Low-speed charging pile and distributed photovoltaic cooperative power supply method

Abstract

The invention discloses a low-speed charging pile and distributed photovoltaic cooperative power supply method, which is applied to a power supply system comprising a distributed photovoltaic power generation system, a plurality of low-speed charging piles and a cooperative controller. The method comprises the following steps: collecting photovoltaic real-time power, a charging pile state, a user charging demand, a power grid net load and environment prediction data; the guarantee power of each charging pile and the total guarantee power demand of the system are calculated, and the dynamic available charging capacity is determined; formulating a day-ahead reference charging power plan curve based on photovoltaic power prediction and charging demand prediction; in a real-time control period, actual photovoltaic power is compared with a planned value, power which can be increased or needs to be reduced is dynamically identified, and real-time optimal distribution is carried out by taking the condition that all guaranteed power is preferentially met as a constraint. According to the method, the photovoltaic local consumption rate is effectively improved, the power impact on the power distribution network is stabilized, and the core charging demand of a user is guaranteed on the premise of not depending on large-scale energy storage.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, and in particular to a method for co-powering low-speed charging piles and distributed photovoltaic power generation. Background Technology

[0002] With the increasing popularity of electric vehicles, low-speed charging stations (typically AC charging stations with a power of 7kW or less) are widely deployed in residential areas, office parks, and public parking lots. Simultaneously, distributed photovoltaic (PV) power generation systems are also being installed extensively in these settings, aiming to utilize clean energy. However, currently, these two systems typically operate independently, lacking effective coordination, leading to a series of technical problems.

[0003] First, distributed photovoltaic (PV) power generation is intermittent and fluctuating, with its output power heavily dependent on sunlight intensity and varying drastically throughout the day. When PV power generation exceeds the local load, the excess power can be fed back into the grid, potentially causing problems such as local voltage exceeding limits and malfunctioning protection systems in the distribution network. When photovoltaic power generation is insufficient, the load of low-speed charging piles is entirely borne by the power grid or energy storage equipment, which increases the burden on the power grid.

[0004] Secondly, charging electric vehicle batteries using low-speed charging stations is a slow process that lasts for several hours, with a relatively flat load curve but a considerable total load. Most existing charging control strategies are based on simple plug-and-charge or fixed-time scheduling, failing to consider real-time photovoltaic power generation capacity and grid conditions. This results in low local consumption rates of clean energy and fails to fully realize the environmental and economic value of photovoltaic power generation.

[0005] Furthermore, some existing collaborative control schemes often focus on the simple logic of "self-generation and self-consumption, with surplus power fed into the grid," or rely on large-scale centralized energy storage systems for power balancing. The former still impacts the grid when photovoltaic power fluctuates rapidly, while the latter increases the huge costs of infrastructure construction and maintenance, making it less economical and difficult to promote on a large scale in distributed scenarios.

[0006] In addition, the carrying capacity of the distribution network is limited, especially during the midday peak of photovoltaic power generation and the evening peak load period. The connection of a large number of uncontrolled charging piles may exacerbate the peak-valley difference of the power grid, leading to transformer overload, increased line losses and deterioration of power quality.

[0007] Therefore, existing technologies lack a low-cost, high-reliability collaborative power supply method that can deeply couple the characteristics of distributed photovoltaic power generation and the load characteristics of low-speed charging piles, achieve dynamic and refined power coordination and allocation without relying on or only relying on minimal energy storage configurations, while ensuring user charging needs, improving the local photovoltaic absorption rate, and reducing the impact on the distribution network. Summary of the Invention

[0008] To achieve the above objectives, this invention provides a method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation, applicable to a power supply system comprising a distributed photovoltaic power generation system, multiple low-speed charging piles, and a coordinated controller; the coordinated controller is communicatively connected to the photovoltaic inverter, the charging pile controllers of each low-speed charging pile, and the measuring equipment at the power grid's point of common connection; the method includes the following steps: Step 1: The collaborative controller collects the real-time output active power of the distributed photovoltaic power generation system, the real-time status and charging data of each low-speed charging pile, the real-time net load power of the grid common connection point, and the future solar irradiance prediction data sequence and temperature prediction data sequence. Step 2: The collaborative controller calculates the guaranteed power of each low-speed charging pile based on its battery type, standard charging curve, charging duration, and user-set expected charging completion time. It then sums the guaranteed power of all low-speed charging piles in operation to obtain the total guaranteed power requirement. The controller also calculates the grid power supply limit based on the distribution network capacity and current base load. Finally, it calculates the photovoltaic power supply capacity based on the rated capacity and real-time active power output of the photovoltaic inverter. The sum of the grid power supply limit and the photovoltaic power supply capacity constitutes the dynamic available charging capacity. Step 3: The collaborative controller generates a photovoltaic power generation prediction curve based on the solar irradiance prediction data sequence and the photovoltaic power generation system model, and generates a charging demand prediction based on historical charging data. With the goal of maximizing the photovoltaic predicted power coverage of charging demand and smoothing the net load power curve at the grid's point of common coupling, a baseline charging power plan curve is formulated. Within a control cycle, the collaborative controller compares the actual photovoltaic output power with the planned value for the corresponding time period in the baseline charging power plan curve to determine the adjustable photovoltaic power or the power that needs to be reduced. With the constraint of prioritizing the overall guaranteed power demand, real-time optimization calculations are performed and power is allocated: when adjustable photovoltaic power exists, it is allocated as bonus power to low-speed charging piles on the basis of meeting the guaranteed power; when power needs to be reduced and the grid power supply limit cannot compensate, low-speed charging piles are sorted according to the urgency of their expected charging completion time, and the charging power of low-speed charging piles ranked lower is reduced. Step 4: The collaborative controller sends the target charging power command to the charging pile controller of each low-speed charging pile according to the power allocation result, and verifies the execution of the command in the next data acquisition cycle. Step 5: The collaborative controller continuously monitors the net load power at the power grid's point of common coupling. When the net load power exceeds the safety limit, it forcibly reduces the charging power of all adjustable low-speed charging piles until the net load power returns to the safe range.

[0009] Preferably, in step one, the collection of real-time status and charging data of each low-speed charging pile specifically includes: obtaining the "idle", "connected but not charging" or "charging" status of the low-speed charging pile by communicating with the charging pile controller. For low-speed charging piles that are in the "charging" state, obtain their real-time charging power, charging time, user-set expected charging completion time, and target state of charge. The standard charging curve for the battery type is a typical constant current-constant voltage charging power curve model pre-existing in the collaborative controller and established according to different electric vehicle battery models.

[0010] Preferably, in step two, the specific process of calculating the guaranteed power of each low-speed charging pile is as follows: the collaborative controller calls the corresponding standard charging curve for the battery type according to the battery model of the electric vehicle connected to the low-speed charging pile. Based on the charging time of the low-speed charging station, locate the current charging stage on the standard charging curve of the battery type; Then, based on the time difference between the user's expected charging completion time and the current time, the minimum average power required to complete the full charging process as specified by the standard charging curve for the battery type from the current time to the expected charging completion time is calculated, and this minimum average power is defined as the guaranteed power of the low-speed charging pile.

[0011] Preferably, in step three, the step of formulating the benchmark charging power plan curve specifically includes: the collaborative controller using the photovoltaic power generation system model to convert the solar irradiance prediction data sequence and temperature prediction data sequence for the next 24 hours into the photovoltaic power generation prediction curve for the next 24 hours; Meanwhile, based on the historical data of the number of low-speed charging piles connected, charging start time, charging duration and charging energy, the range of total guaranteed power demand for each period in the next 24 hours is predicted through a probabilistic statistical model. Using the photovoltaic power generation forecast curve as the power supply side reference and the predicted total guaranteed power demand range as the load side reference, a rolling optimization algorithm is adopted. The primary optimization objective is to maximize the total amount of photovoltaic forecast power used by the charging load in the next 24 hours, and the secondary optimization objective is to minimize the variance of the net load power curve at the grid common connection point. The planned total charging power for each period in the next 24 hours is obtained by solving the algorithm, forming the baseline charging power plan curve.

