Vehicle-network interaction type self-cleaning light energy storage charging pile and working method thereof

Abstract

The application discloses a vehicle-network interaction type self-cleaning light energy storage charging pile and a working method thereof, and relates to the technical field of light energy storage charging piles. The light energy storage charging pile can adjust the angle of a photovoltaic panel in real time and accurately through the angle of light, and most self-cleaning systems clean when the light intensity is low. When the temperature is low, the cleaning water is prone to freeze on the surface of the photovoltaic panel, thereby affecting the service life of the photovoltaic panel. The cleaning control system obtains hourly weather forecasts and historical power generation data, combines real-time illumination sensor data, constructs an LSTM photovoltaic output prediction model, adjusts the angle of the photovoltaic assembly through a photovoltaic assembly angle adjusting telescopic cylinder, improves the annual power generation of the photovoltaic panel, cleans when the predicted light intensity reaches the set cleaning period, adjusts the angle of the photovoltaic assembly through the photovoltaic assembly angle adjusting telescopic cylinder during cleaning, and sprays cleaning water through a cleaning nozzle to achieve cleaning.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic energy storage charging pile technology, and in particular to a vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile and its working method. Background Technology

[0002] The vehicle-to-grid (V2G) interactive self-cleaning photovoltaic energy storage charging pile is a comprehensive energy system integrating photovoltaic power generation, energy storage, vehicle charging, battery monitoring, and V2G interaction, representing the intelligent direction of future transportation energy management. This system converts solar energy into electricity through photovoltaic panels, with a portion used directly for charging and excess energy stored in the energy storage system, achieving energy self-sufficiency and efficient utilization. Its core V2G technology enables electric vehicles to become mobile energy storage units, supplying power in reverse during peak grid periods, participating in peak shaving and valley filling, significantly alleviating grid pressure, and improving energy regulation capabilities. Furthermore, the system integrates battery monitoring and safety monitoring functions, analyzing battery status in real time during charging and providing timely warnings to comprehensively ensure user safety. This type of charging pile not only reduces carbon emissions and promotes the consumption of renewable energy but also generates revenue through peak-valley electricity price differences, bringing users a more convenient and efficient charging experience. Despite challenges such as high technical complexity, large initial investment, and lack of standardized criteria, with policy support and technological iteration, integrated photovoltaic-storage-charging-discharge systems are developing towards large-scale, commercial, and low-cost applications, becoming a key node in building a green and intelligent energy network and promoting the deep integration and transformation of transportation and energy systems.

[0003] The existing technical solutions mentioned above have the following drawbacks: Existing photovoltaic energy storage charging piles rely solely on real-time illumination data, which can accurately adjust the angle of the photovoltaic panel in real time based on the angle of the light. Moreover, most self-cleaning systems clean when the light intensity is low. When the temperature is low, the cleaning water is prone to freeze on the surface of the photovoltaic panel, affecting the service life of the photovoltaic panel. Therefore, we propose a vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile and its working method to solve the problems mentioned above. Summary of the Invention

[0004] The purpose of this invention is to provide a vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile and its working method, in order to solve the problems mentioned in the background art, that existing photovoltaic energy storage charging piles only rely on real-time light data, can accurately adjust the angle of the photovoltaic panel in real time by means of the light angle, and that the self-cleaning system mostly cleans when the light intensity is low, and when the temperature is low, the cleaning water is easy to freeze on the surface of the photovoltaic panel, thus affecting the service life of the photovoltaic panel.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a vehicle-to-grid interactive self-cleaning photovoltaic energy storage charging pile, comprising a photovoltaic panel canopy, a cleaning control system, and a charging pile management and scheduling system. Two sets of photovoltaic modules are symmetrically installed above the photovoltaic panel canopy. Multiple bidirectional charging piles are equidistantly installed below the photovoltaic panel canopy. Three energy storage battery cabinets are installed on one side of the photovoltaic panel canopy. A cleaning mechanism is installed at the upper end of the photovoltaic panel canopy. The four corners below the photovoltaic modules are connected to the photovoltaic panel canopy via photovoltaic module angle adjustment telescopic cylinders. Cleaning nozzle mounting frames are fixedly installed at both ends of the drive frame, and cleaning nozzles are installed inside the cleaning nozzle mounting frames. A light sensor is fixedly installed on one side of the photovoltaic modules. The cleaning control system acquires hourly weather forecasts and historical power generation data. Based on real-time light sensor data, an LSTM photovoltaic output prediction model is constructed. A basic tracking curve is generated according to the sunrise and sunset trajectories to predict the light intensity for the next two hours. The angle of the photovoltaic modules is adjusted using a telescopic cylinder to increase annual photovoltaic power generation. When the predicted hourly light intensity exceeds 6000 lux, cleaning is initiated at the set cleaning cycle time. During cleaning, the angle of the photovoltaic modules is adjusted to be parallel to the angle of the carport mounting frame using the telescopic cylinder, and an external water pump is activated, spraying cleaning water through cleaning nozzles. The charging pile management and scheduling system is built on the Ethereum public chain to construct a charging transaction alliance chain and deploys a load prediction smart contract. When the regional load rate exceeds 85%, a 10% premium for charging piles in neighboring areas is automatically triggered to guide user diversion. Preferably, multiple carport mounting frames are fixedly installed at equal intervals at the lower end of the photovoltaic carport, and a control box is fixedly installed on the outer side of one side of the carport mounting frame. The G module installed inside the control box acquires hourly weather forecasts and historical power generation data.

