A method and system for dynamic scheduling of charging pile power

Abstract

The application relates to the technical field of power load regulation, and discloses a method and system for dynamic scheduling of charging pile power. The method comprises the following steps: obtaining a real-time data sequence containing a basic load and a charging pile load, and obtaining a load fluctuation trend after denoising by using a Kalman filtering algorithm; calculating an instantaneous change rate of the basic load, and determining a power offset amount that needs to be compensated on the charging pile side when the instantaneous change rate exceeds a preset change threshold; inputting the power offset amount into a proportional-integral-derivative control algorithm to generate a preliminary power scheduling instruction for each charging pile; verifying and amplitude correcting the preliminary power scheduling instruction in combination with an upper limit of a total load of a transformer area to obtain an optimized power scheduling instruction; controlling each charging pile to adjust output power, generating a scheduled total load curve and performing stability verification; and when a verification result does not satisfy a preset safety threshold, updating Kalman filtering algorithm parameters and using the updated Kalman filtering algorithm parameters for subsequent real-time data processing. The method can improve the real-time performance, stability and safety of transformer area load regulation.

Description

Technical Field

[0001] This application relates to the field of power load regulation technology, and in particular to a method and system for dynamic power scheduling of charging piles. Background Technology

[0002] With the increasing adoption of electric vehicles, the number of charging stations connected in residential communities, commercial areas, and public power distribution substations is constantly increasing. Charging station loads are characterized by random connection, frequent start-stop cycles, and rapid power fluctuations. When combined with basic loads such as residential and commercial electricity consumption, this can easily cause rapid changes in the total load of the substation transformer, increasing the risk of transformer overload.

[0003] Existing charging pile load control schemes typically employ fixed power limits, timed control, threshold alarms, or current limiting control based on total load exceeding limits. With the development of real-time monitoring and control technologies, some schemes are beginning to combine load acquisition, filtering, and control algorithms to dynamically adjust the charging pile power.

[0004] However, existing technologies still have the following problems: the total load data collected from the transformer side usually contains measurement noise and abnormal disturbances, which can easily lead to misjudgment if used directly for adjustment judgment; at the same time, existing solutions are mostly based on the absolute value of the total load or the over-limit state for control, making it difficult to identify the trend of basic load changes in a timely manner, and thus it is difficult to generate reasonable charging pile power scheduling instructions in advance before the total load approaches the safety boundary; in addition, in the scenario of multiple charging piles operating in parallel, existing solutions do not adequately consider the boundary constraints of scheduling instructions and the stability after scheduling, which affects the adjustment effect.

[0005] Therefore, how to effectively process real-time load data, accurately identify the trend of basic load changes, and generate dynamic power scheduling instructions for charging piles in the operation scenario of charging piles connected to transformer substations, while combining the total load limit of the substation and the stability after scheduling for closed-loop optimization, has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] This application provides a method and system for dynamic power scheduling of charging piles to solve the problems in the prior art where, in the operation scenario of charging piles connected to the transformer in the distribution area, it is difficult to effectively process real-time load data, difficult to identify the trend of basic load changes in a timely manner, difficult to generate reasonable charging pile power scheduling instructions in advance before the total load approaches the safety boundary, and the scheduling instruction boundary constraints and post-scheduling stability are insufficient.

[0007] Firstly, this application provides a method for dynamic power scheduling of charging piles, the method comprising: S1. Collect the total load of the transformer, obtain the real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend. S2. Calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, trigger reverse adjustment to determine the power offset that needs to be compensated on the charging pile side. S3. Input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile. S4. Based on the total load limit of the distribution area, verify whether the initial power dispatch command exceeds the safety boundary. If so, correct the command amplitude and obtain the optimized power dispatch command. S5. Send the optimized power scheduling command to each charging pile controller so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution. S6. Based on the charging pile load distribution and the basic load, generate the total load curve after scheduling and perform stability verification to obtain the verification results; S7. If the verification results show that the total load curve is not stable within the preset safety threshold, then iteratively update the Kalman filter algorithm parameters and use the updated algorithm for subsequent real-time data processing.

[0008] Secondly, this application provides a dynamic power scheduling system for charging piles, the system comprising: The acquisition and filtering module is used to acquire the total load of the transformer, obtain a real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend. The trigger compensation module is used to calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, it triggers reverse adjustment to determine the power offset that needs to be compensated on the charging pile side. The instruction generation module is used to input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile. The safety verification module is used to verify whether the initial power dispatch command exceeds the safety boundary based on the total load limit of the distribution area. If so, the command amplitude is corrected to obtain the optimized power dispatch command. The instruction execution module is used to send the optimized power scheduling instruction to each charging pile controller so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution. The verification and evaluation module is used to generate the total load curve after scheduling based on the charging pile load distribution and basic load, and to perform stability verification to obtain the verification results. The parameter update module is used to iteratively update the Kalman filter algorithm parameters when the verification results show that the total load curve has not stabilized within the preset safety threshold, and then use the updated algorithm for subsequent real-time data processing.

[0009] The beneficial effects of this application are at least as follows: By collecting the total transformer load and using a Kalman filter algorithm to denoise the real-time data sequence, a smoother load fluctuation trend can be obtained even in the presence of measurement noise and abnormal disturbances, improving the accuracy of load change identification. Based on the instantaneous rate of change of the base load, reverse regulation is triggered, and the power offset to be compensated at the charging pile side is determined. This allows for timely dynamic compensation of the charging pile power when the base load changes abruptly, improving the predictability and responsiveness of scheduling and reducing the operational risks caused by rapid increases in the total load of the transformer in the distribution area. A proportional-integral-derivative (PID) control algorithm is used to generate preliminary power scheduling commands for each charging pile. These commands are then checked for safety boundaries and their amplitude corrected based on the upper limit of the total load in the distribution area. This ensures that the power values ​​of each charging pile are within a safe operating range while improving the rationality and executability of power scheduling. After each charging pile executes the optimized power scheduling commands, a total load curve after scheduling is generated based on the adjusted charging pile load distribution and base load. Stability verification is then performed, enabling closed-loop evaluation of the scheduling effect and improving the stability of total load control. When the verification results show that the total load curve is not stable within the preset safety threshold, the Kalman filter algorithm parameters are further updated, and the updated Kalman filter algorithm is used for subsequent real-time data processing. This can enhance the adaptive capability of subsequent load identification and scheduling control, thereby improving the real-time performance, stability and safety of transformer load regulation in the distribution area. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of a method for dynamic power scheduling of charging piles according to this application; Figure 2 This is a comparison chart of the Kalman filter noise reduction effects of embodiments of this application; Figure 3 This is a schematic diagram showing the relationship between the power offset and the PID control output response in this application; Figure 4 This is a schematic diagram comparing the total load curve after scheduling with the safety threshold in this application. Figure 5This is a schematic diagram of the structure of a charging pile power dynamic scheduling system according to this application. Detailed Implementation

[0012] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0013] For ease of understanding, the specific process of the embodiments of this application is described below. Figure 1 The diagram shows a flowchart of a method for dynamic power scheduling of charging piles provided by the present invention. The flowchart specifically includes the following steps: S1. Collect the total load of the transformer, obtain the real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend.

[0014] In one specific embodiment, the process of performing step S1 may specifically include the following steps: The total load of the transformer is monitored in real time by sensors, and load data is collected at a preset frequency to form a real-time data sequence that includes the base load and the charging pile load. The Kalman filter algorithm is applied to the real-time data sequence for state estimation. The actual load value is used as the state variable. The prior estimate is calculated through the prediction step, and the posterior estimate is obtained by fusing the measured value and the prior estimate through the update step. The smoothed load fluctuation trend is generated based on the posterior estimation.

[0015] Specifically, a total load acquisition unit is installed at the main outgoing line location on the low-voltage side of the transformer in the distribution area. Each charging pile controller uploads real-time output power data for its corresponding charging pile. The total load acquisition unit collects the total active power on the transformer's output side, and each charging pile controller uploads the charging power of a single pile or single charging gun. The transformer's total load sampling value and the power values ​​of each charging pile are timestamped according to a unified time base, and data from multiple sampling periods are continuously received at a preset sampling frequency to form a real-time data sequence. The real-time data sequence includes at least the sampling time, the transformer's total load value, and the power values ​​of each charging pile. The total load of the charging piles is obtained by summing the power values ​​of each charging pile at the same sampling time.

[0016] Considering potential sensor measurement errors, communication jitter, duplicate uploads, and transient spikes in the on-site sampling link, the real-time data sequence undergoes preprocessing before Kalman filtering. Preprocessing includes invalid value removal, missing sampling point completion, and abnormal transition marking. Abnormal transition marking identifies sampling points that meet one of the following conditions: the change in the transformer's total load at the current sampling time relative to the adjacent sampling time exceeds a preset abnormal change threshold, and there is no matching change in the charging pile's total load change at the corresponding time; or the deviation between the current sampling point and the adjacent sampling point exceeds a preset statistical deviation threshold. Marked sampling points retain their original records, and their observation weight is reduced during subsequent filtering updates.

[0017] When applying the Kalman filter algorithm to state estimation on preprocessed real-time data sequences, the total active power load state of the transformer is set as the state to be estimated, and the sampled value of the total transformer load is set as the observation input. Within each sampling period, based on the posterior estimate of the previous sampling period and the continuously changing state transition relationships within adjacent sampling periods, prediction processing is performed to obtain the prior estimate and corresponding prior estimate error covariance for the current sampling period. Then, the measured value of the total transformer load for the current sampling period is read, the observation residual between the measured value and the prior estimate is calculated, and the filter gain for the current sampling period is calculated based on the process noise covariance and measurement noise covariance. The filter gain is used to weight and correct the observation residual to obtain the posterior estimate for the current sampling period, and the posterior estimate error covariance is updated. The updated posterior estimate error covariance is used for prediction processing in the next sampling period.

