Expressway photovoltaic system energy efficiency management method based on cloud computing

Abstract

The invention discloses an expressway photovoltaic system energy efficiency management method based on cloud computing, and relates to the technical field of photovoltaic power generation. The method is executed by a cloud computing platform, historical operation, weather and traffic data are obtained firstly, the daily power generation loss rate caused by photovoltaic panel pollution is accurately quantified accordingly, engineering characteristics related to pollution accumulation are extracted, then historical characteristics and the loss rate serve as training samples, a machine learning prediction model is constructed and periodically updated, and the power generation capacity of the photovoltaic panel is predicted. Future weather and traffic prediction data are utilized, corresponding features are extracted and input into the model, the power generation loss rate caused by pollution in several days in the future is predicted, and finally on the premise of judging that no effective rainfall cleaning exists recently, if the predicted loss rate mean value exceeds a threshold value, a photovoltaic panel cleaning task list is automatically generated and issued to an operation and maintenance terminal to be executed. Scientific evaluation, intelligent prediction and active operation and maintenance of the pollution loss of the highway photovoltaic system can be realized, and the system energy efficiency and the operation and maintenance economy are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power generation technology, specifically relating to a cloud computing-based energy efficiency management method for highway photovoltaic systems. Background Technology

[0002] With the widespread application of renewable energy, constructing distributed photovoltaic (PV) systems along highways (such as service areas, tollbooth roofs, and slopes) has become an important measure to achieve green and low-carbon operation of transportation infrastructure. However, the unique environmental characteristics of highways pose a severe challenge to the energy efficiency degradation of their PV systems. Exhaust emissions from continuous vehicle traffic, tire wear particles, and road dust easily adhere to the surface of PV panels, forming a dense layer of dirt that is difficult to remove by natural wind, severely blocking sunlight and causing a significant decrease in the conversion efficiency of the PV panels—the so-called "pollution-induced power generation loss." Compared to ordinary PV power plants, highway PV panels experience faster pollution rates and more complex compositions (often mixed with oily substances), making their energy efficiency management issues particularly prominent.

[0003] Currently, the operation and maintenance management of photovoltaic panel pollution mainly relies on traditional methods. One is periodic cleaning and maintenance, which involves scheduling manual cleaning operations according to fixed time cycles (such as quarterly or semi-annually). This method lacks precise perception of the rate of pollution accumulation and actual power generation loss, potentially leading to unnecessary cleaning when pollution is not yet severe, resulting in resource waste; or failing to intervene in time after pollution has caused significant power generation losses, resulting in economic losses. The second is responsive cleaning based on simple meteorological thresholds, such as assuming that natural precipitation has completed part of the cleaning effect after monitoring a certain amount of rainfall. However, this method ignores the significant differences in the cleaning effect of different rainfall intensities on oil stains and adhesive dust, and completely ignores the impact of traffic flow, a key pollution source factor, resulting in low decision-making accuracy. Third, although some photovoltaic monitoring systems can monitor power generation decline, they typically only issue fault alarms, making it difficult to distinguish whether the power decline is caused by equipment failure, shading, or panel contamination, and even less capable of proactively predicting future energy efficiency losses caused by pollution.

[0004] In addition, it is worth noting that in order to solve the problem of photovoltaic panel cleaning, some photovoltaic systems equipped with automatic cleaning components (such as robotic sweepers or automatic sprinkler systems) have also appeared on the market. However, these solutions mainly focus on the "automatic execution" of the cleaning operation itself, and do not fundamentally solve the problem of "intelligent decision-making" for cleaning. They usually adopt timed start or simple triggering mechanisms, and the cleaning action is seriously out of touch with the actual energy efficiency status of the system, the economic value of pollution loss, and future weather changes. Specifically, this manifests as: (1) blind execution, that is, regardless of whether the photovoltaic panels are dirty enough to affect power generation, cleaning may be started regularly, consuming water and electricity resources and causing equipment wear and tear; (2) delayed response, that is, only passive response after pollution has formed and caused power generation loss, unable to achieve "preventive" or "loss minimization" proactive intervention; (3) poor economic efficiency, that is, cleaning operations may be arranged during periods of low power generation revenue (such as night or rainy days), or before the upcoming effective rainfall, causing unnecessary waste of resources and opportunity costs. Therefore, existing simple automated cleaning equipment has not improved the overall economic efficiency and intelligence level of energy efficiency management.

[0005] In summary, existing technologies for managing the energy efficiency of highway photovoltaic systems generally suffer from lag, blind spots, and inefficiency. Their core deficiency lies in the lack of a central energy efficiency management system capable of integrating multi-source dynamic data to accurately quantify current pollution losses and make forward-looking intelligent decisions based on economic forecasts. Therefore, the industry urgently needs an energy efficiency management method that can accurately assess, intelligently predict, and proactively optimize pollution cleanup strategies for highway photovoltaic systems to improve system power generation efficiency, reduce operation and maintenance costs, and ensure a return on investment. Summary of the Invention

[0006] The purpose of this invention is to provide a cloud computing-based method, device, and computer program product for managing the energy efficiency of highway photovoltaic systems, in order to solve the problems of lag, blindness, and inefficiency that are common in the existing technologies for managing the energy efficiency of highway photovoltaic systems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a cloud-based energy efficiency management method for highway photovoltaic systems is provided, executed by a cloud computing platform, including: The system acquires the operational data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the same M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the same M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified cleanliness reference date of the photovoltaic panels, and M represents a positive integer greater than or equal to 2. For each day in the continuous M days, based on the operating data and meteorological data of the corresponding day, determine the power generation loss rate of the highway photovoltaic system on the corresponding day and due to photovoltaic panel pollution. Based on the meteorological data and traffic data for the consecutive M days, engineering features related to the photovoltaic panel pollution accumulation phenomenon for each day are extracted; Using the engineering characteristics of each day as model input and the power generation loss rate of each day as model output, a power generation loss prediction model based on machine learning algorithm is trained or periodically updated. Based on the weather forecast data for the area where the highway photovoltaic system is located for the next N days and the traffic forecast data for the road section where the highway photovoltaic system is located for the next N days, new engineering features related to the accumulation of photovoltaic panel pollution for the next N days are extracted and imported into the power generation loss prediction model. The output is the power generation loss rate of the highway photovoltaic system due to photovoltaic panel pollution for the next N days, where N represents a positive integer. Based on the meteorological forecast data for the next N days, determine whether there are any effective days for cleaning up solar panels with rainfall within the next N days. If not, then determine whether the average power generation loss rate of the highway photovoltaic system in the current near future N consecutive days due to photovoltaic panel pollution exceeds a preset threshold. If the number exceeds the limit, a photovoltaic panel cleaning task order will be generated for the aforementioned highway photovoltaic system. The photovoltaic panel cleaning task order is sent to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the cleaning task of the photovoltaic panels of the highway photovoltaic system.

