Photovoltaic cleaning equipment intelligent operation and maintenance system based on Internet of Things

Abstract

The invention provides an intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things. The intelligent operation and maintenance system comprises the following steps: acquiring operation state data and operation environment data of target photovoltaic cleaning equipment; calculating a fault occurrence probability value and a reachable operation distance of the target photovoltaic cleaning equipment based on the operation state data; performing operation and maintenance reachable analysis on the target photovoltaic cleaning equipment according to the fault occurrence probability value and the reachable operation distance, and generating a reachable operation and maintenance area of the target photovoltaic cleaning equipment under the current operation condition; all candidate operation and maintenance points located in the reachable operation and maintenance area are screened out, and an optimal operation and maintenance target point is recognized according to the path complexity between the candidate operation and maintenance points and the current position of the target photovoltaic cleaning equipment and the service queue state of the candidate operation and maintenance points; and controlling the target photovoltaic cleaning equipment to go to the optimal operation and maintenance target point to execute preventive maintenance. According to the technical scheme provided by the invention, the preventive maintenance of the photovoltaic cleaning equipment can be realized under the double constraints that the electric quantity of the battery is limited and the fault risk exists at the same time.

Description

Technical Field

[0001] This application relates to the field of equipment operation and maintenance technology, and more specifically, to an intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things. Background Technology

[0002] With the deep development of intelligent manufacturing, production equipment is increasingly evolving towards larger scale, precision, and intelligence. Stable equipment operation has become a core element for enterprises to ensure production capacity and control costs. The traditional operation and maintenance model that relies on manual inspection and post-event repair has pain points such as delayed response, insufficient fault prediction, and high operation and maintenance costs, making it difficult to adapt to the continuous needs of modern production. With the development of IoT, sensor, artificial intelligence, and big data analysis technologies, equipment operation and maintenance is shifting from passive repair to predictive maintenance, achieving precise allocation of operation and maintenance resources.

[0003] In existing equipment operation and maintenance, the process begins by collecting equipment operating parameters through sensors and comparing them with preset thresholds and historical benchmark data to assess the equipment's health status. Next, a fault diagnosis model identifies the root causes of anomalies, distinguishing between potential and existing faults, and developing targeted operation and maintenance strategies. Finally, operation and maintenance are executed. However, in the intelligent operation and maintenance of photovoltaic cleaning equipment, traditional photovoltaic cleaning equipment operation and maintenance are mostly based on fixed cycles or passive responses after fault alarms. They lack the ability to jointly perceive and make decisions regarding the real-time health degradation trend and remaining battery life. This leads to a decision-making conflict between the timing and spatial accessibility of preventative maintenance due to dynamic changes in operating conditions and environment, resulting in wasted maintenance resources. Therefore, how to achieve preventative maintenance of photovoltaic cleaning equipment under the dual constraints of limited battery power and the risk of failure has become a challenge for the industry. Summary of the Invention

[0004] This application provides an intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things, which can realize preventive maintenance of photovoltaic cleaning equipment under the dual constraints of limited battery power and failure risk.

[0005] This application provides an intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things, including the following steps: The data acquisition module is used to collect the operating status data and operating environment data of the target photovoltaic cleaning equipment through the Internet of Things (IoT) sensor unit deployed in the target photovoltaic cleaning equipment. The operation status analysis module is used to calculate the probability of failure of the target photovoltaic cleaning equipment under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning equipment with the remaining power based on the operation status data. The maintenance reachability analysis module is used to perform maintenance reachability analysis on the target photovoltaic cleaning equipment based on the failure probability value and the reachable operating distance, and then generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions. The target identification module is used to filter out all candidate maintenance points located within the reachable maintenance area from a pre-set maintenance resource point location information database, and identify the optimal maintenance target point of the target photovoltaic cleaning equipment based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning equipment and the service queue status of the candidate maintenance point itself. The execution module is used to send collaborative scheduling instructions to the target photovoltaic cleaning equipment and the optimal operation and maintenance target point, and control the target photovoltaic cleaning equipment to go to the optimal operation and maintenance target point to perform preventive maintenance based on the collaborative scheduling instructions.

[0006] In some embodiments, calculating the probability of failure of the target photovoltaic cleaning device under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning device with remaining power based on the operating status data specifically includes: The running status data is normalized and preprocessed to obtain preprocessed running status data; Extract fault characteristic parameters and remaining battery power of the target photovoltaic cleaning equipment from the preprocessed operating status data; Based on the fault characteristic parameters and the pre-trained fault probability prediction model, the probability value of the target photovoltaic cleaning equipment under the current operating conditions is determined. Based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning equipment under the current operating conditions, the achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is identified.

[0007] In some embodiments, determining the probability of failure of the target photovoltaic cleaning equipment under the current operating conditions based on the fault characteristic parameters and the pre-trained fault probability prediction model specifically includes: Obtain a pre-trained fault probability prediction model; The fault characteristic parameters are input into the fault probability prediction model, and the fault probability prediction model outputs the fault occurrence probability value of the target photovoltaic cleaning equipment under the current operating conditions.

[0008] In some embodiments, identifying the achievable operating distance of the target photovoltaic cleaning device under the remaining battery power, based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning device under its current operating conditions, specifically includes: The energy consumption per unit distance of the target photovoltaic cleaning equipment is calculated based on the energy consumption model under the current operating conditions of the target photovoltaic cleaning equipment. The achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is determined based on the remaining battery power and the energy consumption per unit distance.