[0018] When charging piles in the transformer substation experience frequent start-ups and shutdowns or when residential load fluctuates significantly, a higher level of process noise covariance is used to improve the ability of state estimation to follow actual load changes. Conversely, when sensor jitter, communication interference, or significant single-point spikes occur, a higher level of measurement noise covariance is used to reduce the impact of abnormal observations on the estimation results. For example, if a single point of sudden increase in the total transformer load occurs within a sampling period, and the power of each charging pile and the line current do not show corresponding changes in adjacent sampling periods, this sampling point can be marked as an abnormal observation point. In the update processing of this sampling period, the measurement noise covariance is increased to maintain the continuity of the posterior estimation results with the preceding and following sampling periods.

[0019] In the output stage, the posterior estimates corresponding to each sampling period are arranged in chronological order to generate a smoothed load fluctuation trend, and the total load sequence of charging piles corresponding to each sampling time is output synchronously. In subsequent steps, the total load of charging piles at the corresponding time is subtracted from the smoothed load fluctuation trend under the same time reference to obtain the basic load sequence.

[0020] Figure 2This is a comparison chart of the Kalman filtering noise reduction effect in a specific embodiment of the present invention. The horizontal axis represents time, and the vertical axis represents the load value. The blue solid line represents the original acquired transformer total load data sequence, the red solid line represents the load trend curve obtained after Kalman filtering, and the green "×" marks are schematically marked abnormal jump points. As can be seen from the figure, Kalman filtering can suppress high-frequency noise and local abnormal spikes while preserving the overall trend of load changes, thereby improving the stability and accuracy of subsequent calculations of the base load change rate and determination of power regulation.

[0021] S2. Calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, trigger reverse adjustment to determine the power offset that needs to be compensated on the charging pile side.

[0022] In one specific embodiment, the process of performing step S2 may specifically include the following steps: Extract time series points from the load fluctuation trend, calculate the difference between the base load at the current moment and the previous moment, divide by the time interval, and obtain the instantaneous rate of change at the current moment; Determine whether the instantaneous rate of change exceeds a preset threshold; if so, activate the reverse adjustment mechanism. Based on the portion of the current total load that exceeds the base load, the power offset that needs to be compensated on the charging pile side is determined. The power offset represents the power value that needs to be reduced on the charging pile side.

[0023] Specifically, the estimated total load values ​​for the current and previous sampling times are extracted from the smoothed total load fluctuation trend. The total load value of the charging piles at the corresponding time is then read. The base load value for each time is obtained by subtracting the estimated total load value from the corresponding total load value. The instantaneous rate of change of the base load at the current sampling time is then calculated based on the difference between the base load value at the current sampling time and the base load value at the previous sampling time, and the corresponding time interval. When the sampling period is fixed, the instantaneous rate of change is obtained by dividing the difference in base load between two adjacent sampling times by the fixed sampling period. When the sampling period is not fixed, the actual time interval with timestamps is used for calculation to avoid distortion of the rate of change.

[0024] The preset change threshold is pre-set based on the transformer capacity level, historical base load fluctuation characteristics, sampling period, and charging pile access scale to determine whether the base load has entered a rapid growth state. When the instantaneous change rate does not exceed the preset change threshold, no flag or zero compensation request is triggered, and the current scheduling state is maintained; when the instantaneous change rate exceeds the preset change threshold, a reverse adjustment trigger signal is generated, and the power offset calculation stage begins. To reduce false triggering caused by residual noise at a single sampling point, an execution layer judgment condition can be added, requiring that the instantaneous change rate exceeds the preset change threshold in multiple consecutive sampling periods, or requiring that the instantaneous change rate at the current sampling moment exceeds the preset change threshold and the remaining capacity between the current total load and the safe operation threshold is lower than a preset margin value, before the reverse adjustment mechanism is activated; this additional judgment condition does not change the main logic of triggering scheduling based on the instantaneous change rate of the base load.

[0025] In the power offset calculation phase, the estimated total load and base load value at the current sampling time are read. The current load occupancy on the charging pile side is obtained by subtracting the base load value from the estimated total load, and this is taken as the resource scale that can participate in reverse compensation at the current time. Since the current load occupancy on the charging pile side is not necessarily equal to the power that needs to be reduced, the difference between the current total load and the target safe load is further calculated. This difference is used to determine the initial compensation demand that the charging pile side needs to yield at the current time; when the current total load is not higher than the target safe load, the initial compensation demand is recorded as 0. Then, the initial compensation demand is compared with the current load occupancy on the charging pile side and the adjustable total capacity of the current online charging pile cluster. The value that is not greater than the current load occupancy on the charging pile side, not greater than the adjustable total capacity, and not less than 0 is taken as the power offset corresponding to the current sampling time. The power offset represents the active power value that the charging pile side needs to reduce in the current control cycle.

[0026] The adjustable total capacity of the current online charging pile cluster is obtained by summing the differences between the current operating power and the minimum allowable operating power of each online charging pile that can participate in scheduling. When the current operating power of a charging pile is less than or equal to its minimum allowable operating power, the adjustable capacity of that charging pile is recorded as 0. After the above processing, S2 outputs the reverse adjustment trigger flag and the power offset.

[0027] In another optional embodiment, to avoid missing the problem of a sudden surge in total load due to a stable base load but a large number of charging piles starting up simultaneously, which would only use the instantaneous rate of change of the base load as the triggering condition, S2 adopts a dual-threshold hierarchical triggering mechanism, the execution process of which can specifically include the following steps: Extract the total load estimate for the current time and the previous time from the smoothed load fluctuation trend, and combine it with the total load of the charging piles at the corresponding time to obtain the basic load time series data for the current time and the previous time. Calculate the difference between the basic load for the current time and the previous time and divide it by the corresponding time interval to obtain the instantaneous change rate of the basic load at the current time. Read the estimated total load and the real-time output power of each online charging pile at the current moment, calculate the remaining capacity between the current total load and the preset safety threshold, and obtain the instantaneous rate of change of the total load at the current moment based on the difference between the estimated total load at the current moment and the previous moment and the corresponding time interval. The reverse adjustment mechanism is activated when any of the following triggering conditions are met: The basic load triggering condition is that the instantaneous change rate of the basic load exceeds the preset change threshold and the remaining capacity is less than the preset safety margin threshold, or the predicted total load for the next period, obtained based on the current basic load value, the instantaneous change rate of the basic load, and the preset prediction duration, exceeds the preset warning threshold. The total load trigger condition is that the instantaneous rate of change of the total load exceeds the preset total load change threshold, or the current total load exceeds the preset safety threshold. After the reverse adjustment mechanism is activated, when the current total load is higher than the target safe load, the difference between the current total load and the target safe load is used as the initial compensation power demand, and the target safe load is not greater than the preset safe threshold. The system counts the number of currently online charging piles that can participate in scheduling, the current operating power of each charging pile, and the minimum allowable operating power. It calculates the difference between the current operating power and the minimum allowable operating power of each charging pile, and sums the differences to obtain the total adjustable capacity of all charging piles that can participate in scheduling. The initial compensation power requirement is compared with the adjustable total capacity, and the smaller value between the two, which is not less than 0, is taken as the target power offset that needs to be compensated on the charging pile side.

[0028] Specifically, the preset safety threshold is set in advance based on the rated capacity of the transformer in the distribution area, the allowable load rate, and the operation strategy of the distribution area. The preset change threshold is used to identify whether the base load has entered a rapid growth state. The preset safety margin threshold is used to identify whether the remaining capacity is lower than the short-term risk protection requirements. The preset warning threshold is used to determine whether the predicted total load in the next period is close to the overload boundary. The preset total load change threshold is used to determine whether the total load has risen abnormally and rapidly. All of the above thresholds are used as judgment conditions and do not replace the power compensation amount itself.

[0029] After calculating the instantaneous change rate of base load, the instantaneous change rate of total load, and remaining capacity, the system enters the trigger judgment phase. The base load trigger condition is as follows: if the instantaneous change rate of base load exceeds a preset change threshold and the remaining capacity is less than a preset safety margin threshold, it indicates that the base load is rapidly increasing and the remaining capacity of the transformer area is insufficient; or, based on the current base load value, the instantaneous change rate of base load, and a preset prediction duration, the base load value for the next time period is predicted, and this value is added to the current total load value of the charging piles to obtain the predicted total load for the next time period. When the predicted total load for the next time period exceeds a preset warning threshold, the reverse adjustment mechanism is activated. The total load trigger condition is as follows: if the instantaneous change rate of total load exceeds a preset total load change threshold, it indicates that the current total load has experienced an abnormally rapid increase; or, if the current total load exceeds a preset safety threshold, it indicates that the transformer area has entered a state requiring immediate reduction of controllable load. By setting the base load trigger condition and the total load trigger condition in parallel, the system can identify capacity compression caused by rapid base load growth and sudden increases in total load caused by concentrated startup, resumption of charging, or increased output power of charging piles.

