Energy storage type intelligent micro-grid power management method for frequency modulation and peak regulation

Abstract

The invention relates to an energy storage type intelligent micro-grid power management method for frequency modulation and peak regulation, and relates to the technical field of power grid connection, and the method comprises the steps: obtaining the user demand power consumption of a first node, the user demand power consumption of a second node until the user demand power consumption of an Nth node; obtaining first transformer area large power grid pre-distribution output power time sequence information, inputting the information into a power grid frequency trend prediction model, and outputting a first power grid frequency trend prediction result; when the frequency deviates from the reference power grid frequency, configuring the first node energy storage type intelligent micro-grid to be merged into the output power, and configuring the second node energy storage type intelligent micro-grid to be merged into the output power until the Mth node energy storage type intelligent micro-grid to be merged into the output power; obtaining a second power grid frequency trend prediction result; and when the reference power grid frequency is not deviated, executing power grid-connected management. The technical problems that in the prior art, due to the fact that the power grid frequency is monitored in real time for frequency modulation and peak regulation control, time delay is large, and power grid frequency changes cannot be responded in time are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid connection, and in particular to an energy storage type smart microgrid power management method for frequency and peak regulation. Background Art

[0002] In modern power systems, smart microgrids, as a new type of power supply and management method, have attracted widespread attention due to their flexibility and reliability. Smart microgrids can achieve self-control, protection and management, and can be connected to the external power grid or operate in an isolated island when the external power grid is unavailable.

[0003] After the smart microgrid is connected to the large power grid, the grid frequency and peak value can be monitored in real time, thereby realizing adaptive frequency and peak regulation. However, the traditional smart microgrid frequency and peak regulation strategy relies on real-time monitoring of grid frequency deviation. This strategy has time delay and cannot respond to changes in grid frequency in time, resulting in unsatisfactory frequency and peak regulation effects, affecting the stability of the grid and the quality of power supply. Summary of the invention

[0004] The present invention aims to solve the technical problem that in the prior art, due to the use of real-time monitoring of the grid frequency for frequency and peak regulation control, there is a large time delay and it is impossible to respond to the changes in the grid frequency in time. The present invention provides an energy storage type smart microgrid power management method for frequency and peak regulation to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: The present invention provides an energy storage type smart microgrid power management method for frequency and peak regulation, comprising: traversing the user distribution nodes of the first substation area to perform inertial power consumption analysis, obtaining the user demand power consumption of the first node, the user demand power consumption of the second node, and the user demand power consumption of the Nth node; obtaining the pre-allocated output power timing information of the large power grid of the first substation area; inputting the pre-allocated output power timing information of the large power grid of the first substation area, the user demand power consumption of the first node, the user demand power consumption of the second node, and the user demand power consumption of the Nth node into a grid frequency trend prediction model, and outputting the first grid frequency trend prediction result; when the first grid frequency trend prediction result deviates from the reference grid frequency, configuring the first substation area's energy storage type smart microgrid to be incorporated into the node's first node energy storage type smart microgrid incorporated output power, the second node energy storage type smart microgrid incorporated output power, and the Mth node energy storage type smart microgrid incorporated output power. point energy storage type smart microgrid is integrated into the output power; the pre-allocated output power timing information of the first substation large power grid, the first node energy storage type smart microgrid is integrated into the output power, the second node energy storage type smart microgrid is integrated into the output power until the Mth node energy storage type smart microgrid is integrated into the output power, and the first node user demand power consumption, the second node user demand power consumption until the Nth node user demand power consumption are input into the power grid frequency trend prediction model, and the second power grid frequency trend prediction result is output; when the second power grid frequency trend prediction result does not deviate from the reference power grid frequency, the power grid connection management of the energy storage type smart microgrid integration node is performed according to the first node energy storage type smart microgrid is integrated into the output power, the second node energy storage type smart microgrid is integrated into the output power until the Mth node energy storage type smart microgrid is integrated into the output power.

[0006] The beneficial effects of the present invention are as follows: by traversing user distribution nodes to perform inertial power consumption analysis, the power demand on the user side can be predicted in advance; by combining the large power grid's pre-allocated output power timing information and the user-side demand power consumption, the power grid frequency trend prediction model is used to predict the power grid frequency trend in advance and reduce latency; when the prediction result shows that the power grid frequency will deviate from the reference frequency, the energy storage type smart microgrid is timely configured to incorporate the output power, realize rapid adjustment of the power grid frequency, and improve the timeliness of frequency and peak regulation, thereby achieving the technical effect of reducing the frequency and peak regulation delay of the smart microgrid. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic diagram of a flow chart of a method for power management of an energy storage type smart microgrid for frequency and peak regulation provided by the present invention; Figure 2 A schematic diagram of the structure of an electronic device provided by the present invention; Figure 3 A schematic diagram of the structure of a computer-readable storage medium provided by the present invention.

