Light commercial electric vehicle photovoltaic centralized charging control method, device and medium

Abstract

This application discloses a method, equipment, and medium for centralized photovoltaic charging control of light commercial electric vehicles, belonging to the technical field of new energy charging. The method includes: acquiring weather data, vehicle battery state of charge, and task scheduling data; processing photovoltaic module parameters and environmental data using an LSTM neural network model to generate a photovoltaic power generation prediction curve; constructing a multi-dimensional evaluation index system including task urgency, charging efficiency, and grid interaction factors, and generating a charging priority sequence using the entropy-weighted TOPSIS algorithm; solving a real-time power allocation scheme based on a mixed integer programming algorithm to dynamically adjust the charging pile output power; monitoring the system operating status and triggering a multi-level charging mechanism when photovoltaic power is insufficient or low; updating battery status and photovoltaic power in real time, and correcting the charging strategy through a rolling optimization algorithm. This application achieves centralized charging control of light commercial electric vehicles through the above method to meet the needs of a fleet.

Description

Technical Field

[0001] This application relates to the technical field of new energy charging, and in particular to a photovoltaic centralized charging control method, equipment and medium for light commercial electric vehicles. Background Technology

[0002] In the field of new energy charging technology, centralized charging systems for light commercial electric vehicles have become crucial infrastructure in scenarios such as logistics parks and bus depots. Existing technologies generally employ an architecture primarily powered by the power grid, supplemented by fixed photovoltaic (PV) and energy storage devices, using AC-DC conversion to supply electricity. In typical solutions, PV arrays and energy storage systems are configured in a fixed ratio, and a centralized inverter converts DC power to AC power before supplying it to the charging piles. Energy storage systems often use lithium iron phosphate battery packs with static capacity allocation, and charging pile power allocation is based on a first-come, first-served principle, supplemented by a simple time-based electricity pricing strategy.

[0003] While existing technologies have achieved some degree of utilization of photovoltaic energy, significant shortcomings remain. First, energy utilization efficiency is limited by the AC-DC multi-stage conversion architecture, and insufficient direct photovoltaic power supply leads to overall system inefficiency. Second, energy storage systems employ a rigid configuration model, failing to dynamically respond to the spatiotemporal changes in fleet charging demand, resulting in redundant storage capacity. Finally, existing charging scheduling strategies lack multi-dimensional decision-making capabilities, allocating power solely based on vehicle SOC or simple time rules, leading to frequent charging interruptions or grid replenishment delays in scenarios with fluctuating photovoltaic power generation.

[0004] Therefore, how to centrally control the charging of light commercial electric vehicles to meet the needs of the fleet has become an urgent technical problem to be solved. Summary of the Invention

[0005] This application provides a method, device, and medium for centralized photovoltaic charging control of light commercial electric vehicles to solve the following technical problem: how to centrally control the charging of light commercial electric vehicles to meet fleet needs.

[0006] In a first aspect, embodiments of this application provide a photovoltaic centralized charging control method for light commercial electric vehicles, applied to a photovoltaic charging system. The photovoltaic charging system is used to charge light commercial electric vehicles. The photovoltaic charging system includes a charging device, which includes multiple charging piles. The charging device is connected to a photovoltaic power generation device, an energy storage device, and an external power grid. The method is characterized by: acquiring weather data and vehicle data of at least one light commercial electric vehicle; wherein the vehicle data includes battery state of charge and a preset task schedule; processing the weather data and photovoltaic power generation device data based on a preset photovoltaic power generation prediction algorithm to generate a photovoltaic power generation prediction curve; and calculating based on the battery state of charge, the preset task schedule, and the photovoltaic power generation prediction curve. The system prioritizes the charging of multiple light commercial electric vehicles; determines the real-time photovoltaic power generation and processes the charging priority sequence and real-time photovoltaic power generation through a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles; controls multiple charging piles to perform charging operations according to the real-time power allocation scheme and monitors the operating status of the photovoltaic charging system. When insufficient photovoltaic power generation from the photovoltaic power generation device and / or the energy storage capacity of the energy storage device is detected to be lower than the power threshold, supplementary power is provided based on the external power grid; during the charging process, the battery state of charge and real-time photovoltaic power generation of multiple light commercial electric vehicles are updated in real time. If the expected result exceeds the charging threshold value, the charging priority sequence is recalculated and the real-time power allocation scheme is updated.

[0007] In one implementation of this application, a photovoltaic power generation prediction curve is generated by processing weather data and photovoltaic power generation device data based on a preset photovoltaic power generation prediction algorithm. Specifically, this includes: acquiring photovoltaic module parameters and environmental monitoring data of the photovoltaic power generation device; wherein the photovoltaic module parameters include the module efficiency degradation curve and the photovoltaic array tilt angle adjustment range, and the environmental monitoring data includes the module surface temperature and ambient irradiance; performing time-series analysis on the weather data using a preset LSTM neural network model, the weather data including cloud cover change data for a certain future time period and historical seasonal solar intensity distribution data; inputting the photovoltaic module parameters into the LSTM neural network model to generate an initial photovoltaic power generation prediction curve; and correcting the initial prediction curve based on the module surface temperature and ambient irradiance by comparing it with a preset temperature-power reduction coefficient lookup table to generate the final photovoltaic power generation prediction curve.

[0008] In one implementation of this application, a charging priority sequence for multiple light commercial electric vehicles is calculated based on the battery state of charge, a preset task schedule, and a photovoltaic power generation prediction curve. Specifically, this includes: constructing a multi-dimensional evaluation index system based on the battery state of charge and the preset task schedule; wherein the index system includes a task urgency factor, a charging efficiency factor, and a grid interaction factor; processing the multi-dimensional evaluation index system using a preset entropy-weighted TOPSIS algorithm to generate dynamic priority scores for multiple light commercial electric vehicles; performing spatiotemporal matching of the dynamic priority scores with the photovoltaic power generation prediction curve to determine the photovoltaic priority charging period and the energy storage compensation charging period; and generating a charging priority sequence based on the time period matching results; wherein the charging priority sequence includes a time period allocation table and power fluctuation tolerance parameters for each charging pile.

[0009] In one implementation of this application, a multi-dimensional evaluation index system is constructed based on the battery state of charge and a preset task schedule. Specifically, this includes: determining the available charging time based on the planned departure timestamp and current time in the preset task schedule, and deducting a preset safety margin to determine the actual available charging time; obtaining a preset task type weight table; wherein the task type weight table includes at least the urgency coefficients corresponding to cold chain transportation tasks, general freight tasks, and standby vehicle tasks; processing the actual available charging time, battery state of charge, and preset task type weight table based on a preset exponential decay model to generate a task urgency factor; calculating the maximum allowable charging current based on battery pack temperature distribution data, and determining the charging efficiency factor based on the ratio of the actual output power to the rated power of the charging pile; obtaining the time-of-use electricity price signal and regional load factor of the external power grid to determine the grid interaction factor; and constructing a multi-dimensional evaluation index system based on the task urgency factor, charging efficiency factor, and grid interaction factor.

[0010] In one implementation of this application, the real-time photovoltaic power generation is determined, and the charging priority sequence and real-time photovoltaic power generation are processed by a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles. Specifically, this includes: real-time acquisition of voltage and current data of the photovoltaic power generation device, and determination of the real-time photovoltaic power generation through a preset moving average filtering algorithm; establishment of a dynamic power allocation matrix; wherein the dynamic power allocation matrix includes charging pile number, real-time photovoltaic power generation, charging priority sequence, and charging efficiency curve; solving for the optimal solution of the dynamic power allocation matrix based on a preset mixed integer programming algorithm; wherein the objective function of the mixed integer programming algorithm is to minimize the total charging completion time, and the constraints include the maximum power limit of the charging pile and the battery safe charging current limit; generating a real-time power allocation scheme based on the optimal solution, and calculating the expected results for multiple light commercial electric vehicles based on the real-time power allocation scheme.