[0006] Preferably, the carport fixing frame has a Y-shaped structure, and cleaning screws are installed on the front and rear sides of the carport fixing frame. A slide is driven to the lower end of the cleaning screw, and the lower end of the slide is fixedly connected to the lower end of the drive frame. The two cleaning screws run synchronously to drive the slide to slide synchronously.

[0007] Preferably, a scraper holder is installed at the upper end of the drive frame, and a silicone scraper is installed at the lower end of the scraper holder, with the lower end of the silicone scraper adhering to the upper surface of the photovoltaic module.

[0008] Preferably, the water outlet of the cleaning nozzle is directed toward the edge of one side of the photovoltaic module, and two reflectors are symmetrically installed on the front and rear sides of the photovoltaic module, with the upper end of the reflectors tilted away from the photovoltaic module.

[0009] Limiting rod brackets are symmetrically installed on both sides of the middle position of the carport fixing frame, and a limiting rod is fixedly installed above the limiting rod bracket. The middle position of the drive frame is slidably connected to the limiting rod.

[0010] Preferably, the front and rear sides of the drive frame are cleaned towards the photovoltaic panel canopy. When not cleaning, the drive frame is located on both sides above the photovoltaic modules, and the cleaning nozzles are connected to an external water pump via corrugated hoses.

[0011] Preferably, the upper and lower ends of the photovoltaic module angle adjustment telescopic cylinder are rotatably connected to the photovoltaic module and the photovoltaic panel canopy via rotating shafts, and a drive frame is installed on the outer side of the upper end of the photovoltaic panel canopy.

[0012] Preferably, the charging pile management and scheduling system needs to acquire twelve dynamic parameters, including real-time photovoltaic output, predicted photovoltaic output, real-time grid load factor, peak-valley electricity price period identifier, real-time charging pile utilization rate, remaining battery capacity of the energy storage system, charging module health, average user waiting time, charging demand priority, user V2G participation willingness, ambient temperature, and historical peak charging load. The multi-objective weighted reward calculation formula for the charging pile management and scheduling system is as follows: .

[0013] Incentives for charging station utilization As a reward for grid volatility, Incentives for green electricity consumption rate Rewards for user satisfaction; The weights are dynamic, and the sum is 1.