[0008] In the accompanying drawings, the components represented by the reference numerals are described as follows: Electronic device 500 , memory 510 , processor 520 , first computer program 511 , computer-readable storage medium 600 , second computer program 611 . DETAILED DESCRIPTION

[0009] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0010] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0011] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0012] Embodiment 1: Figure 1 As shown, an embodiment of the present invention provides a method for managing power of an energy storage type smart microgrid for frequency and peak regulation, comprising the steps of: S10: traverse the user distribution nodes of the first substation to perform inertial power consumption analysis, and obtain the power consumption required by the user at the first node, the power consumption required by the user at the second node, and so on until the power consumption required by the user at the Nth node; Further, the user distribution nodes of the first substation are traversed to perform inertial power consumption analysis, and the power consumption required by the user of the first node, the power consumption required by the user of the second node, and the power consumption required by the user of the Nth node are obtained. Step S10 includes the following steps: S11: Obtain a first node user of the user distribution node of the first station area, wherein the first node user has a user type label and a user area label; S12: Obtaining time zone portrait information of the inertial power consumption analysis time zone, wherein the time zone portrait information includes a month portrait label, a week portrait label, and a clock portrait label; S13: performing inertial sample fitting according to the user type label, the user area label, the month portrait label, the week portrait label, and the clock portrait label to obtain the power consumption required by the user at the first node; S13: traverse the second node users until the Nth node user and perform inertial power consumption analysis respectively to obtain the power consumption required by the second node users until the Nth node user.

[0013] Further, inertial sample fitting is performed according to the user type label, the user area label, the month portrait label, the week portrait label and the clock portrait label to obtain the power consumption required by the user at the first node. Step S13 includes the following steps: S131: constructing a fixed background condition based on the user type label, the user area label, and the month portrait label; S132: configuring a binary background condition for the week portrait label, wherein the binary background condition is Monday to Friday or Saturday to Sunday; S133: configuring a dynamic background condition for the clock portrait tag based on a time deviation threshold; S134: performing primary normal sample collection according to the fixed background condition, the binary background condition and the dynamic background condition to obtain a primary electricity consumption sample, wherein the primary electricity consumption sample includes a primary sample user type label, a primary sample user area label, a primary sample month portrait label, a primary sample week portrait label and a primary sample clock portrait label; S135: Collecting secondary normal samples according to the primary sample user type label, the primary sample user area label, the primary sample month portrait label, the primary sample week portrait label, and the primary sample clock portrait label to obtain secondary electricity consumption samples; S136: Perform sample fitting on the secondary power consumption sample and the primary power consumption sample to obtain the power consumption required by the user of the first node.

[0014] Furthermore, sample fitting is performed on the secondary power consumption sample and the primary power consumption sample to obtain the power consumption required by the user of the first node. Step S136 includes the following steps: S1361: Grouping the second-level power usage samples according to the first-level power usage samples to obtain multiple groups of power usage samples; S1362: traversing the plurality of groups of electricity consumption samples to perform a centralized trend assessment, and obtaining a plurality of center demand electricity consumptions; S1363: Perform a central trend assessment on the multiple center demand power consumption and the first-level power consumption samples to obtain the first node user demand power consumption.

[0015] S20: Obtaining the timing information of the pre-allocated output power of the large power grid in the first substation area; Specifically, the first substation refers to the power generation substation, which is an area where one or more transformers supply power, usually including the range where one or more transformers supply power to users; the user distribution node refers to the physical location or connection point of the user's electrical equipment in the power system; the user demand power consumption of the first node, the user demand power consumption of the second node, and the user demand power consumption of the Nth node refer to the user demand power consumption in the future time zone obtained after inertial power consumption analysis of each user at the user distribution node of the first substation. The future time zone is the future time zone preset by the user for which energy storage-type smart microgrid power management is required. N is an integer representing the total number of user nodes, and N≥1.

[0016] Furthermore, the pre-allocated output power timing information of the large power grid in the first substation area refers to the output power timing information preset by the large power grid in the first substation area in the future time zone.

[0017] By combining the electricity demand forecast on the user side and the pre-allocated output power timing information on the grid side, the actual trend of the grid frequency can be predicted more accurately, thus providing a scientific basis for the grid-connected management of energy storage-type smart microgrids and ensuring the stable operation of the grid and the quality of power supply.

[0018] Furthermore, the first node user refers to any user of the user distribution node. The process of inertial power consumption analysis is described by taking the first node user as an example: The user type label and the user area label are two attribute identifiers used to describe the characteristics of the user; the user type label includes the industry category to which the user belongs, while the user area label refers to the physical space occupied by the user; the inertial power consumption analysis time zone refers to a specific time period used to analyze the user's power consumption habits, which is customized by the power generation management end, i.e. the aforementioned future time zone; the time zone portrait information represents the time characteristic identifier of the inertial power consumption analysis time zone, including a month portrait label representing the month of the inertial power consumption analysis time zone distribution, a week portrait label representing the day of the week of the inertial power consumption analysis time zone distribution, and a clock portrait label representing the time of the inertial power consumption analysis time zone distribution. For example, if the inertial power consumption analysis time zone is from 1:00 to 5:00 a.m. every Thursday in March, then the month portrait label is March, the week portrait label is March Thursday, and the clock portrait label is from 1:00 to 5:00 a.m.

[0019] Using the user type label, the user area label, the month portrait label, the week portrait label and the clock portrait label as constraints, collect the power consumption record data of users with the same constraints, and perform mode statistics on the user power consumption record data to obtain the power consumption demand of the first node user; use the same method to process the second node user up to the Nth node user to obtain the power consumption demand of the second node users up to the Nth node user.

[0020] By collecting the user type label, the user area label, the month portrait label, the week portrait label and the clock portrait label, multi-factor constraint conditions are constructed to ensure that the extracted user electricity consumption record data is highly integrated with the first node user, thereby ensuring the accuracy of the analysis of demand electricity consumption.

