A charging pile power control method and system

Abstract

A charging pile power control method and system relate to the technical field of charging pile power control. A power interaction prediction model and a power grid load prediction model are constructed based on deep learning to obtain an estimated power interaction size time series sequence of each charging pile in the current interaction cycle and an estimated power grid load power time series sequence of the current interaction cycle of a charging station; a preset time scale is used to divide the current interaction cycle into a number of time slices, and a state division is performed on the several time slices, and a power planning operation of the current interaction cycle of the charging station is performed according to the state division result; the real-time power grid load power time series sequence and the real-time power interaction size time series sequence in each time slice of the current interaction cycle are monitored in real time, and model updates, rapid response power control of a number of charging piles or monitoring of energy storage battery groups of charging stations are performed based on real-time monitoring results to ensure the accuracy and real-time performance of charging pile power control.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile power regulation, and in particular to a charging pile power regulation method and system. Background Art

[0002] The prior art CN108075536A "Flexible charging control method and charging pile system for charging piles" uses self-organizing network communication between charging piles to monitor the charging power of charging piles in real time, automatically control the charging power of the pile body, and make the sum of the power values ​​of charging piles in the area less than the rated power of the transformer, so as to achieve orderly charging and flexible control, and reduce the impact of electric vehicle charging behavior on the distribution network;

[0003] The prior art CN106712166A "A method and system for automatic power control of a charging station" includes: S1, real-time monitoring of the operating status of the charging piles. If a charging pile starts charging, execute S2; if a charging pile stops charging, execute S3; S2, calculate the total charging power of the charging piles that are already charging and starting charging. If the total charging power is within the control power threshold, allocate the rated power of the corresponding gun to the charging pile that starts charging and go to S1, otherwise pre-allocate the charging power of all the charging piles that are already charging to the charging piles that are already charging and starting charging, adjust the pre-allocated power according to the minimum power of the guns of each charging pile and the total available rated power, and then output it, and go to S1; S3, pre-allocate the charging power of the charging pile that stops charging to the remaining charging piles that are charging, and adjust the pre-allocated power according to the rated power of the guns of the charging piles that are charging, and then output it, and go to S1.

[0004] As an important means of low-carbon transportation, electric vehicles have become the focus of attention of domestic and foreign scientific research institutions due to their many advantages such as high efficiency and low pollution. The charging of a few electric vehicles will not have much impact on the power grid, but large-scale access to the power grid will bring challenges to the normal operation and control of the power grid. The access of a large number of electric vehicles to the power grid will lead to an increase in load, an increase in the peak-to-valley difference in the power grid load, an increase in the difficulty of power grid peak regulation, an increase in the burden on the power system, and a serious impact on the safety and reliability of the power grid. Summary of the invention

[0005] In order to solve the above technical problems, the object of the present invention is to provide a charging pile power control method, comprising the following steps:

[0006] Step s1: construct a power interaction prediction model and a grid load prediction model based on deep learning, obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated grid load power time series sequence of the charging station in the current interaction cycle;

[0007] Step s2: preset a time scale, divide the current interaction cycle into a number of time slices, perform state division on the time slices, and perform power planning operation of the current interaction cycle of the charging station according to the state division result;

[0008] Step s3: Monitor the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model update, fast response power regulation of several charging piles, or monitoring of charging station energy storage battery groups based on the real-time monitoring results.

[0009] Furthermore, the process of constructing a power interaction prediction model and a power grid load prediction model based on deep learning to obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated power grid load power time series sequence of the charging station in the current interaction cycle includes:

[0010] Construct a parallel linked list of historical data, the length of which is s, obtain the power interaction records of several historical interaction cycles of each charging pile in the charging station in the target area and the power time series of the power load, the power interaction record includes the power interaction time and power interaction size between the charging pile and the new energy vehicle, obtain the start timestamp of each historical interaction cycle, extract the time span between the start timestamp of each historical interaction cycle and the current time, extract the power interaction records of the first s historical interaction cycles with the smallest time span and the power time series of the power load of the power grid, and enter them into the parallel linked list of historical data in sequence;

[0011] Based on deep learning, a power interaction prediction model and a power grid load prediction model are constructed. Training data is obtained through the power interaction records contained in the parallel linked list of historical data. The day-ahead power interaction prediction model and the power grid load prediction model are trained using the training data. The estimated power interaction size time series sequence of each charging pile in the current interaction cycle is obtained by completing the trained day-ahead power interaction prediction model. The estimated power grid load power time series sequence in the current interaction cycle is obtained by completing the trained power grid load prediction model.