[0014] A method for operating a vehicle-to-grid interactive, self-cleaning photovoltaic energy storage charging pile includes the following steps: Step 1: The cleaning control system integrates hourly weather forecasts, historical power generation data within the day, and real-time light sensor data to construct an LSTM photovoltaic output prediction model. A basic tracking curve is generated based on sunrise and sunset trajectories, and corrected every 10 minutes using real-time data. The system predicts the future hourly light intensity and adjusts the angle of the photovoltaic modules using a telescopic cylinder to increase annual power generation. When the predicted future hourly light intensity exceeds 6000 lux, cleaning is initiated at the set cleaning cycle time. During cleaning, the photovoltaic module angle adjustment cylinder adjusts the module angle to be parallel to the carport mounting frame. The cleaning screw is activated, causing the slide to slide, which in turn causes the drive frame to slide to the other side. The drive frame then synchronously slides the cleaning nozzle and scraper seat. An external water pump is turned on, and cleaning water is sprayed from the cleaning nozzle to rinse the photovoltaic module surface. Simultaneously, the scraper seat, along with the silicone scraper, removes dust from the photovoltaic module surface along with the cleaning water, achieving cleaning. Step 2: When the weather is sunny and cleaning is not required, the photovoltaic module angle adjustment telescopic cylinder adjusts the angle of the photovoltaic module, allowing the photovoltaic module to adjust its angle according to the change of the sun's angle. At the same time, the reflector reflects the light onto the photovoltaic module, thereby improving the photovoltaic module's power generation efficiency.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention integrates multi-dimensional data into a photovoltaic power output prediction model through a clean control system, and dynamically corrects the angle of the photovoltaic module by combining real-time data. This enables the photovoltaic module to always adapt to the sunlight trajectory, maximize the capture of light energy, effectively improve the power generation efficiency of the photovoltaic module, provide more sufficient green power support for vehicle-grid interactive self-cleaning photovoltaic energy storage charging piles, and ensure the energy supply foundation of the charging and energy storage system. The cleaning mechanism designed in this invention, when cleaning conditions are met, first adjusts the photovoltaic modules to an angle parallel to the carport mounting frame using a telescopic cylinder. Then, with the help of a cleaning screw, drive frame, and other structures, it drives the cleaning nozzle and silicone scraper to work synchronously. This achieves comprehensive cleaning of the photovoltaic module surface, preventing dust and dirt from obstructing light transmission and power generation efficiency. Furthermore, the orderly cleaning action reduces damage to the photovoltaic module surface, extends the photovoltaic module's lifespan, and lowers equipment maintenance costs. The coordinated control of the cleaning action and photovoltaic module angle adjustment, along with the precise control of photovoltaic output prediction for cleaning timing, avoids power generation loss caused by cleaning operations during insufficient sunlight. This ensures that cleaning operations are only carried out during periods of suitable sunlight, minimizing the impact of the cleaning process on power generation efficiency while maintaining the cleanliness of the photovoltaic modules, thus achieving a highly efficient balance between cleaning and power generation functions. Under non-clean operating conditions, this invention relies on the synergistic effect of the photovoltaic module angle adjustment telescopic cylinder and the light sensor to dynamically adjust the angle of the photovoltaic module according to the changes in the sun angle. This ensures that the photovoltaic module is in the optimal position for receiving light at different times, avoiding insufficient light reception caused by the sun's position shift. It continuously maintains a high-efficiency power generation state, providing a stable energy input for the stable charging service of bidirectional charging piles and the energy storage needs of energy storage battery cabinets, and ensuring the continuity of energy supply for the entire vehicle-to-grid interaction system.

[0016] 2. By optimizing the power generation efficiency and stability of photovoltaic modules, this invention can improve the energy self-sufficiency of vehicle-grid interactive self-cleaning photovoltaic energy storage charging piles and reduce dependence on external power grids. At the same time, the green electricity generated by photovoltaic modules can be prioritized for charging and energy storage, reducing dependence on fossil fuel power generation and helping the entire vehicle-grid interactive system achieve low-carbon operation, which is in line with the current infrastructure development needs under the energy transition and dual-carbon goals. Attached Figure Description

[0017] Figure 1 This is the front view of the present invention; Figure 2This is a side view of the photovoltaic panel carport in this invention; Figure 3 In this invention Figure 2 A magnified view of a portion of area A; Figure 4 This is a top view of the photovoltaic panel carport in this invention; Figure 5 In this invention Figure 4 A magnified view of a portion of area B.

[0018] In the diagram: 1. Photovoltaic panel carport; 2. Two-way charging pile; 3. Energy storage battery cabinet; 4. Photovoltaic module; 5. Carport mounting frame; 6. Control box; 7. Drive frame; 8. Cleaning screw; 9. Slide seat; 10. Cleaning nozzle mounting frame; 11. Cleaning nozzle; 12. Scraper seat; 13. Silicone scraper; 14. Photovoltaic module angle adjustment telescopic cylinder; 15. Limit rod bracket; 16. Limit rod; 17. Cleaning mechanism; 18. Reflector; 19. Light sensor. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] Please see Figure 1-5This invention provides an embodiment of a vehicle-to-grid interactive self-cleaning photovoltaic energy storage charging pile, comprising a photovoltaic panel canopy 1, a cleaning control system, and a charging pile management and scheduling system. Two sets of photovoltaic modules 4 are symmetrically installed above the photovoltaic panel canopy 1. Multiple bidirectional charging piles 2 are equidistantly installed below the photovoltaic panel canopy 1. Three energy storage battery cabinets 3 are installed on one side of the photovoltaic panel canopy 1. A cleaning mechanism 17 is installed at the upper end of the photovoltaic panel canopy 1. The four corners below the photovoltaic modules 4 are connected to the photovoltaic panel canopy 1 via photovoltaic module angle adjustment telescopic cylinders 14. Cleaning nozzle mounting brackets 10 are fixedly installed at both ends of the drive frame 7. Cleaning nozzles 11 are installed inside the cleaning nozzle mounting brackets 10. A light sensor 19 is fixedly installed on one side of the photovoltaic modules 4. The cleaning control system acquires 72-hour weather forecasts and historical power generation data and combines them for practical application. The system uses data from the solar sensor 19 to construct an LSTM photovoltaic power output prediction model. Based on the sunrise and sunset trajectories, it generates a basic tracking curve to predict the solar intensity for the next two hours. The angle of the photovoltaic module 4 is adjusted by the telescopic cylinder 14 to increase the annual power generation of the photovoltaic system. When the predicted solar intensity for the next two hours is greater than 6000 lux, cleaning is performed when the set cleaning cycle time is reached. During cleaning, the angle of the photovoltaic module 4 is adjusted by the telescopic cylinder 14 to be parallel to the angle of the carport fixing frame 5. The external water pump is turned on, and the cleaning water is sprayed out through the cleaning nozzle 11 to achieve cleaning. The charging pile management and scheduling system is based on the Ethereum public chain to build a charging transaction alliance chain and deploys a load prediction smart contract. When the regional load rate is greater than 85%, it automatically triggers a 10% premium for charging piles in the neighboring area to guide users to divert traffic.

