Operation decision-making method for highway photovoltaic panel cleaning equipment

Abstract

The invention provides an expressway photovoltaic panel cleaning equipment operation decision-making method, which comprises the following steps of: firstly, judging whether the current-day energy storage residual power generation amount is smaller than the total power consumption of photovoltaic panel cleaning facilities or not, if so, acquiring a photovoltaic panel geographical deployment basic data set and a photovoltaic-load time sequence interaction historical data set of a county-level expressway photovoltaic energy system, and if not, acquiring the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time sequence interaction historical data set of the county-level expressway photovoltaic energy system; performing subsequent operation on the basis of the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time sequence interaction historical data set so as to determine a micro-grid pre-scheduling scheme taking the highest cleaning income as an objective function under different weathers of the next day; and randomly determining whether the photovoltaic panel needs to be cleaned in the next day according to the micro-grid pre-scheduling scheme and the occurrence probabilities of different weathers. By optimizing the cleaning time period and utilizing photovoltaic spontaneous power, the cleaning energy consumption can be reduced, the power generation loss can be reduced, the power generation efficiency can be improved, and precise intelligent maintenance can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of photovoltaic, in particular to a highway photovoltaic panel cleaning equipment operation decision method. BACKGROUND

[0002] With the development of highway intelligence and greenization, photovoltaic panels are widely used in service areas, sound barriers and roadside power generation systems to provide clean energy for lighting, monitoring and other facilities. However, photovoltaic panels are exposed to complex highway environments for a long time, and the surface is easy to accumulate dust, exhaust particles, snow-melting agent residues and other pollutants, resulting in a 15%-40% decrease in power generation efficiency. Existing photovoltaic panel cleaning mainly relies on manual cleaning, which has the problems of high labor consumption, low efficiency and difficulty in maintaining the best power generation state of photovoltaic panels. Especially in deserts or remote areas, manual cleaning is costly and cannot achieve high-frequency maintenance. Studies have shown that regular cleaning can increase power generation by 20%-30%, so relevant research focuses on the design of cleaning facilities. The document "A photovoltaic panel cleaning device" relates to a photovoltaic panel cleaning device, which includes a mounting assembly, a cleaning assembly and a limiting assembly, aiming to solve the problem of dust accumulation on the surface of photovoltaic panels and improve power generation efficiency; the document "A solar photovoltaic panel cleaning vehicle" describes a solar photovoltaic panel cleaning vehicle suitable for photovoltaic panel cleaning. Existing cleaning facilities mostly use a monthly fixed cleaning electricity strategy, only considering the cleaning equipment power consumption, without dynamically adjusting the cleaning frequency combined with environmental conditions, resulting in a conflict between the cleaning period and the photovoltaic power generation peak, and the technical problems of energy waste and power generation efficiency loss. SUMMARY

[0003] In order to overcome the above technical defects, the present application provides a highway photovoltaic panel cleaning equipment operation decision method, in order to achieve the above purpose, the present application realizes the following technical scheme:

[0004] The present application provides a highway photovoltaic panel cleaning equipment operation decision method, comprising:

[0005] Step S101: obtaining the remaining power generation capacity of the energy storage for the day;

[0006] Step S102: determining whether the remaining power generation capacity of the energy storage for the day is less than the total power consumption of the photovoltaic panel cleaning facility, the total power consumption of the photovoltaic panel cleaning facility being the total power consumption for completing a cleaning operation on all photovoltaic panels in the highway area;

[0007] If yes, step S103 is executed;

[0008] Step S103: Collecting a photovoltaic panel geographical deployment basic data set and a photovoltaic-load time series interaction historical data set of a county-level expressway photovoltaic energy system, wherein the photovoltaic-load time series interaction historical data set includes photovoltaic historical output data and historical load power consumption data of different periods;

[0009] Based on the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time series interaction historical data set, the maximum output value of the photovoltaic panel under different weather conditions in a plurality of historical days is determined.

[0010] Step S104: Obtaining real-time photovoltaic output efficiency data in a plurality of historical days;

[0011] Based on the maximum output value of the photovoltaic panel under different weather conditions in a plurality of historical days and the real-time photovoltaic output efficiency data in a plurality of historical days, the output prediction value of the photovoltaic panel under different weather conditions in the next day is determined.

[0012] Step S105: Obtaining the occurrence probability of different weather conditions in the next day;

[0013] Based on the output prediction value and the load power consumption data, a micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions in the next day is determined.

