A step-up integrated power transformation system and control method

Abstract

The present application relates to the technical field of voltage increasing and power transformation, and particularly relates to a voltage increasing and power transformation integrated system and a control method. The power generation of renewable energy power stations such as wind power and photovoltaic power is predicted by using a pre-constructed output prediction model to obtain predicted power generation. Then, the predicted power generation and the actual load of the power grid are subjected to feature extraction, and the change features are extracted to reflect the fluctuation or change trend of the data. Then, the change features are matched with a pre-constructed dispatching precursor feature template library. If the match is successful, the change features are taken as the precursor features of the dispatching instruction. The energy storage unit is subjected to charging and discharging and temperature regulation control to make the energy storage unit in a response ready state. The energy storage unit in the response ready state is only further limited in the charging and discharging range, and the capacity of charging and the amount of discharging are sufficient, so that the energy storage unit has both fast response and does not affect the peak-valley electricity arbitrage.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voltage boosting power transformation, and particularly relates to a voltage boosting integrated power transformation system and a control method. BACKGROUND

[0002] The energy storage type voltage boosting integrated power transformation system is a comprehensive electrical system that boosts the low-voltage (such as 690V, 1000V, etc.) electrical energy generated by a wind turbine generator set or a photovoltaic power generation unit to a higher voltage level (such as 10kV, 35kV, etc.) suitable for long-distance transmission through integrated equipment. It is widely used in new energy power generation projects such as wind farms and photovoltaic power stations, and is a key link connecting the power generation unit and the power grid.

[0003] The existing energy storage type voltage boosting integrated power transformation system is usually designed with a highly integrated prefabricated cabin or container, involving the integration of battery packs, transformers, inverters, control equipment, fire extinguishing systems, and other modules. The main function of the existing energy storage type voltage boosting integrated power transformation system is to respond to grid connection according to the dispatching information of the power grid, and the grid connection response is passive. Passive response usually only focuses on meeting the dispatching instructions, and cannot plan the optimal charging and discharging time in advance. In addition, passive response can only respond after a problem occurs, and often has the problem of slow response speed. Active response mainly depends on the prediction results of the power station processing, and when the prediction fails, it often causes great losses, for example, if the output prediction is too optimistic, the system may still plan to discharge when the expected power generation is insufficient, resulting in failure to meet its own or the grid demand; conversely, if the prediction is too pessimistic, it may miss the profit opportunity of charging or discharging. SUMMARY

[0004] Therefore, the present application aims to provide a voltage boosting integrated power transformation system and a control method to solve the problems in the background art.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] The voltage boosting integrated power transformation system of the present application comprises:

[0007] An acquisition module is configured to acquire meteorological prediction data at multiple time points in a current time period and, after authorization, acquire actual load of the power grid in the current time period, wherein the current time period is a time period of a target duration before a current time point;

[0008] A prediction module is configured to predict predicted power generation in a future time period based on the meteorological prediction data and a pre-constructed output prediction model, wherein the future time period is a time period of a target duration after the current time point;

[0009] The feature matching module is configured to extract a load change feature of an actual load of a current time period and a power generation change feature of a predicted power generation of a future time period; match the load change feature and the power generation change feature with a pre-constructed scheduling precursor feature template library, and determine that the load change feature or the power generation change feature is a scheduling precursor feature when the load change feature or the power generation change feature matches any one feature template in the pre-constructed scheduling precursor feature template library.

[0010] The pre-response module is configured to perform charging and discharging and temperature adjustment control on the energy storage unit based on the scheduling precursor feature, so that the energy storage unit is in a response ready state, wherein the charging and discharging range of the energy storage unit in the response ready state is further limited.

[0011] The application also provides a booster integrated substation control method, comprising:

[0012] Obtain meteorological prediction data of multiple time points in a current time period, and obtain an actual load of a power grid in the current time period after authorization, wherein the current time period is a time period of a target length before a current time point;

[0013] Predict a predicted power generation of a future time period based on the meteorological prediction data and a pre-constructed output prediction model, wherein the future time period is a time period of a target length after the current time point;

[0014] Extract a load change feature of an actual load of a current time period and a power generation change feature of a predicted power generation of a future time period; match the load change feature and the power generation change feature with a pre-constructed scheduling precursor feature template library, and determine that the load change feature or the power generation change feature is a scheduling precursor feature when the load change feature or the power generation change feature matches any one feature template in the pre-constructed scheduling precursor feature template library;

[0015] Perform charging and discharging and temperature adjustment control on the energy storage unit based on the scheduling precursor feature, so that the energy storage unit is in a response ready state, wherein the charging and discharging range of the energy storage unit in the response ready state is further limited.