[0021] The photovoltaic panel carport 1 serves as the core integrated carrier, enabling the orderly integration of photovoltaic modules 4, bidirectional charging piles 2 and 3, energy storage battery cabinets 3, and cleaning mechanisms 17. This achieves integrated operation of energy production, storage, charging, and module cleaning, reducing the space occupied by dispersed equipment and optimizing the energy flow path. The cleaning control system integrates 72-hour weather forecasts, historical power generation data, and real-time monitoring data from the light sensor 19 to construct an LSTM photovoltaic output prediction model to generate a light tracking curve. Then, it uses the photovoltaic module angle adjustment telescopic cylinder 14 to dynamically adjust the angle of the photovoltaic module 4, ensuring that the photovoltaic module 4 always adapts to the solar illumination path. To improve the efficiency of light energy utilization and accurately control the cleaning timing, the photovoltaic module 4 is adjusted to a parallel angle with the carport mounting frame 5 during cleaning. The cleaning nozzle 11, supported by the cleaning nozzle mounting frame 10, achieves efficient cleaning of the surface of the photovoltaic module 4, avoiding dirt from obstructing the power generation effect. The charging pile management and scheduling system is based on the Ethereum public chain to build a charging transaction alliance chain and deploy load prediction smart contracts. When the regional load rate is too high, it automatically triggers the premium of the bidirectional charging pile 2 in the neighboring area, effectively guiding users to divert traffic, preventing local grid overload, and ensuring the stability of the entire vehicle-grid interaction system and the orderly operation of user charging services.

[0022] Please see Figure 1-4 Multiple carport mounting frames 5 are fixedly installed at equal intervals at the lower end of the photovoltaic panel carport 1. A control box 6 is fixedly installed on the outer side of one side of the carport mounting frame 5. The 4G module installed inside the control box 6 acquires 72-hour weather forecasts and historical power generation data. The carport mounting frame 5 has a Y-shaped structure. Cleaning screws 8 are installed on the front and rear sides of the carport mounting frame 5. The lower end of the cleaning screw 8 is driven to install a slide 9. The lower end of the slide 9 is fixedly connected to the lower end of the drive frame 7. The two cleaning screws 8 operate synchronously to drive the slide 9 to slide synchronously.

[0023] The equidistant carport mounting frames 5 at the lower end of the photovoltaic panel carport 1 provide stable support for the overall structure. Their Y-shaped structure further enhances the load-bearing stability, creating a safe and reliable installation and operation environment for the photovoltaic modules 4 above and the equipment below. The control box 6 fixed on the outside of one side of the carport mounting frame 5 efficiently acquires 72-hour weather forecasts and historical power generation data through an internal 4G module, providing accurate data support for the photovoltaic output prediction model of the cleaning control system, ensuring the accuracy of the photovoltaic module angle adjustment telescopic cylinder 14 in adjusting the angle of the photovoltaic module 4 and making cleaning decisions. The cleaning screws 8 installed on the front and rear sides of the carport mounting frame 5 form a transmission cooperation with the slide 9. The synchronous operation of the two cleaning screws 8 can drive the drive frame 7 to slide smoothly, ensuring that the cleaning mechanism 17 can fully cover the surface of the photovoltaic module 4, avoiding cleaning dead corners, and significantly improving the comprehensiveness and reliability of the cleaning operation.