[0030] After activating the reverse adjustment mechanism, first determine whether the current total load is higher than the target safe load. If the current total load is higher than the target safe load, use the difference between the current total load and the target safe load as the initial compensation power demand. If the current total load is not higher than the target safe load, the initial compensation power demand is 0. Next, count the number of currently online charging piles that can participate in scheduling, and read the current operating power and minimum allowable operating power of each charging pile. Calculate the adjustable capacity of each charging pile and sum them to obtain the total adjustable capacity. Compare the initial compensation power demand with the total adjustable capacity, and take the smaller value that is not less than 0 as the target power offset.

[0031] In another preferred embodiment, step S2 employs a compensation amount calculation method based on predicted total load and risk correction, and its execution process may specifically include the following steps: Calculate the instantaneous rate of change of the base load and the instantaneous rate of change of the total load at the current moment, respectively; When the instantaneous rate of change of the base load exceeds the preset change threshold, or the instantaneous rate of change of the total load exceeds the preset total load change threshold, or the current total load exceeds the transformer safe operation threshold, the reverse adjustment mechanism is activated. When the reverse adjustment mechanism is triggered by the instantaneous change rate of the base load, the predicted increase of the base load in the next adjustment cycle is calculated based on the instantaneous change rate of the base load and the preset prediction time window. Combined with the current total load of the charging piles, the predicted total load in the next adjustment cycle is obtained. When the reverse regulation mechanism is triggered by the instantaneous rate of change of total load or the current total load, the predicted increase of total load in the next regulation cycle is calculated based on the instantaneous rate of change of total load and the preset prediction time window, and the predicted total load in the next regulation cycle is obtained by combining the current transformer total load. The predicted total load is compared with the transformer's safe operating threshold to obtain the expected overload. A trend risk coefficient is generated based on the instantaneous rate of change corresponding to the triggering of the reverse adjustment mechanism, and a margin risk coefficient is generated based on the remaining capacity between the current total load and the transformer's safe operation threshold. The expected overload is corrected based on the trend risk coefficient and the margin risk coefficient to obtain the initial compensation power requirement. The initial compensation power requirement is compared with the total adjustable capacity of the currently online and dispatchable charging piles. The smaller value between the two, which is not less than zero, is taken as the power offset that needs to be compensated on the charging pile side. The power offset represents the active power value that needs to be reduced on the charging pile side.

[0032] Specifically, after obtaining the instantaneous change rate of the base load and the instantaneous change rate of the total load, the trigger judgment stage begins. A preset change threshold is used to identify a rapid increase in the base load, a preset total load change threshold is used to identify an abnormally rapid increase in the total load, and the transformer safe operation threshold characterizes the upper limit of the transformer's permissible safe operation. These three types of triggering conditions correspond to scenarios such as rigid growth of the base load, a surge in total load due to concentrated startup or resumption of charging at charging piles, and a situation where the current total load has exceeded the safety boundary. This avoids overlooking sudden load surges at the charging pile side when judging solely based on the instantaneous change rate of the base load.

[0033] When the reverse adjustment mechanism is triggered by the instantaneous rate of change of the base load, the predicted increase in base load for the next adjustment cycle is calculated based on the instantaneous rate of change of the base load at the current moment and a preset prediction time window. The preset prediction time window corresponds to the current adjustment cycle or the next control cycle. For example, the predicted increase in base load can be obtained by multiplying the instantaneous rate of change of the base load by the preset prediction time window. The current base load is then added to the predicted increase to obtain the predicted base load for the next adjustment cycle. Finally, the predicted base load is added to the current total charging pile load to obtain the predicted total load for the next adjustment cycle. This path corresponds to the scenario of reduced transformer capacity due to a continuous increase in base load, ensuring that the compensation amount is calculated based on the predicted load for the next adjustment cycle rather than solely on the current load.

[0034] When the reverse adjustment mechanism is triggered by the instantaneous rate of change of total load or the current total load, the predicted increase in total load for the next adjustment cycle is calculated based on the instantaneous rate of change of total load at the current moment and a preset prediction time window. For example, the predicted increase in total load can be obtained by multiplying the instantaneous rate of change of total load by the preset prediction time window. The current transformer total load is then added to the predicted increase in total load to obtain the predicted total load for the next adjustment cycle. This path corresponds to scenarios where rapid and concentrated changes in power on the charging pile side lead to a sudden surge in total load. It is used to avoid inconsistencies between the prediction chain and the trigger source caused by predicting only the change in base load when the total load triggering condition is met.

[0035] When the predicted total load exceeds the transformer's safe operating threshold, the difference between the two is taken as the expected overload; when the predicted total load does not exceed the transformer's safe operating threshold, the expected overload is recorded as 0. The trend risk coefficient characterizes the strength of the load's upward trend and is positively correlated with the absolute value of the corresponding instantaneous rate of change. The margin risk coefficient characterizes the current tightness of the remaining capacity and is negatively correlated with the remaining capacity; both are dimensionless coefficients. Preferably, the trend risk coefficient and margin risk coefficient can be determined according to a preset interval mapping rule. For example, the instantaneous rate of change can be divided into multiple intervals of low, medium, and high rate of change, corresponding to the first trend risk level, the second trend risk level, and the third trend risk level, respectively; the remaining capacity can be divided into multiple capacity intervals of sufficient, tight, and strained, corresponding to the first margin risk level, the second margin risk level, and the third margin risk level, respectively. The trend risk coefficient is selected based on the range of instantaneous rate of change, and the margin risk coefficient is selected based on the range of remaining capacity. Then, a combined correction coefficient is determined based on the trend risk coefficient and the margin risk coefficient using a preset weighted summation or product method. The expected overload is multiplied by the combined correction coefficient to obtain the initial compensation power demand. To avoid abnormally amplified compensation demand, a preset upper limit is set for the combined correction coefficient. When the combined correction coefficient exceeds the preset upper limit, the initial compensation power demand is calculated based on the preset upper limit.

[0036] After obtaining the initial compensation power demand, the set of currently online charging piles that can participate in scheduling is acquired, and the current operating power and minimum allowable operating power of each charging pile are read. The adjustable capacity of each charging pile is calculated and summed to obtain the total adjustable capacity of the current charging pile cluster. The initial compensation power demand is compared with the total adjustable capacity, and the smaller value, which is not less than 0, is taken as the power offset to be compensated on the charging pile side. Thus, the obtained power offset corresponds to both the predicted overload level and the current load increase trend and remaining margin status.

[0037] In another preferred embodiment, the process of performing step S2 may specifically include the following steps: The power offset is corrected based on historical data; When the instantaneous rate of change continuously exceeds the preset change threshold, the corrected power offset is accumulated to form an accumulated compensation requirement, and the accumulated compensation requirement is used as the input for step S3.

[0038] Specifically, after obtaining the power offset corresponding to the current adjustment cycle, power offset correction and cumulative compensation calculations are further performed. Historical load adjustment data is read, which includes at least the power offset setpoint within the historical adjustment cycle, the corresponding instantaneous change rate of the base load and the instantaneous change rate of the total load, the adjustment command issued, the actual power reduction, and the change in total load after adjustment. The actual power reduction can be obtained from the difference between the total load of the charging pile before and after adjustment, and the change in total load after adjustment can be obtained from the difference between the total load before and after adjustment. Based on the deviation between the historical power offset and the actual power reduction, or based on the deviation between the historical power offset and the change in total load after adjustment, a historical compensation deviation sequence is constructed, and the correction amount for the current cycle is determined based on the historical compensation deviation sequence. The correction amount can be determined by the average or weighted average of the historical compensation deviations within the sliding time window, with higher weights assigned to historical compensation deviations closer to the current time.

[0039] The power offset of the current control cycle is superimposed with the correction amount to obtain the corrected power offset. The corrected power offset is then subjected to a limiting process to ensure that it is not less than 0 and not greater than the adjustable total capacity of the online charging pile cluster within the current control cycle.

[0040] When the instantaneous rate of change corresponding to the trigger channel continuously exceeds the corresponding preset threshold for multiple consecutive sampling periods, the accumulated compensation demand retained in the previous effective adjustment period is read, and the corrected power offset of the current control period is added to the accumulated compensation demand of the previous effective adjustment period to obtain the accumulated compensation demand of the current period. To prevent the accumulated compensation demand from increasing indefinitely over multiple consecutive adjustment periods, an upper limit constraint is set on it. The upper limit constraint is taken as the total adjustable capacity of online and dispatchable charging piles in the current period, or as the preset maximum compensation power upper limit. At the same time, a minimum effective compensation threshold can also be set. Only when the corrected power offset is greater than the preset minimum effective compensation amount will it be included in the accumulation, so as to suppress the cumulative drift caused by small fluctuations. When the instantaneous rate of change no longer continuously exceeds the corresponding preset threshold in the current sampling period, the accumulation stops, and the accumulated compensation demand is cleared to zero, or reduced according to the preset release rule.

[0041] In one embodiment, the power offset can first be corrected based on historical adjustment data to obtain the corrected power offset. Then, a trend enhancement coefficient is determined based on the interval of consecutive over-threshold counts, and the corrected power offset is used as the base compensation amount. The corrected power offset is then enhanced based on the trend enhancement coefficient to obtain the cumulative compensation requirement. The trend enhancement coefficient is a dimensionless coefficient not less than 1, and it increases with the number of consecutive over-threshold counts (e.g., when the number of consecutive over-threshold counts is in the first number interval, the trend enhancement coefficient takes the first enhancement level; when the number of consecutive over-threshold counts is in the second number interval higher than the first number interval, the trend enhancement coefficient takes the second enhancement level higher than the first enhancement level; and so on), and is subject to a preset upper limit constraint. Afterwards, the corrected power offset is used as the base compensation amount, and the corrected power offset is enhanced based on the trend enhancement coefficient to obtain the cumulative compensation requirement. To avoid the cumulative compensation requirement exceeding the on-site execution capacity, the cumulative compensation requirement is not greater than the current adjustable total capacity of the online charging pile. After the above processing, S2 outputs the corrected power offset and / or cumulative compensation requirement for S3 to call.