[0008] Based on the above-mentioned invention, a new energy efficiency management solution is provided that can integrate multi-source dynamic data to accurately quantify current pollution losses and make forward-looking intelligent decisions based on economic predictions. This solution is executed by a cloud computing platform, first acquiring historical operational, meteorological, and traffic data to accurately quantify the daily power generation loss rate caused by photovoltaic panel pollution, and extracting engineering features related to pollution accumulation. Then, using historical features and loss rates as training samples, a machine learning prediction model is constructed and periodically updated. Furthermore, future meteorological and traffic prediction data are used to extract corresponding features and input them into the model to predict the power generation loss rate caused by pollution in the coming days. Finally, assuming no effective rainfall for cleaning in the near future, if the average predicted loss rate exceeds a threshold, a photovoltaic panel cleaning task order is automatically generated and sent to the operation and maintenance terminal for execution. This enables scientific assessment, intelligent prediction, and proactive operation and maintenance of pollution losses in highway photovoltaic systems, significantly improving system energy efficiency and operation and maintenance economics, and facilitating practical application and promotion.

[0009] In one possible design, for each day within the M consecutive days, based on the operational data and meteorological data for the corresponding day, the power generation loss rate of the highway photovoltaic system due to photovoltaic panel pollution on the corresponding day is determined, including: For a specific day within the M consecutive days, based on the operational data of the corresponding day, the actual power generation and electrical characteristics of the highway photovoltaic system on the corresponding day are extracted, and based on the meteorological data of the corresponding day, multiple meteorological parameters for calculating the theoretical power generation of the highway photovoltaic system on the corresponding day are extracted, wherein the multiple meteorological parameters include solar irradiance, ambient temperature and wind speed. Based on the electrical characteristics, the power generation loss rate of the highway photovoltaic system on a certain day due to a fault is identified and quantified. Based on the aforementioned meteorological parameters, the theoretical power generation of the highway photovoltaic system on a certain day is calculated. Based on the theoretical and actual power generation of the highway photovoltaic system on a certain day, the power generation loss rate of the highway photovoltaic system on that day is calculated, and the power generation loss rate caused by the failure on that day is subtracted to obtain the power generation loss rate of the highway photovoltaic system on the corresponding day caused by photovoltaic panel pollution.

[0010] In one possible design, the solar irradiance is obtained by the following steps: Multiple solar irradiance influencing parameters are extracted from meteorological data, including cloud conditions, cloud cover, cloud height, visibility, and humidity. Based on the aforementioned multiple solar irradiance influence parameters, and combined with the known mapping relationship between the geographical location and date of the highway photovoltaic system and the solar altitude angle, the solar irradiance is estimated using a validated atmospheric radiative transfer model.

[0011] In one possible design, the engineering features include the cumulative number of days from the most recent effective day for cleaning up solar panels due to rainfall, the cumulative traffic volume, the cumulative fine particulate matter concentration, and the cumulative number of dust storm events, wherein the most recent effective day for cleaning up solar panels due to rainfall refers to the most recent rainfall day where rainfall exceeds a rainfall threshold, as determined based on meteorological data.

[0012] In one possible design, the rainfall threshold is obtained in advance according to the following steps: Based on the power generation loss rate of the highway photovoltaic system during the consecutive M days and due to photovoltaic panel pollution, the change in loss rate for each non-first day during the consecutive M days is calculated. For each non-first day, if the corresponding change in loss rate is lower than a preset negative threshold, then the corresponding day will be regarded as a candidate effective cleaning day. For each of the candidate clean effective days, based on the power generation loss rate of the highway photovoltaic system during the M consecutive days due to photovoltaic panel pollution, the average power generation loss rate for the K most recent consecutive days prior to the corresponding day is calculated. and the average power generation loss rate over the most recent K consecutive days after the corresponding date And calculate the corresponding degree of recovery. , where K represents a positive integer greater than or equal to 2; For each of the candidate clean-up effective days, if the corresponding recovery rate is greater than or equal to a preset recovery rate threshold, and the average power generation loss rate of all the above-mentioned days is... If the standard deviation is less than the preset standard deviation threshold, the corresponding day is determined as the final effective cleaning day; Summarize all the final effective cleaning days and remove the photovoltaic panel cleaning task execution days of the highway photovoltaic system to obtain at least one effective cleaning day for photovoltaic panels due to rainfall; Based on the meteorological data of the M consecutive days, the rainfall of the at least one effective day for rain cleanup of the photovoltaic panels is extracted, and the minimum rainfall from the extraction results is selected as the rainfall threshold.

[0013] In one possible design, the cumulative traffic flow is extracted according to the following steps: Based on the traffic data of the M consecutive days, extract the traffic flow of various vehicles from the current most recent effective date of solar panel rainfall cleanup to the current day. Based on the traffic flow of various types of vehicles on each day, the cumulative traffic flow is calculated using the following weighted formula. :

[0014] In the formula, This indicates the total number of days from the date when the most recent effective cleaning of the photovoltaic panels by rainfall occurred until the current day. Indicates less than or equal to positive integers, This represents the total number of the various vehicle types. Indicates less than or equal to positive integers, Indicates the first The preset weighting coefficients for this type of vehicle This indicates the date on which the most recent effective date for cleaning photovoltaic panels due to rainfall is [date missing]. The first day mentioned Traffic flow of this type of vehicle Represents the time decay factor function and is related to Negative correlation.

[0015] In one possible design, based on the weather forecast data for the next N days, it is determined whether there are any effective rain-cleaning days for photovoltaic panels within the next N days, including: Based on the meteorological forecast data for the next N days, the rainfall for each day in the next N days is extracted. Determine whether there is any day within the current near future consecutive N days where the rainfall exceeds a preset rainfall threshold. If so, determine whether there is an effective day for photovoltaic panel cleaning due to rainfall within the current near future consecutive N days; otherwise, determine whether there is an effective day for photovoltaic panel cleaning due to rainfall within the current near future consecutive N days.

[0016] In one possible design, the photovoltaic operation and maintenance execution terminal is a drone equipped with photovoltaic panel cleaning components, and the photovoltaic panel cleaning task sheet contains the geographical location of the highway photovoltaic system, cleaning path planning information, and cleaning instructions.