[0024] Please see Figure 2-5 A scraper seat 12 is installed at the upper end of the drive frame 7, and a silicone scraper 13 is installed at the lower end of the scraper seat 12. The lower end of the silicone scraper 13 is in contact with the upper surface of the photovoltaic module 4. The water outlet direction of the cleaning nozzle 11 is towards one edge of the photovoltaic module 4. Two reflectors 18 are symmetrically installed on the front and rear sides of the photovoltaic module 4, and the upper ends of the reflectors 18 are inclined away from the photovoltaic module 4. Limiting rod brackets 15 are symmetrically installed on both sides at the middle position of the carport fixing frame 5. A limiting rod 16 is fixedly installed above the limiting rod bracket 15, and the middle position of the drive frame 7 is slidably connected to the limiting rod 16.

[0025] The silicone scraper 13, mounted on the lower end of the drive frame 7 via the scraper seat 12, fits tightly against the upper surface of the photovoltaic module 4. With the water outlet direction facing the edge of the photovoltaic module 4, it can thoroughly scrape away dirt and cleaning water from the surface of the photovoltaic module 4, preventing residue from affecting the light transmittance of the photovoltaic module 4 and further optimizing the cleaning effect. Reflectors 18, symmetrically installed on the front and rear sides of the photovoltaic module 4 and tilted away from the photovoltaic module 4 at their upper ends, can reflect scattered light from the surrounding area onto the surface of the photovoltaic module 4, increasing the total amount of light received by the photovoltaic module 4 and helping to improve photovoltaic power generation efficiency. Limiting rods 16, mounted on both sides of the middle position via limiting rod brackets 15, are slidably connected to the middle position of the drive frame 7, strictly limiting the sliding trajectory and preventing the drive frame 7 from shifting during cleaning, causing misalignment or damage to the photovoltaic module 4, effectively ensuring the stability of the cleaning action and the operational safety of the photovoltaic module 4.

[0026] Please see Figure 2-5 The front and rear sides of the drive frame 7 are tilted towards the photovoltaic panel canopy 1. When not cleaning, the drive frame 7 is located on both sides above the photovoltaic module 4. The cleaning nozzle 11 is connected to an external water pump through a corrugated hose. The upper and lower ends of the photovoltaic module angle adjustment telescopic cylinder 14 are rotatably connected to the photovoltaic module 4 and the photovoltaic panel canopy 1 through a rotating shaft, respectively.

[0027] The drive frames 7, tilted towards the photovoltaic panel canopy 1 on both sides at the front and rear, are positioned above the photovoltaic modules 4 on both sides when not in a cleaning state. This avoids blocking the light-receiving area of ​​the photovoltaic modules 4 and prevents additional impact on photovoltaic power generation efficiency. The cleaning nozzles 11 are connected to an external water pump via corrugated hoses, which can flexibly adapt to the sliding movements of the drive frames 7, ensuring a continuous and stable water supply during cleaning and preventing cleaning interruptions due to pipeline limitations. The upper and lower ends of the photovoltaic module angle adjustment telescopic cylinder 14 are rotatably connected to the photovoltaic modules 4 and the photovoltaic panel canopy 1 via rotating shafts, making the angle adjustment of the photovoltaic modules 4 more flexible and smooth. It can be precisely adjusted to the optimal light-receiving angle or horizontal cleaning angle, ensuring the reliability of the angle adjustment function and further improving the power generation efficiency of the photovoltaic modules 4 and the convenience of cleaning operations.

[0028] The charging pile management and scheduling system needs to acquire twelve dynamic parameters, including real-time photovoltaic output, predicted photovoltaic output, real-time grid load factor, peak-valley electricity price period indicators, real-time charging pile utilization rate, remaining battery capacity of the energy storage system, charging module health, average user waiting time, charging demand priority, user V2G participation willingness, ambient temperature, and historical peak charging load. The multi-objective weighted reward calculation formula for the charging pile management and scheduling system is as follows: .

[0029] Incentives for charging station utilization As a reward for grid volatility, Incentives for green electricity consumption rate Rewards for user satisfaction; The weights are dynamic, and the sum is 1.