[0016] The application has the beneficial effects that: the integrated voltage boosting power transformation system and the control method, the power generation of the renewable energy power station such as wind power / photovoltaic is predicted through the output prediction model constructed in advance, and the predicted power generation is obtained. Then the predicted power generation and the actual load of the power grid are extracted, and the change characteristics are extracted, so as to reflect the fluctuation or change trend of the data. Then the change characteristics are matched with the dispatching precursor characteristic template library constructed in advance, if matched, it is regarded as the precursor characteristic of the dispatching instruction. Once the precursor characteristic is obtained, the charging and discharging and temperature regulation control of the energy storage unit are carried out, so that the energy storage unit is in the response ready state, and once the dispatching instruction arrives, the fast response can be carried out. In addition, the energy storage unit in the response ready state only further limits the charging and discharging range, and leaves enough capacity for charging / discharging. The valley electricity storage and peak electricity grid connection arbitrage can be continued, so as to have the fast response and not affect the peak valley electricity arbitrage. BRIEF DESCRIPTION OF DRAWINGS

[0017] The application will be further described below in combination with the drawings and embodiments:

[0018] Figure 1 is a hardware structure diagram of an integrated voltage boosting power transformation system shown in an embodiment of the application;

[0019] Figure 2 is a layout structure diagram of an integrated voltage boosting power transformation system shown in an embodiment of the application;

[0020] Figure 3 is a logic structure diagram of an integrated voltage boosting power transformation system shown in an embodiment of the application;

[0021] Figure 4 is a change characteristic matching flowchart shown in an embodiment of the application;

[0022] Figure 5 is a change characteristic extraction flowchart shown in an embodiment of the application;

[0023] Figure 6 is a flowchart of an integrated voltage boosting power transformation control method shown in an embodiment of the application. DETAILED DESCRIPTION

[0024] The embodiments of the application will be described below in combination with specific examples, and those skilled in the art can easily understand other advantages and effects of the application from the disclosure. The application can also be implemented or applied by different specific embodiments, and the details in the specification can be modified or changed based on different views and applications without departing from the spirit of the application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0025] It should be noted that the diagrams provided in the following embodiments only schematically illustrate the basic concepts of the present application, and only the layers related to the present application are shown in the diagrams, rather than being drawn according to the number, shape and size of the layers in actual implementation. The actual implementation of each layer can be a random change, and the layer layout pattern can be more complex.

[0026] In the following description, a large number of details are discussed to provide a more thorough explanation of the embodiments of the present application. However, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details.

[0027] Figure 1 is a hardware structure diagram of a voltage boosting integrated power transformation system shown in an embodiment of the present application, as Figure 1 shown, a voltage boosting integrated power transformation system in the present application includes a low-voltage end input module 110, an energy storage unit 120, a bidirectional energy storage converter 130, a transformer 140, a switch cabinet 150, an energy management system 160, a centralized control center 170, and an environment monitoring / heat dissipation control system 180. The specific connection relationship and principles are as follows:

[0028] The low-voltage end input module 110 is used to access the low-voltage end power grid, wherein the low-voltage end is connected to a new energy power generation unit (such as a photovoltaic inverter) or an energy storage system to provide a direct current input interface.

[0029] The energy storage unit 120 is connected to the low-voltage end input module and is used to store or release direct current. The energy storage unit is composed of a plurality of lithium iron phosphate batteries and a battery management system (BMS) and is used to store electrical energy.

[0030] The bidirectional energy storage converter 130 is connected to the energy storage unit and the low-voltage end input module and is used to convert the direct current input by the energy storage unit or the low-voltage end input module into low-voltage alternating current or to convert low-voltage alternating current into direct current.

[0031] The transformer 140 is connected to the bidirectional energy storage converter on the primary side and is connected to the power grid through a grid-connected circuit breaker on the secondary side. The transformer is used to boost the alternating current to obtain medium-voltage alternating current and to connect the medium-voltage alternating current to the power grid or to step down the medium-voltage alternating current in the power grid to obtain low-voltage alternating current.

[0032] The alternating current end of the transformer is connected to the power grid through a grid-connected circuit breaker in the switch cabinet.