[0017] Secondly, a cloud computing-based energy efficiency management device for highway photovoltaic systems is provided, which is suitable for deployment in a cloud computing platform. It includes a multi-source data acquisition unit, a power generation loss determination unit, an engineering feature extraction unit, a prediction model training unit, a prediction model application unit, a rainfall condition judgment unit, a loss condition judgment unit, a cleaning task generation unit, and a cleaning task dispatch unit. The multi-source data acquisition unit is used to acquire the operation data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified photovoltaic panel cleanliness reference date, and M represents a positive integer greater than or equal to 2. The power generation loss determination unit is communicatively connected to the multi-source data acquisition unit and is used to determine the power generation loss rate of the highway photovoltaic system on each day of the continuous M days, based on the operating data and meteorological data on the corresponding day and the situation of photovoltaic panel pollution. The engineering feature extraction unit is communicatively connected to the multi-source data acquisition unit and is used to extract engineering features related to the photovoltaic panel pollution accumulation phenomenon for each day based on the meteorological data and traffic data for the consecutive M days. The prediction model training unit is communicatively connected to the power generation loss determination unit and the engineering feature extraction unit, respectively, and is used to train or periodically update the power generation loss prediction model constructed based on the machine learning algorithm, using the engineering features of each day as the model input and the power generation loss rate of each day as the model output. The prediction model application unit is communicatively connected to the prediction model training unit. It is used to extract new engineering features related to the accumulation of photovoltaic panel pollution over the current N consecutive days based on the meteorological forecast data of the area where the highway photovoltaic system is located and the traffic forecast data of the road section where the highway photovoltaic system is located over the current N consecutive days. It then imports these features into the power generation loss prediction model and outputs the power generation loss rate of the highway photovoltaic system over the current N consecutive days due to photovoltaic panel pollution, where N represents a positive integer. The rainfall condition judgment unit is communicatively connected to the prediction model application unit and is used to determine, based on the meteorological forecast data for the next N consecutive days, whether there are effective days for cleaning up photovoltaic panels with rainfall within the next N consecutive days. The loss condition judgment unit is communicatively connected to the rainfall condition judgment unit and is used to determine, when the rainfall condition does not exist, whether the average power generation loss rate of the highway photovoltaic system in the current near future consecutive N days caused by photovoltaic panel pollution exceeds a preset threshold. The cleaning task generation unit is communicatively connected to the loss condition judgment unit, and is used to generate a photovoltaic panel cleaning task order for the highway photovoltaic system when the condition is exceeded. The cleaning task dispatching unit is communicatively connected to the cleaning task generation unit and is used to send the photovoltaic panel cleaning task order to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the cleaning task of the photovoltaic panels of the highway photovoltaic system.

[0018] Thirdly, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the energy efficiency management method for a highway photovoltaic system as described in the first aspect or any possible design in the first aspect.

[0019] The beneficial effects of the above scheme are: (1) This invention creatively provides a new energy efficiency management scheme that can integrate multi-source dynamic data to accurately quantify current pollution loss and make forward-looking intelligent decisions based on economic prediction. It is executed by a cloud computing platform and first obtains historical operation, meteorological and traffic data. Based on this, it accurately quantifies the daily power generation loss rate caused by photovoltaic panel pollution and extracts engineering features related to pollution accumulation. Then, it uses historical features and loss rate as training samples to construct and periodically update a machine learning prediction model. It also uses future meteorological and traffic prediction data to extract corresponding features and input them into the model to predict the power generation loss rate caused by pollution in the next few days. Finally, under the premise that there is no effective rainfall for cleaning in the near future, if the average predicted loss rate exceeds the threshold, it automatically generates a photovoltaic panel cleaning task order and sends it to the operation and maintenance terminal for execution. This enables scientific assessment, intelligent prediction and proactive operation and maintenance of pollution loss of highway photovoltaic systems, significantly improving system energy efficiency and operation and maintenance economy. (2) It can quantify and accurately measure pollution loss, that is, by separating the impact of equipment failure, it can accurately calculate the power generation loss rate caused solely by photovoltaic panel pollution, providing a reliable data benchmark for subsequent analysis and decision-making; (3) It can have intelligent and forward-looking pollution prediction, that is, it integrates multi-dimensional engineering features (such as cumulative rainless days, weighted traffic flow and fine particulate matter concentration, etc.), and uses machine learning models to establish a dynamic mapping relationship between pollution accumulation and power generation loss. It can accurately predict the pollution loss trend in the next few days and realize the transformation from "post-event response" to "pre-event prediction". (4) It can have clean decision optimization and economy, that is, the decision mechanism comprehensively considers the possibility of future rainfall cleaning and the predicted degree of pollution loss. It only triggers the artificial cleaning task when it is confirmed that there is no natural cleaning opportunity and the predicted loss exceeds the economic threshold, thus avoiding ineffective or premature cleaning operations, significantly reducing operation and maintenance costs and improving the cleaning input-output ratio. (5) It can have automated and efficient operation and maintenance execution. That is, by generating task orders containing geographic and route information and driving automated terminals such as drones to execute them, it realizes remote, accurate and efficient operation and maintenance of photovoltaic facilities along the highways, which greatly improves the response speed and reduces the safety risks of human intervention. (6) A closed-loop energy efficiency management system integrating "precise perception, intelligent prediction, optimized decision-making and automatic execution" has been constructed, which systematically improves the power generation efficiency, operation and maintenance economy and management intelligence of highway photovoltaic power stations, making it convenient for practical application and promotion. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating the cloud computing-based energy efficiency management method for highway photovoltaic systems provided in this application embodiment.

[0022] Figure 2 A schematic diagram of the structure of a cloud computing-based highway photovoltaic system energy efficiency management device provided in this application embodiment. Detailed Implementation

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these embodiments without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0024] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.

[0025] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0026] Example like Figure 1 As shown, the cloud computing-based highway photovoltaic system energy efficiency management method provided in the first aspect of this embodiment can be executed, but is not limited to, by a cloud computing platform with certain computing resources. The cloud computing platform possesses powerful data aggregation, storage, and high-performance computing capabilities, providing an ideal technical foundation for the fusion processing of massive heterogeneous operational data, meteorological data, and dynamic traffic data, and for running complex machine learning prediction models in this embodiment. Figure 1 As shown, the energy efficiency management method for the highway photovoltaic system includes, but is not limited to, the following steps S1 to S9.

[0027] S1. Obtain the operating data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified photovoltaic panel cleanliness reference date, and M represents a positive integer greater than or equal to 2.