[0012] Preferably, in step three, the specific method for determining whether the photovoltaic power can be increased or needs to be reduced is to set a power deviation threshold. The collaborative controller calculates the difference between the actual photovoltaic output power and the planned value for the corresponding time period in the baseline charging power planned curve within the current control cycle; If the difference is greater than zero and exceeds the power deviation threshold, it is determined that there is adjustable photovoltaic power, and the value of adjustable photovoltaic power is equal to the difference minus the power deviation threshold. If the difference is less than zero and its absolute value exceeds the power deviation threshold, it is determined that there is a need to reduce power, and the amount of power to be reduced is equal to the absolute value of the difference minus the power deviation threshold.

[0013] Preferably, in step three, the adjustable photovoltaic power is allocated as bonus power to the low-speed charging piles. The specific allocation strategy is as follows: the collaborative controller first confirms that the guaranteed power of all low-speed charging piles in the charging state has been met. Then, the urgency of the charging demand for each low-speed charging pile is calculated, where the urgency of the charging demand is the expected charging completion time and the remaining time at the current moment. The low-speed charging stations are sorted in order of increasing urgency of charging demand, i.e., the shorter the remaining time, the higher the urgency. Bonus power is preferentially allocated to low-speed charging piles with low charging demand urgency (i.e., long remaining time). The allocation weight is inversely proportional to the charging demand urgency, and the sum of the bonus power allocated to any low-speed charging pile and the guaranteed power of that low-speed charging pile shall not exceed the maximum allowable charging power of that low-speed charging pile.

[0014] Preferably, in step three, the process of sorting the low-speed charging piles according to the urgency of their expected charging completion time and reducing the charging power of the low-speed charging piles that are ranked lower is as follows: when there is a need to reduce the power and the upper limit of the power grid supply has been reached, the collaborative controller sorts all the low-speed charging piles that are charging in order of their expected charging completion time from farthest to nearth, that is, the later the expected charging completion time, the lower the ranking. Starting with the lowest-speed charging pile in the order, the collaborative controller sends a power reduction request to the user of that low-speed charging pile and waits for confirmation. After obtaining user confirmation, the current charging power of the low-speed charging pile is reduced by one step, where the step is a fixed proportion of the guaranteed power of the low-speed charging pile. For each low-speed charging station removed, the system power deficit is recalculated. If the power deficit is still greater than zero, the same reduction process is continued for the next low-speed charging station in the order until the power deficit is eliminated or all adjustable low-speed charging stations have been reduced to the lower limit threshold of the guaranteed power.

[0015] Preferably, in step four, the verification of the command execution status specifically includes: in the next data acquisition cycle after the collaborative controller issues the target charging power command, reading the actual output power of each low-speed charging pile through the charging pile controller; The actual output power is compared with the target charging power command issued to calculate the power tracking error. If the power tracking error of a low-speed charging pile continues to exceed the preset allowable error range for a set number of times, the collaborative controller will determine that the low-speed charging pile has a communication or equipment abnormality, and will remove the low-speed charging pile from the current optimization allocation list. Its power demand will no longer participate in subsequent real-time optimization calculations until the abnormality is eliminated.

[0016] Preferably, in step five, the safety limits include a forward overload limit and a reverse feed limit; The forced reduction of the charging power of all adjustable low-speed charging piles is specifically achieved by adopting a proportional load reduction strategy: the coordinating controller calculates the proportion of the net load power at the grid common connection point that exceeds the safety limit. Based on the proportion of the excessive power, calculate a uniform power reduction coefficient; Multiply the charging power of all low-speed charging piles that are currently charging and not locked by the power constraint in step three by the power reduction coefficient to obtain a set of temporary safe power commands and issue them immediately. The low-speed charging pile that is not locked by the guaranteed power constraint refers to a low-speed charging pile whose current charging power is higher than its guaranteed power.

[0017] Preferably, the method further includes a system initialization and parameter self-learning phase: after the power supply system is put into operation for the first time or after major equipment changes, the coordinating controller performs a self-learning cycle for a set number of days; During the self-learning cycle, the collaborative controller operates in plug-and-charge mode and fully records the daily photovoltaic power generation curve, the charging behavior curve of each low-speed charging pile, and the net load power curve of the grid common connection point. After the self-learning cycle ends, the collaborative controller uses recorded historical data to train and correct the parameters in the photovoltaic power generation system model, the parameters in the charging demand prediction probability statistics model, and the initial value of the power deviation threshold through machine learning algorithms, so that the prediction and control parameters in subsequent steps one to five are adapted to the specific local operating environment.

[0018] The beneficial effects of this invention are: 1. This invention utilizes a multi-timescale collaborative control mechanism combining a "day-ahead benchmark charging power plan curve" and "real-time fine-tuning" to proactively and dynamically match fluctuating photovoltaic (PV) power generation with flexible charging pile loads. This method creatively introduces a "bonus power" allocation strategy. When actual PV output exceeds the planned value, the surplus PV power is automatically allocated preferentially to charging piles, encouraging them to increase charging power during peak PV periods. This maximizes the local use of PV power for charging and significantly reduces the backflow of surplus power to the grid. Compared to the traditional "plug-and-charge" mode, this method proactively guides the charging load to track the PV output curve at the system level, solving the core problem of time-series mismatch between the two. 2. This invention aims to smooth the net load curve of the power grid as one of its optimization objectives. It treats the charging pile group as a whole, adjustable load through coordinated control. During the day-ahead planning phase, the optimization model strives to make the sum of the planned total charging power, the predicted photovoltaic power, and the base load as stable as possible. In the real-time phase, a "power deviation threshold" mechanism and a "guaranteed power" hard constraint prevent frequent jumps in total charging power caused by short-term drastic fluctuations in photovoltaic power. When the grid power exceeds the limit, the protective intervention mechanism can act quickly. This makes the net load power curve observed from the grid side smooth and stable, avoiding the dual impact of peak photovoltaic backfeeding and traditional charging load peaks on the distribution transformer, thus improving the safety and economy of the distribution system. 3. This invention defines a "guaranteed power" based on the user's expected charging completion time and the battery's standard charging curve, and treats it as a mandatory hard constraint in real-time control. This ensures that the user's basic charging progress is prioritized under any operating condition. In extreme cases of insufficient power, an orderly reduction strategy based on the urgency of charging is superior to simple random power outages or staggered charging, demonstrating the fairness and rationality of the control. Most importantly, the aforementioned superior technical effects are achieved through advanced predictive and software control algorithms, primarily relying on information acquisition and communication. This eliminates the need for expensive, large-scale centralized energy storage battery systems, thus achieving near-"photovoltaic-storage-charging" performance while significantly reducing system investment and maintenance costs, making large-scale application feasible. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the steps of the method of the present invention; Figure 2This is a flowchart of the step in step two of the method of the present invention to calculate the guaranteed power of each low-speed charging pile; Figure 3 This is a flowchart illustrating the specific allocation strategy for allocating adjustable photovoltaic power as bonus power to low-speed charging piles in step three of the method of the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0022] Please see Figures 1-3 This invention provides a method for low-speed charging piles and distributed photovoltaic power supply in coordination. The main body of this method is a coordination controller, which first establishes a stable communication link with the photovoltaic inverter, each charging pile controller and the power grid common connection point measurement equipment.

[0023] The implementation of step one is as follows: the collaborative controller polls and collects the real-time AC output active power of the photovoltaic inverter every minute through the standard communication protocol; obtains the real-time working status, charging power, session duration and user-preset expected completion time of each charging pile through the charging pile management protocol; reads the net active power of the grid common connection point through the electricity meter communication protocol; and obtains the solar irradiance and temperature prediction data sequence at 15-minute intervals for the next 24 hours from an authoritative meteorological service provider through the network interface.

[0024] Step two is implemented as follows: The collaborative controller embeds a battery charging characteristic database. For each charging pile in the charging state, it calls the corresponding standard charging power-time curve model based on its associated battery model. Combining the current charging time and the user's expected completion time, it calculates the minimum average power required to complete charging on time, which is the guaranteed power of the pile. The guaranteed power of all charging piles is summed to obtain the total guaranteed power requirement of the system. At the same time, based on the rated capacity of the distribution transformer and deducting the real-time estimated peak value of the park's non-charging basic load, the upper limit of the current grid's maximum charging power is calculated. The photovoltaic power supply capacity is directly taken as the current total photovoltaic output power value. The sum of the two is the dynamic available charging capacity.