[0012] Furthermore, the process of presetting a time scale, dividing the current interaction cycle into a number of time slices, dividing the states of the time slices, and performing a power planning operation of the current interaction cycle of the charging station according to the state division result includes:

[0013] A preset time scale is used to divide the current interaction cycle into several time slices of the same time scale, and the full load value of the grid load power is obtained. The estimated grid load power time series sequence and the estimated power interaction size time series sequence in each time slice are compared with the grid load power full load value. If there is a moment in the time slice when the sum of the estimated grid load power time series sequence and the estimated power interaction size series sequence is greater than the grid load power full load value, the time slice is marked as an overload state time slice, and the power planning operation of the current interaction cycle of the charging station is performed. If the sum of the estimated grid load power time series sequence and the estimated power interaction size series sequence in the time slice is less than or equal to the grid load power full load value, the time slice is marked as a surplus state time slice.

[0014] Furthermore, the process of performing the power planning operation of the current interaction cycle of the charging station includes:

[0015] Determine whether the charging station is equipped with an energy storage battery pack. If the charging station is equipped with an energy storage battery pack, obtain the storage capacity of the energy storage battery pack, obtain the cumulative time when the sum of the estimated grid load power time series and the estimated power interaction size time series of all overload state time slices is greater than the full load value of the grid load power, and obtain the overload capacity according to the cumulative time;

[0016] Determine whether the overload power is greater than the stored power, and if the overload power is greater than the stored power, obtain the power difference between the overload power and the stored power, select a surplus state time slice, and perform power control on the surplus state time slice;

[0017] If the charging station is not equipped with a storage battery pack, power regulation is performed on the overload state time slice.

[0018] Furthermore, the process of selecting the headroom state time slice and performing power control on the headroom state time slice includes:

[0019] Respectively obtain the time spans of the start timestamp of each overload state time slice and the start timestamp of the surplus state time slice and the start timestamp of the current interaction cycle, obtain the overload state time slice with the largest time span, filter out several surplus state time slices whose time span is smaller than the time span of the overload state time slice, and mark the several surplus state time slices as the first surplus state time slices;

[0020] According to the estimated grid load power time series sequence, the estimated power interaction size time series sequence and the power difference in several first margin state time slices, the predetermined charge and discharge power time series sequence of the energy storage battery group in each first margin state time slice is obtained, and the charge and discharge power upper limit of the energy storage battery group in each first margin state time slice is obtained, and it is judged whether there is a predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence of the energy storage battery group. If so, the predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence is replaced with the charge and discharge power upper limit, and then the predetermined power interaction size time series sequence in several first margin state time slices is obtained according to the predetermined charge and discharge power time series sequence, the estimated grid load power time series sequence and the power difference.

[0021] Furthermore, the process of power regulation for the overload state time slice includes:

[0022] According to the estimated grid load power time series sequence in a number of overload state time slices and the grid load power full load value, a predetermined power interaction size time series sequence in the overload state time slice is obtained.

[0023] Furthermore, the process of real-time monitoring of the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle, and performing model update, rapid response power regulation of several charging piles, or monitoring of the energy storage battery group of the charging station based on the real-time monitoring results includes:

[0024] Obtain the real-time grid load power time series sequence and the real-time power interaction size time series sequence within each time slice of the current interaction cycle, compare the real-time grid load power time series sequence and the real-time power interaction size series sequence with the estimated grid load power time series sequence and the estimated power interaction size series sequence respectively, obtain the grid load power similarity and the power interaction size similarity, preset the grid load power similarity threshold and the power interaction size similarity threshold, if the grid load power similarity is less than the grid load power similarity threshold or the power interaction size similarity is less than the power interaction size similarity threshold, then update the power interaction prediction model or the grid load prediction model in real time, and perform rapid response power regulation of several charging piles;

[0025] If the grid load power similarity is less than or equal to the grid load power similarity threshold and the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the charging station energy storage battery group is monitored.

[0026] Furthermore, the process of updating the power interaction prediction model or the power grid load prediction model in real time includes:

[0027] If the power grid load power similarity is less than the power grid load power similarity threshold, the real-time power grid load power time series sequence collected in the current interaction period is inserted into the first position of the historical data parallel linked list, and the power grid load power time series sequence of the last historical interaction period in the historical data parallel linked list is removed. Then, the training data is obtained through the historical data parallel linked list, and the power grid load prediction model is trained using the training data to obtain a trained power grid load prediction model;

[0028] Similarly, if the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the trained power interaction prediction model is obtained.