[0028] In step S1, the highway photovoltaic system refers to a distributed photovoltaic system constructed along the highway (such as service areas, toll station roofs, or slopes), and serves as the energy efficiency management object in this embodiment. Since the highway photovoltaic system is brand new on the installation and commissioning date, it can be directly used as the initialization date. The most recently verified photovoltaic panel cleanliness baseline date is a dynamic and data-verified key time node; it is not simply a cleaning record date, but rather the date on which the photovoltaic panel surface has been clearly confirmed by power generation performance data to have recovered to near the theoretical clean power generation efficiency. Therefore, it can also be used as the initialization date. The operational data specifically includes, but is not limited to, the real-time power generation of the photovoltaic panels, power generation, the real-time current-voltage characteristic curve of the photovoltaic strings, the inverter's AC output power, conversion efficiency, equipment operating temperature, and alarm status logs, etc., to directly reflect the instantaneous operating status and health condition of the photovoltaic system. The operational data can be, but is not limited to, remotely reported by the highway photovoltaic system. The meteorological data specifically includes, but is not limited to, meteorological parameters used to derive solar irradiance (such as cloud conditions, cloud cover, cloud height, visibility, and humidity), ambient temperature, wind speed, wind direction, relative humidity, rainfall, and duration, as well as atmospheric pollution parameters such as air quality index, PM2.5 concentration, and PM10 concentration, and dust storm events; the meteorological data can be obtained, but is not limited to, through routine queries on a meteorological server. The traffic data specifically includes, but is not limited to, hourly traffic flow, daily average traffic flow, vehicle type classification data (such as the respective proportions of passenger cars and large trucks), real-time average vehicle speed, and congestion index for the corresponding road segments, serving as key inputs for subsequent quantification of the intensity and characteristics of traffic pollution sources; the traffic data can be obtained, but is not limited to, through routine queries on a traffic server. Furthermore, for example, if the initialization date is November 1st and the current date is December 4th, then the operational data, meteorological data, and traffic data for 34 consecutive days can be obtained (i.e., M=34).

[0029] S2. For each day in the continuous M days, based on the operating data and meteorological data of the corresponding day, determine the power generation loss rate of the highway photovoltaic system on the corresponding day and due to photovoltaic panel pollution.

[0030] In step S2, the power generation loss rate is used to positively and synchronously reflect the pollution status of the photovoltaic panels in the highway photovoltaic system; the higher the loss rate, the more severe the photovoltaic panel pollution. Specifically, for each day within the consecutive M days, based on the operational data and meteorological data for the corresponding day, the power generation loss rate of the highway photovoltaic system on the corresponding day due to photovoltaic panel pollution is determined, including but not limited to the following steps S21 to S24.

[0031] S21. For a certain day in the continuous M days, based on the operating data of the corresponding day, extract the actual power generation and electrical characteristics of the highway photovoltaic system on the corresponding day, and based on the meteorological data of the corresponding day, extract multiple meteorological parameters for calculating the theoretical power generation of the highway photovoltaic system on the corresponding day, wherein the multiple meteorological parameters include, but are not limited to, solar irradiance, ambient temperature and wind speed.

[0032] In step S21, the actual power generation can be directly obtained. The electrical characteristics can be obtained by scanning the real-time current-voltage characteristic curve of the photovoltaic string to extract key characteristic parameters such as the offset of the voltage and current at the maximum power point, the rate of change of the fill factor, and the relative changes of the open-circuit voltage and short-circuit current. The solar irradiance refers to the radiant energy per unit area per unit time that reaches the solid Earth surface after the solar radiation has been absorbed, scattered, and reflected by the atmosphere. Considering that it is generally not directly presented in meteorological data and is closely related to the location of the photovoltaic panel and the solar altitude angle (i.e., the angle between the sunlight and the ground plane, ranging from 0° to 90°; this angle is used to quantify the position of the sun in the sky, for example, the solar altitude angle reaches its maximum value at noon), the solar irradiance is specifically extracted according to the following steps S211 to S212.

[0033] S211. Extract multiple solar irradiance influencing parameters from meteorological data, wherein the multiple solar irradiance influencing parameters include, but are not limited to, cloud conditions, cloud cover, cloud height, visibility, and humidity.

[0034] In step S211, the cloud conditions may include, for example, sunny, cloudy, or overcast; the cloud cover may be, for example, 60%; the cloud height may be, for example, 3000 meters; and so on.

[0035] S212. Based on the multiple solar irradiance influence parameters and combined with the known mapping relationship between the geographical location and date of the highway photovoltaic system and the solar altitude angle, the solar irradiance is estimated using a validated atmospheric radiative transfer model.

[0036] In step S212, the atmospheric radiative transfer model is a physical and mathematical model describing the propagation process of solar radiation in the Earth's atmosphere. It comprehensively considers the absorption, scattering, and transmission effects of solar radiation by atmospheric molecules, aerosols, and clouds, and quantifies radiation attenuation. In the photovoltaic field, this model is often based on geographical location, time, and real-time meteorological parameters (such as cloud cover and visibility) to invert or predict the spectrum and total solar irradiance reaching the Earth's surface, providing high-precision input for theoretical power generation calculation. It is a key technical means to make up for the lack of direct observation data, and therefore it can be applied to this step.

[0037] S22. Based on the electrical characteristics, identify and quantify the power generation loss rate of the highway photovoltaic system on a certain day due to a fault.

[0038] In step S22, the specific process involves matching the electrical characteristics with a pre-stored equipment fault feature library containing features such as component hot spot fault characteristics (manifested as a step in the IV curve within a specific voltage range), bypass diode failure characteristics (manifested as an abnormal drop in string voltage), inverter MPPT (Maximum Power Point Tracking) fault characteristics (manifested as a sudden drop in maximum power point tracking efficiency), and cable connection fault characteristics (manifested as an abnormal increase in line loss rate). Then, for each identified fault type, the corresponding pre-calibrated loss quantification model is called according to its fault mode to calculate the power loss rate caused by the fault. Finally, the power generation loss rate caused by the fault is calculated conventionally based on the power loss rate.

[0039] S23. Based on the multiple meteorological parameters, calculate the theoretical power generation of the highway photovoltaic system on a certain day.

[0040] In step S23, the specific process is as follows: First, based on the installation tilt angle and orientation of the photovoltaic panel, the solar irradiance is converted into the effective irradiance received by the inclined surface of the photovoltaic panel using an inclined irradiance model. Then, combined with the ambient temperature and wind speed, the operating temperature of the photovoltaic cell is estimated using a thermal balance model (such as a simplified formula based on NOCT parameters; NOCT stands for Nominal Operating Cell Temperature). Finally, the effective irradiance of the inclined surface and the operating temperature of the cell are input into the photovoltaic module performance model (this model is usually constructed using standard test conditions and temperature coefficients provided by the manufacturer) to calculate the theoretical power of a single photovoltaic cell. Then, the power is summed according to the array series and parallel configuration to finally obtain the theoretical daily power generation.