[0009] In some embodiments, the maintenance reachability analysis of the target photovoltaic cleaning equipment is performed based on the failure probability value and the reachable operating distance, thereby generating the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions, specifically including: The failure occurrence probability value is compared and analyzed with the preset failure warning probability threshold to determine the failure risk level of the target photovoltaic cleaning equipment. Obtain geographic information data of the photovoltaic field where the target photovoltaic cleaning equipment is located, and construct the maximum operating space range of the target photovoltaic cleaning equipment with the current location as the center based on the geographic information data and the achievable operating distance; Based on the fault risk level, the maximum operating space range is modulated with risk constraints to obtain candidate space regions that meet the fault handling priority. The feasibility of the candidate spatial areas is verified to generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions.

[0010] In some embodiments, selecting all candidate maintenance points located within the reachable maintenance area from a pre-set maintenance resource point location information database specifically includes: Obtain a pre-set database of operation and maintenance resource point locations, and obtain the geographic coordinate data of all operation and maintenance resource points from the database of operation and maintenance resource point locations; Extract the set of geographic boundary coordinates of the reachable maintenance area; A spatial location matching algorithm is used to determine the spatial affiliation of each operation and maintenance resource point's geographic coordinate data with the geographic boundary coordinate set of the reachable operation and maintenance area; The operation and maintenance resource points whose spatial ownership determination results are located within the reachable operation and maintenance area are selected as candidate operation and maintenance points.

[0011] In some embodiments, identifying the optimal maintenance target point for the target photovoltaic cleaning device based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning device, as well as the service queue status of the candidate maintenance point itself, specifically includes: Calculate the path complexity between each candidate maintenance point and the current location of the target photovoltaic cleaning equipment based on the geographic information data of the photovoltaic field where the target photovoltaic cleaning equipment is located; Obtain the service queue status of each candidate operation and maintenance point, and generate the service queue congestion index of each candidate operation and maintenance point. A weighted evaluation model of path complexity and service queue congestion index is constructed, and the path complexity and service queue congestion index of each candidate operation and maintenance point are substituted into the weighted evaluation model to calculate the operation and maintenance confidence score, thus obtaining the operation and maintenance confidence score of each candidate operation and maintenance point. The maintenance confidence scores of all candidate maintenance points are sorted in descending order, and the candidate maintenance point with the highest maintenance confidence score is selected as the optimal maintenance target point for the target photovoltaic cleaning equipment.

[0012] In some embodiments, sending a collaborative scheduling instruction to the target photovoltaic cleaning device and the optimal maintenance target point specifically includes: Obtain the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal maintenance target point; Based on the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal operation and maintenance target point, a collaborative scheduling instruction containing the equipment navigation path and the preparation requirements for operation and maintenance resources is generated. The edge communication module of the target photovoltaic cleaning device and the cloud communication interface of the optimal operation and maintenance target point are identified, and collaborative scheduling instructions are sent to the target photovoltaic cleaning device and the optimal operation and maintenance target point respectively through the edge communication module and the cloud communication interface.

[0013] In some embodiments, controlling the target photovoltaic cleaning equipment to proceed to the optimal maintenance target point to perform preventive maintenance based on the collaborative scheduling command specifically includes: The target photovoltaic cleaning equipment receives the collaborative scheduling instruction through the edge communication module and extracts the navigation path and maintenance resource preparation requirements from the collaborative scheduling instruction; The target photovoltaic cleaning equipment is controlled to move towards the optimal operation and maintenance target point according to the navigation path. When the target photovoltaic cleaning equipment arrives at the optimal operation and maintenance target point, the preset preventive maintenance operation is started according to the operation and maintenance resource preparation requirements.

[0014] In some embodiments, the IoT sensing unit collects operating status data and operating environment data of the target photovoltaic cleaning device at a fixed sampling frequency.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The IoT-based intelligent operation and maintenance system for photovoltaic cleaning equipment provided in this application first collects the operating status data and operating environment data of the target photovoltaic cleaning equipment through IoT sensing units deployed within the equipment. Second, based on the operating status data, it calculates the probability of failure of the target photovoltaic cleaning equipment under the current operating conditions and the achievable operating distance of the equipment with remaining power. Further, it performs an operation and maintenance reachability analysis on the target photovoltaic cleaning equipment based on the probability of failure and the achievable operating distance, thereby generating an achievable operation and maintenance area for the target photovoltaic cleaning equipment under the current operating conditions. Then, it filters all candidate operation and maintenance points located within the achievable area from a pre-set operation and maintenance resource point location information database, and identifies the optimal operation and maintenance target point for the target photovoltaic cleaning equipment based on the path complexity between the candidate operation and maintenance point and the current location of the target photovoltaic cleaning equipment, as well as the service queue status of the candidate operation and maintenance point itself. Finally, it sends a collaborative scheduling command to both the target photovoltaic cleaning equipment and the optimal operation and maintenance target point, and controls the target photovoltaic cleaning equipment to proceed to the optimal operation and maintenance target point to perform preventative maintenance based on the collaborative scheduling command.

[0016] Therefore, this application can achieve preventative maintenance of photovoltaic cleaning equipment under the dual constraints of limited battery power and the risk of failure. First, the data acquisition module collects operational status and environmental data of the target photovoltaic cleaning equipment through IoT sensing units, obtaining comprehensive operational information and providing objective and comprehensive data support for subsequent failure risk assessment and maintenance planning. Second, the operational status analysis module calculates the probability of failure and the achievable operating distance, quantifying the current failure risk level and endurance limit of the target photovoltaic cleaning equipment. This enables early prediction of equipment failures and precise definition of the operating range, avoiding downtime losses due to sudden failures and resource waste caused by blind maintenance. Furthermore, the maintenance reachability analysis module generates reachable maintenance areas, filtering out areas that combine the necessity of failure handling with the possibility of extended operating range through the dual constraints of the probability of failure and the achievable operating distance. The system provides a feasible maintenance space, excluding low-priority areas to define precise boundaries for subsequent maintenance point selection. This avoids decision-making conflicts between the timing of preventative maintenance and spatial accessibility caused by a lack of joint perception and decision-making capabilities regarding the real-time health degradation trend and remaining battery life of the equipment. Then, the target identification module filters candidate maintenance points and determines the optimal maintenance target point, effectively avoiding arrival delays caused by complex paths and excessively long maintenance wait times due to service queue congestion. Finally, the execution module sends collaborative scheduling instructions and controls the equipment to perform preventative maintenance, achieving coordinated linkage between the target photovoltaic cleaning equipment and the optimal maintenance target point. This ensures that the target photovoltaic cleaning equipment arrives at the maintenance point as planned to complete preventative maintenance, effectively reducing the failure rate of the target photovoltaic cleaning equipment. In summary, the technical solution provided in this application can achieve preventative maintenance of photovoltaic cleaning equipment under the dual constraints of limited battery power and the risk of failure. Attached Figure Description