[0029] Furthermore, the process of performing rapid response power control of a plurality of charging piles includes:

[0030] Obtain the adjusted scheduled power interaction size time series sequence of the charging pile according to the real-time grid load power time series sequence, the scheduled charging and discharging power, and the grid load power full load value within the time slice;

[0031] The power similarity of the grid loads and the power interaction size similarity within the time slice are used as evaluation indicators, the indicator weight matrix and the time scale level of the evaluation indicators are set, and the membership matrix of the evaluation indicators for different time scale levels is obtained through fuzzy comprehensive evaluation;

[0032] The time scale level of the evaluation index is obtained through the membership matrix and the index weight matrix, and the predetermined power interaction size time series sequences in other time slices with a time span greater than the time slice are eliminated. The other time slices are divided into secondary time scales according to the time scale level to generate a number of identical secondary time slices. The several time slices are predicted according to the power interaction prediction model or the power grid load prediction model that has been updated in real time to obtain the estimated power grid load power time series sequence and the estimated power interaction size time series sequence in the several secondary time slices. According to the estimated power grid load power time series sequence and the estimated power interaction size series sequence, the predetermined power interaction size series sequence in the several secondary time slices is obtained to complete the rapid response power regulation of the time slices.

[0033] A charging pile power control system includes a monitoring center, wherein the monitoring center is communicatively connected to a data mining module, a power planning module and a real-time monitoring module;

[0034] The data mining module is used to construct a power interaction prediction model and a power grid load prediction model based on deep learning, and obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated power grid load power time series sequence of the charging station in the current interaction cycle;

[0035] The power planning module is used to preset a time scale, divide the current interaction cycle into a number of time slices, perform state division on the time slices, and perform power planning operations for the current interaction cycle of the charging station according to the state division results;

[0036] The real-time monitoring module is used to monitor the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model updates, rapid response power regulation of several charging piles, or monitoring of charging station energy storage battery groups based on the real-time monitoring results.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. Build a power interaction prediction model and a grid load prediction model based on deep learning to obtain the estimated power interaction size time series of each charging pile in the current interaction cycle and the estimated grid load power time series of the charging station in the current interaction cycle. By predicting the changing trend of the grid load and the power interaction between each charging pile and new energy vehicles in the future, ensure that the charging pile adjusts the charging strategy in advance to cope with the expected changes in grid load and power interaction.

[0039] 2. Monitor the real-time grid load power time series and real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model updates, rapid response power regulation of several charging piles, or monitoring of energy storage battery groups at charging stations based on real-time monitoring results. Utilize the characteristics that the prediction accuracy of the power interaction prediction model and the grid load prediction model increases as the time scale decreases. In the real-time monitoring link, a time scale adjustment mechanism can be introduced in combination with ultra-short-term power generation prediction to continuously revise the predetermined power interaction size formulated according to the power interaction prediction model and the grid load prediction model to ensure the accuracy and real-time performance of the power regulation of the charging piles. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a schematic diagram of a charging pile power control method according to an embodiment of the present application.

[0041] Figure 2 This is a schematic diagram of a charging pile power control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0043] like Figure 1 As shown, a charging pile power control method includes the following steps:

[0044] Step s1: construct a power interaction prediction model and a grid load prediction model based on deep learning, obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated grid load power time series sequence of the charging station in the current interaction cycle;

[0045] Step s2: preset a time scale, divide the current interaction cycle into a number of time slices, perform state division on the time slices, and perform power planning operation of the current interaction cycle of the charging station according to the state division result;

[0046] Step s3: Monitor the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model update, fast response power regulation of several charging piles, or monitoring of charging station energy storage battery groups based on the real-time monitoring results.

[0047] It should be further explained that, in the specific implementation process, the process of constructing a power interaction prediction model and a power grid load prediction model based on deep learning, and obtaining the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated power grid load power time series sequence of the charging station in the current interaction cycle includes:

[0048] Construct a parallel linked list of historical data, the length of which is s, obtain power interaction records and power grid load power time series of several historical interaction cycles of each charging pile in the charging station in the target area, the power interaction record includes the power interaction time and power interaction size between the charging pile and the new energy vehicle, obtain the start timestamp of each historical interaction cycle, extract the time span between the start timestamp of each historical interaction cycle and the current time, extract the power interaction records and power grid load power time series of the first s historical interaction cycles with the smallest time span, and enter them into the parallel linked list of historical data in sequence, one of the parallel linked lists of historical data is used to store the power interaction records of the historical interaction cycle, and the other is used to store the power grid load power time series of the same historical interaction cycle, the first bits of the two linked lists are used to store the power interaction records of the historical interaction cycle with the smallest time span and the power grid load power time series of the historical interaction cycle with the smallest time span, and the last bits are used to store the power interaction records of the historical interaction cycle with the largest time span and the power grid load power time series of the historical interaction cycle with the largest time span;

[0049] A power interaction prediction model and a power grid load prediction model are constructed based on deep learning. Training data is obtained through the power interaction records contained in the historical data parallel linked list. The day-ahead power interaction prediction model and the power grid load prediction model are trained using the training data. The training data of the day-ahead power interaction prediction model are the power interaction moments and power interaction sizes between charging piles and new energy vehicles within several historical interaction cycles contained in the historical data parallel linked list. The training data of the power grid load prediction model are the power time series sequences of the power grid load within several historical interaction cycles contained in the historical data parallel linked list. The estimated power interaction size time series sequence of each charging pile within the current interaction cycle is obtained by completing the trained day-ahead power interaction prediction model. The estimated power grid load power time series sequence within the current interaction cycle is obtained by completing the trained power grid load prediction model.

