;

[0077] S225: Use the formula Calculate and obtain the scatter K1 of the path consumption times J1, J2, ..., JE1 corresponding to the cleaning paths of all photovoltaic cleaning robots in cleaning strategy F1, and compare the magnitudes of K1 and K. Here, J is the mean value of Jk at this time, and K is the preset deviation threshold;

[0078] If K1 ≥ K, then select the corresponding Jk in descending order of |Jk - J| one by one, delete each selected one, recalculate the scatter K1 of the remaining Jk after deletion, and compare the magnitudes of K1 and K again until K1 < K. Then, re - calibrate the mean value of the remaining Jk at this time as the average path time corresponding to the cleaning paths of all photovoltaic cleaning robots in cleaning strategy F1, and mark it as L1;

[0079] S226: Use the formula M1 = |L1 - C1| to calculate and obtain the path deviation time M1 corresponding to cleaning strategy F1;

[0080] S23: Calculate and obtain the path foreign object balance value U1 corresponding to cleaning strategy F1 according to a certain calculation rule, specifically as follows:

[0081] S231: Obtain the foreign object types that exist in all the foreign object labeled image data on the photovoltaic panel surfaces included in the cleaning paths of all photovoltaic cleaning robots in cleaning strategy F1, and mark them as N1, N2, ..., Nn, where n ≥ 1;

[0082] S232: Obtain the other foreign object types except N1, N2, ..., Nn in the foreign object labeled image data on the photovoltaic panel surfaces included in the cleaning path of one photovoltaic cleaning robot in cleaning strategy F1, and mark them as O1, O2, ..., Oo, where o ≥ 1, and obtain the number of red circles marked corresponding to them in all the foreign object labeled image data on the photovoltaic panel surfaces included in the cleaning path of this photovoltaic cleaning robot, and mark them as Q1, Q2, ..., Qo;

[0083] S233: Obtain the number of red circles marked in all photovoltaic panel surface foreign object annotation image data in the cleaning path of the photovoltaic cleaning robot in the cleaning strategy F1, corresponding to the foreign object types N1, N2, ..., Nn. These numbers are labeled as R1, R2, ..., Rn.

[0084] S234: Utilize formula For t = 1, 2, ..., n, and q = 1, 2, ..., o, calculate and obtain the path foreign object feature value T1 of the cleaning path of the photovoltaic cleaning robot in cleaning strategy F1. It should be noted that the path foreign object feature value is artificially defined to represent the quantity feature of foreign object types in the photovoltaic panel surface foreign object annotation image data corresponding to each path node in the path. μ1 and μ2 are the preset proportion coefficients of different foreign object types and the same foreign object type, respectively. λ1, λ2, ..., λn are the preset proportion coefficients within the group of different foreign object types, respectively. δ1, δ2, ..., δq are the preset proportion coefficients within the group of the same foreign object type, respectively.

[0085] S235: Calculate and obtain the path foreign object characteristic values ​​T1, T2, ..., TE1 of all photovoltaic cleaning robot cleaning paths in cleaning strategy F1 according to S231 to S234, and calculate their mean value using the summation and averaging formula. Mark the mean value as the path foreign object equilibrium value corresponding to cleaning strategy F1, and label it as U1.

[0086] S24: Calculate the path rating score V1 corresponding to cleaning strategy F1 using the formula V1=M1 / U1;

[0087] S25: Calculate and obtain the path evaluation scores V1, V2, ..., Vf corresponding to cleaning strategies F1, F2, ..., Ff in sequence according to S21 to S25, and obtain the cleaning strategy corresponding to the maximum value Vmax, and recalibrate it as the optimal cleaning strategy for this inspection and collection cycle;

[0088] The path generation and selection unit, based on the cleaning strategy corresponding to Vmax, for each photovoltaic cleaning robot cleaning path corresponding to the photovoltaic panel with foreign objects on the photovoltaic panel, inputs the serial number of all photovoltaic panels with foreign objects on the photovoltaic panel with foreign objects, the foreign object labeling image data of all photovoltaic panels with foreign objects on the photovoltaic panel with foreign objects, and their corresponding foreign object types into the corresponding photovoltaic cleaning robot according to its corresponding cleaning path.

[0089] In the description of this specification, the references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0090] The above description is merely an example and illustration of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the invention or exceed the scope defined in the claims, they should all fall within the protection scope of the present invention.