[0025] Step three is implemented as follows: At the start of each day, the collaborative controller uses the photovoltaic system power conversion model to convert the irradiance and temperature prediction sequence for the next 24 hours into a photovoltaic power generation prediction curve. Combined with the historical charging statistics model predicting the concurrent charging demand, and aiming to maximize photovoltaic utilization and minimize grid net load fluctuations, an optimization algorithm is run to generate a baseline charging power plan curve for every 15 minutes of the next 24 hours. During real-time operation, the actual photovoltaic power is compared with the corresponding value in the plan curve every 5 minutes. If the positive deviation exceeds a set threshold, adjustable photovoltaic power is generated; if the negative deviation exceeds the threshold, power needs to be reduced. During power allocation, the guaranteed power of all charging piles is strictly ensured. When adjustable photovoltaic power is available, it is used as bonus power and weighted according to the urgency of each pile's charging demand (the reciprocal of the remaining charging time). When power needs to be reduced and the grid's supplementary capacity is exhausted, the charging power of the last-ranked charging pile is gradually reduced to a level slightly below its guaranteed power, based on the expected completion time of each pile (later times are ranked later), after obtaining user consent.

[0026] The implementation of step four is as follows: the collaborative controller converts the allocation result into a target power command for each charging pile and issues it. In the next acquisition cycle, it verifies the consistency between the actual power and the command, and marks and isolates charging piles that continuously exceed the tolerance.

[0027] The implementation of step five is as follows: the collaborative controller continuously monitors the power of the grid's point of common coupling at high speed. Once the power value exceeds the preset forward overload limit or reverse feed limit, protection is immediately activated, and the power of all charging piles whose current power is higher than their guaranteed power is forcibly reduced by a uniform ratio until the grid power is restored to safety.

[0028] Through a complete, closed-loop, multi-timescale collaborative control process, the deep temporal matching of photovoltaic power generation and charging load is systematically achieved, thereby ensuring a high proportion of local photovoltaic consumption, smooth grid interaction power, and reliable satisfaction of users' core charging needs at the technical level, thus constituting a complete and feasible technical solution.

[0029] In one possible implementation, the acquisition of real-time status and charging data of each low-speed charging pile relies on a standard protocol communication established between the collaborative controller and the charging pile controller. The collaborative controller periodically sends status query commands to each charging pile controller. The status information returned by the charging pile controller is encoded into three explicit enumerated values: "Idle" indicates that no vehicle is connected to the charging pile physical connector; "Connected but not charging" indicates that the vehicle is physically connected but the charging session has not yet started or has been paused; and "Charging" indicates that the charging session is in progress and electrical energy is flowing from the charging pile to the vehicle.

[0030] For charging piles in the "Charging" state, the collaborative controller further sends a data read command to obtain a set of real-time parameters, including: real-time charging power, which is the current instantaneous output active power value measured by the charging pile's internal measurement module, usually in kilowatts; charging time, which is accumulated from the start of this charging session, in hours or minutes; the user-set expected charging completion time, which is set by the user before or during charging via a mobile application, charging pile touchscreen, or website portal, and transmitted to the collaborative controller via the backend server or the charging pile controller itself; and the target state of charge, which is the percentage of charge the user expects their vehicle's battery to eventually reach, usually used as a charging stop condition.

[0031] The standard charging curves for the battery type are a series of data models pre-existing in the collaborative controller database. These models were obtained through laboratory testing, publicly available data from vehicle manufacturers, and historical charging big data analysis. For different brands and models of electric vehicle battery packs, curves were established showing the changes in charging power with battery state of charge or charging time during typical charging processes.

[0032] The curve typically includes a constant current stage and a constant voltage stage. The collaborative controller uses the vehicle identification code reported by the charging pile or the vehicle model manually selected by the user to call the corresponding curve model for subsequent calculations.

[0033] Comprehensive and accurate dynamic data of the charging process were obtained through standardized communication interfaces. In particular, the key requirement parameter of expected completion time set by the user was introduced. Combined with the pre-stored battery physical characteristic model, a solid data foundation was laid for the subsequent accurate quantification of the core concept of "guaranteed power". This enabled the collaborative control to change from scheduling fuzzy loads to scheduling specific tasks with clear time and energy constraints, which greatly improved the accuracy and rationality of the control.

[0034] In one possible implementation, calculating the guaranteed power of each low-speed charging station is a dynamic extrapolation process based on models and demand. First, when an electric vehicle begins charging, the collaborative controller needs to determine its battery model. This can be achieved by the charging station or vehicle automatically reporting its vehicle identification number (VIN) via a communication protocol, and the collaborative controller converting it to the battery model by querying a mapping database; or by the user manually selecting the vehicle model on the mobile application interface when charging begins.

[0035] After determining the battery model, the collaborative controller retrieves the corresponding standard charging curve for that battery type from its local database. This curve, stored as a data table or piecewise function, describes the suggested or typical charging power variation trajectory from 0% to 100% state of charge. Next, the collaborative controller reads the charging duration of the current session for that charging station. Using this duration, it performs interpolation or table lookup operations on the standard charging curve to pinpoint the theoretically appropriate charging stage of the battery at the current moment and its corresponding power reference value, and can estimate the accumulated energy absorbed by the battery. Then, the collaborative controller obtains the user-set desired charging completion time and calculates the difference between this and the current system time to determine the remaining available charging time.

[0036] Finally, the core calculation is performed: starting from the current location on the standard charging curve, a complete charging path is planned from the current state to the user's target state of charge. This path follows the shape of the standard charging curve, and the total energy required to complete the remaining charging is calculated. Dividing the required total energy by the remaining available charging time yields the minimum average power that must be maintained from the current moment to complete charging within the user's expected time. The cooperative controller formally defines this minimum average power value as the "guaranteed power" of the low-speed charging station. If the calculated guaranteed power value exceeds the physical maximum power of the charging station, the maximum power is used as its guaranteed power. By combining the user's subjective time desire (expected completion time) with the objective charging law of the battery (standard charging curve), a rigorous calculation is performed to transform an objective, quantitative, and dynamic lower limit value of power demand.

[0037] This "guaranteed power" is not a fixed value or a simple percentage, but changes in real time according to the charging progress and user settings. It scientifically represents the minimum power resources that must be provided to the charging task while respecting the battery charging characteristics and the user's time requirements. This establishes a clear and fair baseline for the entire collaborative power supply system, ensuring that optimized scheduling does not infringe on the user's most basic charging rights.

[0038] In one possible implementation, the development of the baseline charging power plan curve is a complex offline computational process involving prediction, modeling, and optimization. The co-controller initiates this process at a fixed time each day. First, it acquires a high-precision 24-hour forecast sequence of solar irradiance and ambient temperature. This environmental data is then input into the photovoltaic power generation system power prediction model.

[0039] This model is a simulation model calibrated using local historical data. Its calculation logic is as follows: First, the horizontal irradiance is converted to inclined plane irradiance based on the installation tilt and azimuth angle of the photovoltaic array. Then, considering the rated power of the photovoltaic modules under standard test conditions and the ratio of the current irradiance to the standard irradiance, a preliminary calculation of the DC power is performed. Next, a temperature-related correction coefficient is introduced, reflecting the characteristic that the output power of the photovoltaic modules decreases with increasing temperature. Finally, a fixed comprehensive efficiency coefficient covering factors such as inverter efficiency, line losses, and dust obstruction is multiplied to obtain the final predicted AC power value. This model generates a photovoltaic power generation prediction curve for the next 24 hours with a resolution of 15 minutes.