[0017] Figure 1 This is a modular structure diagram of an IoT-based intelligent operation and maintenance system for photovoltaic cleaning equipment, as shown in some embodiments of this application. Figure 2 This is an exemplary flowchart illustrating the determination of reachable maintenance areas according to some embodiments of this application; Detailed Implementation

[0018] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] refer to Figure 1 This figure is a modular structure diagram of an IoT-based intelligent operation and maintenance system for photovoltaic cleaning equipment, according to some embodiments of this application. The system includes: a data acquisition module 100, an operation status analysis module 200, an operation and maintenance reachability analysis module 300, a target identification module 400, and an execution module 500, which are described below: The data acquisition module 100 is used to collect the operating status data and operating environment data of the target photovoltaic cleaning equipment through the Internet of Things sensing unit deployed in the target photovoltaic cleaning equipment.

[0020] In practical implementation, the operating status and environmental data of the target photovoltaic cleaning equipment are collected by IoT sensing units deployed within the equipment. These IoT sensing units include, but are not limited to, vibration sensors, current sensors, temperature sensors, BeiDou modules, dust sensors, radiometers, and temperature and humidity sensors. Specifically, multiple types of dedicated sensors (such as vibration sensors, current sensors, temperature sensors, and BeiDou modules) are deployed on the key moving parts (e.g., walking motors, roller brush motors), power systems, and main structures of the target photovoltaic cleaning equipment, and environmental sensing modules (such as dust sensors, radiometers, and temperature and humidity sensors) are integrated. This constitutes a distributed Internet of Things (IoT) sensing unit. This IoT sensing unit continuously collects operational status data, which characterizes the health status of the target photovoltaic cleaning device, and operational environment data, at a fixed sampling frequency. The operational status data includes, but is not limited to, motor three-phase current, motor temperature, drive wheel speed, battery resistance, and remaining battery power. The operational environment data includes, but is not limited to, on-site wind speed, dust thickness on the component surface, geographical coordinates, and terrain slope. This completes the collection of operational status data and operational environment data of the target photovoltaic cleaning device. The timestamps of the collection of the operational status data and the operational environment data are consistent.

[0021] It should be noted that, in this application, the operational status data refers to a set of parameters reflecting the operational status of the target photovoltaic cleaning equipment. These data together constitute the ontological information source for evaluating the real-time operational performance, wear level, and potential failure risk of the target photovoltaic cleaning equipment. In this application, the operational environment data refers to a set of physical parameters describing the external operating conditions and background information of the target photovoltaic cleaning equipment. These data together constitute the external dynamic constraints affecting the equipment's energy consumption mode, failure rate, and navigation accessibility, providing key environmental context for accurately assessing operating conditions and predicting performance.

[0022] The operation status analysis module 200 is used to calculate the failure probability value of the target photovoltaic cleaning equipment under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning equipment with the remaining power based on the operation status data.

[0023] In some embodiments, the following steps are used to calculate the failure probability value of the target photovoltaic cleaning device under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning device with remaining power, based on the operating status data: The running status data is normalized and preprocessed to obtain preprocessed running status data; Extract fault characteristic parameters and remaining battery power of the target photovoltaic cleaning equipment from the preprocessed operating status data; Based on the fault characteristic parameters and the pre-trained fault probability prediction model, the probability value of the target photovoltaic cleaning equipment under the current operating conditions is determined. Based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning equipment under the current operating conditions, the achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is identified.

[0024] In specific implementation, firstly, min-max normalization is used to map each data point in the operating status data to the [0,1] interval to obtain preprocessed operating status data. Specifically, the difference between each data point and the minimum value of the data dimension is divided by the difference between the maximum and minimum values ​​of the dimension to obtain the preprocessed operating status data. Here, min-max normalization refers to a numerical preprocessing method that scales the data to a specified interval through linear transformation. Secondly, based on the fault correlation dimension of the core components of the target photovoltaic cleaning equipment (e.g., motor temperature corresponds to motor fault, battery internal resistance corresponds to battery fault), fault feature parameters and remaining battery power are extracted from the preprocessed operating status data. The fault feature parameters include motor temperature and battery internal resistance, which are operating status indicators that are correlated with the occurrence of photovoltaic cleaning equipment faults. Then, based on the fault feature parameters and a pre-trained fault probability prediction model, the probability value of the target photovoltaic cleaning equipment's fault occurrence under the current operating conditions is determined. Finally, based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning equipment under the current operating conditions, the achievable operating distance of the target photovoltaic cleaning equipment under the remaining power is identified.

[0025] In some embodiments, the following steps are used to determine the probability of failure of the target photovoltaic cleaning equipment under the current operating conditions based on the fault feature parameters and the pre-trained fault probability prediction model: Obtain a pre-trained fault probability prediction model; The fault characteristic parameters are input into the fault probability prediction model, and the fault probability prediction model outputs the fault occurrence probability value of the target photovoltaic cleaning equipment under the current operating conditions.