[0011] In one implementation of this application, multiple charging piles are controlled to perform charging operations according to a real-time power allocation scheme, and the operating status of the photovoltaic charging system is monitored. When insufficient photovoltaic power generation from the photovoltaic power generation device and / or the energy storage capacity of the energy storage device is detected to be lower than the energy threshold, supplementary power is provided based on the external power grid. Specifically, this includes: real-time acquisition of the output power data of the photovoltaic power generation device, and determining the current total charging demand by accumulating the power allocation values ​​of all charging piles in the real-time power allocation scheme; calculating the difference between the photovoltaic power generation and the total charging demand based on a preset power difference algorithm; when the difference continues to be negative for more than a first preset time, it is determined that the photovoltaic power generation is insufficient, and the energy storage device is called to provide supplementary power; periodically acquiring the remaining power of the energy storage device. The remaining power percentage is compared with a power threshold; when the remaining power percentage is lower than the power threshold, a grid power replenishment command is triggered to enter the power replenishment mode; time-of-use electricity price data of the external grid is obtained; based on the time-of-use electricity price data, the power allocation weight of the external grid is calculated, and a grid power replenishment power allocation scheme is generated; the grid power replenishment power allocation scheme includes the grid power supply ratio and phase synchronization parameters of each charging pile; the charging pile is controlled to switch to grid power supply mode, power allocation is performed based on the grid power replenishment power allocation scheme, and the grid load rate is monitored in real time to see if it exceeds the preset safety range; when it is detected that the photovoltaic power generation power recovers to the total charging demand or the energy storage device power recovers to the safety threshold, the grid power supply ratio is gradually reduced until the power replenishment mode is exited.

[0012] In one implementation of this application, the battery state of charge (SOC) and real-time photovoltaic (PV) power generation of multiple light commercial electric vehicles are updated in real time during the charging process. If the expected result exceeds the charging threshold, the charging priority sequence is recalculated and the real-time power allocation scheme is updated. Specifically, this includes: collecting incremental SOC data of multiple light commercial electric vehicles and actual output power data of the charging pile at preset intervals; inputting the incremental SOC data into a preset charging efficiency evaluation model to generate an actual charging rate curve; comparing the actual charging rate curve with the expected result and calculating the time integral difference between the two; when the time integral difference exceeds the charging threshold, reconstructing a multi-dimensional evaluation index system based on the latest collected SOC and the corrected PV power generation prediction curve; reordering the dynamic priority scores based on a preset rolling optimization algorithm to generate an updated charging priority sequence; and updating the real-time power allocation scheme according to the updated charging priority sequence and constraints.

[0013] In one implementation of this application, the method further includes: collecting historical operation data of a complete charging cycle to construct a charging performance evaluation dataset; processing the charging performance evaluation dataset through a preset feature extraction algorithm to extract a set of key performance indicators; and constructing a random forest regression model based on the set of key performance indicators to analyze the impact of parameters in the set of key performance indicators on the photovoltaic charging system and generate charging strategy optimization suggestions.

[0014] Secondly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes a computer program to implement the photovoltaic centralized charging control method for light commercial electric vehicles as described above.

[0015] Thirdly, embodiments of this application also provide a computer-readable storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements the photovoltaic centralized charging control method for light commercial electric vehicles as described above.

[0016] This application provides a photovoltaic centralized charging control method, equipment, and medium for light commercial electric vehicles. By utilizing an LSTM neural network-based photovoltaic power generation prediction model, combined with a dynamic correction mechanism for photovoltaic module parameters and environmental monitoring data, the spatiotemporal accuracy of power generation prediction is effectively enhanced, providing a reliable basis for energy dispatch. A multi-dimensional evaluation index system incorporating task urgency, charging efficiency, and grid interaction is constructed, and an entropy-weighted TOPSIS algorithm is used to dynamically generate a charging priority sequence. This achieves precise matching between vehicle task scheduling, battery status, and photovoltaic power generation characteristics, improving the rationality of resource allocation and the charging guarantee capability of high-priority vehicles. The dynamic power allocation algorithm is optimized using a mixed integer programming model, shortening the total charging completion time and improving the utilization rate of charging facilities while meeting charging pile power constraints and battery safety. The real-time monitoring and feedback mechanism dynamically adjusts the charging strategy through moving average filtering, power difference calculation, and rolling optimization algorithms, ensuring seamless switching to grid-supplied power mode during photovoltaic fluctuations or insufficient energy storage, enhancing the system's operational stability and environmental adaptability. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A flowchart illustrating a photovoltaic centralized charging control method for a light commercial electric vehicle, provided in an embodiment of this application;

[0019] Figure 2 This is a schematic diagram of the internal structure of a photovoltaic centralized charging control device for a light commercial electric vehicle, provided as an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] This application provides a method, device, and medium for centralized photovoltaic charging control of light commercial electric vehicles to solve the following technical problem: how to centrally control the charging of light commercial electric vehicles to meet fleet needs.

[0022] The technical solutions proposed in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0023] Figure 1 This document provides a flowchart of a photovoltaic centralized charging control system for a light commercial electric vehicle, as illustrated in an embodiment of this application. Figure 1 As shown in the figure, the photovoltaic centralized charging control method for a light commercial electric vehicle provided in this application embodiment specifically includes the following steps:

[0024] Step 1: Obtain weather data and vehicle data for at least one light commercial electric vehicle; the vehicle data includes battery state of charge and preset task schedule.

[0025] Weather data refers to real-time and forecasted weather information obtained through meteorological service interfaces or environmental sensors deployed at charging stations, including parameters such as light intensity, cloud cover, ambient temperature, and humidity.

[0026] Vehicle data: refers to vehicle operation data obtained through in-vehicle terminals (such as TBOX) or logistics management systems, including battery state of charge (SOC, i.e., remaining battery percentage) and preset task schedules (such as planned departure time, task type and priority label).

[0027] In one specific case, a photovoltaic charging system deployed in a logistics park is required to provide charging services for 30 light commercial electric vehicles.

[0028] Call a third-party meteorological service API to obtain cloud cover change data (e.g., cloud cover fluctuating between 20% and 60%), ambient temperature (28℃-32℃), and light intensity (peak 800W / m²) for the next 6 hours. 2 ).

[0029] The battery SOC of each vehicle is collected in real time through the vehicle-mounted TBOX (e.g., the SOC of vehicle A is 40%, and that of vehicle B is 65%).

[0030] Synchronize the task scheduling data of the logistics management system and extract key parameters (such as vehicle C needing to perform a cold chain transportation task in 2 hours, with the priority marked as "urgent").

[0031] Step 2: Process weather data and photovoltaic power generation device data based on the preset photovoltaic power generation prediction algorithm to generate a photovoltaic power generation prediction curve.

[0032] Step 2.1: Obtain photovoltaic module parameters and environmental monitoring data of the photovoltaic power generation device; wherein, photovoltaic module parameters include module efficiency degradation curve and photovoltaic array tilt angle adjustment range, and environmental monitoring data include module surface temperature and environmental irradiance.

[0033] Photovoltaic module parameters: inherent property parameters of photovoltaic panels, including the module efficiency degradation curve (describing the degradation law of photovoltaic panel output power with the years of use) and photovoltaic array tilt angle adjustment range (the range of tilt angles allowed by the mechanical structure of photovoltaic panels).

[0034] Environmental monitoring data: Real-time collected physical environmental parameters, including component surface temperature (real-time temperature value of photovoltaic panel surface) and environmental irradiance (solar radiation intensity received per unit area).

[0035] In one instance of this application:

[0036] Temperature sensors (such as infrared temperature measurement modules) and irradiance sensors (such as photodiode arrays) are installed in the photovoltaic array, and the sampling frequency is set to once per minute.

[0037] Read the component efficiency degradation curve from the photovoltaic system monitoring terminal. The format is a year-degradation percentage comparison table (e.g., 2% degradation in year 1, 3.5% degradation in year 2).

[0038] Obtain tilt adjustment range parameters from the photovoltaic support control system (e.g., the mechanical structure supports tilt adjustment from 15° to 60°).

[0039] Step 2.2: Perform time series analysis on weather data using a pre-set LSTM neural network model. The weather data includes cloud cover changes over a certain period of time and historical seasonal light intensity distribution data.