[0021] Furthermore, the detailed steps of sampling with the user type label, the user area label, the month portrait label, the week portrait label and the clock portrait label as constraint conditions are as follows: Based on the user type label, the user area label, and the month portrait label, a fixed background condition is constructed, and the fixed background condition represents a constraint condition that cannot have the slightest deviation; a binary background condition is configured for the week portrait label, wherein the binary background condition is Monday to Friday or Saturday to Sunday. Preferably, if the week portrait label belongs to Monday to Friday, it is marked as 1, and if it belongs to Saturday to Sunday, it is marked as 0. The binary background condition is used to convert the feature values ​​into 0 and 1. If the values ​​are the same, the binary background condition is met, and if the values ​​are different, the binary background condition is not met; based on the time deviation threshold, the clock The portrait tag configures the dynamic background conditions. In simple terms, based on the clock portrait tag, the forward time deviation threshold and the backward time deviation threshold are used to obtain a new time zone. Samples that meet the time zone meet the dynamic background conditions. For example, if the inertial power consumption analysis time zone is from 1:00 to 5:00 in the morning every Thursday in March, and the time deviation threshold is 20 minutes and hours, then the dynamic background conditions configured by the time deviation threshold are from 0:40 to 5:20 in the morning. If the collection time zone of the new sample user is from 0:40 to 5:20 in the morning, it is considered to meet the dynamic background conditions. If it does not belong to 0:40 to 5:20 in the morning, it is considered to not meet the dynamic background conditions.

[0022] The first-level electricity consumption sample refers to the electricity consumption record data of users without abnormal electricity consumption that simultaneously meets the fixed background conditions, the binary background conditions and the dynamic background conditions; the first-level sample user type label, the first-level sample user area label, the first-level sample month portrait label, the first-level sample week portrait label and the first-level sample clock portrait label refer to the constraint condition characteristics of the first-level electricity consumption sample; the first-level sample user type label, the first-level sample user area label, the first-level sample month portrait label, the first-level sample week portrait label and the first-level sample clock portrait label are used to construct the first-level sample fixed background conditions, the first-level sample binary background conditions and the first-level sample dynamic background conditions, and then the second-level electricity consumption samples that simultaneously meet the first-level sample fixed background conditions, the first-level sample binary background conditions and the first-level sample dynamic background conditions are collected.

[0023] Furthermore, the power consumption demanded by the first-node user refers to the characteristic data obtained by performing sample fitting on the power consumption of the secondary power consumption samples and the primary power consumption samples. Since the primary power consumption samples that meet the requirements may have insufficient data volume, in order to ensure that the collected sample data are representative, the embodiment of the present application continues to sample based on the primary power consumption samples to obtain secondary power consumption samples and expand the amount of data collected. It should be particularly noted that in order to avoid the sample data deviating from the first-node user, only secondary sampling is allowed to take into account the advantages of both data accuracy and increased data volume.

[0024] The detailed process of performing sample fitting on the secondary power usage sample and the primary power usage sample proposed in the embodiment of the present application is as follows: First, since any one sample of the first-level electricity consumption samples and multiple samples of the second-level electricity consumption samples correspond to each other, the second-level electricity consumption samples are grouped according to the first-level electricity consumption samples to obtain multiple groups of electricity consumption samples; then, the electricity consumption record data of each group of electricity consumption samples are subjected to a central trend analysis to calculate multiple center demand electricity consumptions; further, a central trend evaluation is performed on the multiple center demand electricity consumptions and the first-level electricity consumption samples to obtain the first node user demand electricity consumption.

[0025] Preferably, the calculation steps of the central trend analysis mentioned in the embodiment of the present application are illustrated by taking any set of electricity consumption record data as an example: Calculate the pairwise power consumption deviation modulus of power consumption record data and store them as several power consumption deviation modulus values; based on several power consumption deviation modulus values, filter a preset number of adjacent power consumption record data from near to far for each power consumption record data, and construct a characteristic neighborhood for each power consumption record data, and each power consumption record data is the central demand power consumption of the characteristic neighborhood; calculate the inverse of the mean value of the deviation modulus between the power consumption record data and the central demand power consumption in each characteristic neighborhood, and set it as the density parameter of the central demand power consumption of each characteristic neighborhood; extract the central demand power consumption with the maximum value of the density parameter, and set it as the selected central demand power consumption of the corresponding group for output. By default, the preset number is set to 0.25 times the total number of power consumption record data. The more representative characteristic data can be extracted through the central tendency evaluation, avoiding the interference of accidental data, and providing a data basis for the subsequent accurate analysis of the power grid frequency.