[0040] Simultaneously, the collaborative controller loads the historical charging database. Using historical data from the same period, it analyzes the statistical patterns of the number of charging pile connections, charging start time, charging duration, and total charging energy. Employing probabilistic statistical methods, such as time-segmented Poisson distributions or Gaussian mixture models, it predicts the probability distribution of the number of charging pile connections and their average charging demand for each time period in the next 24 hours, thus obtaining a predicted total charging demand power range with confidence intervals. Subsequently, the collaborative controller establishes a multi-objective optimization model with a 24-hour period. The optimization variable is the planned total charging power at the next 96 time points. The first optimization objective is to maximize the utilization rate of the photovoltaic predicted power, i.e., to maximize the sum of the smaller of the planned charging power and the photovoltaic predicted power.

[0041] The second optimization objective is to smooth the net grid load, i.e., minimize the volatility of the sequence obtained by subtracting the predicted photovoltaic power and the predicted base load from the planned charging power. Optimization constraints include: the planned power must not exceed the total physical power limit of all charging piles; the planned power must meet a certain high-confidence lower bound of the probability distribution of predicted charging demand to ensure the feasibility of the plan. Solving this optimization problem yields the optimal planned total charging power sequence for each time point in the next 24 hours, which is the "baseline charging power planned curve".

[0042] By integrating high-precision photovoltaic (PV) output forecasts with charging demand forecasts based on historical statistics, a forward-looking and globally optimized power allocation plan is formulated before the operation date. This plan not only guides charging loads to proactively track PV power generation trends on a macro level, laying the foundation for maximizing local consumption, but also considers smoothing the grid's net load in advance, suppressing grid impacts from a planning perspective. This provides a scientific and reasonable reference benchmark for subsequent real-time control, ensuring that real-time adjustments are no longer blind or reactive, but rather fine-tuned within an optimized framework.

[0043] In one possible implementation, determining whether the adjustable photovoltaic power can be increased or the power needs to be reduced is a key judgment step in the real-time fine-tuning stage. Its core lies in introducing a "power deviation threshold" to avoid overreacting to minor fluctuations in photovoltaic power. During system initialization, the co-controller sets a fixed power deviation threshold, for example, 10 kilowatts, based on the historical fluctuation characteristics of local photovoltaic output and the grid's capacity.

[0044] Within each real-time control cycle, the collaborative controller performs the following operations: First, it obtains the total active power output of the actual photovoltaic power generation system at the current moment. Simultaneously, based on the current system time, it queries the "benchmark charging power plan curve" generated in step three to find the planned total charging power value corresponding to that moment. Next, it calculates the algebraic difference between the actual photovoltaic power and the planned power value. Then, it compares the absolute value of this difference with a preset power deviation threshold. The judgment logic is divided into three cases: In the first scenario, if the actual photovoltaic power exceeds the planned power, and the difference exceeds the power deviation threshold, then the system is deemed to have "adjustable photovoltaic power." The specific value of adjustable photovoltaic power is equal to the actual power minus the planned power, and then minus the power deviation threshold. This portion of power is considered as photovoltaic surplus exceeding the planned expectations and that can be additionally utilized by the charging load.

[0045] In the second scenario, if the actual photovoltaic power is less than the planned power, and the absolute value of the difference exceeds the power deviation threshold, then the system is deemed to have a "power reduction requirement." The specific value of the power reduction requirement is equal to the planned power minus the actual power, and then minus the power deviation threshold. This power represents the power gap caused by insufficient photovoltaic output relative to the plan.

[0046] In the third scenario, if the absolute value of the difference between the actual photovoltaic power and the planned power is less than or equal to the power deviation threshold, the photovoltaic output is considered to be basically in line with the plan. No adjustable power increase or power reduction is required, and the system operates according to the original plan or current state. By setting a reasonable threshold, a "dead zone" or "buffer zone" is established. This effectively filters out high-frequency, small-amplitude random fluctuations in photovoltaic power caused by natural factors such as passing clouds, preventing the control system from frequently acting due to this noise, thus enhancing the system's stability and robustness. Only when the photovoltaic power shows a sustained and significant deviation from the planned trend will the system activate the power redistribution or reduction mechanism, making control actions more purposeful and reducing wear and tear on communication and execution equipment.

[0047] In one possible implementation, allocating adjustable photovoltaic power as bonus power to low-speed charging piles is a process of optimizing the allocation of remaining resources after basic needs are met. When the collaborative controller determines that "adjustable photovoltaic power" exists, it first confirms a prerequisite: all charging piles currently in a "charging" state have reached or exceeded their respective calculated "guaranteed power." This forms the basis for bonus allocation, ensuring fairness.

[0048] The system then calculates the "charging demand urgency" for each charging station. Charging demand urgency is defined as the reciprocal of the remaining time between the user-set expected charging completion time and the current system time. The shorter the remaining time, the higher the urgency value, indicating that the charging task is more time-sensitive. The collaborative controller calculates this urgency value for all eligible charging stations. Next, all charging stations are sorted in ascending order of charging demand urgency. This means that the charging station with the longest remaining time (lowest urgency) is at the top of the list, and the one with the shortest remaining time (highest urgency) is at the bottom. Bonus power allocation prioritizes meeting charging demands with lower urgency.

[0049] A weighted allocation algorithm is used, where the bonus power weight for each charging station is inversely proportional to the urgency of its charging need. That is, the longer the remaining time for a charging station, the higher the proportion of bonus power it receives. The allocation process can be iterative: starting with the top-ranked charging station, a theoretical allocation value is calculated based on its weight, but the sum of this value and the guaranteed power cannot exceed the maximum allowable charging power supported by the charging station's hardware. If the upper limit is not reached, the full amount is allocated; if the upper limit is reached, allocation continues until the upper limit is reached, and the remaining bonus power is allocated to the next charging station. This cycle continues until all adjustable photovoltaic power is allocated, or all charging stations reach their maximum allowable charging power. This creates an incentive-compatible allocation mechanism. It encourages users with less urgent charging needs (e.g., users whose vehicles are parked for long periods) to make better use of the available photovoltaic power. This maximizes the use of clean energy and optimizes the charging process, allowing vehicles with ample time to store more energy in advance. Simultaneously, this strategy has minimal impact on users with urgent charging needs, as they have already received the guaranteed power, and their basic charging progress remains unaffected.

[0050] This allocation method not only improves the overall photovoltaic absorption rate but also takes into account the differentiated needs of different user groups, reflecting the intelligence and humanization of the control.

[0051] In one possible implementation, prioritizing low-speed charging piles according to their expected charging completion time and reducing the power of lower-ranked charging piles is an orderly management measure to address severe power shortages. When the system determines that "power reduction is necessary," and even using the grid's maximum power supply capacity cannot meet the total guaranteed power of all charging piles, the system enters a power shortage state. At this time, the collaborative controller initiates this reduction process.

[0052] First, the controller sorts all charging stations in the "charging" state in ascending order according to their user-defined expected charging completion time, with the earliest completion time at the top and the latest at the bottom. This sorting reflects the urgency of the charging task; the later the expected completion time, the more generous the charging time window given by the user, and the lower the urgency. The decrementing process begins from the very end of the sorted list, starting with the charging station with the lowest urgency.

[0053] For the selected target charging station, the collaborative controller will not directly force a power reduction. Instead, it will send a clear power reduction request notification to the user through the associated user mobile application or the charging station screen. The notification includes an explanation of the current system power shortage, the suggested power reduction value, or an estimated delay in charging completion time, and requests the user's authorization.

[0054] The collaborative controller will only execute the power reduction command after receiving online confirmation from the user. The reduction is gradual, done in fixed "steps." This step is defined as a fixed percentage of the charging station's guaranteed power, such as 20% or 30%. The initial reduction lowers the current charging power by one step. After the reduction, the collaborative controller recalculates the difference between the total power demand and the total power supply capacity of the entire system, i.e., updates the power deficit.

[0055] If the power deficit is still greater than zero, the process continues, selecting the second-to-last charging station in the sorted list and repeating the above request and reduction steps. This process is repeated cyclically, like a "negotiation" process that gradually reduces power demand, with each reduction only made to a limited extent for the individual charging stations with the lowest urgency.

[0056] The cycle terminates under two conditions: first, the total power deficit in the system is eliminated, restoring balance; second, all adjustable charging piles have been reduced to a preset lower threshold, such as 80% of their guaranteed power. This provides a fair, transparent, and relatively user-friendly solution in extreme situations where all demand cannot be met. It adheres to the principle of "prioritizing those with more time constraints" and respects users' right to know and choose through a user confirmation mechanism, avoiding abrupt forced power outages and improving user experience. This orderly reduction strategy minimizes the social impact of power shortages while ensuring the safe operation of the power grid.