[0050] It should be further explained that, in the specific implementation process, the time scale is preset, the current interaction cycle is divided into a number of time slices, the state of the time slices is divided, and the power planning operation of the current interaction cycle of the charging station is performed according to the state division result. The process includes:

[0051] A preset time scale is used to divide the current interaction cycle into several time slices of the same time scale, and the full load value of the grid load power is obtained. The estimated grid load power time series sequence and the estimated power interaction size time series sequence in each time slice are compared with the grid load power full load value. If there is a moment in the time slice when the sum of the estimated grid load power time series sequence and the estimated power interaction size series sequence is greater than the grid load power full load value, the time slice is marked as an overload state time slice, and the power planning operation of the current interaction cycle of the charging station is performed. If the sum of the estimated grid load power time series sequence and the estimated power interaction size series sequence in the time slice is less than or equal to the grid load power full load value, the time slice is marked as a surplus state time slice.

[0052] It should be further explained that, in the specific implementation process, the process of performing the power planning operation of the current interaction cycle of the charging station includes:

[0053] Determine whether the charging station is equipped with an energy storage battery pack. If the charging station is equipped with an energy storage battery pack, obtain the storage power of the energy storage battery pack, obtain the cumulative time when the sum of the estimated grid load power time series and the estimated power interaction size time series of all overload state time slices is greater than the full load value of the grid load power, and obtain the overload power according to the cumulative time. The calculation formula of the overload power is:

[0054] ;

[0055] in, represents the estimated grid load power at time t, represents the estimated power interaction size at time t, Indicates the cumulative moment when the sum of the estimated grid load power time series and the estimated power interaction size time series of all overload state time slices is greater than the full load value of the grid load power, Indicates the full load value of the power grid. Indicates overload power;

[0056] Determine whether the overload power is greater than the stored power, and if the overload power is greater than the stored power, obtain the power difference between the overload power and the stored power, select a surplus state time slice, and perform power control on the surplus state time slice;

[0057] If the charging station is not equipped with a storage battery pack, power regulation is performed on the overload state time slice.

[0058] It should be further explained that, in a specific implementation process, the process of selecting a headroom state time slice and performing power control on the headroom state time slice includes:

[0059] Respectively obtain the time spans of the start timestamp of each overload state time slice and the start timestamp of the surplus state time slice and the start timestamp of the current interaction cycle, obtain the overload state time slice with the largest time span, filter out several surplus state time slices whose time span is smaller than the time span of the overload state time slice, and mark the several surplus state time slices as the first surplus state time slices;

[0060] According to the estimated grid load power time series sequence, the estimated power interaction size time series sequence and the power difference in several first margin state time slices, the predetermined charge and discharge power time series sequence of the energy storage battery group in each first margin state time slice is obtained, and the charge and discharge power upper limit of the energy storage battery group in each first margin state time slice is obtained, and it is judged whether there is a predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence of the energy storage battery group. If so, the predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence is replaced with the charge and discharge power upper limit, and then the predetermined power interaction size time series sequence in several first margin state time slices is obtained according to the predetermined charge and discharge power time series sequence, the estimated grid load power time series sequence and the power difference.

[0061] It should be further explained that, in the specific implementation process, the power constraint formula between the grid load power, the power interaction size, the power difference, and the charging and discharging power of the energy storage battery group is:

[0062] (Power interaction size + power difference power + charge and discharge power) = (grid load power full load value - grid load power);

[0063] Electricity difference power = electricity difference / total time of a number of first margin state time slices;

[0064] It should be further explained that, in the specific implementation process, the process of power regulation for the overload state time slice includes:

[0065] According to the estimated grid load power time series sequence and the grid load power full load value in several overload state time slices, the predetermined power interaction size time series sequence in the overload state time slice is obtained. The constraint formula between the grid load power, the power interaction size and the grid load power full load value is: grid load power + power interaction size = grid load power full load value.