[0040] LSTM neural network model: a deep learning model suitable for time series forecasting that can capture long-term dependencies.

[0041] Weather data includes cloud cover variation data for a future period (usually set to 6 hours) (cloud cover forecast every 30 minutes) and historical seasonal light intensity distribution data (average hourly irradiance for the same period over the past three years).

[0042] In one instance of this application:

[0043] Obtain cloud cover data for the next 6 hours from the meteorological service interface, in the format of a time-cloud cover percentage series (e.g., 12:00-30%, 12:30-50%).

[0044] Load the historical light intensity database and extract the typical irradiance curve for the current season (e.g., average irradiance of 200W / m2 in winter and 600W / m2 in summer).

[0045] The model uses a pre-trained LSTM model with the input dimension being the time step (6 hours × 2 data points / hour) and the output dimension being the predicted power sequence.

[0046] By inputting weather data into the model, the system automatically analyzes the correlation between abrupt changes in cloud cover and sunlight intensity, and outputs the initial power generation forecast every 30 minutes for the next 6 hours.

[0047] Step 2.3: Input the photovoltaic module parameters into the LSTM neural network model to generate the initial photovoltaic power generation prediction curve.

[0048] Initial photovoltaic power generation prediction curve: The theoretical prediction value without environmental factor correction, reflecting the power generation capacity under ideal operating conditions.

[0049] In one instance of this application:

[0050] Convert the component efficiency degradation factor to the degradation percentage corresponding to the current year (e.g., 5% degradation rate for the 3rd year after installation).

[0051] The real-time tilt angle value (e.g., 35°) is normalized to a value in the range of 0-1 (35 / 60≈0.58).

[0052] The encoded component parameters are combined with weather data to form a multidimensional input vector (such as cloud cover, historical irradiance, efficiency decay, and tilt angle).

[0053] The model outputs an initial prediction curve, with the data format being time-power key-value pairs (e.g., 12:00-50kW, 12:30-45kW).

[0054] Step 2.4: Based on the component surface temperature and ambient irradiance, the initial prediction curve is corrected by comparing it with the preset temperature power reduction coefficient table to generate a photovoltaic power generation prediction curve.

[0055] Temperature Power Reduction Factor Comparison Table: A reference table describing the output power loss of photovoltaic panels at different temperatures. For example, for every 1°C increase in temperature, the power decreases by 0.5%.

[0056] In one instance of this application:

[0057] To find the derating rate corresponding to the current component surface temperature (e.g., 45℃): (45-25)×0.5%=10% (based on the standard temperature of 25℃).

[0058] The correction range is adjusted based on the actual irradiance: if the actual irradiance is 20% lower than the predicted value, the compensation coefficient is increased by 5%.

[0059] The initial predicted value is corrected point by point: for example, the initial predicted power at 12:00 is 50kW, which is reduced to 45kW after a 10% temperature reduction; after a 5% irradiance compensation, the final power is 47.25kW.

[0060] Output the corrected time-power sequence and label the reason for the correction (e.g., "Temperature correction -10%" "Irradiance compensation +5%").

[0061] In a specific example, a photovoltaic charging system in a logistics park needs to generate a power generation forecast curve for the next day:

[0062] The efficiency decay curve shows a 3% decay rate in the second year, with the current tilt angle set at 40°.

[0063] The component surface temperature is 38℃ and the ambient irradiance is 480W / m2.

[0064] The weather forecast shows that the cloud cover will increase from 20% to 60% between 10:00 and 14:00 the next day, and the historical average solar irradiance for the same period is 550W / m2.

[0065] The initial predicted values ​​output by the LSTM model are: 10:00-120kW, 12:00-100kW, and 14:00-80kW.

[0066] Temperature correction: At 38℃, the power derating rate is 6.5% according to the table.

[0067] Irradiance compensation: The actual irradiance is 12.7% lower than the historical average, so the compensation coefficient is increased by 3%.

[0068] Corrected power: 12:00 Original 100kW to 100×(1-6.5%+3%)=96.5kW.

[0069] Step 3: Calculate the charging priority sequence of multiple light commercial electric vehicles based on the battery state of charge, preset task schedule, and photovoltaic power generation prediction curve.

[0070] Step 3.1: Construct a multi-dimensional evaluation index system based on the battery state of charge and the preset task schedule; the index system includes task urgency factor, charging efficiency factor and grid interaction factor.

[0071] Step 3.1.1: Based on the planned departure timestamp in the preset task schedule and the current time, determine the available charging time and deduct the preset safety margin to determine the actual available charging time.

[0072] Available charging time: The difference between the vehicle's scheduled departure time and the current time.

[0073] Safety margin: A buffer time (e.g., 30 minutes) reserved to prevent unexpected delays.

[0074] In one instance of this application:

[0075] Parse the planned departure time (e.g., 14:00) in the vehicle task schedule and obtain the current system time (e.g., 11:30).

[0076] Calculate the initial available time: 14:00 - 11:30 = 2.5 hours.

[0077] Deducting safety margin: 2.5 hours - 0.5 hours = 2 hours.

[0078] Step 3.1.2: Obtain the preset task type weight table; wherein, the task type weight table shall at least include the urgency coefficients corresponding to cold chain transportation tasks, general freight tasks and backup vehicle tasks.

[0079] Task Type Weight Table: Defines the urgency of different task types; the higher the weight, the higher the charging priority.

[0080] In one instance of this application:

[0081] Load the preset weight table:

[0082] Cold chain transportation task: Weight 1.5 (requires maintaining a low-temperature environment)

[0083] General freight transport task: Weight 1.0 (standard timeliness requirement)

[0084] Backup vehicle mission: Weight 0.7 (no urgent need)

[0085] The weight value is matched according to the vehicle task type (e.g., the weight of cold chain transportation task is 1.5).

[0086] Step 3.1.3: Based on the preset exponential decay model, process the actual available charging time, battery state of charge and preset task type weight table to generate task urgency factor.

[0087] Exponential decay model: A mathematical model in which the urgency level decreases exponentially over time.

[0088] In one instance of this application:

[0089] Enter the actual available time (2 hours), current SOC (e.g., 30%), and task weight (1.5).

[0090] Calculate the urgency factor: Urgency = Weight × Exponential decay function (available time) × (1-SOC / 100);

[0091] For example, when the attenuation coefficient k = 0.2 and SOC = 30%, 1.5 × e^(-0.2 × 2) × 0.7 ≈ 0.94.

[0092] Step 3.1.4: Determine the maximum allowable charging current based on the battery data of multiple light commercial electric vehicles, and determine the charging efficiency factor based on the ratio of the actual output power to the rated power of the charging pile.

[0093] Maximum permissible charging current: The upper limit of safe current determined by battery temperature and state of health (SOH).

[0094] In one instance of this application:

[0095] Obtain the battery temperature (e.g., 35°C) and SOH (e.g., 90%) from the BMS.

[0096] Consult the battery manufacturer's safety charging curve to determine the maximum allowable current (e.g., 100A).

[0097] Calculate the power utilization rate: Actual output power (45kW) / Rated power (50kW) = 0.9.

[0098] Overall efficiency factor: 0.6 × (actual current / maximum current) + 0.4 × power utilization rate

[0099] For example, the actual current is 90A, which is 0.6×0.9+0.4×0.9=0.9.

[0100] Step 3.1.5: Obtain the time-of-use electricity price signal and regional load factor of the external power grid to determine the power grid interaction factor.

[0101] Time-of-use pricing: A pricing strategy based on time periods (e.g., 1.2 yuan / kWh during peak hours and 0.5 yuan / kWh during off-peak hours).

[0102] In one instance of this application:

[0103] Access the power grid system to obtain the current electricity price (e.g., 0.8 yuan / kWh for the average section) and the regional load factor (e.g., 70%).

[0104] Calculate the interaction factor:

[0105] Basic cost = Electricity price × Estimated charging amount (e.g., 50kWh to 0.8 × 50 = 40 yuan)

[0106] Interaction factor = 1 / (1 + cost coefficient) × environmental benefit bonus

[0107] For example, the load factor penalty is 0.07 to the interaction factor = 1 / (1+40.07) = 0.024.