[0026] S30: input the first substation area large power grid pre-allocated output power timing information, the first node user demand power consumption, the second node user demand power consumption until the Nth node user demand power consumption into the power grid frequency trend prediction model, and output the first power grid frequency trend prediction result; Furthermore, the grid frequency trend prediction model includes a large grid input channel in the substation area, an energy storage type smart microgrid input channel, a user demand input channel and a grid frequency trend output channel, and inputs the first substation area large grid pre-allocated output power timing information, the first node user demand power consumption, the second node user demand power consumption until the Nth node user demand power consumption into the grid frequency trend prediction model, and outputs the first grid frequency trend prediction result, including: Input the pre-allocated output power timing information of the first large power grid in the substation area into the input channel of the large power grid in the substation area, configure the input of the energy storage type smart microgrid input channel to 0, input the power demand of the first node user, the power demand of the second node user until the power demand of the Nth node user into the user demand input channel, and obtain the first power grid frequency trend prediction result output by the power grid frequency trend output channel; The step of constructing the power grid frequency trend prediction model includes: Obtaining a first area power grid topology, wherein the first area power grid topology includes an area large power grid output power incorporation position, an energy storage type smart microgrid power incorporation position, and a user side incorporation position; Based on the first substation area power grid topology, a graph neural network topology is constructed, wherein the graph neural network topology is the same as the first substation area power grid topology structure, the output power incorporation position of the substation area large power grid, the power incorporation position of the energy storage type smart microgrid and the user side incorporation position are configured as input nodes, and the input nodes are fully connected to be set as output nodes of the graph neural network topology, wherein the input nodes are long short-term memory neural networks; Collecting historical power management data of the first substation, wherein the historical power management data of the first substation includes the timing record information of the pre-allocated output power of the large power grid in the first substation, the output power recorded by the first node energy storage type smart microgrid, the output power recorded by the second node energy storage type smart microgrid until the output power recorded by the M-th node energy storage type smart microgrid, and the power consumption demand recorded by the first node user, the power consumption demand recorded by the second node user until the power consumption demand recorded by the N-th node user, and the timing identification information of the power grid frequency; The grid frequency timing identification information is used as the supervision of the output node, the timing record information of the pre-allocated output power of the first substation large power grid, the record of the output power incorporated by the first node energy storage type smart microgrid, the record of the output power incorporated by the second node energy storage type smart microgrid until the record of the output power incorporated by the Mth node energy storage type smart microgrid, and the record of the required power consumption by the user of the first node, the record of the required power consumption by the user of the second node until the record of the required power consumption by the user of the Nth node are used as the input of the input node, the graph neural network topology is trained to obtain the grid frequency trend prediction model.

[0027] Furthermore, the grid frequency timing identification information is used as the supervision of the output node, the first substation area large grid pre-allocated output power timing record information, the first node energy storage type smart microgrid record and output power, the second node energy storage type smart microgrid record and output power until the Mth node energy storage type smart microgrid record and output power, and the first node user record demand power consumption, the second node user record demand power consumption until the Nth node user record demand power consumption as the input node input, the graph neural network topology is trained to obtain the grid frequency trend prediction model, including: Construct the power grid frequency trend prediction loss function: , , , , in, Characterize the grid frequency trend and predict losses, Characterization selection from arrive The probability of the trend, Characterize the information from the grid frequency time series prediction From the beginning to the elements, and the timing identification information from the power grid frequency From the beginning to the The cumulative cost of the optimal alignment sequence of elements, Characterizing the timing prediction information of power grid frequency The i-th element of Timing identification information of power grid frequency The i-th element of The time alignment cost, Represents the square difference of two elements, Characterize the temperature parameters, >0, The larger the probability distribution More focus on the minimum cost trend, The smaller the probability distribution The smoother, n represents the timing identification information of the power grid frequency The length of m represents the grid frequency timing prediction information Length, Characterize the grid frequency timing prediction information, Characterize the timing identification information of the power grid frequency, Characterization No. Moment elements, Characterization No. Moment element; Based on the grid frequency trend prediction loss function, with the grid frequency timing identification information as the supervision of the output node, with the first substation large grid pre-allocated output power timing record information, the first node energy storage type smart microgrid record integrated output power, the second node energy storage type smart microgrid record integrated output power until the Mth node energy storage type smart microgrid record integrated output power, and the first node user record demand power consumption, the second node user record demand power consumption until the Nth node user record demand power consumption as the input node input, the graph neural network topology is trained to obtain the grid frequency trend prediction model.

[0028] Specifically, the power grid frequency trend prediction model is a functional module for predicting the power grid frequency trend based on the neural network model. It is used to predict the changing trend of the power grid frequency based on the input power grid and user power consumption data. The specific construction process is as follows: The first substation grid topology refers to the organizational structure characteristics of the first substation grid. The first substation grid topology includes the output power integration position of the substation grid, the power integration position of the energy storage type smart microgrid and the user-side integration position. The output power integration position of the substation grid refers to the physical distribution position of the substation distribution station, the power integration position of the energy storage type smart microgrid refers to the physical integration position of the energy storage type smart microgrid, and the user-side integration position refers to the user load distribution position.

[0029] Graph neural network topology refers to a neural network topology that is obtained by performing topology simulation on the first substation power grid topology and has the same structure as the first substation power grid topology. Graph neural network can learn graph structure data. The embodiment of the present application abstracts the frequency analysis problem of the first substation power grid into a graph structure problem, and uses graph neural network topology to build the model topology. Preferably, the output power incorporation position of the substation large power grid, the power incorporation position of the energy storage type smart microgrid and the user-side incorporation position are configured as input nodes, and the input nodes are fully connected to obtain a fully connected layer, which is set as the output node of the graph neural network topology. Preferably, each input node is configured as a long short-term memory neural network, which is more sensitive to data of time series nature. Therefore, a long short-term memory neural network is used here to process the time series data of each input node to improve processing accuracy.