[0066] It should be further explained that, in the specific implementation process, the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle are monitored in real time, and the model is updated based on the real-time monitoring results, and the process of rapid response power regulation of several charging piles or monitoring of the energy storage battery group of the charging station includes:

[0067] Obtain the real-time grid load power time series sequence and the real-time power interaction size time series sequence within each time slice of the current interaction cycle, compare the real-time grid load power time series sequence and the real-time power interaction size series sequence with the estimated grid load power time series sequence and the estimated power interaction size series sequence respectively, obtain the grid load power similarity and the power interaction size similarity, preset the grid load power similarity threshold and the power interaction size similarity threshold, if the grid load power similarity is less than the grid load power similarity threshold or the power interaction size similarity is less than the power interaction size similarity threshold, then update the power interaction prediction model or the grid load prediction model in real time, and perform rapid response power regulation of several charging piles;

[0068] If the grid load power similarity is less than or equal to the grid load power similarity threshold and the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the charging station energy storage battery group is monitored.

[0069] It should be further explained that, in the specific implementation process, the calculation formula for obtaining the power similarity of the power grid load and the power interaction size similarity is:

[0070] ;

[0071] in, represents the estimated grid load power at time t, represents the estimated average power grid load power within the time slice, represents the real-time grid load power at time t, represents the average real-time grid load power within the time slice, n2 represents the total time of the time slice, and q represents the grid load power similarity;

[0072] ;

[0073] in, represents the estimated power interaction size at time t, represents the average value of the estimated power interaction size within the time slice, represents the real-time power interaction size at time t, represents the average value of real-time power interaction within the time slice, n3 represents the total moment of the time slice, and r represents the power similarity of the power grid load;

[0074] It should be further explained that, in the specific implementation process, the process of real-time updating of the power interaction prediction model or the power grid load prediction model includes:

[0075] If the power grid load power similarity is less than the power grid load power similarity threshold, the real-time power grid load power time series sequence collected in the current interaction period is inserted into the first position of the historical data parallel linked list, and the power grid load power time series sequence of the last historical interaction period in the historical data parallel linked list is removed. Then, the training data is obtained through the historical data parallel linked list, and the power grid load prediction model is trained using the training data to obtain a trained power grid load prediction model;

[0076] Similarly, if the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the trained power interaction prediction model is obtained.

[0077] It should be further explained that, in the specific implementation process, the process of rapid response power control of several charging piles includes:

[0078] According to the real-time grid load power time series, the scheduled charging and discharging power and the grid load power full load value in the time slice, the scheduled power interaction size time series of the charging pile adjusted in the current time slice is obtained, wherein the scheduled charging and discharging power + the scheduled power interaction size = the grid load power full load value - the real-time grid load power;

[0079] Then, the power similarity of the grid loads and the power interaction size similarity within the time slice are used as evaluation indicators, the indicator weight matrix and the time scale level of the evaluation indicators are set, and the membership matrix of the evaluation indicators for different time scale levels is obtained through fuzzy comprehensive evaluation;

[0080] The time scale level of the evaluation index is obtained through the membership matrix and the index weight matrix, and the predetermined power interaction size time series sequences in other time slices with a time span greater than the time slice are eliminated. The other time slices are divided into secondary time scales according to the time scale level to generate a number of identical secondary time slices. The several time slices are predicted according to the power interaction prediction model or the power grid load prediction model that has been updated in real time to obtain the estimated power grid load power time series sequence and the estimated power interaction size series sequence in the several secondary time slices. Then, the power planning operation is performed in the secondary time slice of the current interaction cycle of the charging station. According to the estimated power grid load power time series sequence and the estimated power interaction size series sequence, the predetermined power interaction size series sequence in the several secondary time slices is obtained to complete the rapid response power regulation of the time slice.

[0081] It should be further explained that, in the specific implementation process, the process of obtaining the time scale level of the evaluation index according to the membership matrix and the index weight matrix includes:

[0082] The indicator weight matrix and the membership matrix of the evaluation indicator are integrated through a formula to obtain a fuzzy comprehensive evaluation matrix of the evaluation indicator, the membership of the evaluation indicator for different time scale levels is obtained according to the fuzzy comprehensive evaluation matrix, the time scale level with the highest membership corresponding to the evaluation indicator is screened out, and the time scale level with the highest membership corresponding to the evaluation indicator is used as the time scale level of the evaluation indicator;

[0083] Wherein, the formula is:

[0084] ;

[0085] in, is the fuzzy comprehensive evaluation matrix of the evaluation index, is the indicator weight matrix of the evaluation indicator, is the membership matrix, " represents the multiplication of the weight matrix of the evaluation index and the elements at the corresponding positions of the membership matrix, It is a weighted parameter for controlling the balance between the weight matrix and the membership matrix in the fuzzy comprehensive evaluation matrix of the evaluation index.