[0014] Step S106: Based on the micro-grid pre-scheduling scheme and the occurrence probability, it is determined whether the photovoltaic panel needs to be cleaned in the next day.

[0015] Optionally, the historical load power consumption data is subject to a Gaussian process f(GP):

[0016]

[0017] In the formula, P L_u is an uncertain load output prediction; is an uncertain load output historical data; and ε is an independent and identically distributed noise variable, and the load mean value is 0 and the Gaussian distribution variance is σ2.

[0018] Optionally, the photovoltaic panel geographical deployment basic data set includes photovoltaic panel geographical information, and the real-time photovoltaic output efficiency data in a plurality of historical days is obtained in the following manner:

[0019] Based on the photovoltaic panel geographical information, a photovoltaic panel single-day inclination adjustment instruction is generated.

[0020] Based on the photovoltaic panel single-day inclination adjustment instruction, photovoltaic panel single-day real-time output efficiency data is determined.

[0021] Based on the photovoltaic panel single-day real-time output efficiency data, the real-time photovoltaic output efficiency data in a plurality of historical days is determined.

[0022] Optionally, the output prediction value of the uncleaned photovoltaic panel under different weather conditions of the next day is determined based on the historical maximum output value of the photovoltaic panel under different weather conditions of the plurality of days and the historical real-time photovoltaic output efficiency data of the plurality of days, and the output prediction value of the uncleaned photovoltaic panel under different weather conditions of the next day is determined based on the historical maximum output value of the photovoltaic panel under different weather conditions of the plurality of days and the historical real-time photovoltaic output efficiency data of the plurality of days, comprising:

[0023] The average value of the output efficiency of the photovoltaic panel of each day in the historical plurality of days is determined based on the historical real-time photovoltaic output efficiency data of the plurality of days;

[0024] The output prediction value of the uncleaned photovoltaic panel under different weather conditions of the next day is determined by training a neural network with the historical maximum output value of the photovoltaic panel under different weather conditions of the plurality of days and the average value of the output efficiency of the photovoltaic panel of each day in the historical plurality of days as training samples:

[0025]

[0026] In the formula, is the output prediction value of the uncleaned photovoltaic panel under the k weather condition of the next day; α is an activation function Sigmoid; is the maximum output value of the photovoltaic panel under the k weather condition in the historical plurality of days; η n is the average value of the output efficiency of the photovoltaic panel of each day in the historical plurality of days; b is a bias vector; w is a hidden layer; l is the number of hidden layers of the neural network; k is the weather type, η max is the theoretical maximum efficiency of the photovoltaic panel;

[0027] The network error function M is:

[0028]

[0029] In the formula, d represents the expected output of the neural network; t represents time.

[0030] Optionally, the micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions of the next day is determined based on the output prediction value and the load power consumption data, and the micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions of the next day is determined based on the output prediction value and the load power consumption data, comprising:

[0031]

[0032] E represents the highest cleaning benefit of the micro-grid scheduling, i.e. the objective function, T1 is the starting time of the photovoltaic panel cleaning device, T2 represents the ending time of the photovoltaic panel cleaning device, is the total power generation of the photovoltaic panel before cleaning today, is the power generation of the photovoltaic panel after cleaning, represents the power consumption of the charging pile and other uncertain loads under the k weather condition at t time of the next day; P L_c represents the power consumption of the sensor and other fixed loads at t time; P L_t represents the power consumption of the photovoltaic panel cleaning device adjustable load at t time.

[0033] Optionally, the related necessary constraints of the objective function of the highest cleaning benefit include the following:

[0034] The energy constraints of the photovoltaic panel cleaning device tonight, the energy constraints of the cleaning device required tomorrow, the power constraints of the cleaning device tomorrow, the total power balance constraints, and the energy storage device constraints.

[0035] Optionally, the determination of whether to clean the photovoltaic panel the next day based on the micro-grid pre-scheduling scheme and the occurrence probability includes:

[0036] Based on the micro-grid pre-scheduling scheme, the probability of the cleaning facility being put into operation the next day under different weather conditions is determined.

[0037]

[0038] In the formula, P(k) represents the probability of the cleaning facility being put into operation the next day under a certain weather condition, P(k) represents the power required for the cleaning facility being put into operation under a certain weather condition;

[0039] Based on the probability of the cleaning facility being put into operation the next day under different weather conditions and the occurrence probability, it is determined whether to clean the photovoltaic panel the next day.