[0042] S3. Input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile.

[0043] In one specific embodiment, the process of performing step S3 may specifically include the following steps: The power offset is used as the input deviation in the proportional-integral-derivative control algorithm. The proportional-integral-derivative (PID) control algorithm calculates the proportional, integral, and derivative terms based on the input deviation, and generates the control output based on the proportional, integral, and derivative terms. Based on the control output, the power adjustment value of each charging pile is determined according to the preset allocation rules to generate the initial power scheduling instruction for each charging pile.

[0044] Specifically, S3 uses the real-time compensation output from S2 as the input deviation and the operating status data of each online charging pile that can participate in scheduling as the allocation constraint to generate preliminary power scheduling instructions for each charging pile. The operating status data includes at least the current output power, minimum allowable operating power, maximum allowable operating power, connection status, charging task status, user constraint status, and a flag indicating whether participation in scheduling is allowed for each charging pile. When S2 adopts the basic implementation method, the dual-threshold hierarchical triggering implementation method, or the predictive correction implementation method, the real-time compensation amount is the power offset; when adopting the continuous over-threshold accumulation implementation method, the real-time compensation amount is the cumulative compensation requirement. S3 uniformly uses the real-time compensation amount as the input deviation of the proportional-integral-derivative control algorithm.

[0045] A discrete control cycle is established, and the instantaneous compensation quantity corresponding to the current control cycle is used as the input deviation of the proportional-integral-derivative (PID) control algorithm. The input deviation represents the difference between the target power reduction of the charging pile cluster in the current control cycle and the actual power reduction in the previous control state. Specifically, the actual power reduction in the previous control state is obtained by the difference between the baseline total power of the charging pile cluster before the scheduling command was issued in the previous control cycle and the currently collected actual total power of the charging pile cluster. The baseline total power of the charging pile cluster is defined as the actual total output power of the target charging pile set currently online and allowed to participate in scheduling at the beginning of the previous control cycle and before the scheduling command was issued, and is refreshed at the beginning of each control cycle. If the current input deviation is less than or equal to 0, the control output of the PID control algorithm is recorded as 0, or the stable output state of the previous control cycle is maintained to avoid power increase adjustment when compensation demand does not exist. The control cycle is the same as or an integer multiple of the sampling cycle in S1; when the control cycle is greater than the sampling cycle, multiple sampling points within a control cycle can be aggregated or averaged before performing PID calculations.

[0046] After the input deviation is fed into the proportional-integral-derivative (PID) control algorithm, the proportional, integral, and derivative terms are calculated separately within the current control cycle. The proportional term is determined based on the current input deviation, reflecting the direct impact of the compensation gap on the control output at the current moment. The integral term is determined based on the cumulative amount of input deviation in the current and historical control cycles, used to eliminate residual deviations during continuous adjustment. The derivative term is determined based on the change in input deviation between the current and previous control cycles, reflecting the impact of the rate of change in compensation demand on the control output. For example, the difference between the current and previous control cycle input deviations can be divided by the control cycle duration to obtain the deviation change rate, which is then combined with the derivative gain to determine the derivative term output. To avoid excessive control output caused by the continuous accumulation of integral quantities when the field adjustability is limited, an integral separation and / or integral limiting mechanism is set for the integral term; the integral term is only updated when the input deviation is within the preset integral action range, and when the output corresponding to the integral term exceeds the preset integral upper limit, it is truncated according to the preset integral upper limit.

[0047] The proportional, integral, and derivative terms are algebraically summed to obtain the original control output for the current control cycle. A limiting process is applied to the original control output to ensure it is neither less than 0 nor greater than the total adjustable capacity of the currently online and dispatchable charging piles. This avoids generating negative adjustment values ​​or cluster control quantities exceeding the field's executable capabilities. The total adjustable capacity of the charging piles is obtained by summing the differences between the current output power and the minimum allowable operating power of each online and dispatchable charging pile.

[0048] The preset allocation rules are used to distribute the total power reduction at the cluster level to each online charging pile that can participate in scheduling. The allocation criteria include at least one or more of the following: the current output power of each charging pile, the adjustable capacity of each charging pile, the adjustment priority, and the user constraint level. During execution, the target charging pile set that can participate in scheduling is first selected from the online charging piles. Then, the adjustable capacity of each charging pile within the target charging pile set is calculated, and the adjustable capacities of each charging pile are summed to obtain the current total allocable capacity. Next, the allocation weight of each charging pile is determined according to the preset allocation rules; for example, the basic allocation weight can be determined according to the proportion of each charging pile's adjustable capacity to the current total allocable capacity, and then adjusted based on the adjustment priority and user constraint level, so that charging piles with lower adjustment priorities and lower user constraints bear a larger share of the power reduction. The control output is then distributed according to the allocation weight of each charging pile to obtain the initial power adjustment value for each charging pile.

[0049] After obtaining the initial power adjustment value for each charging pile, the adjustable capability of each charging pile is verified. For any charging pile, its current output power is subtracted from the corresponding initial power adjustment value to obtain the candidate target power of the charging pile. If the candidate target power is lower than the minimum allowable operating power of the charging pile, the actual power adjustment value of the charging pile is corrected to the difference between its current output power and the minimum allowable operating power. If the charging pile is out of scheduling, in a faulty state, experiencing communication abnormalities, or in a user-locked state, the actual power adjustment value of the charging pile is recorded as 0. After the adjustable capability verification of each charging pile, if the actual allocated reduction amount at the cluster level is less than the control output, the unallocated remaining reduction amount is redistributed among the remaining charging piles with remaining adjustable capacity according to the same preset allocation rules, until the remaining reduction amount is 0, or there are no charging piles to be allocated.

[0050] S4. Based on the total load limit of the distribution area, verify whether the initial power dispatch command exceeds the safety boundary. If so, correct the command amplitude and obtain the optimized power dispatch command.

[0051] In one specific embodiment, the process of performing step S4 may specifically include the following steps: Get the current total load limit of the transformer area; The expected total load after the execution of the preliminary power dispatch command is calculated based on the preliminary power dispatch command. The expected total load is compared with the upper limit of the total load of the distribution area to determine whether the expected total load exceeds the safety boundary. If so, the amplitude of the preliminary power dispatch command is corrected to obtain the corrected command. The power values ​​of each charging pile corresponding to the initial power scheduling command or the revised command that does not exceed the safety boundary are verified to determine whether the power value of each charging pile is within the safe operating range. The instructions that pass the verification are identified as optimized power scheduling instructions.

[0052] Specifically, S4 is used to perform a safety review of the command execution results before the initial power dispatch command is issued, in order to avoid the total load exceeding the limit or the target power of individual charging piles exceeding the limit due to fluctuations in the base load of the transformer area, execution deviations, or the superposition of single-pile operating boundary constraints. The initial power dispatch command includes at least the charging pile identifier, target output power or target reduction power, power adjustment value, command effective time, and control cycle identifier.

[0053] The upper limit of the total load of the transformer area is the boundary value of the total active power allowed for safe operation of the transformer area within the current control cycle. This can be achieved by using the transformer's safe operation threshold or by deducting a preset safety margin from the transformer's rated active capacity. The system reads the current basic load data, the preliminary power dispatch instructions for each charging pile generated in S3, and the actual operating power of the currently online charging piles; the basic load data maintains the same definition as the basic load in S2. If the preliminary power dispatch instruction uses a target output power format, the target output power of each charging pile is extracted and summed to obtain the expected total load of the charging piles after the instruction is executed. If the preliminary power dispatch instruction uses a target reduction power format, the expected output power of each charging pile is obtained by subtracting the corresponding target reduction power from its current operating power, and then summed to obtain the expected total load of the charging piles after the instruction is executed. The expected total load of the charging piles is added to the current basic load data to obtain the expected total load after the instruction is executed.

[0054] When the expected total load is less than or equal to the upper limit of the total load of the distribution area, the initial power dispatch command is determined to have not exceeded the safety boundary at the level of the total load of the distribution area, and the initial power dispatch command is processed as a command pending individual pile verification. When the expected total load is greater than the upper limit of the total load of the distribution area, the initial power dispatch command is determined to have exceeded the safety boundary at the level of the total load of the distribution area, and the additional reduction amount is determined based on the difference between the expected total load and the upper limit of the total load of the distribution area; the additional reduction amount represents the further increase in cluster power reduction required to bring the expected total load back to within the upper limit of the total load of the distribution area.

[0055] While maintaining the relative distribution relationship among charging piles, the total additional reduction is uniformly adjusted and allocated. If the initial power dispatch instruction adopts a target power reduction, the total additional reduction is allocated to each charging pile according to the proportion of the original target power reduction or original power adjustment value, and the corresponding allocation result is added to the original target power reduction or original power adjustment value to obtain the corrected target power reduction or power adjustment value. If the initial power dispatch instruction adopts a target output power, the original reduction value of each charging pile is first obtained based on the difference between the current operating power of each charging pile and the original target output power. Then, the total additional reduction is allocated according to the proportion of the original reduction value of each charging pile, and the corresponding allocation result is subtracted from the original target output power to obtain the corrected target output power. In this way, the expected total load corresponding to the corrected instruction falls back to within the upper limit of the total load of the distribution area, while maintaining the relative distribution relationship formed in S3 basically unchanged.