[0086] It should be further explained that, in the specific implementation process, the process of monitoring the energy storage battery group of the charging station includes:

[0087] Obtain a time slice in which a predetermined charge and discharge power timing sequence exists in the current interaction cycle, and at the same time collect the real-time charge and discharge power timing sequence in the time slice, compare the deviation rate of the real-time charge and discharge power timing sequence with the predetermined charge and discharge power timing sequence, obtain the charge and discharge power deviation rate, set the deviation rate threshold, and generate abnormal alarm information for the energy storage battery pack if the charge and discharge power deviation rate is greater than the deviation rate threshold.

[0088] like Figure 2 As shown, a charging pile power control system includes a monitoring center, and the monitoring center is communicatively connected to a data mining module, a power planning module and a real-time monitoring module;

[0089] The data mining module is used to construct a power interaction prediction model and a power grid load prediction model based on deep learning, and obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated power grid load power time series sequence of the charging station in the current interaction cycle;

[0090] The power planning module is used to preset a time scale, divide the current interaction cycle into a number of time slices, perform state division on the time slices, and perform power planning operations for the current interaction cycle of the charging station according to the state division results;

[0091] The real-time monitoring module is used to monitor the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model updates, rapid response power regulation of several charging piles, or monitoring of charging station energy storage battery groups based on the real-time monitoring results.

[0092] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A charging pile power control method, characterized in that: The following steps are involved: Step s1: construct a power interaction prediction model and a grid load prediction model based on deep learning, obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated grid load power time series sequence of the charging station in the current interaction cycle; Step s2: preset a time scale, divide the current interaction cycle into a number of time slices, divide the states of the time slices, and perform power planning operations for the current interaction cycle of the charging station according to the state division results, including: Preset the time scale, divide the current interaction cycle into several time slices of the same time scale, obtain the full load value of the power grid load, compare the estimated power grid load power time series and the estimated power interaction size time series in each time slice with the full load value of the power grid load. If there is a moment in the time slice when the sum of the estimated power grid load power time series and the estimated power interaction size time series is greater than the full load value of the power grid load, then mark the time slice as an overload state time slice, and perform the power planning operation of the current interaction cycle of the charging station. If the sum of the estimated power grid load power time series and the estimated power interaction size time series in the time slice is less than or equal to the full load value of the power grid load, then mark the time slice as a surplus state time slice. The process of performing power planning operation for the current interaction cycle of the charging station includes: Determine whether the charging station is equipped with an energy storage battery pack. If the charging station is equipped with an energy storage battery pack, obtain the storage capacity of the energy storage battery pack, obtain the cumulative time when the sum of the estimated grid load power time series and the estimated power interaction size time series of all overload state time slices is greater than the full load value of the grid load power, and obtain the overload capacity according to the cumulative time; Determine whether the overload power is greater than the stored power, and if the overload power is greater than the stored power, obtain the power difference between the overload power and the stored power, select a surplus state time slice, and perform power control on the surplus state time slice; If the charging station is not equipped with an energy storage battery pack, power regulation is performed on the overload state time slice; The process of selecting a headroom state time slice and performing power control on the headroom state time slice includes: Respectively obtain the time spans of the start timestamp of each overload state time slice and the start timestamp of the surplus state time slice and the start timestamp of the current interaction cycle, obtain the overload state time slice with the largest time span, filter out several surplus state time slices whose time span is smaller than the time span of the overload state time slice, and mark the several surplus state time slices as the first surplus state time slices; According to the estimated grid load power time series sequence, the estimated power interaction size time series sequence and the power difference in the first margin state time slices, the predetermined charge and discharge power time series sequence of the energy storage battery group in each first margin state time slice is obtained, and the charge and discharge power upper limit of the energy storage battery group in each first margin state time slice is obtained, and it is judged whether there is a predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence of the energy storage battery group, and if so, the predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence is replaced with the charge and discharge power upper limit, and then the predetermined power interaction size time series sequence in the first margin state time slices is obtained according to the predetermined charge and discharge power time series sequence, the estimated grid load power time series sequence and the power difference; The process of power control for the overload state time slice includes: According to the estimated grid load power time series sequence in a number of overload state time slices and the grid load power full load value, a predetermined power interaction size time series sequence in the overload state time slice is obtained; Step s3: Real-time monitoring of the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle, and model update, rapid response power regulation of several charging piles or monitoring of charging station energy storage battery groups based on the real-time monitoring results, including: Obtain the real-time grid load power time series sequence and the real-time power interaction size time series sequence within each time slice of the current interaction cycle, compare the real-time grid load power time series sequence and the real-time power interaction size series sequence with the estimated grid load power time series sequence and the estimated power interaction size series sequence respectively, obtain the grid load power similarity and the power interaction size similarity, preset the grid load power similarity threshold and the power interaction size similarity threshold, if the grid load power similarity is less than the grid load power similarity threshold or the power interaction size similarity is less than the power interaction size similarity threshold, then update the power interaction prediction model or the grid load prediction model in real time, and perform rapid response power regulation of several charging piles; If the grid load power similarity is less than or equal to the grid load power similarity threshold and the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the charging station energy storage battery group is monitored; The process of updating the power interaction prediction model or the power grid load prediction model in real time includes: If the power grid load power similarity is less than the power grid load power similarity threshold, the real-time power grid load power time series sequence collected in the current interaction period is inserted into the first position of the historical data parallel linked list, and the power grid load power time series sequence of the last historical interaction period in the historical data parallel linked list is removed. Then, the training data is obtained through the historical data parallel linked list, and the power grid load prediction model is trained using the training data to obtain a trained power grid load prediction model; Similarly, if the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the trained power interaction prediction model is obtained; The process of performing fast response power control of several charging piles includes: Obtain the adjusted scheduled power interaction size time series sequence of the charging pile according to the real-time grid load power time series sequence, the scheduled charging and discharging power, and the grid load power full load value within the time slice; The power similarity of the grid loads and the power interaction size similarity within the time slice are used as evaluation indicators, the indicator weight matrix and the time scale level of the evaluation indicators are set, and the membership matrix of the evaluation indicators for different time scale levels is obtained through fuzzy comprehensive evaluation; The time scale level of the evaluation index is obtained through the membership matrix and the index weight matrix, and the predetermined power interaction size time series sequences in other time slices with a time span greater than the time slice are eliminated. The other time slices are divided into secondary time scales according to the time scale level to generate a number of identical secondary time slices. The several time slices are predicted according to the power interaction prediction model or the power grid load prediction model that has been updated in real time to obtain the estimated power grid load power time series sequence and the estimated power interaction size time series sequence in the several secondary time slices. According to the estimated power grid load power time series sequence and the estimated power interaction size series sequence, the predetermined power interaction size series sequence in the several secondary time slices is obtained to complete the rapid response power regulation of the time slices.