[0030] If the utilization rate is ≥85%, =1; for every 1% decrease, Subtract 0.02 (minimum = -0.5) If volatility is ≤ ±5%, =1; for every 1% exceeding, Subtract 0.1 (minimum = -1) If the green electricity consumption rate is ≥90%, =1.2 (bonus); for every 1% lower, Subtract 0.03 (minimum = -0.3) If the user waiting time is ≤10 minutes and the V2G participation rate is ≥60%, =0.8; otherwise, decrease linearly according to the deviation. Weight dynamic adjustment rules: When the power grid is overloaded (load rate > 85%): =0.4 (originally 0.2) =0.2 (0.3 in the original DDPG algorithm), prioritizing grid stability; when photovoltaic output is high (predicted value > 80% of rated power): =0.5 (0.3 in the DDPG algorithm), prioritizing the absorption of green electricity; peak hours (18:00-22:00): =0.3 (0.2 in the original DDPG algorithm), prioritizing user experience improvement. Priority experience replay pool optimization Priority of assigning deviation values ​​to the 12 base parameters :

[0031] For timing difference error, Priority weights (This is the highest priority compensation in history) Results: High-value samples with load factor deviation > 10% and green electricity consumption deviation > 8% are more likely to be selected for training, increasing the algorithm's convergence speed by 40%.

[0032] State space and action space expansion State-space vector: The 12 dynamic parameters are standardized to form a 28-dimensional vector (containing the current value + the historical values ​​of the previous 3 time steps, capturing the trend of parameter changes):

[0033] Action space definition: Output two types of continuous actions, covering pricing and scheduling: Action 1: Price Coefficient (0.8-1.5 times the benchmark price, with a step size of 0.01) Action 2: Power Allocation Command (0-100% rated power, step size 1%) Attention mechanism embedding (distinguishing based on parameter importance) Design logic: Add a parameter attention module to the input layer of the Actor network to calculate the contribution of each parameter to the current decision. : ( (The attention weight matrix is ​​optimized through training iterations.) Example: Power grid load factor Contribution The value can reach 0.3 under overload (other parameters average 0.03). The algorithm prioritizes adjusting power allocation based on load rate to avoid decision bias.

[0034] Explore noise adaptive adjustment Noise formula: Exploratory noise is generated using the Ornstein-Uhlenbeck process, but the noise intensity (sigma) dynamically changes with parameter stability.

[0035] ( Initial noise, The attenuation coefficient is... (The average rate of change of 12 parameters) Effect: When parameters are stable (e.g., in the valley at night). Reduce the value to 0.05 to avoid ineffective exploration; when parameters fluctuate greatly (such as during the evening peak period). Increased to 0.15 to improve algorithm robustness.

[0036] The coupled computational logic of dynamic parameters and the improved DDPG algorithm (4-step execution process) Step 1: Parameter Preprocessing and Standardization Data cleaning: Outlier handling was performed on the 12 collected parameters (e.g., if the photovoltaic output is greater than 120% of the rated power, it is determined to be a sensor fault, and the average value of the previous 5 minutes is used as a replacement). Standardized formula: (Map the parameters to the [0,1] interval to eliminate magnitude differences, such as 0.5 after standardization for a grid load factor of 50% and 0.6 after standardization for a SOC of 60%) Output: 28-dimensional normalized state vector (Including current value + historical value from the previous 3 moments) Step 2: Improve the forward computation (action output) of DDPG Attention weight calculation: The contribution of 28 parameters is generated through the attention module. The attention state vector is obtained by weighting.

[0037] Actor Network Reasoning: Input an Actor network (3 fully connected layers, 256 / 128 hidden nodes, ReLU activation function), output an action vector. (Price coefficient + power allocation instructions) Action constraint verification: like And the grid load factor is less than 70%, so it is mandatory to Lowered to 1.2 (to avoid excessive premium) like And the energy storage SOC < 30%, Reduce to 80% (to avoid excessive discharge of stored energy). Step 3: Triggering and Executing Blockchain Smart Contracts Contract parameter writing: This involves writing the algorithm's output. (Price coefficient) The power allocation and 12 original parameters are written into the Ethereum consortium blockchain smart contract (developed using Solidity 0.8.17, automatically updated every 5 minutes). Automatic triggering logic: when (Grid overload): Contract automatically executed. (10% premium for neighboring areas), and traffic diversion notifications will be pushed through the app. when (High-output solar power): The contract will automatically increase the V2G subsidy from 0.3 yuan / kWh to 0.35 yuan / kWh, incentivizing users to discharge power. Transaction voucher generation: After charging is complete, the contract is based on... Calculate the green electricity ratio (accuracy ±2%) based on actual green electricity consumption, and generate an immutable blockchain certificate (including user ID, charging duration, green electricity ratio, and revenue amount). Reward value calculation: Based on actual operating results, such as charging pile utilization rate of 88%, grid fluctuation rate of 3%, and green electricity consumption rate of 92%, the reward value is calculated using a multi-objective reward function.