[0056] The safe operating range is limited at least by the minimum and maximum allowable operating power of each charging pile, and may also be combined with the rated capacity of the charging module, the temperature rise constraint of the charging gun, the allowable receiving power on the vehicle side, and the communication status constraint when necessary. For any charging pile, the corresponding target power is determined according to its verification command; if the target power is less than the minimum allowable operating power of the charging pile, the target power of the charging pile is corrected to the minimum allowable operating power; if the target power is greater than the maximum allowable operating power of the charging pile, the target power of the charging pile is corrected to the maximum allowable operating power; if the charging pile is in a fault, offline, out of scheduling, or user-locked state, the target power of the charging pile is corrected to the currently allowed safe power, or the original power is maintained. If the verification command adopts the form of target power reduction, the above verification is converted into judging whether the reduced target power falls within the safe operating range, and the corresponding target power reduction or power adjustment value is calculated after correction.

[0057] After completing the single-pile safe operating range verification, the target output power, target reduction power, power adjustment value, and expected total load for each charging pile are updated synchronously. If the expected total load after single-pile correction is still greater than the upper limit of the total load of the distribution area, charging piles that still have remaining reduction capacity and are allowed to participate in scheduling are further screened to form a set of remaining adjustable charging piles. Within the set of remaining adjustable charging piles, the undigested remaining additional reduction amount is redistributed according to the preset allocation rule adopted in S3 or a compatible allocation rule. After redistribution, the single-pile safe range verification and expected total load verification are re-executed until the expected total load does not exceed the upper limit of the total load of the distribution area and the target power of each charging pile is within the safe operating range, or there are no charging piles that can be further allocated. If there are no charging piles that can be further allocated, the current verified instruction combination is used as the output, and the undigested remaining additional reduction amount is recorded as execution restriction information for subsequent control cycle updates of the historical adjustment records in S2 and the control deviation in S3.

[0058] In a preferred embodiment, the process of performing step S4 may specifically include the following steps: Obtain the current total load limit of the transformer area, and calculate the expected total load after the execution of the initial power dispatch instruction based on the initial power dispatch instruction; When the total load is expected to exceed the upper limit of the total load of the distribution area, the total reduction target is determined based on the excess portion; The system obtains the current operating power, historical response speed, adjustment priority, and user constraint level of each online charging pile, and constructs the corresponding correction weight for each online charging pile based on the current operating power, historical response speed, adjustment priority, and user constraint level. Based on the correction weight corresponding to each online charging pile, the total reduction target is allocated to obtain the target correction amount for each online charging pile; The initial power scheduling command is differentiated and modified based on the target correction amount to obtain the modified command; The power values ​​of each charging pile corresponding to the corrected instruction are checked for safety range, and the corrected instruction that passes the check is determined as the optimized power scheduling instruction.

[0059] Specifically, this step is used to perform a safety review and differential correction on the execution results after the initial power dispatch command has been formed. This is to avoid situations where the total load of the distribution area still exceeds the upper limit of the total load of the distribution area after the command is executed, or where the resource allocation is reduced in an unreasonable manner, due to fluctuations in the basic load of the distribution area, differences in the response capability of individual piles, differences in execution priorities, and differences in user constraints. The method for obtaining the expected total load is the same as the aforementioned implementation method, and will not be repeated here.

[0060] When the expected total load is greater than the upper limit of the total load of the transformer area, the total reduction target is determined based on the difference between the expected total load and the upper limit of the total load of the transformer area. The total reduction target is characterized by the additional active power reduction required to bring the expected total load back to within the upper limit of the total load of the transformer area. It is not directly used as the correction amount for any single pile, but as the total input for subsequent differentiated allocation.

[0061] The current operating power represents the basic capacity of the charging pile to handle additional adjustments; the adjustment priority is a pre-set dimensionless hierarchical parameter that represents the order of different charging piles in the area scheduling; the user constraint level is a pre-set dimensionless hierarchical parameter that represents the degree of constraint of different charging tasks on power reduction.

[0062] Historical response speed characterizes how quickly the output power of a charging pile changes towards the target power after receiving a command during historical adjustment. In one embodiment, multiple effective control cycles can be selected within a preset historical time window. The ratio of the actual power change to the corresponding response time in each effective control cycle is calculated. After removing abnormal samples, each ratio is weighted and averaged according to the time frame (from most recent to oldest) to obtain the historical response speed of the charging pile. Abnormal sample removal can be achieved by removing samples exceeding a preset deviation threshold or by removing samples exceeding a preset statistical interval. In this way, the historical response speed reflects recent execution characteristics while reducing the impact of abnormal records on weight construction.

[0063] To ensure that different factors can participate in the weight calculation, each factor is first standardized. Specifically, current operating power, historical response speed, and adjustment priority are converted into standardized values ​​positively correlated with the correction weight, while user constraint level is converted into a standardized value negatively correlated with the correction weight. Then, the basic weight value for each charging pile is determined according to preset combination rules. Finally, all basic weight values ​​are normalized so that the sum of the correction weights for each online charging pile is 1. For example, the basic weight value for each charging pile is obtained by multiplying the standardized current operating power, historical response speed, adjustment priority, and user constraint level by their corresponding preset influence coefficients and then summing them. The proportion of each charging pile's basic weight value to the total sum of all basic weight values ​​is then used as the corresponding correction weight. The preset influence coefficients are pre-set according to the area dispatching strategy.

[0064] The total reduction target is allocated according to the correction weights corresponding to each online charging pile to obtain the initial target correction amount for each online charging pile. Then, the initial target correction amount for each online charging pile is compared with the remaining adjustable capacity that can be further reduced for that charging pile. The remaining adjustable capacity is determined by the difference between the target power corresponding to the current command and the minimum allowable operating power for that charging pile. If the initial target correction amount for a charging pile is greater than its remaining adjustable capacity, its remaining adjustable capacity is used as the actual target correction amount for that charging pile, and the excess is recorded as unallocated correction amount. If the initial target correction amount for a charging pile is less than or equal to its remaining adjustable capacity, the initial target correction amount is used as the actual target correction amount for that charging pile. If there is unallocated correction amount, it is redistributed among the remaining online charging piles that still have remaining adjustable capacity according to the updated correction weights, until the total reduction target is fully allocated, or there are no more online charging piles that can be allocated. The target correction amount represents the differential reduction amount that needs to be applied additionally relative to the initial power scheduling command of S3.

[0065] If the initial power scheduling instruction adopts the form of target power reduction, the target correction amount for the corresponding charging pile is added to the original target power reduction to obtain the corrected target power reduction. If the initial power scheduling instruction adopts the form of target output power, the original target output power is subtracted from the corresponding target correction amount to obtain the corrected target output power, and the corresponding power adjustment value is updated synchronously. Then, based on the corrected target power of each charging pile, the expected total load of the charging piles and the expected total load are recalculated, and it is verified whether the corrected expected total load is not greater than the upper limit of the total load of the distribution area. If it is still greater than the upper limit of the total load of the distribution area, and there are still online charging piles with remaining adjustable capacity, the correction weight update and target correction amount redistribution continue to be performed until the upper limit of the total load of the distribution area is met, or there are no online charging piles that can be further allocated. If there are no online charging piles that can be further allocated, the remaining undigested reduction demand is recorded as execution restricted information.

[0066] The safe operating range is limited at least by the minimum and maximum allowable operating power of each charging pile, and may be further constrained by the rated capacity of the charging module, the allowable receiving power on the vehicle side, communication status, and equipment fault status if necessary. For any charging pile, if the corrected target power is lower than the minimum allowable operating power, it is corrected to the minimum allowable operating power; if the corrected target power is higher than the maximum allowable operating power, it is corrected to the maximum allowable operating power; if the charging pile is in a faulty, offline, out-of-scheduling, or user-locked state, its target power is corrected to the currently allowed safe power, or the original power is maintained. After completing the single-pile safe operating range verification, the target reduction power, target output power, power adjustment value, and expected total load for each charging pile are updated synchronously. If, after the single-pile safe operating range verification, the expected total load again exceeds the upper limit of the total load of the distribution area, the remaining demand reduction allocation and differentiated correction will continue among the remaining online charging piles that still have remaining adjustable capacity and are allowed to participate in scheduling, until the expected total load does not exceed the upper limit of the total load of the distribution area and the target power of each charging pile is within the safe operating range, or there are no more online charging piles that can be allocated.

[0067] S5. Send the optimized power scheduling command to each charging pile controller so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution.

[0068] In one specific embodiment, the process of performing step S5 may specifically include the following steps: Extract the power adjustment value corresponding to each charging pile based on the optimized power scheduling command; The power adjustment value corresponding to each charging pile is sent to the corresponding charging pile controller through the communication interface. Each charging pile controller adjusts its output power based on the received power adjustment value to obtain the adjusted charging pile load distribution.

[0069] Specifically, the power adjustment value corresponding to each charging pile is extracted based on the optimized power scheduling command. If the optimized power scheduling command adopts the form of target power reduction, the target power reduction or power adjustment value corresponding to each charging pile is directly extracted; if the optimized power scheduling command adopts the form of target output power, the power adjustment value corresponding to the charging pile is calculated based on the difference between the current operating power and the target output power of the corresponding charging pile. At the same time, the charging pile identifier, command effective time, control cycle identifier, and command validity period are extracted, and the above information is associated with the power adjustment value to form a command issuing unit for the corresponding charging pile controller.