[0026] In specific implementation, firstly, a pre-trained fault probability prediction model is obtained. The pre-training process of the fault probability prediction model includes: acquiring a historical photovoltaic cleaning equipment operation data sample set, dividing the historical photovoltaic cleaning equipment operation data sample set into a fault probability prediction training set and a fault probability prediction test set. The historical photovoltaic cleaning equipment operation data sample set also includes fault labeling corresponding to the historical photovoltaic cleaning equipment operation data samples. The fault labeling includes positive fault labels and negative fault labels. Positive fault labels indicate that the photovoltaic cleaning equipment corresponding to the historical photovoltaic cleaning equipment operation data sample has actually experienced a fault (including a fault occurrence probability reference value), and negative fault labels indicate that the photovoltaic cleaning equipment corresponding to the historical photovoltaic cleaning equipment operation data sample has not experienced a fault (fault occurrence probability is 0). A random forest classifier is used as the fault probability predictor. The basic model initializes a random forest classifier, using sample data from the fault probability prediction training set as input and fault labels as output to train the random forest classifier, resulting in a fault probability prediction model to be validated. A fault probability prediction test set is then used to validate the model, calculating the prediction error between the model's prediction results and the fault labels. The model with a prediction error less than or equal to a preset test error threshold is considered a successfully trained fault probability prediction model. This preset test error threshold can be set according to actual needs and is not limited here. Then, the fault feature parameters are input into the fault probability prediction model, which then outputs the probability of fault occurrence of the target photovoltaic cleaning equipment under the current operating conditions.

[0027] It should be noted that the failure probability value in this application refers to the numerical value that characterizes the likelihood of the target photovoltaic cleaning equipment failing under the current operating conditions. Photovoltaic cleaning equipment operates under complex conditions in outdoor photovoltaic fields for a long time. The failures of its core components (such as drive motors, battery packs, cleaning brush rollers, etc.) are random and hidden. If only post-event maintenance or fixed-cycle preventive maintenance is relied upon, it is impossible to identify high-risk equipment in advance to avoid sudden downtime, and it is also easy to waste resources due to blind operation and maintenance. Therefore, by quantifying the failure probability value, the degree of failure risk of the equipment under the current operating conditions can be accurately characterized, providing a quantifiable core basis for subsequent operation and maintenance decisions.

[0028] In some embodiments, the achievable operating distance of the target photovoltaic cleaning device under the remaining battery power and the energy consumption model of the target photovoltaic cleaning device under the current operating conditions is identified by the following steps: The energy consumption per unit distance of the target photovoltaic cleaning equipment is calculated based on the energy consumption model under the current operating conditions of the target photovoltaic cleaning equipment. The achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is determined based on the remaining battery power and the energy consumption per unit distance.

[0029] In specific implementation, firstly, when calculating the energy consumption per unit distance based on the energy consumption model of the target photovoltaic cleaning equipment under its current operating conditions, this energy consumption model is constructed using a multiple linear regression algorithm based on the historical operating data of the target photovoltaic cleaning equipment (including power consumption, travel speed, load rate, and environmental resistance under different operating conditions). The multiple linear regression algorithm is a statistical method used to predict the value of the dependent variable by establishing a mathematical model of the linear relationship between multiple independent variables and the dependent variable. In the construction process, historical operating data (such as travel speed, load rate, and environmental resistance) are used as independent variables, and energy consumption per unit distance is used as the dependent variable. The coefficients of the energy consumption model are obtained by fitting using the least squares method, thus constructing the energy consumption model. The least squares method refers to minimizing the average error... The mathematical optimization method for finding the best function match for the data will not be elaborated here. Subsequently, the current operating parameters (i.e., driving speed, load rate, and environmental resistance) extracted from the preprocessed operating status data are input into the energy consumption model. The energy consumption model outputs the energy consumption per unit distance of the target photovoltaic cleaning device under the current operating conditions. The energy consumption per unit distance refers to the electricity consumed by the target photovoltaic cleaning device to travel a unit length distance under the current operating conditions. Then, the remaining battery power extracted from the preprocessed operating status data is used as the dividend and the calculated energy consumption per unit distance as the divisor to perform a division operation. The quotient obtained from the division operation is taken as the achievable operating distance of the target photovoltaic cleaning device under the remaining battery power.

[0030] It should be noted that the reachable operating distance in this application refers to the maximum distance that the target photovoltaic cleaning equipment can continuously travel under the current remaining power and operating conditions. In the operation and maintenance of the target photovoltaic cleaning equipment, the photovoltaic cleaning equipment relies on battery power and needs to move over a large area of ​​the photovoltaic field. Its endurance directly determines whether the equipment can reach the maintenance point to complete the maintenance. If there is a lack of quantitative determination of the reachable operating distance, the equipment may stop on the way to the maintenance point due to insufficient power. This will not only prevent the completion of the maintenance task, but also incur additional equipment rescue costs. Therefore, determining the reachable operating distance can provide a spatial boundary basis for screening candidate maintenance points and avoid selecting maintenance points that exceed the equipment's endurance, which would lead to maintenance failure.

[0031] The maintenance reachability analysis module 300 is used to perform maintenance reachability analysis on the target photovoltaic cleaning equipment based on the failure probability value and the reachable operating distance, and then generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions.