[0108] Step 3.1.6: Construct a multi-dimensional evaluation index system based on task urgency factor, charging efficiency factor and grid interaction factor.

[0109] In one instance of this application:

[0110] Integrate all factors to generate an evaluation matrix:

[0111] Vehicles; urgency factor; efficiency factor; interaction factor;

[0112] V001; 0.94; 0.9; 0.024;

[0113] V002; 0.41; 0.85; 0.015.

[0114] Step 3.2: Based on the preset entropy weight TOPSIS algorithm, process the multidimensional evaluation index system to generate dynamic priority scores for multiple light commercial electric vehicles.

[0115] Entropy-weighted TOPSIS: A decision-making algorithm that uses information entropy to objectively assign weights and combines them with the distance ranking of ideal solutions.

[0116] In one instance of this application:

[0117] Data standardization: normalize urgency, efficiency, and interaction factors to the [0,1] interval.

[0118] Entropy weight calculation: Calculate the information entropy of each factor. The smaller the entropy value, the higher the discrimination and the greater the weight.

[0119] For example, the weighting is: urgency 0.5, efficiency 0.3, interaction 0.2.

[0120] TOPSIS sorting: Determine the positive ideal solution (0.94, 0.9, 0.024) and the negative ideal solution (0.41, 0.85, 0.015).

[0121] Calculate the Euclidean distance between each vehicle and the positive / negative ideal solution, and generate a priority score:

[0122] Score = Distance to negative ideal solution / Distance to positive ideal solution + Distance to negative ideal solution;

[0123] For example, V001 has a score of 0.82 (high priority) and V002 has a score of 0.65 (medium priority).

[0124] Step 3.3: Perform spatiotemporal matching between the dynamic priority score and the photovoltaic power generation prediction curve to determine the photovoltaic priority charging period and the energy storage compensation charging period.

[0125] Spatiotemporal matching: dynamically aligning vehicle charging demand with the spatiotemporal distribution characteristics of photovoltaic power generation.

[0126] In one instance of this application:

[0127] Photovoltaic power segmentation: Discretize the prediction curve in 30-minute intervals (e.g., predict 50kW from 12:00 to 12:30).

[0128] Establish matching rules:

[0129] High-priority vehicles will be allocated to periods with sufficient solar power (e.g., 12:00-13:30).

[0130] Low- and medium-priority vehicles are allocated to periods of photovoltaic degradation or energy storage power supply.

[0131] Capacity verification: Total demand for the period ≤ 110% of the photovoltaic forecast (allowing 10% fluctuation redundancy).

[0132] For example, if the total demand from 12:00 to 12:30 is 45kW ≤ 50kW × 110% = 55kW, then the match is successful.

[0133] Step 3.4: Generate a charging priority sequence based on the time period matching results; wherein, the charging priority sequence includes the time period allocation table and power fluctuation tolerance parameters for each charging pile.

[0134] In one instance of this application:

[0135] Generate the scheduling matrix:

[0136] Charging station; time period; vehicle; power; tolerance;

[0137] P01;12:00-13:30;V001;50kW;±10%;

[0138] P02; 13:30-14:30; V002; 30kW; ±15%.

[0139] In a specific example, a logistics park needs to schedule charging for two light electric vehicles:

[0140] V001: Cold chain transportation task, SOC=30%, departure time 14:00;

[0141] V002: General freight transport mission, SOC=45%, departure time 16:30;

[0142] Photovoltaic forecast curve: 60kW power from 12:00 to 13:00, 40kW power from 13:00 to 14:00;

[0143] V001 urgency factor = 0.94, efficiency factor = 0.9, interaction factor = 0.024;

[0144] V002 urgency factor = 0.41, efficiency factor = 0.85, interaction factor = 0.015;

[0145] The entropy weight TOPSIS scores are V001 = 0.82 and V002 = 0.65.

[0146] V001 is allocated from 12:00 to 13:30 (PV 60kW to 40kW);

[0147] V002 is allocated from 13:30 to 14:30 (40kW photovoltaic + 10kW energy storage);

[0148] Generate a scheduling instruction: V001 is charged at P01 at 50kW ± 10%, and V002 is charged at P02 at 30kW ± 15%.

[0149] Step 4: Determine the real-time photovoltaic power generation and process the charging priority sequence and real-time photovoltaic power generation through a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles.

[0150] Step 4.1: Collect voltage and current data of the photovoltaic power generation device in real time, and determine the real-time photovoltaic power generation power through a preset moving average filtering algorithm.

[0151] Moving average filtering algorithm: a data processing method that eliminates instantaneous noise by calculating the average value of data within a continuous time window.

[0152] In one instance of this application:

[0153] Voltage and current sensors are deployed on the photovoltaic DC bus to collect raw data (e.g., voltage 600V, current 80A) at a frequency of once per second.

[0154] The data format is a timestamp-voltage-current triplet (e.g., 12:00:00, 600V, 80A).

[0155] Set a 5-minute (300 data points) sliding window and calculate the average voltage and current using a rolling calculation:

[0156] Average voltage = the arithmetic mean of the most recent 300 voltage data points;

[0157] Average current = the arithmetic mean of the most recent 300 current data points;

[0158] Real-time power = average voltage × average current × efficiency coefficient (taken as 0.95);

[0159] For example, if the average voltage is 600V and the average current is 80A, then the real-time power = 600 × 80 × 0.95 = 45.6kW.

[0160] Step 4.2: Establish a dynamic power allocation matrix; wherein, the dynamic power allocation matrix includes charging pile number, real-time photovoltaic power generation, charging priority sequence and charging efficiency curve.

[0161] Dynamic power allocation matrix: A multi-dimensional data structure that integrates charging pile status, available power, vehicle priority, and charging efficiency.

[0162] In one instance of this application, the matrix dimension is defined as follows:

[0163] Row dimension: charging pile number (e.g., P01, P02);

[0164] Column dimension: Real-time photovoltaic power generation (e.g., 45.6kW);

[0165] Charging priority sequence (e.g., V001 priority 0.9, V002 priority 0.7);

[0166] Charging efficiency curve (dynamic efficiency table based on SOC and temperature, e.g., efficiency is 90% when SOC = 30%);

[0167] Data population: Load the charging pile's rated power (e.g., P01 rated power 50kW) and vehicle battery parameters (e.g., V001 maximum allowable current 100A) from the database.

[0168] For example: charging piles; available power; priority sequence; efficiency curves;

[0169] P01; 45.6kW; V001(0.9); SOC-efficiency mapping table.

[0170] Step 4.3: Solve for the optimal solution of the dynamic power allocation matrix based on the preset mixed integer programming algorithm; wherein, the objective function of the mixed integer programming algorithm is to minimize the total charging completion time, and the constraints include the maximum power limit of the charging pile and the battery safe charging current limit.

[0171] Mixed-integer programming: a mathematical optimization method that simultaneously handles continuous variables (power values) and discrete variables (charging pile allocation).

[0172] In one instance of this application:

[0173] Objective function setting: Minimize the maximum charging completion time of all vehicles (i.e., minimize the total time).

[0174] Constraint definition: Charging pile power limit: The output power of each charging pile shall not exceed its rated power (e.g., P01≤50kW).

[0175] Battery safety current limit: The vehicle charging current must not exceed its maximum permissible value (e.g., V001≤100A).

[0176] Power allocation continuity: Charging tasks should be allocated according to priority order for the same charging station.

[0177] The branch and bound method is used to traverse the feasible solution space and select power allocation schemes that satisfy all constraints.

[0178] For example, the solution is: P01 allocates 45kW to V001, P02 allocates 30kW to V002, and the total completion time is 2 hours.

[0179] Step 4.4: Generate a real-time power allocation scheme based on the optimal solution, and calculate the expected results for multiple light commercial electric vehicles based on the real-time power allocation scheme.

[0180] In one instance of this application:

[0181] Convert the optimal solution into executable charging instructions, for example:

[0182] P01 charges V001 at 45kW from 12:00 to 14:00;

[0183] P02 charges V002 at 30kW from 13:00 to 14:30;

[0184] Predicted charging completion time: V001 requires 70kWh of electricity (from 30% to 90%), and the time is approximately 70 / 45 ≈ 1.56 hours.