[0030] Furthermore, the historical power management data of the first substation refers to the historical power operation record data of the first substation, including the timing record information of the pre-allocated output power of the large power grid in the first substation, the record of the output power incorporated by the first node energy storage type smart microgrid, the record of the output power incorporated by the second node energy storage type smart microgrid until the record of the output power incorporated by the M-th node energy storage type smart microgrid, and the record of the required power consumption by the first node user, the record of the required power consumption by the second node user until the record of the required power consumption by the N-th node user, and the timing identification information of the power grid frequency. Preferably, the power grid frequency timing identification information is used as the supervision of the output node, and the pre-allocated output power of the large power grid in the first substation is used. The time series record information, the first node energy storage type smart microgrid record integrated output power, the second node energy storage type smart microgrid record integrated output power until the M-th node energy storage type smart microgrid record integrated output power, and the first node user record demand power consumption, the second node user record demand power consumption until the N-th node user record demand power consumption are input to the input node, and a grid frequency trend prediction model construction data set is constructed. The grid frequency trend prediction model construction data set is divided into a ratio of 8:2 to obtain a grid frequency trend prediction model training data set with a ratio of 8 and a grid frequency trend prediction model training data set with a ratio of 2.

[0031] Furthermore, a power grid frequency trend prediction loss function is constructed: , , , The loss function constructed through the embodiment of the present application combines the time alignment capability of DTW and the differentiability of probability distribution, which is suitable for training deep learning models for time series prediction. It not only focuses on local deviations, but also pays more attention to the overall deviation of the power grid frequency trend, thereby improving the stability of the model.

[0032] Based on the grid frequency trend prediction loss function, the grid frequency trend prediction model training data set is retrieved, and the graph neural network topology is trained. Whenever the loss value of the grid frequency trend prediction loss function is trained for a preset number of consecutive times is less than or equal to the loss value threshold preset by the user, the grid frequency trend prediction model verification data set is retrieved for verification. When the loss value verified for a preset number of consecutive times is less than or equal to the loss value threshold, the model is deemed to have converged, and the grid frequency trend prediction model is output. The input node of the energy storage type smart microgrid power integration position is set as the energy storage type smart microgrid input channel, the input node of the substation area large power grid output power integration position is set as the substation area large power grid input channel, and the input node of the user-side integration position is set as the user demand input channel.

[0033] When the grid frequency trend prediction model is scheduled for the first time, the pre-allocated output power timing information of the first substation large grid is input into the substation large grid input channel, the input of the energy storage type smart microgrid input channel is configured to 0, and the user demand power consumption of the first node, the user demand power consumption of the second node, and the user demand power consumption of the Nth node are input into the user demand input channel to obtain the first grid frequency trend prediction result output by the grid frequency trend output channel, which is the grid frequency trend prediction result of the inertial analysis time zone.

[0034] S40: When the first grid frequency trend prediction result deviates from the reference grid frequency, configuring the first node energy storage smart microgrid incorporation output power of the energy storage smart microgrid incorporation node of the first transformer substation, the second node energy storage smart microgrid incorporation output power until the Mth node energy storage smart microgrid incorporation output power; Specifically, the proportion of the time when the deviation between the first grid frequency trend prediction result and the reference grid frequency is greater than or equal to the frequency deviation threshold is calculated, and is set as the grid frequency trend deviation. When the grid frequency trend deviation is ≥ the grid frequency trend deviation threshold, the first grid frequency trend prediction result is considered to deviate from the reference grid frequency. The grid frequency trend deviation threshold is a threshold value preset by the user, and the default value is 0.2. When the first grid frequency trend prediction result does not deviate from the reference grid frequency, there is no need to perform energy storage type smart microgrid grid connection configuration. If the first grid frequency trend prediction result deviates from the reference grid frequency, it is necessary to perform energy storage type smart microgrid grid connection configuration to control the grid frequency. Therefore, the energy storage type smart microgrid connection output power of the first energy storage type smart microgrid connection node of the first substation is set, and the energy storage type smart microgrid connection output power of the second energy storage type smart microgrid is set until the energy storage type smart microgrid connection output power of the Mth energy storage type smart microgrid. The output power refers to the power generation output power. M represents the number of energy storage type smart microgrid nodes connected. M≥1. The output power is determined based on the random value of the rated interval of each connected energy storage type smart microgrid node.

[0035] S50: input the pre-allocated output power timing information of the first substation area large power grid, the output power of the first node energy storage type smart microgrid, the output power of the second node energy storage type smart microgrid until the output power of the M-th node energy storage type smart microgrid, and the power consumption required by the user at the first node, the power consumption required by the user at the second node until the power consumption required by the user at the N-th node into the power grid frequency trend prediction model, and output the second power grid frequency trend prediction result; Specifically, in the second call of the grid frequency trend prediction model, the pre-allocated output power timing information of the first substation large power grid, the output power of the first node energy storage type smart microgrid, the output power of the second node energy storage type smart microgrid until the output power of the Mth node energy storage type smart microgrid, and the user demand power consumption of the first node, the user demand power consumption of the second node until the Nth node user demand power consumption are input into the grid frequency trend prediction model, and the second grid frequency trend prediction result is output.

[0036] S60: When the second grid frequency trend prediction result does not deviate from the reference grid frequency, the power grid connection management of the energy storage smart microgrid connection node is performed according to the first node energy storage smart microgrid connection output power, the second node energy storage smart microgrid connection output power, and so on until the Mth node energy storage smart microgrid connection output power.

[0037] Specifically, when the second grid frequency trend prediction result does not deviate from the benchmark grid frequency, it means that the current power grid-connected strategy is feasible. Then, during the inertial power consumption analysis time zone, the power grid-connected management of the energy storage type smart microgrid connection node can be performed according to the first node energy storage type smart microgrid connection output power, the second node energy storage type smart microgrid connection output power, and so on until the Mth node energy storage type smart microgrid connection output power.