[0032] In some embodiments, reference Figure 2As shown in the figure, this is an exemplary flowchart for determining an accessible maintenance area according to some embodiments of this application. In this embodiment, the maintenance accessibility analysis of the target photovoltaic cleaning equipment is performed based on the fault occurrence probability value and the accessible operating distance, thereby generating the accessible maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions. This can be achieved by the following steps: In step S31, the failure occurrence probability value is compared and analyzed with the preset failure warning probability threshold to determine the failure risk level of the target photovoltaic cleaning equipment. In step S32, the geographic information data of the photovoltaic field where the target photovoltaic cleaning device is located is obtained, and the maximum operating space range of the target photovoltaic cleaning device with the current location as the center is constructed based on the geographic information data and the achievable operating distance. In step S33, the maximum operating space range is modulated with risk constraints according to the fault risk level, thereby obtaining candidate space regions that meet the fault handling priority. In step S34, the feasibility of the candidate spatial area is verified to generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions.

[0033] In specific implementation, firstly, the fault occurrence probability value is compared with a preset fault warning probability threshold. The preset fault warning probability threshold includes a high-risk warning threshold and a low-risk warning threshold. If the fault occurrence probability value is higher than the high-risk warning threshold, it is determined to be a high fault risk level; if it is between the high-risk and low-risk warning thresholds, it is determined to be a medium fault risk level; and if it is lower than the low-risk warning threshold, it is determined to be a low fault risk level. Finally, the fault risk level of the target photovoltaic cleaning equipment is determined. The preset fault warning probability threshold can be set according to actual needs or based on expert knowledge; no limitation is made here. The fault risk level refers to the level based on the fault occurrence probability. The system prioritizes emergency response to equipment malfunctions based on probability values. Next, it acquires geographic information data of the photovoltaic field where the target photovoltaic cleaning device is located. This geographic information data refers to the topographical data of the photovoltaic field. Using the current location of the target photovoltaic cleaning device as the center and the reachable operating distance as the radius, a circular spatial range is drawn on the corresponding layer of the geographic information data and cropped to obtain the maximum operating space range of the target photovoltaic cleaning device. This maximum operating space range refers to the physical spatial boundary that the target photovoltaic cleaning device can reach based on its current range. Then, based on the determined fault risk level, risk constraint modulation is applied to the maximum operating space range, specifically for high fault rates. The risk level retains the entire maximum operating space range. For medium fault risk, the maximum operating space range is reduced by a first preset ratio, and for low fault risk, it is further reduced by a second preset ratio. This yields candidate space areas that meet the priority of fault handling. It should be noted that high-fault-risk equipment has a high probability of failure and urgent handling, requiring maximum coverage of its reachable operating space range to avoid missing urgent maintenance needs. Therefore, the entire maximum operating space range is retained. The urgency of fault handling for medium and low fault risk equipment decreases sequentially. Covering the entire reachable range would disperse limited maintenance resources, leading to delays in maintenance response in high-risk areas. Therefore, the maximum operating space range is reduced by a second preset ratio. The spatial range is gradually reduced by a certain ratio, prioritizing the maintenance of high-risk areas. The first and second preset ratios can be determined by obtaining maintenance cases of similar photovoltaic cleaning equipment in historical photovoltaic fields under different fault risk levels, statistically analyzing the maintenance completion rate under different preset ratios, and using the preset ratio corresponding to the highest maintenance completion rate as the corresponding first and second preset ratios. For example, the first and second preset ratios can be set to 75% and 50%, respectively, without further elaboration. The risk constraint modulation refers to the optimization method of adjusting the equipment maintenance space range based on the urgency of the fault. The candidate space area refers to the maintenance space after being screened by risk priority.Finally, existing network analysis algorithms are used to verify the traversability of the candidate spatial regions. The connectivity between each location within the candidate spatial region and the current location of the target photovoltaic cleaning equipment is checked. After eliminating inaccessible sub-regions, the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions is generated. The network analysis algorithm refers to an algorithm used to analyze the connectivity of nodes and paths in geographic space, which will not be elaborated upon here.

[0034] It should be noted that, in this application, the reachable maintenance area refers to the maintenance space range that meets the priority of fault handling and is passable. In the maintenance of the target photovoltaic cleaning equipment, it is impossible to determine whether the target photovoltaic cleaning equipment's battery life can support it to reach the maintenance area based solely on the fault occurrence probability value. That is, there may be situations where there is a high fault risk but the equipment is unable to reach the area due to insufficient power. Based solely on the reachable operating distance, the equipment would be included in the low fault risk area, which does not require emergency maintenance, resulting in a waste of maintenance resources. Therefore, identifying the reachable maintenance area by combining the fault occurrence probability value and the reachable operating distance can provide a precise spatial boundary for the subsequent selection of candidate maintenance points that combines the necessity of fault handling and the feasibility of equipment battery life. This excludes inaccessible areas that the equipment cannot reach due to insufficient power and also eliminates invalid areas with low fault risk that do not require emergency maintenance, ensuring that the focus is only on the maintenance range with high fault risk and reachable equipment.

[0035] The target identification module 400 is used to filter out all candidate maintenance points located within the reachable maintenance area from a pre-set maintenance resource point location information database, and identify the optimal maintenance target point of the target photovoltaic cleaning equipment based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning equipment and the service queue status of the candidate maintenance point itself.

[0036] In some embodiments, the following steps are used to filter all candidate maintenance points located within the reachable maintenance area from a pre-set maintenance resource point location information database: Obtain a pre-set database of operation and maintenance resource point locations, and obtain the geographic coordinate data of all operation and maintenance resource points from the database of operation and maintenance resource point locations; Extract the set of geographic boundary coordinates of the reachable maintenance area; A spatial location matching algorithm is used to determine the spatial affiliation of each operation and maintenance resource point's geographic coordinate data with the geographic boundary coordinate set of the reachable operation and maintenance area; The operation and maintenance resource points whose spatial ownership determination results are located within the reachable operation and maintenance area are selected as candidate operation and maintenance points.