[0185] System energy efficiency index: Photovoltaic utilization rate = actual power used / available power = 45 / 45.6 ≈ 98.7%;

[0186] In a specific example, a logistics park needs to dynamically allocate charging power to two light electric vehicles:

[0187] Real-time photovoltaic power: 45.6kW (after filtering)

[0188] Charging priority sequence: V001 (0.9) V002 (0.7)

[0189] V001: SOC = 30% to 90%, battery capacity 100kWh, maximum current 100A;

[0190] V002: SOC = 45% to 80%, battery capacity 80kWh, maximum current 80A;

[0191] The photovoltaic power was calculated to be 45.6 kW using a moving average filter.

[0192] Establish a dynamic matrix containing P01(45.6kW, V001) and P02(45.6kW, V002);

[0193] The optimal allocation scheme is solved using mixed integer programming: P01 allocates 45kW to V001, and P02 allocates 30kW to V002; the charging time for V001 is approximately 1.33 hours, and for V002 it is approximately 0.93 hours, for a total time of 1.33 hours.

[0194] Step 5: Control multiple charging piles to perform charging operations according to the real-time power distribution scheme, and monitor the operating status of the photovoltaic charging system. When it is detected that the photovoltaic power generation of the photovoltaic power generation device is insufficient and / or the energy storage capacity of the energy storage device is lower than the power threshold, supplement the power based on the external power grid.

[0195] Step 5: Collect the output power data of the photovoltaic power generation device in real time, and determine the current total charging demand by accumulating the power allocation values ​​of all charging piles in the real-time power allocation scheme.

[0196] Step 5.1: Calculate the difference between photovoltaic power generation and total charging demand based on the preset power difference algorithm.

[0197] Total charging demand: The total power consumption of all charging stations currently implementing the power allocation scheme.

[0198] In one instance of this application:

[0199] The power allocation values ​​of each charging station are obtained in real time through the charging station controller (e.g., P01 = 42kW, P02 = 3.2kW).

[0200] The data format is [timestamp, charging pile ID, power value (kW)], and it is updated every 5 seconds.

[0201] Adding up the allocated power values ​​of all charging stations: Total demand = 42 + 3.2 = 45.2kW.

[0202] If an offline charging station or abnormal power fluctuations (such as ±20% sudden changes) are detected, an alarm will be triggered and abnormal data will be removed.

[0203] Step 5.2: When the difference continues to be negative for more than the first preset time, it is determined that the photovoltaic power generation is insufficient, and the energy storage device is called to supplement the power supply.

[0204] Power difference algorithm: Real-time comparison of the difference between photovoltaic power generation and total charging demand.

[0205] In one instance of this application:

[0206] The difference ΔP is calculated every 10 seconds as follows: Photovoltaic power (45.2kW) - Total demand (45.2kW) = 0kW.

[0207] If ΔP remains <0 for a duration of ≥5 minutes (e.g., if the photovoltaic power suddenly drops to 40kW and the total demand is 45.2kW, then ΔP = -5.2kW), it is determined that the photovoltaic power generation is insufficient.

[0208] Send a command to the energy storage controller to replenish the power deficit by the absolute value of the difference (5.2kW).

[0209] For example, when the energy storage battery discharges at a power of 5kW, the remaining 0.2kW gap is automatically adjusted by the system to allocate power to the charging pile.

[0210] Step 5.3: Periodically obtain the remaining percentage of energy in the energy storage device and compare it with the energy threshold.

[0211] Energy threshold: The minimum remaining energy percentage (e.g., 30%) that an energy storage device is allowed to discharge.

[0212] In one instance of this application: the remaining energy storage capacity (e.g., current capacity 35%) is read every 5 minutes via the BMS (Battery Management System).

[0213] Compare with the preset threshold: if the remaining power is ≤30%, trigger the grid replenishment preparation command.

[0214] When the energy level approaches the threshold (e.g., 31%), reduce the energy storage discharge power to a safe level (e.g., 3kW) in advance.

[0215] Step 5.4: When the remaining power percentage is lower than the power threshold, trigger the grid power replenishment command to enter the power replenishment mode.

[0216] Power replenishment mode: The system automatically switches to the operating state of being powered by the external power grid.

[0217] In one instance of this application:

[0218] Command triggering conditions: insufficient photovoltaic power and energy storage capacity ≤30%.

[0219] For example, when the energy storage capacity drops to 28%, a power replenishment request is immediately sent to the power grid dispatch system.

[0220] Disconnect the energy storage power supply link and synchronize the charging pile communication protocol (such as ModbusTCP) to the grid interface.

[0221] Step 5.5: Obtain time-of-use electricity price data from the external power grid.

[0222] Time-of-use pricing: Differentiated electricity pricing standards set by power companies for different time periods.

[0223] In one instance of this application:

[0224] Obtain 24-hour electricity price data (e.g., 1.2 yuan / kWh during peak hours and 0.6 yuan / kWh during off-peak hours) through the power company's API.

[0225] Data is stored in the format of [time period, electricity price] (e.g., 13:00-15:00, 1.0 yuan / kWh).

[0226] Mark the current time period type (e.g., 14:30 is a peak period), calculate the electricity price trend for the next 2 hours, and prioritize replenishing electricity during low-price periods.

[0227] Step 5.6: Calculate the power allocation weight of the external power grid based on the time-of-use electricity price data, and generate a power grid supplementary power allocation scheme; wherein, the power grid supplementary power allocation scheme includes the power grid supply ratio and phase synchronization parameters of each charging pile.

[0228] Power supply ratio: The proportion of power supplied by the power grid to the total demand for each charging station.

[0229] In one instance of this application:

[0230] Weighting calculation: Priority is set according to electricity price: 0.9 weight is allocated to off-peak hours and 0.3 weight to peak hours.

[0231] For example, during peak hours, the power grid is only handling 30% of the required power.

[0232] Scheme generation: Power allocation to the grid = total demand × weight = 45.2 × 0.3 ≈ 13.6kW.

[0233] Based on charging pile priority allocation: V001 (high priority) is allocated 10kW, and V002 is allocated 3.6kW.

[0234] Step 5.7: Control the charging pile to switch to grid power supply mode, perform power allocation based on the grid power supplementation power distribution scheme, and monitor in real time whether the grid load rate exceeds the preset safety range.

[0235] Phase synchronization parameters: Current phase adjustment parameters for each charging pile when powered by the grid, to avoid three-phase imbalance.

[0236] In one instance of this application:

[0237] Control the charging pile relay to switch to the grid power supply line and set the initial power to the allocated value (e.g., P01 = 10kW).

[0238] The phase angle of the synchronous three-phase current (e.g., phase A 0°, phase B 120°, phase C 240°) is controlled within ±5°. Real-time monitoring of grid-side current (e.g., 50A) and voltage (220V ± 5%) is performed, and the load factor is calculated as actual load / line capacity. If the load factor is ≥85% (e.g., line capacity 60A, actual current 55A), the power of low-priority charging piles is automatically reduced.

[0239] Step 5.8: When it is detected that the photovoltaic power generation has recovered to the total charging demand or the energy storage device power has recovered to the safe threshold, gradually reduce the grid power supply ratio until the power supply mode is exited.

[0240] Safety threshold: The critical value of energy storage capacity at which grid-connected power can be stopped (e.g., 40%).

[0241] In one instance of this application:

[0242] Exit condition check:

[0243] Periodically (every 2 minutes) check whether the photovoltaic power has recovered (e.g., ΔP≥0) or the energy storage capacity is ≥40%.

[0244] If the photovoltaic power recovers to 45kW:

[0245] Phase 1: The grid power supply ratio is reduced to 50% (6.8kW), with photovoltaic power accounting for 38.4kW.

[0246] Phase 2 (10 minutes later): Completely disconnect grid power supply and switch back to direct photovoltaic power supply mode.

[0247] In a specific example, within a logistics park setting, intelligent power replenishment is performed during rainy weather:

[0248] Photovoltaic power output: fluctuates continuously during cloudy and rainy days (from 35kW to 28kW to 32kW).