[0040]

[0041] In the formula, p represents the total probability of the cleaning facility being put into operation the next day; p k P(k) represents the occurrence probability of different weather conditions;

[0042] P(k) represents the probability of the cleaning facility being put into operation the next day under a certain weather condition;

[0043] If p is higher than 50%, the photovoltaic panel is cleaned the next day.

[0044] The present application has the following beneficial effects:

[0045] The method provided by the present application can reduce the cleaning energy consumption, reduce the loss of power generation capacity, improve the power generation efficiency, and realize precise and intelligent maintenance by optimizing the cleaning period and utilizing the photovoltaic self-generation power.

[0046] In addition to the purposes, features and advantages described above, the present application has other purposes, features and advantages. The present application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0047] The accompanying drawings, which form a part of the present application, are intended to provide further understanding of the present application, and the illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute improper limitations on the present application. In the drawings:

[0048] Figure 1 is a flowchart of a highway photovoltaic panel cleaning device operation decision method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0049] The embodiments of the present application are described in detail below with reference to the accompanying drawings, but the present application can be implemented in various different ways as limited and covered by the claims.

[0050] Therefore, in order to solve the above problems, as shown in the present application, a highway photovoltaic panel cleaning device operation decision method is proposed, comprising: Figure 1

[0051] Step S101: Obtain the daily energy storage remaining available power generation;

[0052] The daily energy storage remaining available power generation refers to the total amount of electrical energy that the energy storage system can theoretically release to the grid or load within the remaining time period of the day based on the current remaining capacity, system efficiency and operation limit. After the photovoltaic panel works every day, the daily energy storage remaining available power generation needs to be obtained according to the subsequent application.

[0053] Step S102: Determine whether the daily energy storage remaining available power generation is less than the total power consumption of the photovoltaic panel cleaning facility, and the total power consumption of the photovoltaic panel cleaning facility is the total power consumption for completing a cleaning operation on all photovoltaic panels in the highway area;

[0054] If yes, step S103 is executed.

[0055] Since the photovoltaic panel is beside the highway, the photovoltaic panel may form a covering due to various reasons such as vehicle emissions and weather, and the increased covering will affect the efficiency of photovoltaic power generation. After the photovoltaic panel performs photoelectric conversion every day, the photovoltaic panels in the highway area are generally cleaned the next day. The power required for cleaning also needs to consider whether the actual stored power meets the total power required for cleaning the photovoltaic panels. Therefore, first, it is necessary to determine whether the daily energy storage remaining available power generation is less than the total power consumption of the photovoltaic panel cleaning facility. If it is less, it means that the dispatching scheme needs to be executed, and step S103 can be executed. If not, it means that the daily energy storage remaining available power generation is sufficient to clean all photovoltaic panels in the area, and the cleaning operation on all photovoltaic panels can be performed tonight.

[0056] Step S103: Collect the photovoltaic panel geographic deployment basic data set and the photovoltaic-load time sequence interaction historical data set of the county-level highway photovoltaic energy system, and the photovoltaic-load time sequence interaction historical data set includes photovoltaic historical output data and each load historical power consumption data in different periods;

[0057] ​determine historical multi-day different weather photovoltaic maximum output value based on the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time series interaction historical data set;

[0058] At this time, the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time series interaction historical data set of the county-level highway photovoltaic energy system are collected.

[0059] The photovoltaic panel geographical deployment basic data set of the county-level highway photovoltaic energy system generally includes photovoltaic panel parameters and geographical information, and the photovoltaic panel parameters generally include output upper limit, theoretical maximum efficiency, etc.

[0060] The photovoltaic-load time series interaction historical data set includes photovoltaic historical output data and historical power consumption data of each load in different periods, and the historical power consumption data of each load includes historical data of non-adjustable loads such as sensors and uncertain loads such as charging piles, and power consumption demand of adjustable loads such as photovoltaic panel cleaning facilities.

[0061] Then, the historical multi-day different weather photovoltaic maximum output value is determined based on the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time series interaction historical data set. Since the historical data has reference value, the corresponding historical multi-day different weather photovoltaic maximum output value can be found one by one based on the photovoltaic panel geographical deployment basic data set and the photovoltaic-load time series interaction historical data. For example, the photovoltaic maximum output value in overcast weather is 30, and the photovoltaic maximum output value in sunny weather is 80 in a certain area of the highway.