[0033] The function of the bidirectional energy storage converter PCS in the above structure is to realize bidirectional energy conversion, control battery charging and discharging, photovoltaic grid connection, reactive power compensation, etc. The low-voltage end input module 110 and the energy storage unit 120 can be connected to the DC side of the PCS, and the PCS realizes MPPT and battery voltage matching through the DC-DC converter respectively. The PCS output is connected to the low-voltage AC bus, and is connected to the medium-voltage power grid through a step-up transformer. All protection signals (overvoltage, overcurrent, temperature) are fed back to the energy management system EMS. The BMS system is also connected to the energy management system EMS through a data line, and all control tasks are generated by the centralized control center 170 and sent to the energy management system EMS for control.

[0034] Figure 2 A layout structure diagram of a booster integrated power transformation system shown in an embodiment of the present application is adopted by the box-type booster integrated power transformation system, and the layout structure is as shown in Figure 2 The left half is an electrical room, and the inside is arranged with an energy storage bidirectional converter, an on-site monitoring cabinet, an air conditioner indoor unit, etc. The right half is a transformer room, and a main transformer and an isolation transformer are arranged. The isolation transformer is used to access AC power from the power grid to supply power to the equipment (such as an air conditioner).

[0035] Figure 3 A logic structure diagram of a booster integrated power transformation system shown in an embodiment of the present application, the logic structure of the present application runs in a centralized control center, as shown in Figure 3 The booster integrated power transformation system of the embodiment includes an acquisition module 310, a prediction module 320, a feature matching module 330, and a pre-response module 340, and the principles are as follows.

[0036] The acquisition module 310 is used to acquire meteorological prediction data at multiple time points in a current time period, and to acquire actual load of the power grid in the current time period after authorization, wherein the current time period is a time period of a target length before the current time point.

[0037] The meteorological prediction data includes temperature, solar intensity, wind speed, precipitation, etc. These data can be obtained from a related meteorological release platform. The actual load needs to be obtained from the dispatching system of the power grid, and authorization needs to be obtained before acquisition. The corresponding data is requested to enter the semi-active response model to quickly respond to the dispatching instruction of the power grid dispatching system.

[0038] The prediction module 320 is used to predict predicted power generation in a future time period based on the meteorological prediction data and a pre-constructed output prediction model, wherein the future time period is a time period of a target length after the current time point.

[0039] The output prediction model is constructed by using the historical output sample data of the photovoltaic power generation or the wind power generation, and the historical output sample data includes the values of the multiple meteorological factors and the power generation power.

[0040] (1) Obtain multiple historical time point output sample data, wherein the output sample data includes values of multiple meteorological factors and power generation power.

[0041] The multiple meteorological factors (such as light intensity, temperature, wind speed, humidity, etc.) and corresponding power generation power values are collected through the historical time point output sample data (such as photovoltaic power generation or wind power generation data).

[0042] (2) Extract multiple meteorological factor value sequences and power generation power value sequences from the output sample data, and calculate the correlation of the multiple meteorological factor value sequences and the power generation power value sequences , wherein the correlation is mathematically expressed as:

[0043]

[0044] In the formula, represents the value in the meteorological factor value sequence , represents the value in the power generation power value sequence , represents the joint probability distribution of the meteorological factor value sequence and the power generation power value sequence , represents the marginal probability distribution of the meteorological factor value sequence , represents the marginal probability distribution of the power generation power value sequence .

[0045] In the present application, mutual information is used as the correlation, and in the above calculation formula, the ratio of the joint probability distribution to the marginal distribution can capture the nonlinear relationship between variables (such as the nonlinear response of light intensity to power generation power).

[0046] (3) The meteorological factor with a correlation greater than a preset correlation threshold is taken as a target meteorological factor .

[0047] The correlation threshold (for example, 0.8) is set according to the field experience or statistical significance (such as p value), the weakly correlated meteorological factors (such as the weak influence of humidity on photovoltaic power generation) are removed by quantifying the correlation, and the noise interference is reduced.

[0048] (4) Screen out the samples containing the target meteorological factor from the output sample data. The value and corresponding target sample data of power generation;

[0049] The target sample data includes a feature matrix (target meteorological factors) and a label vector (power generation). By filtering, the multidimensional data is concentrated on key variables, reducing the curse of dimensionality problem in subsequent modeling.

[0050] (5) Cluster the target sample data to obtain multiple clusters;

[0051] K-means, DBSCAN, or hierarchical clustering are used to divide the target sample data into multiple clusters. Samples under similar meteorological conditions (such as high sunshine and high temperature combinations) may form clusters, reflecting local power generation patterns.