[0041] S24. Based on the theoretical and actual power generation of the highway photovoltaic system on a certain day, calculate the power generation loss rate of the highway photovoltaic system on that day, and subtract the power generation loss rate caused by the fault on that day to obtain the power generation loss rate of the highway photovoltaic system on the corresponding day caused by photovoltaic panel pollution.

[0042] In step S24, the power generation loss rate is specifically equal to the theoretical power generation minus the actual power generation, divided by the theoretical power generation and multiplied by 100%.

[0043] S3. Based on the meteorological data and traffic data of the consecutive M days, extract the engineering features of each day that are related to the photovoltaic panel pollution accumulation phenomenon.

[0044] In step S3, the engineering features are used as input items for training samples of the subsequent power generation loss prediction model. Since power generation loss is positively correlated with photovoltaic panel pollution, they need to be related to the accumulation of photovoltaic panel pollution. Taking into account factors across time accumulation, traffic pollution, atmospheric environment, and special events, the engineering features preferably include, but are not limited to, the cumulative number of days from the most recent effective date for cleaning the photovoltaic panels due to rainfall, cumulative traffic flow, cumulative fine particulate matter concentration, and the cumulative number of dust storm events. The most recent effective date for cleaning the photovoltaic panels due to rainfall refers to the most recent rainfall day where rainfall exceeds a threshold, as determined by meteorological data. The cumulative number of days represents the influencing factor for photovoltaic panel pollution accumulation in the time accumulation dimension; the cumulative traffic flow represents the influencing factor in the traffic pollution dimension; the cumulative fine particulate matter concentration represents the influencing factor in the atmospheric environment dimension; and the cumulative number of dust storm events represents the influencing factor in the special events dimension. Since sufficient rainfall (e.g., more than 10 mm) has a significant and useful natural cleaning effect on photovoltaic panels, it is necessary to collect cumulative factors from the most recent effective day for cleaning photovoltaic panels due to rainfall (if such a day does not exist, the initial log can be set to that day) to the current day. The rainfall threshold can be obtained based on historical experience, for example, 10 mm; however, considering that the photovoltaic panel orientation and environment of different highway photovoltaic systems may vary significantly, this will lead to different effects of rainfall cleaning photovoltaic panels. Therefore, it is necessary to personalize the rainfall threshold used as the criterion. To achieve this personalization, it is further preferred that the rainfall threshold can be obtained in advance according to, but is not limited to, the following steps S301 to S306.

[0045] S301. Based on the power generation loss rate of the highway photovoltaic system during the consecutive M days and due to photovoltaic panel pollution, calculate the change in the loss rate for each non-first day during the consecutive M days.

[0046] S302. For each non-first day, if the corresponding change in loss rate is lower than a preset negative threshold, then the corresponding day is selected as a candidate effective cleaning day.

[0047] In step S302, the negative threshold may be, for example, -8%.

[0048] S303. For each of the candidate clean effective days, based on the power generation loss rate of the highway photovoltaic system during the consecutive M days due to photovoltaic panel pollution, calculate the average power generation loss rate for the most recent consecutive K days prior to the corresponding day. and the average power generation loss rate over the most recent K consecutive days after the corresponding date And calculate the corresponding degree of recovery. , where K represents a positive integer greater than or equal to 2.

[0049] In step S303, K is 5 for example; if a candidate effective clean-up day is November 20, then the average power generation loss rate from November 15 to 19 can be calculated. and the average power generation loss rate from November 21st to 25th .

[0050] S304. For each of the candidate clean effective days, if the corresponding recovery rate is greater than or equal to a preset recovery rate threshold, and the average power generation loss rate of all the above-mentioned days is... If the standard deviation is less than the preset standard deviation threshold, then the corresponding day is determined as the final effective cleaning day.

[0051] In step S304, the recovery threshold is a "significance of effect" criterion used to quantify the performance improvement brought about by the cleaning event. It measures the percentage decrease in the power generation loss rate after cleaning relative to before cleaning. The threshold (e.g., 30%) ensures that only events that bring substantial efficiency improvements are recognized, filtering out minor changes caused by daily weather fluctuations. The standard deviation threshold is a "stability of effect" criterion used to assess the degree of performance fluctuation after cleaning. It measures the dispersion of the loss rate over a period of time after cleaning. The threshold (e.g., 2%) ensures that power generation efficiency remains high and stable after cleaning, excluding "pseudo-cleaning" events that temporarily improve but rebound rapidly due to incomplete pollution removal or weather interference.

[0052] S305. Summarize all the final effective cleaning days and remove the photovoltaic panel cleaning task execution days of the highway photovoltaic system to obtain at least one effective cleaning day for photovoltaic panels due to rainfall.

[0053] S306. Based on the meteorological data of the M consecutive days, extract the rainfall of the at least one effective day for rain cleanup of the photovoltaic panels, and select the minimum rainfall from the extraction results as the rainfall threshold.

[0054] In step S3, it is considered that the contribution of different vehicle types to photovoltaic panel pollution varies significantly. For example, heavy-duty trucks / diesel vehicles emit more particulate matter with stronger adhesion, and their pollution effect per vehicle far exceeds that of small passenger cars. Therefore, simply accumulating the total number of vehicles would severely underestimate the impact of high-polluting vehicles and overestimate the role of low-polluting vehicles. It is necessary to assign pollution weight coefficients based on actual emissions or deposition studies to different vehicle types to weight the cumulative traffic flow calculation, thereby more accurately reflecting the pollution load caused by traffic sources and improving the accuracy of the pollution prediction model. Furthermore, it is considered that earlier pollutants may partially dissipate due to weathering, weak rainfall, or natural shedding, and their current residual impact on power generation efficiency will weaken over time. Therefore, when calculating the cumulative traffic flow, it is also necessary to assign historical data a weight that decays exponentially or linearly over time, so that the contribution of recent pollution data to the current pollution state is greater than that of long-term data. To address the aforementioned two points, a further preferred method is to extract the cumulative traffic flow using the following steps: First, based on the traffic data from the consecutive M days, extract the traffic flow data for various types of vehicles from the current effective date of the recent solar panel rainfall cleanup to the current day; then, based on the traffic flow data for each day, calculate the cumulative traffic flow using the following weighted formula. :

[0055] In the formula, This indicates the total number of days from the date when the most recent effective cleaning of the photovoltaic panels by rainfall occurred until the current day. Indicates less than or equal to positive integers, This represents the total number of the various vehicle types. Indicates less than or equal to positive integers, Indicates the first The preset weighting coefficients for this type of vehicle This indicates the date on which the most recent effective date for cleaning photovoltaic panels due to rainfall is [date missing]. The first day mentioned Traffic flow of this type of vehicle Represents the time decay factor function and is related to Negative correlation.