[0038] Furthermore, it also includes: when the second grid frequency trend prediction result deviates from the reference grid frequency, a cyclic analysis is performed after updating the first node energy storage type smart microgrid integrated output power, the second node energy storage type smart microgrid integrated output power, and the Mth node energy storage type smart microgrid integrated output power.

[0039] Specifically, when the second grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage type smart microgrid integrated output power, the second node energy storage type smart microgrid integrated output power until the Mth node energy storage type smart microgrid integrated output power are updated, and then the process returns to step S50 to execute the loop again until the grid frequency trend prediction result does not deviate from the reference grid frequency, and the corresponding first node to the Mth node smart microgrid integrated output frequency is output.

[0040] Further, when the second grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage type smart microgrid integrated output power, the second node energy storage type smart microgrid integrated output power, and the Mth node energy storage type smart microgrid integrated output power are updated and then a cyclic analysis is performed, including: When the number of cycle analysis meets the population size threshold, the historical grid-connected strategy set of energy storage smart microgrid is obtained; Taking the deviation from the reference grid frequency as the fitness function, the grid connection strategy is optimized based on the historical grid connection strategy set of the energy storage type smart microgrid, and the energy storage type smart microgrid expansion grid connection strategy is obtained to perform a cyclic analysis.

[0041] Specifically, the population size threshold refers to the preset number of cycle analysis, which defaults to 500 groups. If the number of cycle analysis is greater than or equal to the population size threshold, the grid connection strategy that does not deviate from the reference grid frequency has not been obtained, and each subsequent update uses the following method: The smart microgrid incorporation output frequency of the first node to the Mth node updated each time is retrieved and stored as a set of historical grid-connected strategies for energy storage smart microgrids; the set of historical grid-connected strategies for energy storage smart microgrids is sorted from small to large using the degree of deviation from the reference grid frequency as the fitness function to obtain the sorting result of the historical grid-connected strategies for energy storage smart microgrids; the historical grid-connected strategies for energy storage smart microgrids with a preset number of ranked fronts are used as the targets of the historical grid-connected strategies for energy storage smart microgrids with a preset number of ranked backs, and the same node incorporation output power variation is performed to obtain the expansion grid-connected strategies for energy storage smart microgrids, and the return to step S50 is executed in a loop until the grid frequency is When the rate trend prediction result does not deviate from the reference grid frequency, the smart microgrid input and output frequencies of the corresponding first node to the Mth node are output. Preferably, variation mode 1 is: using the input output power of some nodes of the target strategy to replace the input output power of the same dimension nodes of the historical grid-connected strategy of the energy storage type smart microgrid with a later ranking; variation mode 2 is: updating the input output power of the same dimension nodes of the historical grid-connected strategy of the energy storage type smart microgrid with a later ranking to be closer to the input output power of the corresponding dimension nodes of the target strategy. Preferably, the total number of mutated nodes is less than M. Through the group optimization mechanism, the probability of obtaining a converged grid-connected strategy can be increased, and the processing efficiency can be improved.

[0042] An energy storage type smart microgrid power management method for frequency and peak regulation provided by an embodiment of the present invention has at least the following technical effects: By traversing user distribution nodes to perform inertial power consumption analysis, the power demand on the user side can be predicted in advance; by combining the large power grid's pre-allocated output power timing information and the user-side power demand, the power grid frequency trend prediction model can be used to predict the power grid frequency trend in advance and reduce latency; when the prediction results show that the grid frequency will deviate from the reference frequency, the energy storage-type smart microgrid's integrated output power is configured in time to achieve rapid adjustment of the grid frequency and improve the timeliness of frequency and peak regulation, thereby achieving the technical effect of reducing the frequency and peak regulation delay of the smart microgrid.

[0043] Example 2: Please refer to Figure 2 , Figure 2 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 2 As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a first computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the first computer program 511, the following steps are implemented: Traverse the user distribution nodes in the first substation to perform inertial power consumption analysis, and obtain the power consumption required by the user at the first node, the power consumption required by the user at the second node, and finally the power consumption required by the user at the Nth node; Obtain the timing information of the pre-allocated output power of the large power grid in the first substation area; Input the pre-allocated output power timing information of the first substation area large power grid, the power consumption demanded by the first node user, the power consumption demanded by the second node user, and the power consumption demanded by the Nth node user into the power grid frequency trend prediction model, and output the first power grid frequency trend prediction result; When the first grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage smart microgrid incorporation node of the first substation is configured to incorporate the output power of the first node energy storage smart microgrid, the output power of the second node energy storage smart microgrid, and the output power of the Mth node energy storage smart microgrid; Input the pre-allocated output power timing information of the first substation area large power grid, the output power of the first node energy storage type smart microgrid, the output power of the second node energy storage type smart microgrid until the output power of the M-th node energy storage type smart microgrid, and the power consumption required by the user of the first node, the power consumption required by the user of the second node until the power consumption required by the user of the N-th node into the power grid frequency trend prediction model, and output the second power grid frequency trend prediction result; When the second grid frequency trend prediction result does not deviate from the reference grid frequency, the power grid connection management of the energy storage smart microgrid connection node is performed according to the first node energy storage smart microgrid connection output power, the second node energy storage smart microgrid connection output power, and so on until the Mth node energy storage smart microgrid connection output power.