[0057] In one possible implementation, the verification of command execution is a key feedback loop to ensure the reliability and stability of the closed-loop control system. After the collaborative controller calculates the target charging power command for each charging pile according to the optimization algorithm and sends it to the corresponding charging pile controller via the communication network, the system does not assume that the command has been perfectly executed, but actively verifies it.

[0058] In the next preset data acquisition cycle, the collaborative controller will again send data read requests to all charging pile controllers that have just received commands via the communication protocol to obtain their current actual output power values. After obtaining the actual power values, the collaborative controller calculates a "power tracking error" for each charging pile. This error value is the absolute value of the difference between the target power command value and the actual read power value. The system has a preset "allowable error range," for example, ±0.5 kW, which takes into account normal error factors such as communication delay, measurement accuracy, and the response accuracy of the power regulation mechanism. The collaborative controller compares the calculated error of each charging pile with this allowable range. If the power tracking error of a charging pile exceeds the allowable range in a single check, the system will not immediately alarm but will instead activate a continuous monitoring mechanism.

[0059] For example, the system records the number of times a charging pile exceeds the tolerance limit consecutively. If the power tracking error of a charging pile continuously exceeds the allowable error range for three consecutive data acquisition cycles (e.g., three consecutive minute cycles), the collaborative controller determines that the charging pile has a "communication or equipment anomaly." Once an anomaly is determined, the collaborative controller will immediately take isolation measures: removing the charging pile's identifier from the list of active devices currently participating in optimization calculations. This means that in subsequent control cycles, the collaborative controller will no longer consider the power demand and allocation share of this abnormal charging pile when performing power allocation calculations.

[0060] Simultaneously, the collaborative controller generates an operation and maintenance alarm message, notifying management personnel to inspect the charging pile. The original power share of the abnormal charging pile will be temporarily reclaimed by the system and can be redistributed to other normally functioning charging piles according to the strategy. Only after the operation and maintenance personnel have troubleshooted the problem and confirmed through the management interface that the charging pile has returned to normal will the collaborative controller reinstate it to the list of active devices. This adds strong fault tolerance and self-healing capabilities to the entire collaborative control system. Through proactive feedback verification, it can promptly detect "disobedient" execution terminals and quickly isolate them, preventing the failure of a single device (such as communication interruption or power module damage) from affecting the correctness of the entire optimized control algorithm, or even causing grid power runaway. This greatly improves the operational robustness and security of the large-scale distributed charging pile group control system.

[0061] In one possible implementation, the forced reduction of the charging power of all adjustable low-speed charging piles is a high-priority, rapid protection action for grid safety.

[0062] First, the safety limits include two specific values: one is the forward overload limit, which is set based on the long-term overload capacity of the distribution transformer and the upstream line, and is usually slightly lower than its short-term allowable overload value, for example, set at 90% of the transformer's rated capacity; the other is the reverse feed-back limit, which is set based on the local power department's allowable upper limit for the reverse feed-back power of distributed power sources, or the maximum allowable reverse feed-back power from the perspective of distribution network voltage safety, for example, -50 kilowatts.

[0063] The coordinating controller samples the net load power at the point of common coupling in real time at a frequency much higher than the business control cycle (e.g., once per second). Once the sampled value exceeds the forward overload limit or falls below the reverse feed limit, the protection logic is immediately triggered. The protection logic employs a "proportional load reduction strategy." Specifically, the coordinating controller first calculates the overload ratio: for overload conditions, the overload ratio equals (current net load power - forward overload limit) / forward overload limit; for reverse feed conditions, the overload ratio equals (reverse feed limit - current net load power) / the absolute value of the reverse feed limit.

[0064] Then, a uniform "power reduction factor" is calculated based on this ratio. For example, this factor is set to (1 - excess ratio × a gain factor), ensuring that the factor is less than 1. Next, the collaborative controller quickly filters out all currently "adjustable" charging piles. Here, "adjustable" charging piles specifically refer to those whose current charging power is higher than their own "guaranteed power," that is, those charging piles that are enjoying "bonus power" or have a power margin. For these charging piles, their current charging power is multiplied by the calculated power reduction factor to obtain a new, reduced "temporary safe power instruction." This set of instructions is immediately and in parallel sent to all target charging piles, with execution priority exceeding any ongoing optimization allocation algorithm.

[0065] A "firewall" has been established to directly safeguard the safe operation of the distribution network. Responding to grid emergencies at millisecond speeds, it proportionally reduces all non-core charging power (i.e., the portion exceeding the guaranteed power), bringing grid power back to a safe range in the fastest and most direct way. This protection mechanism is independent of complex optimization algorithms, with simple and reliable logic and rapid and decisive action, ensuring that the physical safety of the grid is paramount under any circumstances. It is the cornerstone for the safe implementation of the entire collaborative control method in practice.

[0066] In one possible implementation, the system initialization and parameter self-learning phase is a crucial process for adapting the general control algorithm to specific installation scenarios and improving long-term operational performance. After the initial installation of the power supply system or changes to major equipment (such as large-scale photovoltaic expansion or charging piles), the collaborative controller does not immediately activate the full set of collaborative control functions. Instead, it enters a "self-learning cycle" of a set number of days, such as 14 or 30 days. During this cycle, the controller switches to a pure monitoring mode, meaning it does not issue any optimized scheduling commands to the charging piles, and all charging piles operate in the default "plug and charge" mode.

[0067] The controller's core task during this period is to comprehensively and completely record local operational data. Daily recorded data includes: a complete photovoltaic power generation curve, the start-up and shutdown times and charging power curves for each charging pile, and the net load power curve of the point of common connection (PCC). Simultaneously, the controller also records corresponding meteorological observation data. After the self-learning cycle ends, the co-controller initiates an offline analysis program. The program utilizes a large amount of recorded local historical data to train and correct key internal models and parameters using machine learning algorithms. First, for the photovoltaic power generation system model, the program fits historical meteorological data (irradiance, temperature) with corresponding actual photovoltaic output data, correcting parameters such as the overall efficiency coefficient and temperature coefficient in the model, making the prediction model closer to the actual performance of the local photovoltaic array.

[0068] Secondly, for the probabilistic statistical model of charging demand prediction, the program analyzes the patterns of historical charging behavior, such as charging pile utilization rates under different weekdays and weather conditions, the distribution of charging start times, and the distribution of charging energy demand, thereby optimizing the parameters of the prediction model to make its predictions of future charging demand more accurate. Finally, the program also analyzes the fluctuation characteristics of photovoltaic power and the grid response, optimizing the key control parameter of "power deviation threshold" to find an optimal value that effectively filters noise without ignoring the true trend. After completing the automatic learning and correction of all parameters, the collaborative controller officially switches to the full-function collaborative control mode, giving the control system strong adaptability. This avoids the performance degradation that may result from applying a fixed set of parameters and models to all different scenarios.

[0069] Through a phased self-learning process, the system can proactively "recognize" local resource characteristics (photovoltaics) and load characteristics (charging behavior), thereby "personalizing" its control parameters and models. This significantly improves the accuracy of photovoltaic power prediction, the precision of charging demand prediction, and the rationality of real-time control parameters, enabling the entire collaborative power supply method to maintain excellent performance over long-term operation, truly achieving intelligence and self-adaptation.

[0070] Example: Low-speed charging piles in office parks and distributed photovoltaic power generation work together This embodiment uses a high-tech office park as an application scenario. The park has a rooftop distributed photovoltaic power generation system with a total area of ​​approximately 5,000 square meters, and has deployed 80 AC low-speed charging piles with a power of 7kW in the ground parking lot and underground garage to serve the electric vehicle charging needs of the park's employees.

[0071] 1. System composition and connection relationships; The park's power supply system includes: a distributed photovoltaic power generation array, multiple string photovoltaic inverters (total capacity of 400kW), 80 low-speed charging piles, a multi-functional meter installed at the common connection point of the power grid where the distribution network enters the park's main power distribution room, and a high-performance industrial computer deployed in the park's data center as a collaborative controller.