[0056] In step S3, considering the significant differences in the contribution of different fine particulate matter to photovoltaic panel pollution, and the need to assign exponentially or linearly decaying weights to historical data when calculating the cumulative fine particulate matter concentration, the cumulative fine particulate matter concentration can also be conventionally modified and extracted by referring to the extraction process of the cumulative traffic flow. Furthermore, considering the significant differences in the contribution of different levels of dust storm events to photovoltaic panel pollution, and the need to assign exponentially or linearly decaying weights to historical data when calculating the cumulative number of dust storm events, the cumulative number of dust storm events can also be conventionally modified and extracted by referring to the extraction process of the cumulative traffic flow.

[0057] S4. Using the engineering characteristics of each day as model input and the power generation loss rate of each day as model output, train or periodically update the power generation loss prediction model constructed based on the machine learning algorithm.

[0058] In step S4, the machine learning algorithm is used to construct a predictive model by automatically learning patterns from historical data (engineering features as input and power generation loss rate as output). For example, random forests handle nonlinear relationships and evaluate feature importance through multi-decision tree ensembles; gradient boosting machines (such as XGBoost) optimize predictions iteratively and have a strong ability to capture complex data patterns; long short-term memory networks can effectively model the temporal dependencies of pollution accumulation; these algorithms can all establish an accurate mapping from meteorological and traffic features to power generation loss rate.

[0059] S5. Based on the meteorological forecast data of the area where the highway photovoltaic system is located for the next N days and the traffic forecast data of the road section where the highway photovoltaic system is located for the next N days, extract the new engineering features related to the accumulation of photovoltaic panel pollution for the next N days, import them into the power generation loss prediction model, and output the power generation loss rate of the highway photovoltaic system due to photovoltaic panel pollution for the next N days, where N represents a positive integer.

[0060] In step S5, the weather forecast data and traffic forecast data can be obtained conventionally using existing weather forecasting techniques. The value of N can be 7, for example, to obtain the weather forecast data and traffic forecast data for December 5th to 11th. Details and extraction methods for the new engineering features can be found in the section on engineering features, and will not be repeated here (in this case, extraction needs to be performed by combining the weather data and traffic data from the consecutive M days). Specifically, the new engineering features for each future day are imported into the power generation loss prediction model to output the power generation loss rate of the highway photovoltaic system for that future day due to photovoltaic panel pollution.

[0061] S6. Based on the meteorological forecast data for the next N days in the near future, determine whether there are effective days for cleaning up solar panels with rainfall within the next N days in the near future.

[0062] In step S6, the effective day for cleaning up photovoltaic panels due to rainfall can also be determined based on the comparison between rainfall and a rainfall threshold. Specifically, based on the meteorological forecast data for the next N days, it is determined whether there is an effective day for cleaning up photovoltaic panels due to rainfall within the next N days. This includes, but is not limited to, the following steps: first, extracting the rainfall for each day in the next N days based on the meteorological forecast data; then, determining whether there is any day in the next N days where the rainfall exceeds a preset rainfall threshold. If so, it is determined that there is an effective day for cleaning up photovoltaic panels due to rainfall within the next N days; otherwise, it is determined that there is no effective day for cleaning up photovoltaic panels due to rainfall within the next N days.

[0063] S7. If not, determine whether the average power generation loss rate of the highway photovoltaic system in the current near future N consecutive days due to photovoltaic panel pollution exceeds a preset threshold.

[0064] In step S7, the preset threshold can be determined based on actual cleaning needs, for example, 20%. Additionally, if an opportunity for natural precipitation to clean the photovoltaic panels exists, it can be considered that, to conserve cleaning resources, a photovoltaic panel cleaning task order may not be generated for the highway photovoltaic system at this time.

[0065] S8. If the number of cases exceeds the limit, a photovoltaic panel cleaning task order will be generated for the aforementioned highway photovoltaic system.

[0066] S9. The photovoltaic panel cleaning task order is sent to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the photovoltaic panel cleaning task of the highway photovoltaic system.

[0067] In step S9, the photovoltaic (PV) operation and maintenance (O&M) execution terminal can be an O&M personnel terminal, or an automated O&M execution terminal such as a robotic sweeper or an automated sprinkler system. These terminals can then routinely respond to and complete the cleaning tasks of the PV panels of the highway PV system, improving the energy efficiency of the PV system. Considering that some highway PV systems may not be equipped with automated cleaning components, drones can also be dispatched to perform the PV panel cleaning tasks. Preferably, the PV O&M execution terminal uses a drone equipped with PV panel cleaning components, and the PV panel cleaning task order includes, but is not limited to, the geographical location of the highway PV system, cleaning path planning information, and cleaning instructions.

[0068] Therefore, based on the energy efficiency management method for highway photovoltaic systems described in steps S1 to S9 above, a new energy efficiency management solution is provided that can integrate multi-source dynamic data to accurately quantify current pollution losses and make forward-looking intelligent decisions based on economic predictions. This solution is executed by a cloud computing platform, which first acquires historical operation, meteorological, and traffic data to accurately quantify the daily power generation loss rate caused by photovoltaic panel pollution and extract engineering features related to pollution accumulation. Then, using historical features and loss rates as training samples, a machine learning prediction model is constructed and periodically updated. Furthermore, future meteorological and traffic prediction data are used to extract corresponding features and input them into the model to predict the power generation loss rate caused by pollution in the next few days. Finally, assuming no effective rainfall for cleaning in the near future, if the average predicted loss rate exceeds a threshold, a photovoltaic panel cleaning task order is automatically generated and sent to the operation and maintenance terminal for execution. This enables scientific assessment, intelligent prediction, and proactive operation and maintenance of pollution losses in highway photovoltaic systems, significantly improving system energy efficiency and operation and maintenance economy, and facilitating practical application and promotion.