[0070] After a command unit is generated, its validity is verified. This verification includes at least: whether the command's effective time falls within the current control cycle, whether the command's validity period has expired, and whether the power adjustment value or target output power falls within the allowable adjustment range for the corresponding charging pile. If the verification fails, the corresponding command is marked as invalid and an exception record is generated; the command will not be issued to that charging pile.

[0071] The communication interface can be an in-station Ethernet interface, an industrial serial communication interface, a wireless communication interface, a power line carrier communication interface, or an existing control bus interface of the charging system. If message communication is used, the sent message must include at least the charging pile identifier, power adjustment value, target output power or target reduction power, command effective time, control cycle identifier, and message verification field. The message verification field is used to verify message integrity and avoid control parameter errors due to message loss, truncation, or misalignment during communication. To ensure the reliability of command transmission, after the command is sent, the system waits for the corresponding charging pile controller to return a reception confirmation message within a preset feedback time limit. If no reception confirmation message is received, the corresponding command is retransmitted, with the number of retransmissions not exceeding a preset limit. If no reception confirmation message is received after reaching the preset limit, the charging pile is recorded as having an unconfirmed communication status, and the corresponding abnormal status is written into the execution record.

[0072] The charging pile controller first performs legality and matching checks on the received command. These checks include at least verifying the control cycle identifier, the command's validity period, and whether the target power is within the safe operating range of the charging pile. Upon successful verification, power adjustment is executed. If the controller receives a target power reduction, it subtracts the target power reduction from the current charging output power to obtain the target output power, and adjusts the output power to the target output power according to the controller's internal power adjustment logic. If the controller receives the target output power, it directly sets the charging module's output target to the target output power. Considering that charging pile power adjustment is constrained by module start / stop granularity, current adjustment resolution, vehicle-side allowable received power, and internal closed-loop response characteristics, the charging pile controller can convert the target output power into the corresponding output current setting value, module switching quantity, or module allocation power. Within the execution time window from the command's effective time, it gradually approaches the target output power at a preset adjustment slope, avoiding sudden power changes that could cause charging module protection actions, vehicle-side handshake anomalies, or localized impacts on the charging station area.

[0073] After the command takes effect, each charging pile controller reports the actual output power, execution status identifier, collection timestamp, and abnormal status information at the end of the current control cycle or within the next sampling cycle. The execution status identifier indicates whether the charging pile has completed the execution of the corresponding command, and the abnormal status information indicates execution anomalies such as communication timeout, power tracking failure, vehicle refusal to adjust, module failure, or user interruption of charging. The substation edge controller matches and summarizes the actual output power reported by each charging pile according to the charging pile identifier, and combines it with the collection timestamp to form the adjusted charging pile load distribution. The adjusted charging pile load distribution includes at least the adjusted actual output power of each charging pile and the adjusted total load of the charging pile cluster.

[0074] If a charging pile fails to return an execution result within the preset feedback time limit, the charging pile is marked as having an unconfirmed execution status, and a corresponding statistical confidence flag is added to the charging pile. Within the current control cycle, the most recent effective output power of the charging pile can be used as a temporary estimate to participate in the generation of the current charging pile load distribution. However, the temporary estimate is only used for load statistical completion in the current cycle and is not considered a confirmed execution value. When verifying load stability in subsequent steps, the temporary estimate corresponding to the unconfirmed execution status is processed using estimated value processing or weight reduction processing to distinguish the stability judgment from the confirmed execution result. At the same time, the abnormal record and unconfirmed execution flag of the charging pile are retained for subsequent control cycles to update historical response speed, execution deviation, and scheduling reliability.

[0075] S6. Based on the charging pile load distribution and the total load curve generated and scheduled from the basic load, and to verify the stability, the verification results are obtained.

[0076] In one specific embodiment, the process of performing step S6 may specifically include the following steps: The load distribution of charging piles is superimposed onto the base load to generate a total load curve; Calculate the deviation between the total load curve and the preset safety threshold; If the deviation exceeds the preset deviation threshold, it is determined that the total load curve has not stabilized within the preset safety threshold; otherwise, it is determined that the total load curve has stabilized within the preset safety threshold. Verification results are generated based on the above determination results.

[0077] Specifically, the charging pile load distribution includes at least the adjusted actual output power of each charging pile, execution confirmation status, statistical confidence flag, data collection time, and total load of the charging pile cluster. The charging pile load distribution is time-series aligned with the basic load data. Within the verification time window, for each sampling time, the total load of the charging pile cluster corresponding to that sampling time and the latest estimated basic load value for the same sampling time are read and added together to obtain the total load value corresponding to that sampling time. Then, the total load values ​​corresponding to all sampling times are arranged in chronological order to form the scheduled total load curve. The verification time window can be consistent with the current control cycle or cover the current control cycle and several subsequent sampling cycles.

[0078] For any sampling moment within the verification time window, the difference between the total load value at that sampling moment and the preset safety threshold is taken as the original deviation value for that sampling moment; this forms the original deviation sequence within the verification time window. The original deviation value retains the deviation direction: when the total load value is higher than the preset safety threshold, the original deviation value is positive; when the total load value is lower than the preset safety threshold, the original deviation value is negative or zero. Simultaneously, to perform over-limit determination, a corresponding over-limit deviation sequence is generated based on the original deviation sequence; when the original deviation value is greater than 0, it is taken as the over-limit deviation value for that sampling moment; when the original deviation value is less than or equal to 0, the over-limit deviation value for that sampling moment is recorded as 0. Thus, the original deviation sequence is used to characterize the complete deviation state of the total load relative to the safety threshold, and the over-limit deviation sequence is used to characterize the degree to which the total load exceeds the safety threshold.

[0079] To ensure stability verification reaches the executable layer, deviation statistics are further extracted based on the out-of-limit deviation sequence. These statistics include at least the maximum out-of-limit deviation, and if necessary, the duration for which the out-of-limit deviation is continuously greater than a preset deviation threshold. If the maximum out-of-limit deviation is used as the criterion, the maximum value is extracted from all sampling times. If the continuous duration is used as an auxiliary criterion, the duration for which the out-of-limit deviation is continuously greater than the preset deviation threshold is statistically analyzed. All these deviation statistics are based on a dimensional comparison between the total load curve and the preset safety threshold, without altering the physical meaning of the total load curve itself.

[0080] After obtaining the deviation statistics, they are compared with a preset deviation threshold. The preset deviation threshold is the allowable over-limit boundary for determining whether the total load curve is stable within a preset safety threshold. When continuous duration is used as an auxiliary judgment condition, a corresponding duration threshold is also set. Specifically, if the maximum over-limit deviation exceeds the preset deviation threshold, the total load curve is determined to be unstable within the preset safety threshold; if the maximum over-limit deviation does not exceed the preset deviation threshold, the total load curve is determined to be stable within the preset safety threshold. To avoid misjudgment caused by single sampling fluctuations, a composite judgment rule can also be used. That is, when the over-limit deviation exceeds the preset deviation threshold and the corresponding continuous duration exceeds the duration threshold, the total load curve is determined to be unstable within the preset safety threshold; when individual sampling points show over-limit deviations but do not continuously exceed the duration threshold, the total load curve is still determined to be stable within the preset safety threshold. Thus, short-term disturbances and continuous over-limit states can be distinguished.

[0081] During the verification process, if the adjusted charging pile load distribution includes temporary estimates corresponding to unconfirmed execution states, an unconfirmed execution flag or statistical confidence flag is added to that portion of the data. During stability verification, the corresponding sampling points are processed using estimated values ​​or weighted down to distinguish unconfirmed execution data from confirmed execution data. This reduces the distortion of verification conclusions caused by unconfirmed execution results.

[0082] After completing the above determination, a verification result is generated. The verification result includes at least a verification status identifier, a verification time window, deviation statistics, a total load curve identifier, and the original deviation sequence and / or total load curve time series data. The verification status identifier indicates whether the total load curve is stable within the preset safety threshold. The deviation statistics characterize the degree of exceeding the limit after the current control cycle is executed. The original deviation sequence and / or total load curve time series data are used for subsequent steps to extract the number of oscillations, the rate of change of deviation, and the steady-state deviation. If the total load curve is determined to be stable within the preset safety threshold, the verification status identifier is recorded as verification passed, and the corresponding total load curve, original deviation sequence, deviation statistics, and control cycle identifier are written into the operation record. If the total load curve is determined not to be stable within the preset safety threshold, the verification status identifier is recorded as verification failed, and the corresponding over-limit sampling interval, original deviation sequence, deviation statistics, continuous duration, and control cycle identifier are written into the anomaly record.

[0083] In one implementation, if the verification result is that the total load curve is not stable within the preset safety threshold, the area edge controller will write the corresponding verification result back to the historical execution record and increase the execution deviation correction weight or trigger stricter scheduling constraints in the next control cycle; if the verification result is that the total load curve is stable within the preset safety threshold, the execution result of the corresponding control cycle will be recorded as a valid sample and used to update the historical response speed and execution deviation statistics.

[0084] Figure 3 and Figure 4 This diagram illustrates the PID control response and the overall load stabilization effect after scheduling in a specific embodiment of the present invention. Based on simulated data, the diagram illustrates the dynamic response process of the PID control algorithm after the power offset input, and the closed-loop adjustment effect as the total load in the background area of ​​the charging pile power scheduling execution drops from an over-limit state to below the safe threshold.