[0091] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A photovoltaic power station cleaning intelligent management system, characterized in that, The method comprises the following steps: A patrol collection module collects image data of all photovoltaic panel surfaces in a photovoltaic power station every other patrol collection period and generates panel surface image data of the patrol collection period according to the image data, the image data of the photovoltaic panel surface contains the serial number of the photovoltaic panel, and the serial number of the photovoltaic panel is a string of data or a combination of letters that uniquely identifies the photovoltaic panel; An alien object identification module is used to identify alien objects on the panel surface image data, the alien object identification module stores a photovoltaic panel alien object identification model, and after receiving the transmitted panel surface image data of the patrol collection period, the alien object identification module inputs the panel surface image data into the photovoltaic panel alien object identification model to identify alien objects and obtain panel identification data of the patrol collection period, the panel identification data of the patrol collection period contains the serial numbers of all photovoltaic panels with alien objects on the panel surfaces in the photovoltaic power station in the patrol collection period, panel surface alien object labeled image data of all photovoltaic panels with alien objects on the panel surfaces, and the corresponding alien object types, and the alien objects present on the panel surface in the panel surface alien object labeled image data are labeled with a red circle; The alien object identification module stores preset basic cleaning times corresponding to all alien object types, and the alien object identification module calculates and obtains a predicted cleaning duration A1 of the patrol collection period according to the preset basic cleaning times corresponding to all alien object types stored in the alien object identification module and the number of red circles corresponding to each alien object type in each panel surface alien object labeled image data carried in the panel identification data of the patrol collection period; An illumination prediction module is used to predict the illumination intensity and illumination time of the photovoltaic power station, the illumination prediction module stores an illumination prediction model, the illumination prediction module collects meteorological data of the location of the photovoltaic power station in real time and inputs the meteorological data into the illumination prediction model to obtain illumination prediction data of the photovoltaic power station in the future A1 time, the illumination prediction data of the photovoltaic power station in the future A1 time includes the effective illumination time and the average illumination intensity in the A1 time; A cleaning management module is used to schedule and manage the cleaning of the photovoltaic panel surfaces of the photovoltaic power station, the cleaning management module comprises a scheduling allocation unit and a path generation selection unit; The scheduling allocation unit calculates and obtains the most suitable number E1 of distributed photovoltaic cleaning robots in the current patrol collection period according to the panel identification data of the patrol collection period and the illumination prediction data of the photovoltaic power station in the future A1 time; The path generation selection unit generates all cleaning strategies based on the patrol collection period based on the principle of avoiding collision and repeated cleaning when working cooperatively according to the most suitable number E1 of distributed photovoltaic cleaning robots in the current patrol collection period, the panel identification data, and the illumination prediction data of the photovoltaic power station in the future A1 time, and each cleaning strategy based on the patrol collection period contains the cleaning path of E1 photovoltaic cleaning robots; the cleaning path of one photovoltaic cleaning robot contains a plurality of cleaning nodes, and each cleaning node corresponds to a photovoltaic panel with an alien object on the panel surface. The path generation unit filters all cleaning strategies based on the inspection collection period to obtain an optimal cleaning strategy of the inspection collection period; The path generation selection unit inputs the serial numbers of all photovoltaic panels with foreign matter on the panel surface, the panel surface foreign matter annotation image data of all photovoltaic panels with foreign matter on the panel surface, and the corresponding foreign matter types in the panel surface identification data of the inspection collection period into the corresponding photovoltaic cleaning robot according to the photovoltaic panels with foreign matter on the panel surface corresponding to each photovoltaic cleaning robot cleaning path in the optimal cleaning strategy of the inspection collection period, and the corresponding photovoltaic cleaning robot starts cleaning the panel surface of the corresponding photovoltaic panel in the photovoltaic power station according to the corresponding cleaning path. The scheduling allocation unit calculates a specific calculation allocation rule of the current optimal allocation of photovoltaic cleaning robots E1 in the inspection collection period as follows: S11: Obtain the number of all photovoltaic panels with foreign matter on the panel surface in the panel surface identification data of the inspection collection period, and mark it as B1; Obtain the effective illumination time and the average illumination intensity in the future A1 time, and mark them as C1 and C2, respectively. S12: Utilize the equation Calculate the optimal number of photovoltaic cleaning robots E1 for the current inspection cycle, where α1 is the preset reference cleaning power, α2 is the preset power consumption per unit of reference path, S1 is the surface area of the photovoltaic panel mounted on the photovoltaic cleaning robot, η1 is the photoelectric conversion efficiency of the photovoltaic cleaning robot, and α3 is the preset number of adjustment powers for the idle power of the photovoltaic cleaning robot.

2. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, The path generation selection unit filters all cleaning strategies based on the inspection collection period to obtain an optimal cleaning strategy of the inspection collection period, and the specific filtering evaluation rule is as follows: S21: Mark all cleaning strategies based on the inspection collection period as F1, F2,..., Ff, f≥1 in turn; S22: Calculate the path deviation duration M1 corresponding to the cleaning strategy F1 of the inspection collection period; S23: Calculate the path foreign matter balance value U1 corresponding to the cleaning strategy F1; S24: Calculate the path evaluation score V1 corresponding to the cleaning strategy F1 using the formula V1=M1 / U1; S25: Calculate the path evaluation scores V1, V2,..., Vf corresponding to the cleaning strategies F1, F2,..., Ff in turn according to S21 to S25, and obtain the maximum value Vmax corresponding to the cleaning strategy, and re-mark it as the optimal cleaning strategy of the inspection collection period; The specific calculation rule of calculating the path deviation duration M1 corresponding to the cleaning strategy F1 of the inspection collection period in S22 is as follows: S221: Obtain the path travel duration G1 of one photovoltaic cleaning robot cleaning path in the cleaning strategy F1, and obtain the foreign matter types present in the foreign matter annotation image data of each photovoltaic panel with foreign matter on the panel surface and the number of red circles corresponding to each foreign matter type; S222: Obtain the foreign matter types present in the foreign matter annotation image data of one photovoltaic panel with foreign matter on the panel surface in the cleaning path of the photovoltaic cleaning robot, and mark them as H1, H2,..., Hh, h≥1; Obtain the number of red circles corresponding to the foreign matter types H1, H2,..., Hh in the foreign matter annotation image data of the photovoltaic panel, and mark them as I1, I2,..., Ih, respectively. S223: calculate the path cost time length J1 corresponding to the cleaning path of the photovoltaic cleaning robot by using the formula , g = 1, 2,..., h, wherein β1, β2,..., βg are respectively preset cleaning time lengths corresponding to the foreign matter types H1, H2,..., Hh. S224: Calculate the path time lengths J1, J2,..., JE1 corresponding to all photovoltaic cleaning robot cleaning paths in the cleaning strategy F1 in turn according to S221 to S223; S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula S225: using the formula < If K1≥K, then select the corresponding Jk in turn according to the order from large to small of |Jk-J|, delete each selected one, recalculate the dispersion K1 of the remaining Jk after deletion, and compare K1 and K again, until K1<K, and then re-label the average of the remaining Jk as the average path time length of all photovoltaic cleaning robot cleaning paths in the cleaning strategy F1, and mark it as L1; S226: Calculate the path deviation time length M1 corresponding to the cleaning strategy F1 using the formula M1=|L1-C1|; The specific calculation rule of the path foreign matter balance value U1 corresponding to the cleaning strategy F1 calculated in S23 is as follows: S231: Obtain the foreign matter types existing in all photovoltaic panel surface foreign matter labeled image data contained in all photovoltaic cleaning robot cleaning paths in the cleaning strategy F1, and mark them as N1, N2,..., Nn, n≥1; S232: Obtain other foreign matter types in all photovoltaic panel surface foreign matter labeled image data contained in one photovoltaic cleaning robot cleaning path in the cleaning strategy F1 except N1, N2,..., Nn, mark them as O1, O2,..., Oo, o≥1, and obtain the number of red circles marked in all photovoltaic panel surface foreign matter labeled image data contained in the photovoltaic cleaning robot cleaning path, mark them as Q1, Q2,..., Qo; S233: Obtain the number of red circles marked in all photovoltaic panel surface foreign matter labeled image data contained in the photovoltaic cleaning robot cleaning path in the cleaning strategy F1 corresponding to the foreign matter types N1, N2,..., Nn in the photovoltaic cleaning robot cleaning path, mark them as R1, R2,..., Rn; S234: calculating the path foreign matter characteristic value T1 of the photovoltaic cleaning robot cleaning path in the cleaning strategy F1 by using the formula , t = 1, 2,..., n, q = 1, 2,..., o, where it is to be noted that the path foreign matter characteristic value is artificially defined to represent the number of foreign matter types in the foreign matter labeling image data corresponding to each path node in the path, and μ1, μ2 are respectively the proportion coefficients of different foreign matter types and the same foreign matter type, λ1, λ2,..., λn are respectively the intra-group proportion coefficients of different foreign matter types, and δ1, δ2,..., δq are respectively the intra-group proportion coefficients of the same foreign matter type. S235: Calculate the path foreign matter characteristic values T1, T2,..., TE1 of all photovoltaic cleaning robot cleaning paths in the cleaning strategy F1 according to S231 to S234, and calculate the average value using the sum average formula, and label the average value as the path foreign matter balance value corresponding to the cleaning strategy F1, and mark it as U1.

3. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, The serial number of the photovoltaic panel is composed of 14 digits.

4. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, One said inspection collection interval P1 time, said P1 is a preset time threshold.

5. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, The photovoltaic panel foreign matter recognition model is used to recognize and label the foreign matter on the photovoltaic panel surface.

6. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, The light prediction model is used to predict the light time and light intensity of the future photovoltaic power station based on the meteorological data of the photovoltaic power station location.

7. The intelligent management system for cleaning photovoltaic power stations according to claim 1, characterized in that, The foreign matter types are classified based on their sources, and the foreign matter types existing on the photovoltaic panel surface are divided into dust and particulate matter, plant seedlings, oil stains or adhesive residue, and bird droppings or insects.

Images