[0038] Experience storage: will Store in the priority experience replay pool, according to Priority allocation Network update: Every 100 steps, 32 high-priority samples are drawn from the replay pool, and the Actor / Critic network parameters (learning rate) are updated using gradient descent. Target network soft update coefficient

[0039] Convergence criterion: When 5000 consecutive steps... The algorithm is considered convergent if the duration of time where volatility is less than 5% and all three objectives (utilization ≥ 85%, volatility ≤ 5%, absorption rate ≥ 90%) are simultaneously met is greater than 90%.

[0040] A method for operating a vehicle-to-grid interactive, self-cleaning photovoltaic energy storage charging pile includes the following steps: Step 1: The cleaning control system integrates 72-hour weather forecasts, historical power generation data over 365 days, and real-time light sensor data to construct an LSTM photovoltaic output prediction model. A basic tracking curve is generated based on sunrise and sunset trajectories, and corrected every 10 minutes using real-time data. The system predicts the light intensity for the next 2 hours and adjusts the angle of the photovoltaic module 4 using the photovoltaic module angle adjustment telescopic cylinder 14 to increase annual photovoltaic power generation. When the predicted light intensity for the next 2 hours is >6000 lux, cleaning is initiated when the set cleaning cycle time is reached. During cleaning, the angle of the photovoltaic module 4 is adjusted by the photovoltaic module angle adjustment telescopic cylinder 14 until it is parallel to the angle of the carport fixing frame 5. The cleaning screw 8 is activated, causing the sliding seat 9 to slide. The sliding seat 9 causes the drive frame 7 to slide to the other side. The drive frame 7 causes the cleaning nozzle 11 and scraper seat 12 to slide synchronously. The external water pump is turned on, and cleaning water is sprayed out through the cleaning nozzle 11 to rinse the surface of the photovoltaic module 4. Simultaneously, the scraper seat 12 drives the silicone scraper 13 to scrape away dust from the surface of the photovoltaic module 4 along with the cleaning water, achieving cleaning. Step 2: When the weather is sunny and cleaning is not required, the photovoltaic module angle adjustment telescopic cylinder 14 adjusts the angle of the photovoltaic module 4, allowing the photovoltaic module 4 to adjust its angle according to the change of the sun's angle. At the same time, the reflector 18 reflects the light onto the photovoltaic module 4, thereby improving the power generation efficiency of the photovoltaic module 4.

[0041] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A vehicle-network interaction type self-cleaning light energy storage charging pile, comprising a photovoltaic panel carport (1), a driving frame (7), a cleaning control system and a charging pile management and scheduling system, characterized in that: Two sets of photovoltaic modules (4) are symmetrically installed above the photovoltaic panel carport (1). Multiple bidirectional charging piles (2) are installed at equal intervals below the photovoltaic panel carport (1). Three energy storage battery cabinets (3) are installed on one side of the photovoltaic panel carport (1). A cleaning mechanism (17) is installed at the upper end of the photovoltaic panel carport (1). The four corners below the photovoltaic modules (4) are connected to the photovoltaic panel carport (1) through photovoltaic module angle adjustment telescopic cylinders (14). Cleaning nozzle fixing brackets (10) are fixedly installed at both ends of the drive frame (7). Cleaning nozzles (11) are installed inside the cleaning nozzle fixing brackets (10). A light sensor (19) is fixedly installed on one side of the photovoltaic module (4). The cleaning control system obtains 72-hour weather forecasts and historical power generation data combined with real-time light sensor data. (19) Data, construct an LSTM photovoltaic power output prediction model, generate a basic tracking curve based on the sunrise and sunset trajectory, predict the light intensity in the next 2 hours, and adjust the angle of the photovoltaic module (4) through the photovoltaic module angle adjustment telescopic cylinder (14) to increase the annual photovoltaic power generation. When the predicted light intensity in the next 2 hours is >6000lx, cleaning is carried out when the cleaning cycle is set. During cleaning, the angle of the photovoltaic module (4) is adjusted to be parallel to the angle of the carport fixing frame (5) through the photovoltaic module angle adjustment telescopic cylinder (14), and the external water pump is turned on. The cleaning water is sprayed out through the cleaning nozzle (11) to achieve cleaning. The charging pile management and scheduling system is based on the Ethereum public chain to build a charging transaction alliance chain and deploy a load prediction smart contract. When the regional load rate is >85%, the premium of charging piles in the neighboring area is automatically triggered by 10%. ​ The carport fixing frame (5) has a Y-shaped structure. Cleaning screws (8) are installed on the front and rear sides of the carport fixing frame (5). A slide (9) is installed at the lower end of the cleaning screw (8). The lower end of the slide (9) is fixedly connected to the lower end of the drive frame (7). The two cleaning screws (8) run synchronously to drive the slide (9) to slide synchronously. The water outlet of the cleaning nozzle (11) is directed toward the edge of the photovoltaic module (4). Two reflectors (18) are symmetrically installed on the front and rear sides of the photovoltaic module (4). The upper end of the reflector (18) is inclined away from the photovoltaic module (4). The charging pile management and scheduling system needs to acquire twelve dynamic parameters, including real-time photovoltaic output, photovoltaic output prediction, real-time grid load rate, peak-valley electricity price period identifier, real-time charging pile utilization rate, remaining battery capacity of energy storage system, charging module health, average user waiting time, charging demand priority, user V2G participation willingness, ambient temperature, and historical charging load peak value.