[0062] The historical power consumption data of each load generally conforms to a Gaussian process f(GP):

[0063]

[0064] In the formula, P L_u is the uncertain load output prediction; is the uncertain load output historical data; ε is an independent and identically distributed noise variable, and the load mean is 0 and the variance is σ 2 Gaussian distribution.

[0065] Step S104: Obtain historical multi-day real-time photovoltaic output efficiency data;

[0066] Determine the output prediction value of the uncleaned photovoltaic panel under different weather conditions the next day based on the historical multi-day different weather photovoltaic maximum output value and the historical multi-day real-time photovoltaic output efficiency data.

[0067] The historical multi-day real-time photovoltaic output efficiency data refers to the dynamic change data of the power generation efficiency recorded in real time in units of minutes or hours in continuous multi-day operation of the photovoltaic system.

[0068] The generation process of the historical multi-day real-time photovoltaic output efficiency data is as follows:

[0069] Firstly, the geographical information of the photovoltaic panel is acquired, and then the solar azimuth angle a and the solar elevation angle b are calculated through the geographical information and the time, so as to generate the target inclination angle θt of the photovoltaic panel. Based on the difference Δθ between the current inclination angle θc and the target inclination angle θt of the photovoltaic panel, the pulse control signal of the servo driving mechanism, i.e. the single-day inclination adjustment instruction of the photovoltaic panel, is generated through the PID control algorithm. Then, the single-day real-time output efficiency data η of the photovoltaic panel can be obtained by calculating the single-day inclination adjustment instruction of the photovoltaic panel, i.e.

[0070] η = η max cos (θ c - θ t) (2) max 2 r t

[0071] In the formula, η_max is the theoretical maximum efficiency of the photovoltaic panel.

[0072] The single-day real-time processing efficiency data of the photovoltaic panel in the past multi-day is combined, so as to obtain the historical multi-day real-time photovoltaic output efficiency data. Then, the historical multi-day real-time photovoltaic output efficiency data and the historical multi-day are divided, so as to obtain the average value of the output efficiency of the photovoltaic panel in each day of the historical multi-day.

[0073] Then, the maximum output value of the photovoltaic panel under different weather conditions in the historical multi-day and the average value of the output efficiency of the photovoltaic panel in each day of the historical multi-day are taken as the training samples to train the neural network, so as to determine the output prediction value of the uncleaned photovoltaic panel under different weather conditions of the next day:

[0074]

[0075] In the formula, is the output prediction value of the uncleaned photovoltaic panel under the k weather condition of the next day; a is the activation function Sigmoid; is the maximum output value of the photovoltaic panel under the k weather condition of the historical multi-day; η n is the average value of the output efficiency of the photovoltaic panel in each day of the historical n-day; b is the bias vector; w is the hidden layer; l is the number of hidden layers of the neural network; k is the weather type, η max is the theoretical maximum efficiency of the photovoltaic panel;

[0076] The network error function M is:

[0077]

[0078] In the formula, d represents the expected output of the neural network; t represents the time.

[0079] Step S105: Obtain the occurrence probability of different weather of the next day; ​​​​

[0080] Based on the output prediction value and the load power consumption data, a micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions the next day is determined;

[0081] Since weather prediction is based on certain phenomena, the resulting weather result is only the probability of a certain weather occurring, and does not mean that a certain weather will occur 100%. The probability will be described in detail later. Here, no calculation is performed.

[0082] The goal of cleaning photovoltaic panels is naturally to achieve the highest cleaning benefit. Therefore, based on the output prediction value and the load power consumption data, a micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions the next day is determined;

[0083]

[0084] E represents the maximum cleaning benefit of micro-grid scheduling, i.e., the objective function, T1 is the starting time of the photovoltaic panel cleaning device, T2 represents the end time of the photovoltaic panel cleaning device, is the total power generation of the photovoltaic panel before cleaning today, is the power generation of the cleaned photovoltaic panel, represents the power consumption of the charging pile and other uncertain loads at t time under k weather conditions the next day; P L_c (t) represents the power consumption of the sensor and other fixed loads at t time; P L_t (t) represents the power consumption of the photovoltaic panel cleaning device adjustable load at t time.