[0072] The collaborative controller establishes stable data connections via the industrial Ethernet network within the park to the monitoring interfaces of all photovoltaic inverters, the communication interfaces of all low-speed charging pile controllers, and the communication interface of the multi-function meter at the power grid common connection point. The collaborative controller runs the control software described in this invention to achieve centralized monitoring and collaborative control of all equipment.

[0073] 2. Detailed implementation of the methods and steps; Step 1: The system basic data acquisition and preprocessing co-controller synchronously acquires data from the entire system once per minute.

[0074] Real-time data acquisition of photovoltaic power generation system: The collaborative controller polls each photovoltaic inverter in turn through the ModbusTCP protocol, reads its instantaneous AC output active power value, sums the active power values ​​of all inverters, and obtains the real-time total output active power of the photovoltaic power generation system, denoted as P_pv(t).

[0075] Real-time data acquisition for low-speed charging piles: The collaborative controller communicates with each charging pile controller via the OCPP1.6 protocol. For each charging pile, the controller reads its status, including "Idle," "Connected but not charging," and "Charging." For charging piles in the "Charging" state, the following data is further read: real-time charging power P_ev_i(t) (unit: kW), duration of the current charging session T_charged_i (unit: hours), and the "expected vehicle pickup time" (i.e., expected charging completion time) T_target_i set by the user through the mobile application. The user-set target state of charge, usually 95% by default, is also read.

[0076] Data acquisition at the power grid connection point: The collaborative controller reads the real-time "total active power" from the multi-function meter at the power grid common connection point via the DL / T645 protocol. This power value is the net load power flowing into the park, denoted as P_grid(t). When P_grid(t) is positive, it indicates that the park is drawing power from the power grid; when it is negative, it indicates that the park is feeding power back to the power grid.

[0077] Environmental data acquisition: The collaborative controller accesses the API interface of an authoritative meteorological data service provider via HTTPS protocol to obtain the predicted sequence of horizontal total solar irradiance (unit: W / m²) and ambient temperature (unit: °C) for the park's location at 15-minute intervals for the next 72 hours every morning. This invention primarily uses data from the next 24 hours.

[0078] Step 2: Charging demand analysis and available charging capacity calculation; This step is performed every 5 minutes.

[0079] Guaranteed Power Calculation: The collaborative controller pre-stores a database of electric vehicle battery charging characteristics, containing typical constant current-constant voltage charging power curve models for battery packs of mainstream models on the market. When a vehicle begins charging, the user needs to select the vehicle model on the mobile app (or the charging station automatically identifies the vehicle model), and the collaborative controller then calls the corresponding curve model. For any charging station i in the "charging" state, the process of calculating its guaranteed power P_guarantee_i is as follows: First, based on the charging time T_charged_i, locate the corresponding standard charging curve and determine the energy currently received by the battery. Then, calculate the remaining time T_remain_i from the current moment to the user-set expected charging completion time T_target_i. Finally, based on the standard charging curve, calculate the minimum total energy required to charge the battery from the current state to the target state of charge within the remaining time T_remain_i, and then divide this total energy by the remaining time T_remain_i to obtain the required average power, which is the guaranteed power P_guarantee_i. For example, a vehicle with a 60kWh battery has been charging for 1 hour (in constant current mode), and the goal is to charge it to 95% in 3 hours. The calculated remaining energy requirement is 42kWh. Therefore, the guaranteed power P_guarantee_i = 42kWh / 3h = 14kW. Since the maximum power of the charging station is 7kW, the actual guaranteed power is min(14,7) = 7kW.

[0080] Total guaranteed power requirement calculation: Add the guaranteed power P_guarantee_i of all charging piles in the "charging" state to obtain the current total guaranteed power requirement P_guarantee_total of the system.

[0081] Calculation of the upper limit of grid power supply: The rated capacity of the distribution transformer in the park is 1250kVA. The average power P_base of the park's non-charging basic load (lighting, air conditioning, office equipment, etc.) is approximately 300kW, and the peak power is approximately 500kW. To ensure that the transformer is not overloaded, after reserving a certain safety margin, the maximum power that the grid side can safely use for charging is set, that is, the upper limit of grid power supply P_grid_max is: P_grid_max = transformer capacity × power factor (taken as 0.95) - the currently estimated peak value of the basic load (taken as 400kW as a real-time conservative value) ≈ 1187kW - 400kW = 787kW.

[0082] Photovoltaic power supply capacity calculation: The photovoltaic power supply capacity P_pv_available is the current total output power P_pv(t) of the photovoltaic inverter, but it does not exceed the total rated capacity of the inverter of 400kW.

[0083] Dynamic available charging capacity calculation: The upper limit of the total power currently available for charging, i.e., dynamic available charging capacity P_available_total = P_grid_max + P_pv_available.

[0084] Step 3: Multi-timescale power allocation based on photovoltaic prediction; This step involves operations at two time scales.

[0085] 3.1 Day-ahead coarse allocation phase: At 00:00 each day, the coordinating controller initiates the day-ahead planning procedure. The procedure reads the solar irradiance prediction sequence I_pred(t) and temperature prediction sequence T_pred(t) for the next 24 hours (from 00:00 of the current day to 00:00 of the next day). These are then input into the photovoltaic power generation system simulation model of the park.

[0086] This model is an empirical formula model, and its core formula is: P_pv_pred(t)=P_stc×(I_pred(t) / I_stc)×[1+k×(T_pred(t)-T_stc)]×η. Where P_stc is the peak power under standard test conditions (400kW), I_stc is the standard irradiance (1000W / m²), k is the power temperature coefficient (taken as -0.004 / ℃), T_stc is the standard temperature (25℃), and η is the overall efficiency coefficient (taken as 0.85).

[0087] The model calculates the photovoltaic power generation prediction curve for the next 24 hours at 15-minute intervals. Simultaneously, the program retrieves historical charging data from the past 30 similar workdays (e.g., Tuesdays) and uses a Gaussian mixture model to predict the number of vehicles N_ev_pred(t) likely to be charging at each time point on that day, along with their average guaranteed power, thus obtaining the probability distribution of total charging demand. Subsequently, a multi-objective optimization algorithm is run at 15-minute intervals. The optimization variable is the planned total charging power P_plan(t) for the next 96 time points (24 hours × 4). The first optimization objective is to maximize the utilization rate of the predicted photovoltaic power: MaximizeΣmin(P_pv_pred(t),P_plan(t)). The second optimization objective is to smooth the net grid load: MinimizeΣ[P_plan(t)-P_pv_pred(t)-P_base]². Optimization constraints include: P_plan(t) cannot exceed the upper limit of the total physical power of the charging piles (80 × 7 = 560kW), and must meet the lower confidence limit of the predicted charging demand probability distribution. Solving this optimization problem yields the "baseline charging power plan curve".

[0088] 3.2 Real-time fine-tuning phase: Every 5 minutes (one control cycle), the co-controller executes the following real-time closed-loop control: a. Data Acquisition: Read the current actual P_pv_real, the actual total guaranteed power demand P_guarantee_total_real, the actual grid power supply limit P_grid_max_real (this value will be fine-tuned according to the real-time base load), and the actual photovoltaic power supply capacity P_pv_available_real.

[0089] b. Power Deviation Judgment: Query the baseline planned power value P_plan_now corresponding to the current time point. Calculate the deviation ΔP = P_pv_real - P_plan_now. Set the power deviation threshold ΔP_threshold to 10kW. If ΔP > ΔP_threshold, it is determined that there is "adjustable increase in photovoltaic power" P_add = ΔP - ΔP_threshold. If ΔP < -ΔP_threshold, it is determined that there is "power reduction required" P_cut = |ΔP| - ΔP_threshold. If the deviation is within ±ΔP_threshold, the photovoltaic output is considered to be basically consistent with the plan, and no increase or decrease operation is triggered.