[0249] Total charging demand: 40kW (P01 = 30kW, P02 = 10kW).

[0250] Initial energy storage capacity: 45%, capacity threshold set at 30%.

[0251] Time-of-use electricity price data: The current period (14:00) is the peak electricity price of 1.1 yuan / kWh, and it will enter the off-peak period of 0.7 yuan / kWh after 16:00.

[0252] The total demand is calculated in real-time to be 40kW, with photovoltaic power at 32kW and ΔP = -8kW. If ΔP remains negative for 10 minutes, energy storage is triggered to supplement the 8kW shortfall, and the energy storage capacity decreases to 32% at a rate of 8kW. After 5 minutes, the remaining energy storage capacity is detected at 29% (below 30%), triggering a grid power supply command. Electricity price data is obtained, marking the current period as peak, and a grid weight of 0.4 is set. A power supply plan is generated: the grid supplies 40 × 0.4 = 16kW, with P01 = 12kW and P02 = 4kW allocated. The system switches to grid power supply, and the load factor is monitored to stabilize at 78% (below the 85% safety threshold). After 2 hours, the photovoltaic power recovers to 42kW (ΔP = +2kW), and the grid power supply ratio is gradually reduced to 0.

[0253] Step 6: During the charging process, update the battery state of charge and real-time photovoltaic power generation of multiple light commercial electric vehicles in real time. If the expected result exceeds the charging threshold, recalculate the charging priority sequence and update the real-time power allocation scheme.

[0254] Step 6.1: Collect incremental data of battery state of charge and actual output power data of charging piles for multiple light commercial electric vehicles at preset intervals.

[0255] Preset period: A fixed time interval (e.g., 5 minutes) set by the system for synchronizing and updating data.

[0256] In one instance of this application:

[0257] The vehicle's BMS (Battery Management System) acquires incremental battery SOC data every 5 minutes (e.g., V001 SOC from 30% to 35%).

[0258] Read the actual output power from the charging pile controller (e.g., P01 is nominally allocated 30kW, but the actual output is 28.5kW).

[0259] Data format: [Vehicle ID, SOC change (%), charging pile ID, actual power (kW)].

[0260] If an abnormal increase in SOC (such as ±10% mutation) is detected for three consecutive cycles, a manual review process is triggered.

[0261] For example, if the V002SOC suddenly increases from 45% to 60%, the system will suspend its charging and send a fault alarm.

[0262] Step 6.2: Input the incremental data of battery state of charge into the preset charging efficiency evaluation model to generate the actual charging rate curve.

[0263] Charging efficiency evaluation model: A charging rate prediction algorithm based on historical data, including a temperature compensation factor.

[0264] In one instance of this application:

[0265] Model input processing: Converting SOC incremental data into charging efficiency values:

[0266] Actual efficiency = (SOC increment corresponding to power consumption / charging pile output power consumption) × 100%;

[0267] For example: V001 receives 28.5kW × 5 minutes = 2.375kWh, and the SOC increases by 5% (corresponding to 5kWh), then the efficiency = 5 / 2.375 ≈ 210% (outliers need to be corrected).

[0268] Curve generation: Fit a curve based on data from 6 consecutive cycles and label the effect of temperature (e.g., efficiency is 92% at 25℃ and 88% at 35℃).

[0269] Dynamic correction factor: When the ambient temperature changes by ≥5℃, the efficiency prediction value is automatically adjusted.

[0270] Step 6.3: Compare the actual charging rate curve with the expected result and calculate the time integral difference between the two.

[0271] Time integral difference: The cumulative deviation between the actual charging progress and the expected plan in the time dimension.

[0272] In one instance of this application:

[0273] Difference calculation:

[0274] Calculate the area difference between the planned SOC curve and the actual curve for each vehicle (e.g., graphical integration).

[0275] For example: If the plan is to charge to 80% SOC in 2 hours, but the actual charge to 75% is only achieved in 2 hours, the time integral difference is 5% × 2h = 10%·h.

[0276] Threshold determination:

[0277] The preset charging threshold is 15%·h. If the difference in V001 reaches 10%·h, it is marked as a yellow warning; if it exceeds 15%·h, the scheme will be updated.

[0278] Step 6.4: When the time integral difference exceeds the charging threshold value, reconstruct the multi-dimensional evaluation index system based on the latest collected battery state of charge and the corrected photovoltaic power generation prediction curve.

[0279] Multidimensional evaluation index system: a set of decision parameters including vehicle urgency, charging cost, and system energy efficiency.

[0280] In one instance of this application:

[0281] Index expansion: Add a real-time electricity price influencing factor (e.g., peak period weight +0.2).

[0282] Added vehicle departure time variation rate (originally scheduled to leave at 14:00, but actually needs to be delayed to 14:30, increasing the urgency).

[0283] Dynamic weight adjustment: When the photovoltaic power forecast is reduced by 10%, the system energy efficiency weight increases from 0.6 to 0.8.

[0284] For example, the indicator system:

[0285] Indicators: Original weights; Adjusted weights;

[0286] Task urgency: 0.7; 0.8;

[0287] Charging cost: 0.5; 0.4;

[0288] Photovoltaic utilization rate: 0.6; 0.7.

[0289] Step 6.5: Reorder the dynamic priority scores based on the preset rolling optimization algorithm to generate an updated charging priority sequence.

[0290] Rolling optimization algorithm: an iterative optimization method that dynamically adjusts priorities based on the latest data.

[0291] In one instance of this application:

[0292] Scoring model: Priority score = Task urgency × 0.8 + Charging cost × 0.4 + Photovoltaic utilization rate × 0.7.

[0293] For example: Original fraction V001 = 0.9 × 0.8 + 0.6 × 0.4 + 0.7 × 0.7 = 1.69; Corrected fraction = 0.95 × 0.8 + 0.5 × 0.4 + 0.8 × 0.7 = 1.86.

[0294] Sort by new scores in descending order: V003 (2.01) > V001 (1.86) > V002 (1.55).

[0295] New sequences were generated: V003 (0.95), V001 (0.90), V002 (0.75).

[0296] Step 6.6: Update the real-time power allocation scheme according to the updated charging priority sequence and constraints.

[0297] Constraint loading:

[0298] Maximum power limit of charging pile (e.g., P01≤50kW).

[0299] Minimum charging current requirement for vehicles (e.g., V003 ≥ 20A).

[0300] According to the new priority allocation: V003 is allocated 45kW (P01+P02 combined power supply), V001 is adjusted to 30kW (P03), and V002 is suspended from charging.

[0301] Generate control commands:

[0302] P01: Switch to V003, power = 30kW;

[0303] P02: Switch to V003, power = 15kW;

[0304] P03: Allocate V001, power = 30kW;

[0305] In a specific example, during the midday charging peak at a logistics center, the system dynamically adjusts its strategy to cope with unexpected situations:

[0306] Initial conditions:

[0307] Vehicles: V001 (SOC 40% to 80%), V002 (50% to 85%), V003 (Emergency Mission, 30% to 95%)

[0308] Photovoltaic power: fluctuating at midday (50kW to 42kW to 47kW);

[0309] Original priority sequence: V001(0.9)>V002(0.7)>V003(0.6);

[0310] Data collected in the first cycle: V001's actual charging power was 28kW (expected 30kW), and the SOC only increased by 3%.

[0311] The cumulative difference in the third cycle reached 18%·h, exceeding the threshold of 15%·h.

[0312] Reconstruct the indicator system: the weight of emergency tasks has been increased from 0.7 to 0.9, and the priority score of V003 has been significantly improved.

[0313] New sequence: V003 (0.95) > V001 (0.85) > V002 (0.70).

[0314] Updated solution: Allocate the 45kW of P01+P02 to V003, and V001 will be powered by 30kW from P03.

[0315] This application also includes the following methods:

[0316] A1. Collect historical operation data for the entire charging cycle to construct a charging efficiency evaluation dataset.

[0317] Complete charging cycle: The entire process from the start of vehicle charging to reaching the target SOC, including environmental parameters and equipment status records.