[0085] In practical applications, some variables of the objective function have corresponding constraint conditions, and the constraint conditions corresponding to formula (5) include the following:

[0086] Tonight's photovoltaic panel cleaning device energy constraint:

[0087] ΔE L_t '≤ΔE E -ΔE L_c -ΔE L_u (6)

[0088] In the formula, ΔE L_t ' represents the energy used by the photovoltaic panel cleaning device tonight; ΔE E represents the remaining disposable energy of the energy storage tonight; ΔE L_c represents the power consumption of the sensor and other fixed loads tonight; ΔE L_u represents the power consumption of the charging pile and other uncertain loads tonight.

[0089] Tomorrow's required cleaning device energy constraint:

[0090] ΔEL_t = ΔE L_t '-(ΔE E -ΔE L_c -ΔE L_u ) (7)

[0091] In the formula, Δ E L_t represents the energy required by the photovoltaic panel cleaning device the next day.

[0092] Next-day cleaning device power constraints:

[0093]

[0094] In the formula, P L_t (t) represents the power consumption of the photovoltaic panel cleaning device adjustable load at time t;

[0095] Total power balance constraint:

[0096]

[0097] In the formula, represents the output power of the energy storage device at time t; represents the power consumption of the uncertain load such as charging piles at time t; represents the power consumption of the uncertain load such as charging piles at time t under the k-day weather condition; P L_c (t) represents the power consumption of the fixed load such as sensors at time t; P L_t (t) represents the power consumption of the photovoltaic panel cleaning device adjustable load at time t; represents the input power of the energy storage device at time t.

[0098] Energy storage device constraints:

[0099]

[0100] The input and output states of the energy storage device cannot exist at the same time, so the input and output powers of the energy storage device at each time should be multiplied by 0.

[0101] Step S106: Based on the microgrid pre-scheduling scheme and the occurrence probability, determine whether the photovoltaic panel cleaning is to be performed the next day.

[0102] First, according to the microgrid pre-scheduling scheme, the probability of putting the cleaning facility into operation under different weather the next day is determined, that is, the power consumption of the photovoltaic panel cleaning device adjustable load at time t:

[0103]

[0104] In the formula, represents the probability of putting the cleaning facility into operation under a certain weather the next day, represents the amount of electricity required to put the cleaning facility into operation under a certain weather.

[0105] Then, according to the probability of putting the cleaning facility into operation under different weather of the next day and the probability of occurrence of different weather of the next day, it is determined whether the cleaning of the photovoltaic panel is to be performed tomorrow according to the calculation result, that is:

[0106]

[0107] In the formula, p represents the total probability of dispatching the cleaning facility tomorrow; p k represents the probability of occurrence of different weather; represents the probability of putting the cleaning facility into operation under k type weather of the next day;

[0108] If p is higher than 50%, the photovoltaic panel is cleaned tomorrow.

[0109] That is, if the probability of putting the cleaning facility into operation under different weather of the next day is obtained, and then the probability of different weather of the next day is obtained, they are multiplied respectively, and then the results are added, if the added value is 50% or more, it is determined that the dispatching scheme is to be performed tomorrow, if not, the dispatching scheme is not performed tomorrow, and the remaining electricity is reserved for use when it is determined whether the dispatching scheme is to be performed the day after tomorrow.

[0110] In summary, the method provided in the application can reduce the cleaning energy consumption, reduce the loss of power generation, improve the power generation efficiency, and realize precise and intelligent maintenance by optimizing the cleaning period and using the photovoltaic self-generated power.

[0111] The above only describes the preferred embodiments of the application and is not used to limit the application. For those skilled in the art, the application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A method for operating a highway photovoltaic panel cleaning apparatus, characterized in that, The method comprises the following steps: Step S101: obtaining the remaining available power generation of energy storage on the current day; Step S102: determining whether the remaining available power generation of energy storage on the current day is less than the total power consumption of the photovoltaic panel cleaning device, wherein the total power consumption of the photovoltaic panel cleaning device is the total power consumption for completing one cleaning operation on all photovoltaic panels in the expressway area; If yes, step S103 is performed; Step S103: collecting the photovoltaic panel geographic deployment basic data set and the photovoltaic-load time series interaction historical data set of the county-level expressway photovoltaic energy system, wherein the photovoltaic-load time series interaction historical data set comprises photovoltaic historical output data and load historical power consumption data in different periods; Based on the photovoltaic panel geographic deployment basic data set and the photovoltaic-load time series interaction historical data set, the maximum photovoltaic output value under different weather conditions in the historical days is determined; Step S104: obtaining historical multi-day real-time photovoltaic output efficiency data; Based on the maximum photovoltaic output value under different weather conditions in the historical days and the historical multi-day real-time photovoltaic output efficiency data, the output prediction value of the uncleaned photovoltaic panel under different weather conditions on the next day is determined; Step S105: obtaining the occurrence probability of different weather on the next day; Based on the output prediction value and the load power consumption data, the micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions on the next day is determined; Step S106: determining whether the photovoltaic panel cleaning on the next day is to be performed based on the micro-grid pre-scheduling scheme and the occurrence probability.