2. A charging pile power control method according to claim 1, characterized in that: The process of building a power interaction prediction model and a grid load prediction model based on deep learning to obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated grid load power time series sequence of the charging station in the current interaction cycle includes: Construct a parallel linked list of historical data, the length of which is s, obtain the power interaction records of several historical interaction cycles of each charging pile in the charging station in the target area and the power time series of the power load, the power interaction record includes the power interaction time and power interaction size between the charging pile and the new energy vehicle, obtain the start timestamp of each historical interaction cycle, extract the time span between the start timestamp of each historical interaction cycle and the current time, extract the power interaction records of the first s historical interaction cycles with the smallest time span and the power time series of the power load of the power grid, and enter them into the parallel linked list of historical data in sequence; Based on deep learning, a power interaction prediction model and a power grid load prediction model are constructed. Training data is obtained through the power interaction records contained in the parallel linked list of historical data. The day-ahead power interaction prediction model and the power grid load prediction model are trained using the training data. The estimated power interaction size time series sequence of each charging pile in the current interaction cycle is obtained by completing the trained day-ahead power interaction prediction model. The estimated power grid load power time series sequence in the current interaction cycle is obtained by completing the trained power grid load prediction model.

3. A charging pile power control system, specifically applied to a charging pile power control method according to any one of claims 1 to 2, characterized in that: It includes a monitoring center, which is communicatively connected to a data mining module, a power planning module and a real-time monitoring module; The data mining module is used to construct a power interaction prediction model and a power grid load prediction model based on deep learning, and obtain the estimated power interaction size time series sequence of each charging pile in the current interaction cycle and the estimated power grid load power time series sequence of the charging station in the current interaction cycle; The power planning module is used to preset a time scale, divide the current interaction cycle into a number of time slices, perform state division on the time slices, and perform power planning operations for the current interaction cycle of the charging station according to the state division results, including: Preset the time scale, divide the current interaction cycle into several time slices of the same time scale, obtain the full load value of the power grid load, compare the estimated power grid load power time series and the estimated power interaction size time series in each time slice with the full load value of the power grid load. If there is a moment in the time slice when the sum of the estimated power grid load power time series and the estimated power interaction size time series is greater than the full load value of the power grid load, then mark the time slice as an overload state time slice, and perform the power planning operation of the current interaction cycle of the charging station. If the sum of the estimated power grid load power time series and the estimated power interaction size time series in the time slice is less than or equal to the full load value of the power grid load, then mark the time slice as a surplus state time slice. The process of performing power planning operation for the current interaction cycle of the charging station includes: Determine whether the charging station is equipped with an energy storage battery pack. If the charging station is equipped with an energy storage battery pack, obtain the storage capacity of the energy storage battery pack, obtain the cumulative time when the sum of the estimated grid load power time series and the estimated power interaction size time series of all overload state time slices is greater than the full load value of the grid load power, and obtain the overload capacity according to the cumulative time; Determine whether the overload power is greater than the stored power, and if the overload power is greater than the stored power, obtain the power difference between the overload power and the stored power, select a surplus state time slice, and perform power control on the surplus state time slice; If the charging station is not equipped with an energy storage battery pack, power regulation is performed on the overload state time slice; The process of selecting a headroom state time slice and performing power control on the headroom state time slice includes: Respectively obtain the time spans of the start timestamp of each overload state time slice and the start timestamp of the surplus state time slice and the start timestamp of the current interaction cycle, obtain the overload state time slice with the largest time span, filter out several surplus state time slices whose time span is smaller than the time span of the overload state time slice, and mark the several surplus state time slices as the first surplus state time slices; According to the estimated grid load power time series sequence, the estimated power interaction