[0052] (6) Calculate the average contour coefficient of multiple target sample data based on the multiple clusters, and compare the average contour coefficient with a preset contour coefficient threshold;

[0053] The average silhouette coefficient reflects the effectiveness of clustering. The average silhouette coefficient falls between -1 and 1. A closer average silhouette coefficient to 1 indicates better clustering, making it suitable for fitting models to various scenarios based on clusters, resulting in better fit. Conversely, a lower average silhouette coefficient indicates poor clustering, making cluster-based model fitting unnecessary. In this embodiment, the calculation process for the average silhouette coefficient is as follows:

[0054] (6-1) For any target sample data Calculate the target sample data Compared with other target sample data within the same cluster European distance The average of multiple Euclidean distances is calculated to obtain the intra-cluster similarity. ;

[0055] (6-2) For any target sample data Calculate the target sample data With each target sample data not belonging to the cluster European distance The target sample data is obtained by calculating the average of multiple Euclidean distances. Average distance to each non-cluster And by selecting the minimum average distance, the inter-cluster separation degree is obtained. ;

[0056] (6-3) Based on the intra-cluster similarity and the inter-cluster separation Calculate target sample data contour coefficient , wherein the contour coefficient The mathematical expression is:

[0057]

[0058] (6-4) Calculate the average silhouette coefficient of all target sample data to obtain the average silhouette coefficient.

[0059] In this embodiment, clustering is valid if the average silhouette coefficient is greater than or equal to a preset value (e.g., 0.7); otherwise, it is invalid. The clustering result is dynamically determined based on the silhouette coefficient to avoid inefficient clustering that could degrade model performance. If the data distribution is discrete (e.g., extreme weather samples), univariate regression is triggered when the silhouette coefficient is low to ensure model reliability.

[0060] (7) When the average profile coefficient is greater than or equal to the preset profile coefficient threshold, multiple linear regression is performed on multiple target meteorological factors and power generation based on the target sample data in each cluster to obtain multiple linear regression models for multiple clusters; and a power output prediction model is constructed based on multiple linear regression models.

[0061] For multiple local clusters obtained by clustering, meteorological factors and power generation are more likely to have a linear relationship (such as the linear response between light intensity and photovoltaic power). Therefore, if the clustering is effective, a multiple linear regression model is used for fitting.

[0062] For each cluster of samples, multiple linear regression (MLR) is used to fit the relationship between the target meteorological factor and power generation. The mathematical expression for multiple linear regression is:

[0063]

[0064] In the formula, Indicates power generation capacity. This represents the target meteorological factors from the 1st to the nth. Represents the regression coefficient. The error term is represented by . The regression coefficients are estimated by minimizing the sum of squared residuals.

[0065] (8) When the average profile coefficient is less than the preset profile coefficient threshold, perform univariate multiple regression on each target meteorological factor and power generation based on multiple target sample data to obtain a univariate multiple regression model for multiple target meteorological factors; and the correlation of multiple target meteorological factors. Normalization is performed to obtain the weights of multiple target meteorological factors; an output prediction model is constructed based on the weights of multiple target meteorological factors and a univariate multiple regression model.

[0066] When the average profile coefficient is less than 0.7, it indicates that the data distribution is discrete, resulting in poor clustering effect. When the data distribution is discrete, the polynomial regression can capture complex nonlinear relationships (such as the nonlinear response of temperature and photovoltaic power). Therefore, a multiple regression mathematical formula is constructed for each meteorological factor for fitting, wherein the multiple regression mathematical formula is:

[0067]

[0068] In the formula, represents the regression coefficient, is an error term, is a natural number. Similarly, the regression coefficient is estimated by minimizing the residual sum of squares.

[0069] In addition, if the output model is composed of multiple multiple regression, the weight is introduced by normalizing the correlation degree. The prediction result of the high correlation factor contributes more to the final output, improving the rationality of the model.

[0070] The correlation degree is normalized by using the maximum-minimum normalization, and the calculation formula is:

[0071]

[0072] In the formula, is the maximum correlation degree, is the minimum correlation degree.

[0073] In the above process, the multiple linear regression or weighted multiple regression is automatically switched according to the clustering effect, taking into account the local rule and global trend. Through the profile coefficient and correlation degree threshold, the model complexity is dynamically adjusted to avoid overfitting or underfitting. Traditional methods (such as single multiple regression) are difficult to adapt to complex data distribution, while the present method improves adaptability through clustering and dynamic model selection.