2. The method of claim 1, wherein, The load historical power consumption data is subject to a Gaussian process f(GP): wherein P L_u is the uncertain load power prediction; is the uncertain load power history data; ε is an independent and identically distributed noise variable with mean 0 and variance σ 2 Gaussian distribution.

3. The method of claim 1, wherein, The photovoltaic panel geographic deployment basic data set comprises photovoltaic panel geographic information, and the historical multi-day real-time photovoltaic output efficiency data is obtained in the following manner: Based on the photovoltaic panel geographic information, a photovoltaic panel single-day inclination adjustment instruction is generated; Based on the photovoltaic panel single-day inclination adjustment instruction, a photovoltaic panel single-day real-time output efficiency data is determined; Based on the photovoltaic panel single-day real-time output efficiency data, historical multi-day real-time photovoltaic output efficiency data is determined.

4. The method of claim 1, wherein, The determination of the output prediction value of the uncleaned photovoltaic panel under different weather conditions on the next day based on the maximum photovoltaic output value under different weather conditions in the historical days and the historical multi-day real-time photovoltaic output efficiency data comprises: Based on the historical multi-day real-time photovoltaic output efficiency data, the average value of the photovoltaic panel output efficiency in the historical days is determined; The maximum photovoltaic output value under different weather conditions in the historical days and the average value of the photovoltaic panel output efficiency in the historical days are used as training samples to train a neural network, so as to determine the output prediction value of the uncleaned photovoltaic panel under different weather conditions on the next day: wherein, is the output prediction value of the uncleaned photovoltaic panel under the next day k weather condition; a is the activation function Sigmoid; is the historical maximum output value of the photovoltaic panel under the historical multi-day k weather; η n is the average value of the output efficiency of the photovoltaic panel in the historical n days; b is the bias vector; w is the hidden layer; l is the number of hidden layers of the neural network; k is the weather type, η max is the theoretical maximum efficiency of the photovoltaic panel; The network error function M is: In the formula, d represents the expected output of the neural network, and t represents time.

5. The method of claim 4, wherein, The determination of the micro-grid pre-scheduling scheme with the highest cleaning benefit as the objective function under different weather conditions on the next day based on the output prediction value and the load power consumption data comprises: E represents the maximum clean benefit of the micro-grid scheduling, i.e. the objective function, T1 is the starting time of the photovoltaic panel cleaning device, T2 represents the ending time of the photovoltaic panel cleaning device, is the total power generation of the photovoltaic panel before cleaning today, is the power generation of the cleaned photovoltaic panel, represents the power consumption of the uncertain load such as the charging pile at time t under the k-day weather condition; P L_c (t) represents the power consumption of the sensor and other fixed load at time t; P L_t (t) represents the power consumption of the photovoltaic panel cleaning device adjustable load at time t.

6. The method of claim 5, wherein, The related necessary constraint conditions with the highest cleaning benefit as the objective function comprise the following: Tonight photovoltaic panel cleaning device energy constraint, next day required cleaning device energy constraint, next day cleaning device power constraint, total power balance constraint, energy storage device constraint.

7. The method of claim 5, wherein, The method comprises the following steps: The method comprises the following steps: wherein P (weather tomorrow = weather today + 1) represents the probability of a certain weather tomorrow given the weather today, P (energy needed for cleaning facilities tomorrow = weather today + 1) represents the energy needed for cleaning facilities tomorrow given the weather today. The method comprises the following steps: where p represents the total probability of dispatching to the cleaning facility the next day; p k represents the probability of occurrence of different weather; represents the probability of dispatching to the cleaning facility the next day under k type of weather; If p is higher than 50%, the photovoltaic panel is cleaned the next day.

Images