size time series sequence and the power difference in the first margin state time slices, the predetermined charge and discharge power time series sequence of the energy storage battery group in each first margin state time slice is obtained, and the charge and discharge power upper limit of the energy storage battery group in each first margin state time slice is obtained, and it is judged whether there is a predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence of the energy storage battery group, and if so, the predetermined charge and discharge power greater than the charge and discharge power upper limit in the predetermined charge and discharge power time series sequence is replaced with the charge and discharge power upper limit, and then the predetermined power interaction size time series sequence in the first margin state time slices is obtained according to the predetermined charge and discharge power time series sequence, the estimated grid load power time series sequence and the power difference; The process of power control for the overload state time slice includes: According to the estimated grid load power time series sequence in a number of overload state time slices and the grid load power full load value, a predetermined power interaction size time series sequence in the overload state time slice is obtained; The real-time monitoring module is used to monitor the real-time grid load power time series and the real-time power interaction size time series in each time slice of the current interaction cycle in real time, and perform model update, rapid response power regulation of several charging piles or monitoring of charging station energy storage battery packs based on the real-time monitoring results, including: Obtain the real-time grid load power time series sequence and the real-time power interaction size time series sequence within each time slice of the current interaction cycle, compare the real-time grid load power time series sequence and the real-time power interaction size series sequence with the estimated grid load power time series sequence and the estimated power interaction size series sequence respectively, obtain the grid load power similarity and the power interaction size similarity, preset the grid load power similarity threshold and the power interaction size similarity threshold, if the grid load power similarity is less than the grid load power similarity threshold or the power interaction size similarity is less than the power interaction size similarity threshold, then update the power interaction prediction model or the grid load prediction model in real time, and perform rapid response power regulation of several charging piles; If the grid load power similarity is less than or equal to the grid load power similarity threshold and the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the charging station energy storage battery group is monitored; The process of updating the power interaction prediction model or the power grid load prediction model in real time includes: If the power grid load power similarity is less than the power grid load power similarity threshold, the real-time power grid load power time series sequence collected in the current interaction period is inserted into the first position of the historical data parallel linked list, and the power grid load power time series sequence of the last historical interaction period in the historical data parallel linked list is removed. Then, the training data is obtained through the historical data parallel linked list, and the power grid load prediction model is trained using the training data to obtain a trained power grid load prediction model; Similarly, if the power interaction size similarity is less than or equal to the power interaction size similarity threshold, the trained power interaction prediction model is obtained; The process of performing fast response power control of several charging piles includes: Obtain the adjusted scheduled power interaction size time series sequence of the charging pile according to the real-time grid load power time series sequence, the scheduled charging and discharging power, and the grid load power full load value within the time slice; The power similarity of the grid loads and the power interaction size similarity within the time slice are used as evaluation indicators, the indicator weight matrix and the time scale level of the evaluation indicators are set, and the membership matrix of the evaluation indicators for different time scale levels is obtained through fuzzy comprehensive evaluation; The time scale level of the evaluation index is obtained through the membership matrix and the index weight matrix, and the predetermined power interaction size time series sequences in other time slices with a time span greater than the time slice are eliminated. The other time slices are divided into secondary time scales according to the time scale level to generate a number of identical secondary time slices. The several time slices are predicted according to the power interaction prediction model or the power grid load prediction model that has been updated in real time to obtain the estimated power grid load power time series sequence and the estimated power interaction size time series sequence in the several secondary time slices. According to the estimated power grid load power time series sequence and the estimated power interaction size series sequence, the predetermined power interaction size series sequence in the several secondary time slices is obtained to complete the rapid response power regulation of the time slices.

Images