[0090] c. Power allocation decision: Scenario 1 (Sufficient Solar Power): Assume that at midday on a certain day, P_pv_real = 350kW, P_plan_now = 320kW, ΔP = 30kW > 10kW, then P_add = 20kW. At this time, the total guaranteed power demand P_guarantee_total_real is 200kW (generated by 30 vehicles currently charging). First, ensure that all 30 vehicles charge at their respective guaranteed power, consuming 200kW. Solar power remains at 150kW (350kW - 200kW). Of this, 120kW is used for matching the plan (P_plan_now - guaranteed power demand = 320kW - 200kW), and of the remaining 30kW, 20kW is "adjustable solar power" exceeding the plan. This 20kW "bonus power" will be allocated. The co-controller calculates the urgency of the charging demand of these 30 vehicles (reciprocal of remaining time), sorting them from lowest to highest urgency (i.e., longest remaining time to shortest). Bonus power is preferentially allocated to vehicles ranked higher (lowest urgency). For example, if the first vehicle has 8 hours remaining, it is allocated a bonus power of 2kW, allowing it to charge at 9kW (assuming its guaranteed power is 7kW and its maximum power is 11kW); the second vehicle has 7.5 hours remaining, it is allocated 1.8kW, and so on, until the 20kW bonus power is fully allocated or all vehicles reach their maximum charging power.

[0091] Scenario 2 (Insufficient Photovoltaic Power): Assume that one evening, P_pv_real = 20kW, P_plan_now = 150kW, ΔP = -130kW < -10kW, then P_cut = 120kW. The total guaranteed power requirement P_guarantee_total_real is 300kW. Photovoltaics can only provide 20kW, and the grid power supply limit P_grid_max_real is 787kW. The theoretical maximum power supply capacity is 20 + 787 = 807kW > 300kW, which seems sufficient. However, at this time, the co-controller needs to execute a "power reduction" operation, that is, the total charging power should not exceed P_plan_now (150kW) + ΔP_threshold (10kW) = 160kW. Therefore, the system needs to reduce the total charging power from the current 300kW (guaranteed power requirement) by at least 140kW (300kW - 160kW). This is a serious power shortage. The co-controller immediately sorts all charging vehicles according to their expected charging completion time from latest to earliest. First, the last car owner in the priority list (with the latest expected completion time, e.g., 8 AM the next day) is contacted via a push notification on their mobile app: "Due to insufficient solar power output, to ensure grid safety, would you agree to temporarily reduce the charging power? A one-hour delay is expected." Once the user agrees, the coordinating controller reduces the vehicle's charging power from 7kW to 5kW (in increments of approximately 30% of the guaranteed power). This is a reduction of 2kW. The calculation then shows that a further reduction of 138kW is needed. This process is repeated with the next car owner until the total charging power drops below 160kW. Throughout this process, the charging power of all vehicles will not fall below 80% of their guaranteed power (i.e., 5.6kW) to ensure at least basic charging progress.

[0092] Step 4: Issuance and execution feedback of charging pile instructions; Every 5 minutes, after a power allocation decision is made, the collaborative controller generates a target charging power command P_cmd_i for each charging pile in the "charging" state, and sends it to the corresponding charging pile controller via the "RemoteStartTransaction" or "ChangeConfiguration" command of the OCPP protocol. The charging pile controller parses the command and adjusts its internal power module to achieve the target output power. During the next data acquisition (1 minute later), the collaborative controller reads the actual power P_actual_i of all charging piles and calculates the tracking error ε_i = |P_actual_i - P_cmd_i|. The preset allowable error range is ±0.5kW. If the ε_i of a charging pile is greater than 0.5kW for 3 consecutive cycles (3 minutes), the collaborative controller marks it as "abnormal," isolates it from the current control list, and issues a maintenance alarm. Its original power share will be temporarily released to the system for reallocation.

[0093] Step 5: Grid Interaction and Protection; The co-controller monitors the grid point of common coupling power P_grid(t) once per second. The set forward overload limit P_grid_limit_high is 1000kW (to prevent transformer overload), and the reverse feed limit P_grid_limit_low is -50kW (to limit reverse feed power to comply with local grid connection regulations). Once P_grid(t) > 1000kW or P_grid(t) < -50kW is detected, protection is immediately triggered. For example, if P_grid(t) reaches 1050kW, the over-limit ratio is calculated as (1050-1000) / 1000 = 5%. The co-controller immediately issues an emergency command to all "adjustable" charging piles (i.e., vehicles receiving bonus power) whose current power exceeds their guaranteed power, uniformly reducing their charging power to 95% of the original value (i.e., a reduction factor of 0.95). This operation is global and executed immediately, with higher priority than the optimization algorithm in step three, aiming to bring the grid power back to a safe range as quickly as possible.

[0094] 3. System initialization and parameter self-learning; During the initial deployment of the system in the park, a 14-day self-learning period was conducted. During this period, all charging piles adopted a "plug-and-charge" mode, with the co-controller only recording data and not performing active control. After the learning phase, the controller analyzed the deviation between the actual photovoltaic output and the predicted data, correcting the comprehensive efficiency coefficient η in the photovoltaic model; it analyzed the statistical patterns of charging demand, optimizing the parameters of the Gaussian mixture model; and based on grid fluctuations, it fine-tuned the power deviation threshold ΔP_threshold, adjusting it from the initial 15kW to a more suitable 10kW for the park.

[0095] To verify the effectiveness of the method described in this embodiment, operational data from the same week within the park were compared. Comparative Example 1 shows operational data using the traditional "plug-and-charge" mode; Comparative Example 2 shows operational data using a simple "timed charging" mode (set to charge during off-peak hours at night); Comparative Example 3 shows operational data assuming a 200kWh / 100kW large-scale energy storage system is installed for coordinated "photovoltaic-storage-charging" operation. The method of this invention is the operational data from this embodiment. The comparison results are shown in the table below: Comparison scale explanation: Comparative Example 1 (Plug and Charge): Vehicles can be charged anytime, with a constant charging power. During peak solar power generation at midday, charging demand may be insufficient, leading to a large amount of solar power being fed back into the grid, causing voltage increases and grid fluctuations; at night when there is no sunlight, the peak charging time coincides with the peak load of the industrial park, resulting in high peak power on the grid.

[0096] Comparative Example 2 (Timed Charging): This forces all charging to occur during off-peak hours at night. It completely misses the daytime photovoltaic power generation period, resulting in extremely low photovoltaic absorption rates. Furthermore, it shifts all daytime charging loads to the night, creating extremely high nighttime grid peak loads and exacerbating the nighttime grid burden.

[0097] Comparative Example 3 (Photovoltaic-Storage-Charging): A large-scale energy storage system was installed, which can store energy during periods of high photovoltaic power generation and discharge it during peak charging periods or when photovoltaic power is insufficient. Although the technology is effective, the energy storage system itself is expensive and suffers from issues such as capacity decay and safety maintenance, resulting in poor economic efficiency.

[0098] Conclusion: As shown in the table above, the method described in this invention achieves a high local photovoltaic absorption rate (79.2%), approaching the "photovoltaic-storage-charging" model, without relying on large-scale energy storage. Simultaneously, it effectively mitigates grid power fluctuations (standard deviation 120kW) and controls peak grid power within a reasonable range (650kW). Most importantly, through the "guaranteed power" mechanism, it maximizes the protection of users' core charging needs (99% satisfaction rate) while optimizing operation, achieving the best balance between technical effectiveness and economic efficiency.