[0085] Figure 3 The graph illustrates the relationship between power offset and PID control output over time. The horizontal axis represents time, and the vertical axis represents power. The blue curve represents the power offset, indicating the total power reduction required for the charging pile cluster at the current moment; the red curve represents the PID control output, indicating the total power reduction of the charging pile cluster; the gray vertical dashed line indicates the trigger time for reverse adjustment. After the reverse adjustment is triggered at t=20s, the power offset increases sharply and then gradually decreases. The PID control output responds quickly accordingly, exhibiting a slight dynamic deviation in the initial stage of adjustment, which then generally follows the trend of the power offset and gradually converges. This figure demonstrates that the PID control algorithm can dynamically generate power reduction commands for the charging pile cluster based on the input deviation. Figure 4 The figure illustrates the relationship between the total load curve after scheduling and the preset safety threshold. The green curve represents the total load after scheduling, the orange dashed line represents the preset safety threshold, and the shaded area represents the over-limit region. Before the reverse adjustment is triggered at t=20s, the total load is higher than the preset safety threshold, indicating an over-limit region. After the reverse adjustment is triggered, as the total power reduction of the charging pile cluster gradually takes effect, the total load after scheduling quickly falls back below the preset safety threshold and remains within the limit in subsequent periods. This figure illustrates that the present invention generates the scheduling quantity through PID control, forms the total load curve after scheduling after execution by the charging piles, and achieves the effect of the total load falling back to within the safety boundary in stability verification.

[0086] Figure 3 and Figure 4 The processing steps S3 to S6 of the present invention and their effects are illustrated from the perspectives of control response and load stability.

[0087] S7. If the verification results show that the total load curve is not stable within the preset safety threshold, then iteratively update the Kalman filter algorithm parameters and use the updated algorithm for subsequent real-time data processing.

[0088] In one specific embodiment, the process of performing step S7 may specifically include the following steps: When the total load curve is not stable within the preset safety threshold, update the process noise covariance matrix and measurement noise covariance matrix of the Kalman filter algorithm; Adjust the filter gain based on the updated process noise covariance matrix and measurement noise covariance matrix; The updated Kalman filter algorithm is then used for filtering subsequent real-time data sequences.

[0089] Specifically, the verification results include at least the verification status indicator, the total load curve, the deviation statistics, and the over-limit sampling interval; the deviation statistics characterize the degree to which the total load curve exceeds the preset safety threshold, and the over-limit sampling interval is used to characterize the time range in which instability occurs.

[0090] After triggering the parameter update, an updated dataset for parameter correction is constructed. The updated dataset includes at least the original real-time measured load sequence within the verification time window, the estimated load sequence processed by the current Kalman filter algorithm, the charging pile load distribution returned by S5, the total load curve generated by S6, and the corresponding deviation statistics. The original real-time measured load sequence can be the total load measurement sequence of the distribution area, the base load measurement sequence, or both; the estimated load sequence is the filtered output sequence at the same sampling time. Based on the residual between the original real-time measured load sequence and the estimated load sequence, and the degree of deviation of the total load curve from the preset safety threshold, the changing trends of system state fluctuation intensity and measurement noise intensity within the current control cycle are determined. The residual is the difference between the original real-time measured value and the filtered estimated value at the same sampling time.

[0091] The process noise covariance matrix characterizes the intensity of random disturbances in the load state during state transitions, while the measurement noise covariance matrix characterizes the intensity of fluctuations in sampling errors, communication jitter errors, or measurement errors during the measurement process. In one embodiment, the residual mean square value, the rate of change of the original deviation, and the duration of continuous exceedances can be statistically analyzed within a sliding window, and the update amounts of the process noise covariance matrix and the measurement noise covariance matrix can be determined accordingly. For example, when the residual fluctuation increases and the original deviation continues to increase in the same direction, the value of the corresponding state component in the process noise covariance matrix is ​​increased to enhance the filter's ability to track load changes; when the measured value fluctuates significantly in the short term and the residual mean square value increases, the value of the corresponding measurement component in the measurement noise covariance matrix is ​​increased to reduce the impact of abnormal measurements on state estimation; when both enhanced state changes and enhanced measurement disturbances exist simultaneously, the process noise covariance matrix and the measurement noise covariance matrix are jointly adjusted. If the Kalman filter algorithm uses a one-dimensional load state model, the two covariance matrices mentioned above can be degenerated into scalar parameters. If a multi-dimensional state model is used, the matrix elements are updated according to the state dimension and the measurement dimension, respectively. To avoid parameter updates exceeding a reasonable range, upper and lower limits are set for the process noise covariance matrix and measurement noise covariance matrix after each iteration. The update result of the current cycle is then smoothly fused with the effective parameters of the previous cycle according to a preset fusion ratio, and used as the effective parameters for the next cycle.

[0092] The filter gain is used to balance the weights of the predicted state value and the real-time measurement value in the state correction process. In the prediction step of the Kalman filter algorithm, based on the state estimation result of the previous sampling time, the state transition relationship, and the updated process noise covariance matrix, the predicted state value and prediction error covariance at the current sampling time are obtained. In the correction step, a new filter gain is calculated based on the updated prediction error covariance and measurement noise covariance matrix, and the filter gain is used to fuse the real-time measurement value at the current sampling time with the predicted state value to obtain a new state estimate. Thus, when the process noise covariance matrix increases, the filter's assessment of the model prediction uncertainty improves, and the filter gain increases accordingly to improve the tracking of real-time measurement changes; when the measurement noise covariance matrix increases, the filter's assessment of real-time measurement uncertainty improves, and the filter gain decreases accordingly to reduce the impact of abnormal measurement disturbances.

[0093] After adjusting the filter gain, the updated Kalman filter algorithm is used for filtering subsequent real-time data sequences. Subsequent real-time data sequences include at least one of the following: the real-time measurement sequence of the total load of the transformer substation within the new sampling period, the real-time measurement sequence of the base load, or the real-time load sequence of the charging pile cluster.

[0094] In a preferred embodiment, the process of performing step S7 may specifically include the following steps: When the verification results show that the total load curve is not stable within the preset safety threshold, the stability feature quantity corresponding to the total load curve is extracted. The stability feature quantity includes at least one or more of the following: steady-state deviation, deviation change rate, duration of exceeding limit and number of oscillations. The current instability type is determined based on stability characteristics. The instability type includes at least one of the following: instability due to insufficient trend following, instability due to measurement disturbance, and instability due to regulated oscillation. When the instability is determined to be due to insufficient trend tracking, the process noise covariance matrix parameter of the Kalman filter algorithm is increased. When the instability is determined to be dominated by measurement disturbance, the measurement noise covariance matrix parameter of the Kalman filter algorithm is increased; When the instability is determined to be oscillatory, the process noise covariance matrix parameters and the measurement noise covariance matrix parameters are adjusted together, and the filter gain is updated according to the adjusted parameters. The updated Kalman filter algorithm is then used for real-time data sequence filtering in the next adjustment cycle.

[0095] Specifically, when extracting stability features, the aforementioned features are calculated based on the total load curve and the original deviation sequence within the verification time window. Steady-state deviation characterizes the degree of continuous deviation of the total load curve from the preset safety threshold within the instability zone, and can be obtained using the average, final, or weighted average of the original deviation values ​​at each sampling time within the instability zone. Deviation change rate characterizes how quickly the deviation changes over time, and can be calculated based on the ratio of the difference in original deviation values ​​between adjacent sampling times to the sampling interval. Exceeding limit duration characterizes the duration for which the total load curve is continuously higher than the preset safety threshold. Oscillation count characterizes the number of times the total load curve repeatedly crosses above and below the preset safety threshold or repeatedly oscillates near the target load, and can be obtained by statistically analyzing the number of sign changes in the original deviation sequence or the number of local peak-valley alternations. To avoid distortion of features due to instantaneous noise, local statistics can be performed within a sliding window consistent with the current control cycle.

[0096] In one embodiment, when the steady-state deviation continuously exceeds a preset deviation threshold, the duration of the out-of-limit exceeds a preset duration threshold, the rate of change of deviation continuously changes in the same direction, and the number of oscillations is lower than a preset number threshold, it is determined to be a trend-following instability type. This type of instability indicates that the current filter responds too slowly to changes in load status, resulting in a lag in the estimation of base load or total load. When the rate of change of deviation changes frequently abruptly within a short period of time, the dispersion of the residual sequence increases, the duration of a single out-of-limit is short, and the number of oscillations does not reach the high-frequency oscillation condition, it is determined to be a measurement disturbance-dominated instability type. This type of instability indicates that sampling noise, communication jitter, or metering spikes have increased interference with the filtering results. When the total load curve repeatedly exceeds and falls back near a preset safety threshold, the number of oscillations exceeds a preset oscillation threshold, the absolute value of the steady-state deviation is at a low level, and the rate of change of deviation alternates between positive and negative, it is determined to be a regulation oscillation type instability. This type of instability indicates that there is repeated correction and high-frequency oscillation between the filter output and the scheduling execution feedback. In the above determinations, the stability characteristics are only used to identify the type of instability and determine the direction of parameter updates.

[0097] When the instability is determined to be of the insufficient trend-following type, the process noise covariance matrix parameter is increased to improve the filter's ability to track rapid load changes. In one embodiment, the increment of the process noise covariance matrix parameter can be determined based on the steady-state deviation and the duration of exceeding the limit. For example, a correction amount related to the mean steady-state deviation and the duration of exceeding the limit can be superimposed on the current parameter, thereby reducing the overly strong constraint of the prediction model on state changes and improving the response speed to actual load abrupt changes.