[0037] In specific implementation, firstly, a pre-set location information database of operation and maintenance resource points stored on the server is obtained. This database contains unique identifiers for each operation and maintenance resource point. Geographic coordinate data of all operation and maintenance resource points is extracted from this database; the geographic coordinate data refers to the spatial location data of the operation and maintenance resource points. Secondly, the geographic coordinates of each boundary point in the reachable operation and maintenance area are extracted, and a set of geographic boundary coordinates is obtained by combining the geographic coordinates of all boundary points. This set of geographic boundary coordinates refers to the vertex coordinate sequence that constitutes the polygonal outline of the reachable operation and maintenance area. Then, using the ray-mapping method in existing spatial location matching algorithms, the geographic coordinate data of each operation and maintenance resource point is used as the point to be determined, and the set of geographic boundary coordinates of the reachable operation and maintenance area is used as the target polygon. Rays are emitted from the point to be determined in any direction, and the number of intersections between the ray and the boundary of the target polygon is counted. If the number of intersections is even, the point is outside the target polygon; if it is odd, it is inside the target polygon. This completes the spatial attribution determination of each operation and maintenance resource point. Finally, based on the spatial attribution determination results, operation and maintenance resource points determined to be located within the reachable operation and maintenance area are retained and marked as candidate operation and maintenance points.

[0038] It should be noted that, in this application, candidate maintenance points refer to maintenance resource points that are located within the reachable maintenance area and have maintenance service capabilities.

[0039] In some embodiments, the optimal maintenance target point for the target photovoltaic cleaning device is identified based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning device, as well as the service queue status of the candidate maintenance point itself, using the following steps: Calculate the path complexity between each candidate maintenance point and the current location of the target photovoltaic cleaning equipment based on the geographic information data of the photovoltaic field where the target photovoltaic cleaning equipment is located; Obtain the service queue status of each candidate operation and maintenance point, and generate the service queue congestion index of each candidate operation and maintenance point. A weighted evaluation model of path complexity and service queue congestion index is constructed, and the path complexity and service queue congestion index of each candidate operation and maintenance point are substituted into the weighted evaluation model to calculate the operation and maintenance confidence score, thus obtaining the operation and maintenance confidence score of each candidate operation and maintenance point. The maintenance confidence scores of all candidate maintenance points are sorted in descending order, and the candidate maintenance point with the highest maintenance confidence score is selected as the optimal maintenance target point for the target photovoltaic cleaning equipment.

[0040] In specific implementation, firstly, based on the geographical information data of the photovoltaic field where the target photovoltaic cleaning equipment is located, the Dijkstra algorithm, a common path planning algorithm, is used to plan the shortest path between each candidate maintenance point and the current location of the target photovoltaic cleaning equipment. Simultaneously, feature parameters are extracted from the path, including terrain undulation, number of turns, and number of obstacles. Weights are assigned to each feature parameter and a weighted sum is calculated. The weights of each feature parameter can be set between 0 and 1 according to actual needs (details omitted here). Then, the path complexity corresponding to each candidate maintenance point is obtained through min-max standardization. Path complexity refers to an indicator representing the difficulty of the target photovoltaic cleaning equipment traveling from its current location to the candidate maintenance point. Secondly, the shortest path between each candidate maintenance point is obtained. The service queue status of the selected maintenance points is analyzed. This status includes the number of pending devices and the average maintenance time per device. The ratio of the number of pending devices to the average maintenance time per device is used as the service queue congestion index, which is an indicator representing the busyness of the service queues at the candidate maintenance points. Then, a weighted evaluation model of path complexity and service queue congestion index is constructed, and the weights of path complexity and service queue congestion index are set (for example, the weights of path complexity and service queue congestion index can be set to 0.6 and 0.4, respectively; alternatively, they can be set according to actual needs, which is not limited here). The weighted evaluation model is set as follows: Maintenance Confidence Score = Path Complexity Weight × (1 - Path Complexity Normalization) The formula is calculated by substituting the path complexity (minimum-maximum normalized value) and the service queue congestion index weight into a weighted evaluation model to obtain the operation and maintenance confidence score for each candidate operation and maintenance point. This confidence score is a rating of the merits of the candidate operation and maintenance point. The normalized values ​​of both path complexity and service queue congestion index are in the range [0,1]. The formula uses "1 - normalized value" to achieve a reverse mapping from negative indicators (path complexity, congestion index) to positive scores; that is, the smoother the path (lower complexity) and the more relaxed the service queue (lower congestion index), the higher the corresponding positive score. This formula integrates two key indicators through weighted calculation. Low path complexity reduces equipment travel energy consumption and arrival delay, while low congestion index shortens maintenance waiting time. Together, they improve the time efficiency of the entire operation and maintenance process, avoiding waste of operation and maintenance resources and response delays caused by path complexity or queue congestion. This formula, through positive transformation and weighted balancing, quantifies the core negative factors affecting operation and maintenance effectiveness into directly comparable positive scores, ensuring that the scoring results accurately reflect the comprehensive advantages of the operation and maintenance solution in terms of equipment accessibility and service efficiency, ultimately achieving a balance between maximizing operation and maintenance efficiency and optimizing fault handling. Finally, the operation and maintenance confidence scores of all candidate operation and maintenance points are sorted in descending order, and the candidate operation and maintenance point with the highest operation and maintenance confidence score in the sorting results is selected as the optimal operation and maintenance target point for the target photovoltaic cleaning equipment.

[0041] It should be noted that the optimal maintenance target point in this application refers to the target maintenance resource point from which the target photovoltaic cleaning equipment goes to perform maintenance. Determining the optimal maintenance target point can provide the target photovoltaic cleaning equipment with a clear maintenance destination that combines path feasibility and service efficiency, avoiding the problems of increased equipment energy consumption and arrival delay caused by complex paths, as well as the drawbacks of excessively long maintenance waiting time and untimely fault handling caused by service queue congestion.