2. The vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile according to claim 1, characterized in that: Multiple carport mounting brackets (5) are fixedly installed at equal intervals at the lower end of the photovoltaic carport (1). A control box (6) is fixedly installed on the outer side of one side of the carport mounting bracket (5). The 4G module installed inside the control box (6) acquires 72-hour weather forecasts and historical power generation data.

3. The vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile according to claim 2, characterized in that: The upper end of the drive frame (7) is equipped with a scraper seat (12), and the lower end of the scraper seat (12) is equipped with a silicone scraper (13). The lower end of the silicone scraper (13) is in contact with the upper surface of the photovoltaic module (4).

4. The vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile according to claim 3, characterized in that: Limiting rod brackets (15) are symmetrically installed on both sides of the middle position of the carport fixing frame (5), and a limiting rod (16) is fixedly installed above the limiting rod bracket (15). The middle position of the drive frame (7) is slidably connected to the limiting rod (16).

5. A vehicle-to-grid interactive self-cleaning photovoltaic energy storage charging pile according to claim 4, characterized in that: The front and rear sides of the drive frame (7) are inclined toward the photovoltaic panel canopy (1). When not cleaning, the drive frame (7) is located on both sides above the photovoltaic module (4). The cleaning nozzle (11) is connected to an external water pump through a corrugated hose.

6. The vehicle-grid interactive self-cleaning photovoltaic energy storage charging pile according to claim 5, characterized in that: The upper and lower ends of the photovoltaic module angle adjustment telescopic cylinder (14) are rotatably connected to the photovoltaic module (4) and the photovoltaic panel canopy (1) respectively via rotating shafts.

7. A method for operating a vehicle-to-grid interactive self-cleaning photovoltaic energy storage charging pile as described in claim 6, characterized in that, Includes the following steps: Step 1: The cleaning control system integrates 72-hour weather forecasts, historical power generation data over 365 days, and real-time light sensor data to construct an LSTM photovoltaic output prediction model. Based on the sunrise and sunset trajectories, a basic tracking curve is generated and corrected every 10 minutes using real-time data. The system predicts the light intensity for the next 2 hours and adjusts the angle of the photovoltaic module (4) using the photovoltaic module angle adjustment telescopic cylinder (14) to increase the annual photovoltaic power generation. When the predicted light intensity for the next 2 hours is >6000 lx, cleaning is performed when the set cleaning cycle time is reached. During cleaning, the angle of the photovoltaic module is adjusted. The telescopic cylinder (14) adjusts the angle of the photovoltaic module (4) to be parallel to the angle of the carport fixing frame (5), and starts the cleaning screw (8). The cleaning screw (8) drives the slide (9) to slide, and the slide (9) drives the drive frame (7) to slide to the other side. The drive frame (7) drives the cleaning nozzle (11) and the scraper seat (12) to slide synchronously. The external water pump is turned on, and the cleaning water is sprayed out through the cleaning nozzle (11) to rinse the surface of the photovoltaic module (4). At the same time, the scraper seat (12) drives the silicone scraper (13) to scrape away the dust on the surface of the photovoltaic module (4) along with the cleaning water to achieve cleaning. Step 2: When the weather is sunny and cleaning is not required, the photovoltaic module angle adjustment telescopic cylinder (14) adjusts the angle of the photovoltaic module (4) so ​​that the photovoltaic module (4) adjusts the angle with the change of the sun angle. At the same time, the reflector (18) reflects the light onto the photovoltaic module (4) to improve the power generation efficiency of the photovoltaic module (4).

Images