[0069] like Figure 2 As shown, the second aspect of this embodiment provides a virtual device for implementing the energy efficiency management method of the highway photovoltaic system described in the first aspect. It is suitable for deployment in a cloud computing platform and includes a multi-source data acquisition unit, a power generation loss determination unit, an engineering feature extraction unit, a prediction model training unit, a prediction model application unit, a rainfall condition judgment unit, a loss condition judgment unit, a cleaning task generation unit, and a cleaning task dispatch unit. The multi-source data acquisition unit is used to acquire the operation data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified photovoltaic panel cleanliness reference date, and M represents a positive integer greater than or equal to 2. The power generation loss determination unit is communicatively connected to the multi-source data acquisition unit and is used to determine the power generation loss rate of the highway photovoltaic system on each day of the continuous M days, based on the operating data and meteorological data on the corresponding day and the situation of photovoltaic panel pollution. The engineering feature extraction unit is communicatively connected to the multi-source data acquisition unit and is used to extract engineering features related to the photovoltaic panel pollution accumulation phenomenon for each day based on the meteorological data and traffic data for the consecutive M days. The prediction model training unit is communicatively connected to the power generation loss determination unit and the engineering feature extraction unit, respectively, and is used to train or periodically update the power generation loss prediction model constructed based on the machine learning algorithm, using the engineering features of each day as the model input and the power generation loss rate of each day as the model output. The prediction model application unit is communicatively connected to the prediction model training unit. It is used to extract new engineering features related to the accumulation of photovoltaic panel pollution over the current N consecutive days based on the meteorological forecast data of the area where the highway photovoltaic system is located and the traffic forecast data of the road section where the highway photovoltaic system is located over the current N consecutive days. It then imports these features into the power generation loss prediction model and outputs the power generation loss rate of the highway photovoltaic system over the current N consecutive days due to photovoltaic panel pollution, where N represents a positive integer. The rainfall condition judgment unit is communicatively connected to the prediction model application unit and is used to determine, based on the meteorological forecast data for the next N consecutive days, whether there are effective days for cleaning up photovoltaic panels with rainfall within the next N consecutive days. The loss condition judgment unit is communicatively connected to the rainfall condition judgment unit and is used to determine, when the rainfall condition does not exist, whether the average power generation loss rate of the highway photovoltaic system in the current near future consecutive N days caused by photovoltaic panel pollution exceeds a preset threshold. The cleaning task generation unit is communicatively connected to the loss condition judgment unit, and is used to generate a photovoltaic panel cleaning task order for the highway photovoltaic system when the condition is exceeded. The cleaning task dispatching unit is communicatively connected to the cleaning task generation unit and is used to send the photovoltaic panel cleaning task order to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the cleaning task of the photovoltaic panels of the highway photovoltaic system.

[0070] The working process, working details and technical effects of the aforementioned device provided in the second aspect of this embodiment can be found in the energy efficiency management method for highway photovoltaic systems described in the first aspect, and will not be repeated here.

[0071] This embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the energy efficiency management method for a highway photovoltaic system as described in the first aspect. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0072] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A cloud computing-based energy efficiency management method for highway photovoltaic systems, characterized in that, Performed by a cloud computing platform, including: The system acquires the operational data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the same M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the same M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified cleanliness reference date of the photovoltaic panels, and M represents a positive integer greater than or equal to 2. For each day in the continuous M days, based on the operating data and meteorological data of the corresponding day, determine the power generation loss rate of the highway photovoltaic system on the corresponding day and due to photovoltaic panel pollution. Based on the meteorological data and traffic data for the consecutive M days, engineering features related to the photovoltaic panel pollution accumulation phenomenon for each day are extracted; Using the engineering characteristics of each day as model input and the power generation loss rate of each day as model output, a power generation loss prediction model based on machine learning algorithm is trained or periodically updated. Based on the weather forecast data for the area where the highway photovoltaic system is located for the next N days and the traffic forecast data for the road section where the highway photovoltaic system is located for the next N days, new engineering features related to the accumulation of photovoltaic panel pollution for the next N days are extracted and imported into the power generation loss prediction model. The output is the power generation loss rate of the highway photovoltaic system due to photovoltaic panel pollution for the next N days, where N represents a positive integer. Based on the meteorological forecast data for the next N days, determine whether there are any effective days for cleaning up solar panels with rainfall within the next N days. If not, then determine whether the average power generation loss rate of the highway photovoltaic system in the current near future N consecutive days due to photovoltaic panel pollution exceeds a preset threshold. If the number exceeds the limit, a photovoltaic panel cleaning task order will be generated for the aforementioned highway photovoltaic system. The photovoltaic panel cleaning task order is sent to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the cleaning task of the photovoltaic panels of the highway photovoltaic system.

2. The energy efficiency management method for highway photovoltaic systems according to claim 1, characterized in that, For each day within the M consecutive days, based on the operational data and meteorological data for the corresponding day, determine the power generation loss rate of the highway photovoltaic system on the corresponding day due to photovoltaic panel pollution, including: For a specific day within the M consecutive days, based on the operational data of the corresponding day, the actual power generation and electrical characteristics of the highway photovoltaic system on the corresponding day are extracted, and based on the meteorological data of the corresponding day, multiple meteorological parameters for calculating the theoretical power generation of the highway photovoltaic system on the corresponding day are extracted, wherein the multiple meteorological parameters include solar irradiance, ambient temperature and wind speed. Based on the electrical characteristics, the power generation loss rate of the highway photovoltaic system on a certain day due to a fault is identified and quantified. Based on the aforementioned meteorological parameters, the theoretical power generation of the highway photovoltaic system on a certain day is calculated. Based on the theoretical and actual power generation of the highway photovoltaic system on a certain day, the power generation loss rate of the highway photovoltaic system on that day is calculated, and the power generation loss rate caused by the failure on that day is subtracted to obtain the power generation loss rate of the highway photovoltaic system on the corresponding day caused by photovoltaic panel pollution.

3. The energy efficiency management method for highway photovoltaic systems according to claim 2, characterized in that, The solar irradiance was extracted according to the following steps: Multiple solar irradiance influencing parameters are extracted from meteorological data, including cloud conditions, cloud cover, cloud height, visibility, and humidity. Based on the aforementioned multiple solar irradiance influence parameters, and combined with the known mapping relationship between the geographical location and date of the highway photovoltaic system and the solar altitude angle, the solar irradiance is estimated using a validated atmospheric radiative transfer model.

4. The energy efficiency management method for highway photovoltaic systems according to claim 1, characterized in that, The engineering features include the cumulative number of days from the most recent effective day for cleaning up solar panels due to rainfall, the cumulative traffic volume, the cumulative fine particulate matter concentration, and the cumulative number of sandstorm weather events. The most recent effective day for cleaning up solar panels due to rainfall refers to the most recent rainfall day where the rainfall exceeds the rainfall threshold, as determined by meteorological data.