[0042] The execution module 500 is used to send a collaborative scheduling instruction to the target photovoltaic cleaning device and the optimal operation and maintenance target point, and control the target photovoltaic cleaning device to go to the optimal operation and maintenance target point to perform preventive maintenance based on the collaborative scheduling instruction.

[0043] In some embodiments, sending collaborative scheduling instructions to the target photovoltaic cleaning equipment and the optimal maintenance target point is achieved through the following steps: Obtain the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal maintenance target point; Based on the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal operation and maintenance target point, a collaborative scheduling instruction containing the equipment navigation path and the preparation requirements for operation and maintenance resources is generated. The edge communication module of the target photovoltaic cleaning device and the cloud communication interface of the optimal operation and maintenance target point are identified, and collaborative scheduling instructions are sent to the target photovoltaic cleaning device and the optimal operation and maintenance target point respectively through the edge communication module and the cloud communication interface.

[0044] In specific implementation, firstly, the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal maintenance target point are obtained from the operation and maintenance resource management database. This resource configuration data includes the number of maintenance personnel on duty at the optimal maintenance target point, the number of available maintenance workstations, and the progress of currently pending tasks. This resource configuration data refers to information related to resource assurance at the maintenance point. Secondly, the Dijkstra algorithm is used to plan the optimal navigation path from the equipment's current location to the optimal maintenance target point. Combined with the resource configuration data of the optimal maintenance target point, the operation and maintenance resource preparation requirements are clarified. These requirements include reserving designated available workstations and maintenance tools for corresponding fault types. The optimal navigation path and the operation and maintenance resource preparation requirements are then integrated into a standardized format. The system employs a collaborative scheduling instruction, which is a set of scheduling information including equipment navigation and maintenance point resource preparation requirements. Then, it identifies the edge communication module (e.g., a 5G industrial communication module supporting the MQTT IoT communication protocol) built into the target photovoltaic cleaning equipment and the cloud communication interface (e.g., a RESTful API interface adapted to the HTTP communication protocol) deployed at the optimal maintenance target point. The collaborative scheduling instruction is pushed to the edge communication module of the target photovoltaic cleaning equipment using the MQTT protocol, and sent to the cloud communication interface of the optimal maintenance target point via the HTTP protocol, thus completing the distribution of collaborative scheduling instructions to both the target photovoltaic cleaning equipment and the optimal maintenance target point.

[0045] It should be noted that, in this application, the cloud communication interface refers to the standardized data interaction interface used by the operation and maintenance target point to receive cloud scheduling instructions.

[0046] In some embodiments, controlling the target photovoltaic cleaning equipment to proceed to the optimal maintenance target point to perform preventive maintenance based on the collaborative scheduling command is achieved through the following steps: The target photovoltaic cleaning equipment receives the collaborative scheduling instruction through the edge communication module and extracts the navigation path and maintenance resource preparation requirements from the collaborative scheduling instruction; The target photovoltaic cleaning equipment is controlled to move towards the optimal operation and maintenance target point according to the navigation path. When the target photovoltaic cleaning equipment arrives at the optimal operation and maintenance target point, the preset preventive maintenance operation is started according to the operation and maintenance resource preparation requirements.

[0047] In specific implementation, firstly, the target photovoltaic cleaning device receives the collaborative scheduling instruction through the edge communication module, and extracts the navigation path and maintenance resource preparation requirements from the collaborative scheduling instruction through the instruction parsing module. The instruction parsing module is a program module used to identify and extract key information from the scheduling instruction. Then, the target photovoltaic cleaning device is controlled to move towards the optimal maintenance target point according to the navigation path. When the target photovoltaic cleaning device arrives at the optimal maintenance target point, it calls the preset preventive maintenance process (such as motor performance testing, battery capacity calibration, and brush roller wear inspection) to perform preventive maintenance operations based on the extracted maintenance resource preparation requirements. These details will not be elaborated here.

[0048] It should be noted that the preventive maintenance process in this embodiment refers to a standardized maintenance operation sequence that is executed in advance to reduce the probability of equipment failure. Preventive maintenance of the target photovoltaic cleaning equipment aims to eliminate potential fault hazards, effectively reduce the probability of sudden failure and shutdown of the target photovoltaic cleaning equipment, extend the service life of core components, and avoid the loss of photovoltaic power generation efficiency caused by equipment failure and shading of photovoltaic modules in the photovoltaic field. At the same time, compared with emergency repair after failure, preventive maintenance can significantly reduce maintenance costs and downtime, reduce the emergency occupation of operation and maintenance resources, and ensure the orderly progress of the field operation and maintenance plan.

[0049] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0050] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. An intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things, characterized in that, The system includes the following steps: The data acquisition module is used to collect the operating status data and operating environment data of the target photovoltaic cleaning equipment through the Internet of Things (IoT) sensor unit deployed in the target photovoltaic cleaning equipment. The operation status analysis module is used to calculate the probability of failure of the target photovoltaic cleaning equipment under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning equipment with the remaining power based on the operation status data. The maintenance reachability analysis module is used to perform maintenance reachability analysis on the target photovoltaic cleaning equipment based on the failure probability value and the reachable operating distance, and then generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions. The target identification module is used to filter out all candidate maintenance points located within the reachable maintenance area from a pre-set maintenance resource point location information database, and identify the optimal maintenance target point of the target photovoltaic cleaning equipment based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning equipment and the service queue status of the candidate maintenance point itself. The execution module is used to send collaborative scheduling instructions to the target photovoltaic cleaning equipment and the optimal operation and maintenance target point, and control the target photovoltaic cleaning equipment to go to the optimal operation and maintenance target point to perform preventive maintenance based on the collaborative scheduling instructions.

2. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, The calculation of the failure probability value of the target photovoltaic cleaning equipment under the current operating conditions and the achievable operating distance of the target photovoltaic cleaning equipment with remaining power based on the operating status data specifically includes: The running status data is normalized and preprocessed to obtain preprocessed running status data; Extract fault characteristic parameters and remaining battery power of the target photovoltaic cleaning equipment from the preprocessed operating status data; Based on the fault characteristic parameters and the pre-trained fault probability prediction model, the probability value of the target photovoltaic cleaning equipment under the current operating conditions is determined. Based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning equipment under the current operating conditions, the achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is identified.

3. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 2, characterized in that, The determination of the failure probability value of the target photovoltaic cleaning equipment under the current operating conditions based on the aforementioned fault characteristic parameters and the pre-trained fault probability prediction model specifically includes: Obtain a pre-trained fault probability prediction model; The fault characteristic parameters are input into the fault probability prediction model, and the fault probability prediction model outputs the fault occurrence probability value of the target photovoltaic cleaning equipment under the current operating conditions.

4. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 2, characterized in that, Based on the remaining battery power and the energy consumption model of the target photovoltaic cleaning equipment under its current operating conditions, the achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is identified, specifically including: The energy consumption per unit distance of the target photovoltaic cleaning equipment is calculated based on the energy consumption model under the current operating conditions of the target photovoltaic cleaning equipment. The achievable operating distance of the target photovoltaic cleaning equipment under the remaining battery power is determined based on the remaining battery power and the energy consumption per unit distance.

5. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, Based on the fault occurrence probability value and the reachable operating distance, an operation and maintenance reachability analysis is performed on the target photovoltaic cleaning equipment to generate the reachable operation and maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions. Specifically, this includes: The failure occurrence probability value is compared and analyzed with the preset failure warning probability threshold to determine the failure risk level of the target photovoltaic cleaning equipment. Obtain geographic information data of the photovoltaic field where the target photovoltaic cleaning equipment is located, and construct the maximum operating space range of the target photovoltaic cleaning equipment with the current location as the center based on the geographic information data and the achievable operating distance; Based on the fault risk level, the maximum operating space range is modulated with risk constraints to obtain candidate space regions that meet the fault handling priority. The feasibility of the candidate spatial areas is verified to generate the reachable maintenance area of ​​the target photovoltaic cleaning equipment under the current operating conditions.

6. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, The process of selecting all candidate maintenance points located within the reachable maintenance area from the pre-set maintenance resource point location information database specifically includes: Obtain a pre-set database of operation and maintenance resource point locations, and obtain the geographic coordinate data of all operation and maintenance resource points from the database of operation and maintenance resource point locations; Extract the set of geographic boundary coordinates of the reachable maintenance area; A spatial location matching algorithm is used to determine the spatial affiliation of each operation and maintenance resource point's geographic coordinate data with the geographic boundary coordinate set of the reachable operation and maintenance area; The operation and maintenance resource points whose spatial ownership determination results are located within the reachable operation and maintenance area are selected as candidate operation and maintenance points.

7. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, The optimal maintenance target point for the target photovoltaic cleaning equipment is identified based on the path complexity between the candidate maintenance point and the current location of the target photovoltaic cleaning equipment, as well as the service queue status of the candidate maintenance point itself. Specifically, this includes: Calculate the path complexity between each candidate maintenance point and the current location of the target photovoltaic cleaning equipment based on the geographic information data of the photovoltaic field where the target photovoltaic cleaning equipment is located; Obtain the service queue status of each candidate operation and maintenance point, and generate the service queue congestion index of each candidate operation and maintenance point. A weighted evaluation model of path complexity and service queue congestion index is constructed, and the path complexity and service queue congestion index of each candidate operation and maintenance point are substituted into the weighted evaluation model to calculate the operation and maintenance confidence score, thus obtaining the operation and maintenance confidence score of each candidate operation and maintenance point. The maintenance confidence scores of all candidate maintenance points are sorted in descending order, and the candidate maintenance point with the highest maintenance confidence score is selected as the optimal maintenance target point for the target photovoltaic cleaning equipment.

8. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, Sending collaborative scheduling instructions to the target photovoltaic cleaning equipment and the optimal operation and maintenance target point specifically includes: Obtain the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal maintenance target point; Based on the current location of the target photovoltaic cleaning equipment and the resource configuration data of the optimal operation and maintenance target point, a collaborative scheduling instruction containing the equipment navigation path and the preparation requirements for operation and maintenance resources is generated. The edge communication module of the target photovoltaic cleaning device and the cloud communication interface of the optimal operation and maintenance target point are identified, and collaborative scheduling instructions are sent to the target photovoltaic cleaning device and the optimal operation and maintenance target point respectively through the edge communication module and the cloud communication interface.

9. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, The specific steps of controlling the target photovoltaic cleaning equipment to proceed to the optimal maintenance target point based on the aforementioned collaborative scheduling command include: The target photovoltaic cleaning equipment receives the collaborative scheduling instruction through the edge communication module and extracts the navigation path and maintenance resource preparation requirements from the collaborative scheduling instruction; The target photovoltaic cleaning equipment is controlled to move towards the optimal operation and maintenance target point according to the navigation path. When the target photovoltaic cleaning equipment arrives at the optimal operation and maintenance target point, the preset preventive maintenance operation is started according to the operation and maintenance resource preparation requirements.

10. The intelligent operation and maintenance system for photovoltaic cleaning equipment based on the Internet of Things as described in claim 1, characterized in that, The IoT sensing unit collects the operating status data and operating environment data of the target photovoltaic cleaning equipment at a fixed sampling frequency.

Images