[0044] Example 3: Please refer to Figure 3 , Figure 3 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 3 As shown, this embodiment provides a computer-readable storage medium 600, on which a second computer program 611 is stored. When the second computer program 611 is executed by a processor, the following steps are implemented: Traverse the user distribution nodes in the first substation to perform inertial power consumption analysis, and obtain the power consumption required by the user at the first node, the power consumption required by the user at the second node, and finally the power consumption required by the user at the Nth node; Obtain the timing information of the pre-allocated output power of the large power grid in the first substation area; Input the pre-allocated output power timing information of the first substation area large power grid, the power consumption demanded by the first node user, the power consumption demanded by the second node user, and the power consumption demanded by the Nth node user into the power grid frequency trend prediction model, and output the first power grid frequency trend prediction result; When the first grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage smart microgrid incorporation node of the first substation is configured to incorporate the output power of the first node energy storage smart microgrid, the output power of the second node energy storage smart microgrid, and the output power of the Mth node energy storage smart microgrid; Input the pre-allocated output power timing information of the first substation area large power grid, the output power of the first node energy storage type smart microgrid, the output power of the second node energy storage type smart microgrid until the output power of the M-th node energy storage type smart microgrid, and the power consumption required by the user of the first node, the power consumption required by the user of the second node until the power consumption required by the user of the N-th node into the power grid frequency trend prediction model, and output the second power grid frequency trend prediction result; When the second grid frequency trend prediction result does not deviate from the reference grid frequency, the power grid connection management of the energy storage smart microgrid connection node is performed according to the first node energy storage smart microgrid connection output power, the second node energy storage smart microgrid connection output power, and so on until the Mth node energy storage smart microgrid connection output power.

[0045] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0046] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, methods, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0047] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (methods), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0048] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0049] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0050] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0051] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A method for power management of energy storage type smart microgrid for frequency and peak regulation, characterized in that: include: Traverse the user distribution nodes in the first substation to perform inertial power consumption analysis, and obtain the power consumption required by the user at the first node, the power consumption required by the user at the second node, and finally the power consumption required by the user at the Nth node; Obtain the timing information of the pre-allocated output power of the large power grid in the first substation area; Input the pre-allocated output power timing information of the first substation area large power grid, the power consumption demanded by the first node user, the power consumption demanded by the second node user, and the power consumption demanded by the Nth node user into the power grid frequency trend prediction model, and output the first power grid frequency trend prediction result; When the first grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage smart microgrid incorporation node of the first substation is configured to incorporate the output power of the first node energy storage smart microgrid, the output power of the second node energy storage smart microgrid, and the output power of the Mth node energy storage smart microgrid; Input the pre-allocated output power timing information of the first substation area large power grid, the output power of the first node energy storage type smart microgrid, the output power of the second node energy storage type smart microgrid until the output power of the M-th node energy storage type smart microgrid, and the power consumption required by the user of the first node, the power consumption required by the user of the second node until the power consumption required by the user of the N-th node into the power grid frequency trend prediction model, and output the second power grid frequency trend prediction result; When the second grid frequency trend prediction result does not deviate from the reference grid frequency, the power grid connection management of the energy storage smart microgrid connection node is performed according to the first node energy storage smart microgrid connection output power, the second node energy storage smart microgrid connection output power, and so on until the Mth node energy storage smart microgrid connection output power.

2. The method according to claim 1, characterized in that Traverse the user distribution nodes in the first substation to perform inertial power consumption analysis, and obtain the power consumption required by the user at the first node, the power consumption required by the user at the second node, and finally the power consumption required by the user at the Nth node, including: Obtain a first node user of the user distribution node of the first station area, wherein the first node user has a user type label and a user area label; Obtaining time zone portrait information of the inertial power consumption analysis time zone, wherein the time zone portrait information includes a month portrait label, a week portrait label, and a clock portrait label; Perform inertial sample fitting according to the user type label, the user area label, the month portrait label, the week portrait label, and the clock portrait label to obtain the power consumption required by the user at the first node; Traverse the second node user until the Nth node user and perform inertial power consumption analysis respectively to obtain the power consumption required by the second node user until the Nth node user.

3. The method according to claim 2, characterized in that Performing inertial sample fitting according to the user type label, the user area label, the month portrait label, the week portrait label, and the clock portrait label to obtain the power consumption required by the user at the first node includes: Building a fixed background condition based on the user type label, the user area label, and the month portrait label; Configuring a binary background condition for the week portrait label, wherein the binary background condition is Monday to Friday or Saturday to Sunday; Based on the time deviation threshold, configuring a dynamic background condition for the clock portrait tag; According to the fixed background condition, the binary background condition and the dynamic background condition, a first-level normal sample is collected to obtain a first-level electricity consumption sample, wherein the first-level electricity consumption sample includes a first-level sample user type label, a first-level sample user area label, a first-level sample month portrait label, a first-level sample week portrait label and a first-level sample clock portrait label; Perform secondary normal sample collection according to the primary sample user type label, the primary sample user area label, the primary sample month portrait label, the primary sample week portrait label, and the primary sample clock portrait label to obtain a secondary electricity consumption sample; Sample fitting is performed on the secondary power consumption sample and the primary power consumption sample to obtain the power consumption required by the user of the first node.

4. The method according to claim 3, characterized in that Performing sample fitting on the secondary power consumption sample and the primary power consumption sample to obtain the power consumption required by the user of the first node includes: Grouping the secondary power usage samples according to the primary power usage samples to obtain multiple groups of power usage samples; Traversing the plurality of groups of electricity consumption samples to perform a central trend assessment, and obtaining a plurality of center demand electricity consumptions; A central tendency assessment is performed on the multiple center demand power consumptions and the first-level power consumption samples to obtain the first-node user demand power consumption.