[0098] When the instability is determined to be dominated by measurement disturbance, the measurement noise covariance matrix parameter is increased to reduce the impact of measurement noise on the load trend estimation results. In one embodiment, the increment of the measurement noise covariance matrix parameter can be determined based on the residual mean square value, the degree of abrupt change in the rate of change of deviation, and the density of abnormal sampling points. For example, the residual mean square value between the original real-time measurement value and the filtered estimate is statistically analyzed within a sliding window formed by the most recent sampling points, and mapped to the update amount of the measurement noise covariance matrix to reduce the filter's confidence in abnormal measurement values.

[0099] When an oscillating instability is identified, the process noise covariance matrix parameters and the measurement noise covariance matrix parameters are jointly adjusted, and the filter gain is updated based on the adjusted parameters to suppress high-frequency fluctuations in subsequent real-time data processing. Specifically, the process noise covariance matrix parameters can be adjusted based on the number of oscillations and the degree of alternating positive and negative changes in the deviation rate of change, while the measurement noise covariance matrix parameters can be adjusted based on the dispersion of the residual sequence and the frequency of abnormal fluctuations. After adjustment, in the prediction step of the Kalman filter algorithm, the predicted state value and prediction error covariance are obtained based on the updated process noise covariance matrix. In the correction step, the filter gain is recalculated based on the updated prediction error covariance and the measurement noise covariance matrix, and the updated filter gain is used to fuse the real-time measurement value with the predicted state value to obtain a new state estimate. This reduces high-frequency back-and-forth corrections while maintaining necessary tracking capabilities.

[0100] To ensure the feasibility of parameter updates, upper and lower limits are set for the process noise covariance matrix parameters and measurement noise covariance matrix parameters after each update. The update results of the current cycle are then smoothly fused with the effective parameters of the previous cycle according to a preset fusion ratio, and used as the effective parameters for the next cycle. If the Kalman filter algorithm uses a one-dimensional load state model, the process noise covariance matrix and measurement noise covariance matrix can be degenerated into scalar parameters. If a multi-dimensional state model is used, the diagonal elements of the matrix are updated according to the state dimension and the measurement dimension, and related elements are updated as necessary to avoid excessive parameter mutations caused by a single abnormal cycle, which could lead to filter divergence or estimation distortion.

[0101] In one implementation, if the same instability type is determined in multiple consecutive adjustment cycles, the parameter correction amount in the corresponding direction is allowed to accumulate periodically within the upper and lower limits of the parameters. If the verification result in a subsequent adjustment cycle shows that the total load curve is stable within the preset safety threshold, the currently updated parameters are maintained, and the parameters, filtering results, and verification results of the corresponding adjustment cycle are written into the historical execution record for subsequent parameter initialization or historical response characteristic correction.

[0102] The above describes a method for dynamic power scheduling of charging piles in an embodiment of this application. The following describes a system for dynamic power scheduling of charging piles in an embodiment of this application. Please refer to [link / reference]. Figure 5 This application provides a schematic diagram of a dynamic power scheduling system for charging piles, which includes: The data acquisition and filtering module is used to acquire the total load of the transformer, obtain a real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend.

[0103] The trigger compensation module is used to calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, it triggers reverse adjustment to determine the power offset that needs to be compensated on the charging pile side.

[0104] The instruction generation module is used to input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile.

[0105] The safety verification module is used to verify whether the initial power dispatch command exceeds the safety boundary based on the total load limit of the distribution area. If so, the command amplitude is corrected to obtain an optimized power dispatch command.

[0106] The instruction execution module is used to send the optimized power scheduling instructions to each charging pile controller, so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution.

[0107] The verification and evaluation module is used to generate the total load curve after scheduling based on the charging pile load distribution and basic load, and to perform stability verification to obtain the verification results.

[0108] The parameter update module is used to iteratively update the Kalman filter algorithm parameters when the verification results show that the total load curve has not stabilized within the preset safety threshold, and then use the updated algorithm for subsequent real-time data processing.

[0109] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for dynamic power scheduling of charging piles, characterized in that, The method includes: S1. Collect the total load of the transformer, obtain the real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend. S2. Calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, trigger reverse adjustment to determine the power offset that needs to be compensated on the charging pile side. S3. Input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile. S4. Based on the total load limit of the distribution area, verify whether the initial power dispatch command exceeds the safety boundary. If so, correct the command amplitude and obtain the optimized power dispatch command. S5. Send the optimized power scheduling command to each charging pile controller so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution. S6. Based on the charging pile load distribution and the basic load, generate the total load curve after scheduling and perform stability verification to obtain the verification results; S7. If the verification results show that the total load curve is not stable within the preset safety threshold, then iteratively update the Kalman filter algorithm parameters and use the updated algorithm for subsequent real-time data processing.

2. The method according to claim 1, characterized in that, S1 includes: The total load of the transformer is monitored in real time by sensors, and load data is collected at a preset frequency to form a real-time data sequence that includes the base load and the charging pile load. The Kalman filter algorithm is applied to the real-time data sequence for state estimation. The actual load value is used as the state variable. The prior estimate is calculated through the prediction step, and the posterior estimate is obtained by fusing the measured value and the prior estimate through the update step. The smoothed load fluctuation trend is generated based on the posterior estimation.

3. The method according to claim 1, characterized in that, S2 include: Extract time series points from the load fluctuation trend, calculate the difference between the base load at the current moment and the previous moment, divide by the time interval, and obtain the instantaneous rate of change at the current moment; Determine whether the instantaneous rate of change exceeds a preset threshold; if so, activate the reverse adjustment mechanism. Based on the portion of the current total load that exceeds the base load, the power offset that needs to be compensated on the charging pile side is determined. The power offset represents the power value that needs to be reduced on the charging pile side.

4. The method according to claim 3, characterized in that, S2 also includes: The power offset is corrected based on historical data; When the instantaneous rate of change continuously exceeds the preset change threshold, the corrected power offset is accumulated to form an accumulated compensation requirement, and the accumulated compensation requirement is used as the input for step S3.

5. The method according to claim 1, characterized in that, S3 include: The power offset is used as the input deviation in the proportional-integral-derivative control algorithm. The proportional-integral-derivative (PID) control algorithm calculates the proportional, integral, and derivative terms based on the input deviation, and generates the control output based on the proportional, integral, and derivative terms. Based on the control output, the power adjustment value of each charging pile is determined according to the preset allocation rules to generate the initial power scheduling instruction for each charging pile.

6. The method according to claim 1, characterized in that, S4 includes: Get the current total load limit of the transformer area; The expected total load after the execution of the preliminary power dispatch command is calculated based on the preliminary power dispatch command. The expected total load is compared with the upper limit of the total load of the distribution area to determine whether the expected total load exceeds the safety boundary. If so, the amplitude of the preliminary power dispatch command is corrected to obtain the corrected command. The power values ​​of each charging pile corresponding to the initial power scheduling command or the revised command that does not exceed the safety boundary are verified to determine whether the power value of each charging pile is within the safe operating range. The instructions that pass the verification are identified as optimized power scheduling instructions.

7. The method according to claim 1, characterized in that, S5 include: Extract the power adjustment value corresponding to each charging pile based on the optimized power scheduling command; The power adjustment value corresponding to each charging pile is sent to the corresponding charging pile controller through the communication interface. Each charging pile controller adjusts its output power based on the received power adjustment value to obtain the adjusted charging pile load distribution.

8. The method according to claim 1, characterized in that, S6 include: The load distribution of charging piles is superimposed onto the base load to generate a total load curve; Calculate the deviation between the total load curve and the preset safety threshold; If the deviation exceeds the preset deviation threshold, it is determined that the total load curve has not stabilized within the preset safety threshold; otherwise, it is determined that the total load curve has stabilized within the preset safety threshold. Verification results are generated based on the above determination results.

9. The method according to claim 1, characterized in that, S7 includes: When the total load curve is not stable within the preset safety threshold, update the process noise covariance matrix and measurement noise covariance matrix of the Kalman filter algorithm; Adjust the filter gain based on the updated process noise covariance matrix and measurement noise covariance matrix; The updated Kalman filter algorithm is then used for filtering subsequent real-time data sequences.

10. A system for dynamic power scheduling of charging piles, used to implement the method as described in any one of claims 1 to 9, characterized in that, The system includes: The acquisition and filtering module is used to acquire the total load of the transformer, obtain a real-time data sequence including the base load and the charging pile load, and use the Kalman filter algorithm to remove noise and generate a smoothed load fluctuation trend. The trigger compensation module is used to calculate the instantaneous change rate of the base load based on the load fluctuation trend. If it exceeds the preset change threshold, it triggers reverse adjustment to determine the power offset that needs to be compensated on the charging pile side. The instruction generation module is used to input the power offset into the proportional-integral-derivative control algorithm to generate preliminary power scheduling instructions for each charging pile. The safety verification module is used to verify whether the initial power dispatch command exceeds the safety boundary based on the total load limit of the distribution area. If so, the command amplitude is corrected to obtain the optimized power dispatch command. The instruction execution module is used to send the optimized power scheduling instruction to each charging pile controller so that each charging pile controller can adjust the output power in real time and obtain the adjusted charging pile load distribution. The verification and evaluation module is used to generate the total load curve after scheduling based on the charging pile load distribution and basic load, and to perform stability verification to obtain the verification results. The parameter update module is used to iteratively update the Kalman filter algorithm parameters when the verification results show that the total load curve has not stabilized within the preset safety threshold, and then use the updated algorithm for subsequent real-time data processing.

Images