[0318] In one instance of this application:

[0319] Data collection scope:

[0320] Time span: Data on the operation of the photovoltaic charging system was collected over the past 12 months.

[0321] Data type:

[0322] Environmental parameters: light intensity (W / m2), ambient temperature (°C), and surface contamination of photovoltaic panels (0-1 coefficient).

[0323] Equipment status: charging pile output power deviation rate (actual / nominal value), energy storage battery health (SOH).

[0324] Vehicle data: Battery SOC change curve, number of charging interruptions, BMS alarm records.

[0325] Remove abnormal segments with a charging time of less than 10 minutes (such as accidental plugging and unplugging of the charging gun).

[0326] Correct sensor drift data (e.g., temperature data ±3℃ correction).

[0327] [Record ID, timestamp, light intensity, temperature, charging pile power, SOC start value, SOC end value, charging time];

[0328] For example, entry: C20241105-0853, 2023-11-0508: 53, 850, 25, 45.2kW, 30%, 80%, 2.3h.

[0329] A2. Process the charging performance evaluation dataset using a preset feature extraction algorithm to extract a set of key performance indicators.

[0330] Feature extraction algorithm: Based on Pearson correlation coefficient, select parameter combinations that are strongly correlated with charging efficiency.

[0331] In one instance of this application, the core metrics are:

[0332] Photovoltaic volatility = (maximum power - minimum power) / average power × 100% (e.g., daily volatility 28%).

[0333] Charging efficiency decay slope = (initial efficiency - final efficiency) / charging time (e.g. -1.2% / h).

[0334] Derivative metric: Temperature-efficiency coupling coefficient: the change in charging efficiency caused by each 1°C increase (e.g., -0.5% / °C).

[0335] Feature dimensionality reduction: Principal component analysis (PCA) was used to reduce the dimensionality of the 32 original parameters to 5 principal components, with a cumulative contribution rate of ≥85%.

[0336] For example, the output is: Principal Component 1 (photovoltaic stability, weight 0.52) and Principal Component 2 (battery aging effect, weight 0.23).

[0337] A3. Construct a random forest regression model based on the set of key performance indicators to analyze the impact of parameters in the set of key performance indicators on the photovoltaic charging system and generate suggestions for optimizing the charging strategy.

[0338] Random forest regression model: A machine learning model composed of multiple decision trees, used to analyze multi-parameter nonlinear relationships.

[0339] In one instance of this application:

[0340] Dataset split: 70% training set (approximately 10,000 records), 20% validation set, and 10% test set.

[0341] Number of decision trees = 200, maximum depth = 8, minimum number of leaf samples = 50.

[0342] Feature importance output, for example:

[0343] Feature: Importance;

[0344] Photovoltaic volatility: 0.38;

[0345] Ambient temperature: 0.22°C;

[0346] Initial battery SOC: 0.18.

[0347] Strategy suggestion generation:

[0348] Dynamic priority adjustment: When the photovoltaic volatility is greater than 25%, it is recommended to increase the priority of high SOC vehicles by 0.2 levels.

[0349] Energy storage dispatch optimization: In scenarios with temperatures >35℃, it is recommended to reduce the upper limit of energy storage discharge power to 80% of the rated value.

[0350] Charging pile coordination strategy: Identify charging piles with a power deviation rate >15% and recommend reducing their weight by 50%.

[0351] In a specific example, in a logistics park charging station efficiency optimization project, the system implemented the following improvements:

[0352] Historical dataset: Contains 8,760 hours of runtime data (January 1, 2024 - December 31, 2024).

[0353] Key issues: Summer charging efficiency decreased by 12% year-on-year, and photovoltaic utilization fluctuated by ±35%.

[0354] Execution process:

[0355] Data was collected during the high-temperature period from June to August, and 1,253 charging interruption events caused by temperatures exceeding 40°C were identified.

[0356] The dataset contains 18 parameters, including temperature, photovoltaic power, and charging efficiency.

[0357] Extract key metrics:

[0358] It was found that for every 10°C increase in temperature, the average charging efficiency decreased by 4.7%.

[0359] The midday photovoltaic volatility (11:00-13:00) peaked at 42%.

[0360] Model output suggestions:

[0361] When the temperature is above 35℃, the charging power is dynamically reduced by 10%-15% to maintain efficiency.

[0362] During peak periods of photovoltaic fluctuations, the energy storage buffer power is set to 30% of the maximum fluctuation value.

[0363] The above are embodiments of the method proposed in this application. Based on the same inventive concept, embodiments of this application also provide a photovoltaic centralized charging control device for light commercial electric vehicles, the structure of which is as follows: Figure 2 As shown.

[0364] Figure 2 This is a schematic diagram of the internal structure of a photovoltaic centralized charging control device for a light commercial electric vehicle, provided as an embodiment of this application. Figure 2 As shown, the device includes:

[0365] At least one processor 201;

[0366] And a memory 202 that is communicatively connected to at least one processor;

[0367] The memory 202 stores instructions executable by at least one processor, which are executed by at least one processor 201 to enable at least one processor 201 to:

[0368] The system acquires weather data and vehicle data for at least one light commercial electric vehicle, including battery state of charge (SOC), vehicle location information, preset task schedule, and estimated departure time. It processes the weather data and photovoltaic (PV) power generation data using a preset PV power generation prediction algorithm to generate a PV power prediction curve. Based on the battery SOC, preset task schedule, estimated departure time, and PV power prediction curve, it calculates a charging priority sequence for multiple light commercial electric vehicles. It determines the real-time PV power generation and processes the charging priority sequence and real-time PV power generation using a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles. It controls multiple charging piles to perform charging operations according to the power allocation scheme and monitors the operating status of the PV charging system. When insufficient PV power generation from the PV power generation device and / or the energy storage capacity of the energy storage device is lower than the power threshold is detected, it supplements power based on the external power grid. During the charging process, it updates the battery SOC and real-time PV power of multiple light commercial electric vehicles in real time. If the expected result exceeds the charging threshold, it recalculates the charging priority sequence and updates the power allocation scheme.

[0369] Some embodiments of this application provide corresponding to Figure 1 A non-volatile computer storage medium for photovoltaic centralized charging control of light commercial electric vehicles, storing computer-executable instructions, wherein the computer-executable instructions are configured as follows:

[0370] The system acquires weather data and vehicle data for at least one light commercial electric vehicle, including battery state of charge (SOC), vehicle location information, preset task schedule, and estimated departure time. It processes the weather data and photovoltaic (PV) power generation data using a preset PV power generation prediction algorithm to generate a PV power prediction curve. Based on the battery SOC, preset task schedule, estimated departure time, and PV power prediction curve, it calculates a charging priority sequence for multiple light commercial electric vehicles. It determines the real-time PV power generation and processes the charging priority sequence and real-time PV power generation using a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles. It controls multiple charging piles to perform charging operations according to the power allocation scheme and monitors the operating status of the PV charging system. When insufficient PV power generation from the PV power generation device and / or the energy storage capacity of the energy storage device is lower than the power threshold is detected, it supplements power based on the external power grid. During the charging process, it updates the battery SOC and real-time PV power of multiple light commercial electric vehicles in real time. If the expected result exceeds the charging threshold, it recalculates the charging priority sequence and updates the power allocation scheme.

[0371] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for IoT devices and media are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0372] The systems, media, and methods provided in this application are one-to-one correspondences. Therefore, the systems and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the systems and media will not be repeated here.

[0373] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0374] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0375] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0376] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0377] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0378] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is, for example, a computer-readable medium.