5. The method according to claim 1, characterized in that The grid frequency trend prediction model includes a large grid input channel in the substation area, an energy storage type smart microgrid input channel, a user demand input channel and a grid frequency trend output channel. The pre-allocated output power timing information of the large grid in the substation area, the power demand of the user at the first node, the power demand of the user at the second node until the power demand of the user at the Nth node are input into the grid frequency trend prediction model, and the first grid frequency trend prediction result is output, including: Input the pre-allocated output power timing information of the first large power grid in the substation area into the input channel of the large power grid in the substation area, configure the input of the energy storage type smart microgrid input channel to 0, input the power demand of the first node user, the power demand of the second node user until the power demand of the Nth node user into the user demand input channel, and obtain the first power grid frequency trend prediction result output by the power grid frequency trend output channel; The step of constructing the power grid frequency trend prediction model includes: Obtaining a first area power grid topology, wherein the first area power grid topology includes an area large power grid output power incorporation position, an energy storage type smart microgrid power incorporation position, and a user side incorporation position; Based on the first substation area power grid topology, a graph neural network topology is constructed, wherein the graph neural network topology is the same as the first substation area power grid topology structure, the output power incorporation position of the substation area large power grid, the power incorporation position of the energy storage type smart microgrid and the user side incorporation position are configured as input nodes, and the input nodes are fully connected to be set as output nodes of the graph neural network topology, wherein the input nodes are long short-term memory neural networks; Collecting historical power management data of the first substation, wherein the historical power management data of the first substation includes the timing record information of the pre-allocated output power of the large power grid in the first substation, the output power recorded by the first node energy storage type smart microgrid, the output power recorded by the second node energy storage type smart microgrid until the output power recorded by the M-th node energy storage type smart microgrid, and the power consumption demand recorded by the first node user, the power consumption demand recorded by the second node user until the power consumption demand recorded by the N-th node user, and the timing identification information of the power grid frequency; The grid frequency timing identification information is used as the supervision of the output node, the timing record information of the pre-allocated output power of the first substation large power grid, the record of the output power incorporated by the first node energy storage type smart microgrid, the record of the output power incorporated by the second node energy storage type smart microgrid until the record of the output power incorporated by the Mth node energy storage type smart microgrid, and the record of the required power consumption by the user of the first node, the record of the required power consumption by the user of the second node until the record of the required power consumption by the user of the Nth node are used as the input of the input node, the graph neural network topology is trained to obtain the grid frequency trend prediction model.

6. The method according to claim 5, characterized in that The grid frequency timing identification information is used as the supervision of the output node, the pre-allocated output power timing record information of the first substation large grid, the first node energy storage type smart microgrid record and the second node energy storage type smart microgrid record and the output power record and the Mth node energy storage type smart microgrid record and the first node user record demand power consumption, the second node user record demand power consumption and the Nth node user record demand power consumption are used as the input of the input node, and the graph neural network topology is trained to obtain the grid frequency trend prediction model, including: Construct the power grid frequency trend prediction loss function: ， ， ， ， in, Characterize the power grid frequency trend and predict losses, Characterization selection from arrive The probability of the trend, Characterize the information from the grid frequency time series prediction From the beginning to the elements, and the timing identification information from the power grid frequency From the beginning to the The cumulative cost of the optimal alignment sequence of elements, Characterizing the timing prediction information of power grid frequency The i-th element of Timing identification information of power grid frequency The i-th element of The time alignment cost, Represents the square difference of two elements, Characterize the temperature parameters, >0, The larger the probability distribution More focus on the minimum cost trend, The smaller the probability distribution The smoother, n represents the timing identification information of the power grid frequency The length of m represents the grid frequency timing prediction information Length, Characterize the grid frequency timing prediction information, Characterize the timing identification information of the power grid frequency, Characterization No. Moment elements, Characterization No. Moment element; Based on the grid frequency trend prediction loss function, with the grid frequency timing identification information as the supervision of the output node, with the first substation large grid pre-allocated output power timing record information, the first node energy storage type smart microgrid record integrated output power, the second node energy storage type smart microgrid record integrated output power until the Mth node energy storage type smart microgrid record integrated output power, and the first node user record demand power consumption, the second node user record demand power consumption until the Nth node user record demand power consumption as the input node input, the graph neural network topology is trained to obtain the grid frequency trend prediction model.

7. The method according to claim 1, characterized in that Also includes: When the second grid frequency trend prediction result deviates from the reference grid frequency, a cyclic analysis is performed after updating the first node energy storage type smart microgrid integrated output power, the second node energy storage type smart microgrid integrated output power, and the Mth node energy storage type smart microgrid integrated output power.

8. The method according to claim 7, characterized in that When the second grid frequency trend prediction result deviates from the reference grid frequency, the first node energy storage type smart microgrid integrated output power, the second node energy storage type smart microgrid integrated output power, and the Mth node energy storage type smart microgrid integrated output power are updated and then a cyclic analysis is performed, including: When the number of cycle analysis meets the population size threshold, the historical grid-connected strategy set of energy storage smart microgrid is obtained; Taking the deviation from the reference grid frequency as the fitness function, the grid connection strategy is optimized based on the historical grid connection strategy set of the energy storage type smart microgrid, and the energy storage type smart microgrid expansion grid connection strategy is obtained to perform a cyclic analysis.

Images