5. The energy efficiency management method for highway photovoltaic systems according to claim 4, characterized in that, The rainfall threshold is obtained in advance according to the following steps: Based on the power generation loss rate of the highway photovoltaic system during the consecutive M days and due to photovoltaic panel pollution, the change in loss rate for each non-first day during the consecutive M days is calculated. For each non-first day, if the corresponding change in loss rate is lower than a preset negative threshold, then the corresponding day will be regarded as a candidate effective cleaning day. For each of the candidate clean effective days, based on the power generation loss rate of the highway photovoltaic system during the M consecutive days due to photovoltaic panel pollution, the average power generation loss rate for the K most recent consecutive days prior to the corresponding day is calculated. and the average power generation loss rate over the most recent K consecutive days after the corresponding date And calculate the corresponding degree of recovery. , where K represents a positive integer greater than or equal to 2; For each of the candidate clean-up effective days, if the corresponding recovery rate is greater than or equal to a preset recovery rate threshold, and the average power generation loss rate of all the above-mentioned days is... If the standard deviation is less than the preset standard deviation threshold, the corresponding day is determined as the final effective cleaning day; Summarize all the final effective cleaning days and remove the photovoltaic panel cleaning task execution days of the highway photovoltaic system to obtain at least one effective cleaning day for photovoltaic panels due to rainfall; Based on the meteorological data of the M consecutive days, the rainfall of the at least one effective day for rain cleanup of the photovoltaic panels is extracted, and the minimum rainfall from the extraction results is selected as the rainfall threshold.

6. The energy efficiency management method for highway photovoltaic systems according to claim 4, characterized in that, The cumulative traffic flow is extracted according to the following steps: Based on the traffic data of the M consecutive days, extract the traffic flow of various vehicles from the current most recent effective date of solar panel rainfall cleanup to the current day. Based on the traffic flow of various types of vehicles on each day, the cumulative traffic flow is calculated using the following weighted formula. : In the formula, This indicates the total number of days from the date when the most recent effective rainfall-induced cleaning of the photovoltaic panels occurred until the current day. Indicates less than or equal to positive integers, This represents the total number of the various vehicle types. Indicates less than or equal to positive integers, Indicates the first Preset weighting coefficients for this type of vehicle This indicates the date on which the most recent effective date for cleaning photovoltaic panels due to rainfall is [date missing]. The first day mentioned Traffic flow of this type of vehicle Represents the time decay factor function and is related to Negative correlation.

7. The energy efficiency management method for highway photovoltaic systems according to claim 1, characterized in that, Based on the weather forecast data for the next N days, determine whether there are any effective days for cleaning up solar panels through rainfall within the next N days, including: Based on the meteorological forecast data for the next N days, the rainfall for each day in the next N days is extracted. Determine whether there is any day within the current near future consecutive N days where the rainfall exceeds a preset rainfall threshold. If so, determine whether there is an effective day for photovoltaic panel cleaning due to rainfall within the current near future consecutive N days; otherwise, determine whether there is an effective day for photovoltaic panel cleaning due to rainfall within the current near future consecutive N days.

8. The energy efficiency management method for highway photovoltaic systems according to claim 1, characterized in that, The photovoltaic operation and maintenance execution terminal uses a drone equipped with a photovoltaic panel cleaning component, and the photovoltaic panel cleaning task order includes the geographical location of the highway photovoltaic system, cleaning path planning information, and cleaning instructions.

9. A cloud computing-based energy efficiency management device for a highway photovoltaic system, characterized in that, Suitable for deployment in cloud computing platforms, it includes a multi-source data acquisition unit, a power generation loss determination unit, an engineering feature extraction unit, a prediction model training unit, a prediction model application unit, a rainfall condition judgment unit, a loss condition judgment unit, a cleaning task generation unit, and a cleaning task dispatch unit. The multi-source data acquisition unit is used to acquire the operation data of the highway photovoltaic system for M consecutive days from the initialization date to the current day, the meteorological data of the area where the highway photovoltaic system is located for the M consecutive days, and the traffic data of the road section where the highway photovoltaic system is located for the M consecutive days. The initialization date refers to the installation and commissioning date of the highway photovoltaic system or the current most recently verified photovoltaic panel cleanliness reference date, and M represents a positive integer greater than or equal to 2. The power generation loss determination unit is communicatively connected to the multi-source data acquisition unit and is used to determine the power generation loss rate of the highway photovoltaic system on each day of the continuous M days, based on the operating data and meteorological data on the corresponding day and the situation of photovoltaic panel pollution. The engineering feature extraction unit is communicatively connected to the multi-source data acquisition unit and is used to extract engineering features related to the photovoltaic panel pollution accumulation phenomenon for each day based on the meteorological data and traffic data for the consecutive M days. The prediction model training unit is communicatively connected to the power generation loss determination unit and the engineering feature extraction unit, respectively, and is used to train or periodically update the power generation loss prediction model constructed based on the machine learning algorithm, using the engineering features of each day as the model input and the power generation loss rate of each day as the model output. The prediction model application unit is communicatively connected to the prediction model training unit. It is used to extract new engineering features related to the accumulation of photovoltaic panel pollution over the current N consecutive days based on the meteorological forecast data of the area where the highway photovoltaic system is located and the traffic forecast data of the road section where the highway photovoltaic system is located over the current N consecutive days. It then imports these features into the power generation loss prediction model and outputs the power generation loss rate of the highway photovoltaic system over the current N consecutive days due to photovoltaic panel pollution, where N represents a positive integer. The rainfall condition judgment unit is communicatively connected to the prediction model application unit and is used to determine, based on the meteorological forecast data for the next N consecutive days, whether there are effective days for cleaning up photovoltaic panels with rainfall within the next N consecutive days. The loss condition judgment unit is communicatively connected to the rainfall condition judgment unit and is used to determine, when the rainfall condition does not exist, whether the average power generation loss rate of the highway photovoltaic system in the current near future consecutive N days caused by photovoltaic panel pollution exceeds a preset threshold. The cleaning task generation unit is communicatively connected to the loss condition judgment unit, and is used to generate a photovoltaic panel cleaning task order for the highway photovoltaic system when the condition is exceeded. The cleaning task dispatching unit is communicatively connected to the cleaning task generation unit and is used to send the photovoltaic panel cleaning task order to the photovoltaic operation and maintenance execution terminal so that the photovoltaic operation and maintenance execution terminal can respond and complete the cleaning task of the photovoltaic panels of the highway photovoltaic system.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the energy efficiency management method for highway photovoltaic systems as described in any one of claims 1 to 8.

Images