[0379] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0380] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0381] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A centralized photovoltaic charging control method for light commercial electric vehicles, applied to a photovoltaic charging system, wherein the photovoltaic charging system is used to charge light commercial electric vehicles, the photovoltaic charging system includes a charging device, the charging device includes multiple charging piles, and the charging device is connected to a photovoltaic power generation device, an energy storage device, and an external power grid, characterized in that... The method includes: Acquire weather data and vehicle data of at least one of the light commercial electric vehicles; wherein the vehicle data includes battery state of charge and preset task schedule; The weather data and the data from the photovoltaic power generation device are processed based on a preset photovoltaic power generation prediction algorithm to generate a photovoltaic power generation prediction curve. Based on the battery state of charge, preset task schedule, and photovoltaic power generation prediction curve, a charging priority sequence for multiple light commercial electric vehicles is calculated. The real-time photovoltaic power generation is determined, and the charging priority sequence and real-time photovoltaic power generation are processed by a preset dynamic power allocation algorithm to generate a real-time power allocation scheme and expected results for multiple charging piles. According to the real-time power distribution scheme, multiple charging piles are controlled to perform charging operations, and the operating status of the photovoltaic charging system is monitored. When it is detected that the photovoltaic power generation of the photovoltaic power generation device is insufficient and / or the energy storage capacity of the energy storage device is lower than the power threshold, supplementary power is provided based on the external power grid. During the charging process, the battery state of charge and real-time photovoltaic power generation of multiple light commercial electric vehicles are updated in real time. If the expected result exceeds the charging threshold, the charging priority sequence is recalculated and the real-time power allocation scheme is updated. Based on the battery state of charge, preset task schedule, and photovoltaic power generation prediction curve, a charging priority sequence for multiple light commercial electric vehicles is calculated, specifically including: A multi-dimensional evaluation index system is constructed based on the battery state of charge and the preset task schedule; wherein, the index system includes a task urgency factor, a charging efficiency factor, and a grid interaction factor; A multi-dimensional evaluation index system is constructed based on the battery state of charge and the preset task schedule, specifically including: Based on the planned departure timestamp and the current time in the preset task schedule, the available charging time is determined, and a preset safety margin is deducted to determine the actual available charging time. Obtain a preset task type weight table; wherein, the task type weight table includes at least the urgency coefficients corresponding to cold chain transportation tasks, general freight tasks, and standby vehicle tasks; The actual available charging time, battery state of charge, and preset task type weight table are processed based on a preset exponential decay model to generate a task urgency factor. The maximum allowable charging current is determined based on the battery data of multiple light commercial electric vehicles, and the charging efficiency factor is determined based on the ratio of the actual output power to the rated power of the charging pile. The time-of-use electricity price signal and regional load factor of the external power grid are obtained to determine the power grid interaction factor; A multi-dimensional evaluation index system is constructed based on the task urgency factor, charging efficiency factor, and grid interaction factor.

2. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 1, characterized in that, Based on a preset photovoltaic power generation prediction algorithm, the weather data and the data from the photovoltaic power generation device are processed to generate a photovoltaic power generation prediction curve, specifically including: Acquire photovoltaic module parameters and environmental monitoring data of the photovoltaic power generation device; wherein, the photovoltaic module parameters include the module efficiency degradation curve and the photovoltaic array tilt angle adjustment range, and the environmental monitoring data includes the module surface temperature and environmental irradiance; The weather data is analyzed over time using a pre-defined LSTM neural network model. The weather data includes cloud cover changes over a certain period of time and historical seasonal light intensity distribution data. The photovoltaic module parameters are input into the LSTM neural network model to generate an initial photovoltaic power generation prediction curve; Based on the surface temperature of the component and the ambient irradiance, the initial photovoltaic power generation prediction curve is corrected by comparing it with a preset temperature power reduction coefficient table to generate the photovoltaic power generation prediction curve.

3. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 1, characterized in that, Based on the battery state of charge, preset task schedule, and photovoltaic power generation prediction curve, a charging priority sequence for multiple light commercial electric vehicles is calculated, specifically including: The multidimensional evaluation index system is processed based on the preset entropy weight TOPSIS algorithm to generate dynamic priority scores for multiple light commercial electric vehicles. The dynamic priority score is spatiotemporally matched with the photovoltaic power generation prediction curve to determine the photovoltaic priority charging period and the energy storage compensation charging period. A charging priority sequence is generated based on the time period matching results; wherein, the charging priority sequence includes a time period allocation table and power fluctuation tolerance parameters for each charging pile.

4. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 1, characterized in that, The real-time photovoltaic power generation is determined, and the charging priority sequence and real-time photovoltaic power generation are processed by a preset dynamic power allocation algorithm to generate real-time power allocation schemes and expected results for multiple charging piles, specifically including: The voltage and current data of the photovoltaic power generation device are collected in real time, and the real-time photovoltaic power generation is determined by a preset moving average filtering algorithm. Establish a dynamic power allocation matrix; wherein, the dynamic power allocation matrix includes charging pile number, real-time photovoltaic power generation, charging priority sequence and charging efficiency curve; The optimal solution of the dynamic power allocation matrix is ​​obtained based on a preset mixed integer programming algorithm; wherein, the objective function of the mixed integer programming algorithm is to minimize the total charging completion time, and the constraints include the maximum power limit of the charging pile and the safe charging current limit of the battery; A real-time power allocation scheme is generated based on the optimal solution, and the expected results of multiple light commercial electric vehicles are calculated based on the real-time power allocation scheme.

5. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 4, characterized in that, According to the real-time power allocation scheme, multiple charging piles are controlled to perform charging operations, and the operating status of the photovoltaic charging system is monitored. When insufficient photovoltaic power generation of the photovoltaic power generation device and / or the energy storage capacity of the energy storage device is detected to be lower than the energy threshold, supplementary power is provided based on the external power grid, specifically including: The output power data of the photovoltaic power generation device is collected in real time, and the power allocation values ​​of all charging piles in the real-time power allocation scheme are accumulated to determine the current total charging demand. The difference between the photovoltaic power generation and the total charging demand is calculated based on a preset power difference algorithm. When the difference continues to be negative for more than a first preset time, it is determined that the photovoltaic power generation is insufficient, and the energy storage device is called to supplement the power supply. The remaining percentage of energy in the energy storage device is periodically obtained and compared with the energy threshold. When the remaining power percentage is lower than the power threshold, a grid power replenishment command is triggered to enter the power replenishment mode; Obtain time-of-use electricity price data from the external power grid; Based on the time-of-use electricity price data, the power allocation weight of the external power grid is calculated, and a power grid supplementary power allocation scheme is generated; wherein, the power grid supplementary power allocation scheme includes the power grid supply ratio and phase synchronization parameters of each charging pile; The system controls the charging pile to switch to grid power supply mode, performs power allocation based on the grid supplementary power distribution scheme, and monitors in real time whether the grid load rate exceeds the preset safety range. When it is detected that the photovoltaic power generation has recovered to the total charging demand or the energy storage device power has recovered to the safe threshold, the proportion of grid power supply will be gradually reduced until the power supply mode is exited.

6. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 4, characterized in that, During the charging process, the battery state of charge and real-time photovoltaic power generation of multiple light commercial electric vehicles are updated in real time. If the expected result exceeds the charging threshold, the charging priority sequence is recalculated and the real-time power allocation scheme is updated, specifically including: At preset intervals, incremental data of the battery state of charge of multiple light commercial electric vehicles and actual output power data of the charging piles are collected. The incremental data of the battery state of charge is input into a preset charging efficiency evaluation model to generate the actual charging rate curve; The actual charging rate curve is compared with the expected result, and the time integral difference between the two is calculated. When the time integral difference exceeds the charging threshold, the multi-dimensional evaluation index system is reconstructed based on the latest collected battery state of charge and the corrected photovoltaic power generation prediction curve. The dynamic priority scores are reordered based on a preset rolling optimization algorithm to generate an updated charging priority sequence. The real-time power allocation scheme is updated based on the updated charging priority sequence and constraints.

7. The photovoltaic centralized charging control method for light commercial electric vehicles according to claim 1, characterized in that, The method further includes: Collect historical operation data for the entire charging cycle to construct a charging efficiency evaluation dataset; The charging performance evaluation dataset is processed by a preset feature extraction algorithm to extract a set of key performance indicators. A random forest regression model is constructed based on the set of key performance indicators to analyze the impact of the parameters in the set of key performance indicators on the photovoltaic charging system and generate charging strategy optimization suggestions.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the photovoltaic centralized charging control method for light commercial electric vehicles as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the photovoltaic centralized charging control method for light commercial electric vehicles as described in any one of claims 1-7.

Images