[0099] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0100] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0101] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation, applied to a power supply system comprising a distributed photovoltaic power generation system, multiple low-speed charging piles, and a coordinated controller; the coordinated controller is communicatively connected to the photovoltaic inverter, the charging pile controllers of each low-speed charging pile, and the measuring equipment at the point of common connection of the power grid; characterized in that, The method includes the following steps: Step 1: The collaborative controller collects the real-time output active power of the distributed photovoltaic power generation system, the real-time status and charging data of each low-speed charging pile, the real-time net load power of the grid common connection point, and the future solar irradiance prediction data sequence and temperature prediction data sequence. Step 2: The collaborative controller calculates the guaranteed power of each low-speed charging pile based on its battery type, standard charging curve, charging duration, and user-set expected charging completion time. It then sums the guaranteed power of all low-speed charging piles in operation to obtain the total guaranteed power requirement. The controller also calculates the grid power supply limit based on the distribution network capacity and current base load. Finally, it calculates the photovoltaic power supply capacity based on the rated capacity and real-time active power output of the photovoltaic inverter. The sum of the grid power supply limit and the photovoltaic power supply capacity constitutes the dynamic available charging capacity. Step 3: The collaborative controller generates a photovoltaic power generation prediction curve based on the solar irradiance prediction data sequence and the photovoltaic power generation system model, and generates a charging demand prediction based on historical charging data. With the goal of maximizing the photovoltaic predicted power coverage of charging demand and smoothing the net load power curve at the grid's point of common coupling, a baseline charging power plan curve is formulated. Within a control cycle, the collaborative controller compares the actual photovoltaic output power with the planned value for the corresponding time period in the baseline charging power plan curve to determine the adjustable photovoltaic power or the power that needs to be reduced. With the constraint of prioritizing the overall guaranteed power demand, real-time optimization calculations are performed and power is allocated: when adjustable photovoltaic power exists, it is allocated as bonus power to low-speed charging piles on the basis of meeting the guaranteed power; when power needs to be reduced and the grid power supply limit cannot compensate, low-speed charging piles are sorted according to the urgency of their expected charging completion time, and the charging power of low-speed charging piles ranked lower is reduced. Step 4: The collaborative controller sends the target charging power command to the charging pile controller of each low-speed charging pile according to the power allocation result, and verifies the execution of the command in the next data acquisition cycle. Step 5: The collaborative controller continuously monitors the net load power at the power grid's point of common coupling. When the net load power exceeds the safety limit, it forcibly reduces the charging power of all adjustable low-speed charging piles until the net load power returns to the safe range.

2. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step one, the collection of real-time status and charging data of each low-speed charging pile specifically includes: obtaining the "idle", "connected but not charging" or "charging" status of the low-speed charging pile by communicating with the charging pile controller. For low-speed charging piles that are in the "charging" state, obtain their real-time charging power, charging time, user-set expected charging completion time, and target state of charge. The standard charging curve for the battery type is a typical constant current-constant voltage charging power curve model pre-existing in the collaborative controller and established according to different electric vehicle battery models.

3. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step two, the specific process for calculating the guaranteed power of each low-speed charging pile is as follows: The collaborative controller calls the corresponding standard charging curve for the battery type based on the battery model of the electric vehicle connected to the low-speed charging pile. Based on the charging time of the low-speed charging station, locate the current charging stage on the standard charging curve of the battery type; Then, based on the time difference between the user's expected charging completion time and the current time, the minimum average power required to complete the full charging process as specified by the standard charging curve for the battery type from the current time to the expected charging completion time is calculated, and this minimum average power is defined as the guaranteed power of the low-speed charging pile.

4. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step three, the formulation of the baseline charging power plan curve specifically includes: the collaborative controller using the photovoltaic power generation system model to convert the solar irradiance prediction data sequence and temperature prediction data sequence for the next 24 hours into the photovoltaic power generation prediction curve for the next 24 hours; Meanwhile, based on the historical data of the number of low-speed charging piles connected, charging start time, charging duration and charging energy, the range of total guaranteed power demand for each period in the next 24 hours is predicted through a probabilistic statistical model. Using the photovoltaic power generation forecast curve as the power supply side reference and the predicted total guaranteed power demand range as the load side reference, a rolling optimization algorithm is adopted. The primary optimization objective is to maximize the total amount of photovoltaic forecast power used by the charging load in the next 24 hours, and the secondary optimization objective is to minimize the variance of the net load power curve at the grid common connection point. The planned total charging power for each period in the next 24 hours is obtained by solving the algorithm, forming the baseline charging power plan curve.

5. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step three, the specific method for determining whether the photovoltaic power can be increased or needs to be reduced is to set a power deviation threshold. The collaborative controller calculates the difference between the actual photovoltaic output power and the planned value for the corresponding time period in the baseline charging power planned curve within the current control cycle; If the difference is greater than zero and exceeds the power deviation threshold, it is determined that there is adjustable photovoltaic power, and the value of adjustable photovoltaic power is equal to the difference minus the power deviation threshold. If the difference is less than zero and its absolute value exceeds the power deviation threshold, it is determined that there is a need to reduce power, and the amount of power to be reduced is equal to the absolute value of the difference minus the power deviation threshold.

6. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step three, the adjustable photovoltaic power is allocated as bonus power to low-speed charging piles. The specific allocation strategy is as follows: The collaborative controller first confirms that the guaranteed power of all low-speed charging piles that are charging has been met; Then, the urgency of the charging demand for each low-speed charging pile is calculated, where the urgency of the charging demand is the expected charging completion time and the remaining time at the current moment. The low-speed charging stations are sorted in order of increasing urgency of charging demand, i.e., the shorter the remaining time, the higher the urgency. Bonus power is preferentially allocated to low-speed charging piles with low charging demand urgency (i.e., long remaining time). The allocation weight is inversely proportional to the charging demand urgency, and the sum of the bonus power allocated to any low-speed charging pile and the guaranteed power of that low-speed charging pile shall not exceed the maximum allowable charging power of that low-speed charging pile.

7. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step three, the process of sorting low-speed charging piles according to the urgency of their expected charging completion time and reducing the charging power of low-speed charging piles that are ranked lower is as follows: when there is a need to reduce the power and the upper limit of the power grid has been reached, the collaborative controller sorts all low-speed charging piles that are charging in order of their expected charging completion time from farthest to nearth, that is, the later the expected charging completion time, the lower the ranking. Starting with the lowest-speed charging pile in the order, the collaborative controller sends a power reduction request to the user of that low-speed charging pile and waits for confirmation. After obtaining user confirmation, the current charging power of the low-speed charging pile is reduced by one step, where the step is a fixed proportion of the guaranteed power of the low-speed charging pile. For each low-speed charging station removed, the system power deficit is recalculated. If the power deficit is still greater than zero, the same reduction process is continued for the next low-speed charging station in the order until the power deficit is eliminated or all adjustable low-speed charging stations have been reduced to the lower limit threshold of the guaranteed power.

8. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step four, the verification of the command execution status specifically includes: in the next data acquisition cycle after the collaborative controller issues the target charging power command, reading the actual output power of each low-speed charging pile through the charging pile controller; The actual output power is compared with the target charging power command issued to calculate the power tracking error. If the power tracking error of a low-speed charging pile continues to exceed the preset allowable error range for a set number of times, the collaborative controller will determine that the low-speed charging pile has a communication or equipment abnormality, and will remove the low-speed charging pile from the current optimization allocation list. Its power demand will no longer participate in subsequent real-time optimization calculations until the abnormality is eliminated.

9. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, In step five, the safety limits include forward overload limits and reverse feed limits; The forced reduction of the charging power of all adjustable low-speed charging piles is specifically achieved by adopting a proportional load reduction strategy: the coordinating controller calculates the proportion of the net load power at the grid common connection point that exceeds the safety limit. Based on the proportion of the excessive power, calculate a uniform power reduction coefficient; Multiply the charging power of all low-speed charging piles that are currently charging and not locked by the power constraint in step three by the power reduction coefficient to obtain a set of temporary safe power commands and issue them immediately. The low-speed charging pile that is not locked by the guaranteed power constraint refers to a low-speed charging pile whose current charging power is higher than its guaranteed power.

10. The method for coordinated power supply of low-speed charging piles and distributed photovoltaic power generation according to claim 1, characterized in that, The method also includes a system initialization and parameter self-learning phase: after the power supply system is put into operation for the first time or after major equipment changes, the coordinating controller performs a self-learning cycle for a set number of days; During the self-learning cycle, the collaborative controller operates in plug-and-charge mode and fully records the daily photovoltaic power generation curve, the charging behavior curve of each low-speed charging pile, and the net load power curve of the grid common connection point. After the self-learning cycle ends, the collaborative controller uses recorded historical data to train and correct the parameters in the photovoltaic power generation system model, the parameters in the charging demand prediction probability statistics model, and the initial value of the power deviation threshold through machine learning algorithms, so that the prediction and control parameters in subsequent steps one to five are adapted to the specific local operating environment.

Images