[0074] Based on the above output prediction model, the present application predicts in the following way:

[0075] (1) Extract the values of the target meteorological factors at multiple time points from the meteorological prediction data;

[0076] (2) Determine the model type;

[0077] (2-1) If it is a multiple linear regression model corresponding to multiple power generation conditions, construct a meteorological feature vector at multiple time points based on the values of multiple target meteorological factors at multiple time points, and perform cosine similarity calculation on the meteorological feature vector and the center vector of each cluster to obtain the target cluster with the highest similarity. Substitute the values of the multiple target meteorological factors at the multiple time points into the multiple linear model of the target cluster to obtain the predicted power generation power at the multiple time points ;

[0078] (2-2) If it is a univariate multiple regression model with multiple target meteorological factors, then substitute the predicted values ​​of each target meteorological factor at multiple time points into the corresponding univariate multiple regression model to obtain the predicted power generation of each target meteorological factor at multiple time points. The predicted power generation at each time point is weighted based on the weights of multiple target meteorological factors to obtain the final predicted power generation at multiple time points. , .

[0079] The feature matching module 330 is used to extract the load change features of the actual load in the current time period and the power generation change features of the predicted power generation in the future time period; to match the load change features and the power generation change features with a pre-built scheduling precursor feature template library; and to determine that the load change features or the power generation change features are scheduling precursor features when the load change features or the power generation change features match any feature template in the pre-built scheduling precursor feature template library.

[0080] In order to maintain the stable operation of the power grid, balance supply and demand, and ensure the reliability of power supply, when the power grid load suddenly rises, falls, or fluctuates, or when the power generation of photovoltaic power plants fluctuates or decreases, the power grid dispatching system will issue dispatching instructions to the storage-type step-up integrated substation system to use the energy storage unit to achieve the function of peak shaving and valley filling.

[0081] The above process analyzed the predicted power generation of photovoltaic power plants. Since load fluctuations, increases, and decreases in the power grid are generally caused by sudden events and are difficult to predict, this application analyzes load data for the current time period (e.g., the last half hour) to extract any abnormal increases, decreases, or fluctuations. The same pattern of changes is applied to the predicted power generation at multiple future time points. These features are then matched with a pre-built precursor feature template library. If a match is found, it indicates a certain probability that a dispatch command will be issued. Entering a pre-response state at this point will improve the response speed.

[0082] Specifically, the methods for constructing the precursor feature template library include:

[0083] (1) Obtain scheduling sample data with scheduling type of sudden scheduling, wherein the scheduling sample data includes the power generation and grid load values ​​of multiple historical time points within a preset time period before the arrival time of the scheduling instruction;

[0084] Specifically, sample data of "sudden dispatch" events are screened from the dispatch history of the power grid. Each sample contains time series data of current photovoltaic power plant power generation and power grid load values within a preset time period (e.g., 6 hours) before the time point when the dispatch instruction arrives.

[0085] (2) A sliding window with a target length is constructed, and the window is slid forward from the time point when the dispatch instruction arrives. At each sliding, the load change feature and the power generation change feature of the dispatch sample data in the sliding window are extracted;

[0086] The window is slid forward (e.g., once every 5 minutes) from the time when the dispatch instruction arrives, and the window width is the target length (e.g., 30 minutes). The load change feature and the power generation change feature of the dispatch sample data in each window are extracted. Through dynamic adjustment of the sliding window, the response time requirements of different sudden dispatches are adapted.

[0087] The extraction of the load change feature and the power generation change feature is mentioned several times in this application. In order to avoid redundancy, this application will be introduced in detail later.

[0088] (3) The load change features of multiple sliding windows are matched with a preset load change feature template, and the power generation change features of multiple sliding windows are matched with a preset power generation change feature template, to obtain matching results of multiple sliding windows;

[0089] The load change feature template is the change feature of the load data sequence and the power generation data sequence under normal operation. The construction of the load change feature template can be achieved by collecting a large number of normal operation samples to extract the change feature, and then calculating the average value and the standard deviation of each dimension parameter in the change feature (essentially a vector) to construct a reference value range of three standard deviations, so as to obtain the load change feature template.

[0090] When matching, if the values of all dimension parameters fall within the reference value range corresponding to the load change feature template, it is determined to be matched, otherwise it is determined to be not matched.

[0091] (4) When the load change features and the power generation change features of multiple sliding windows are matched with the corresponding feature templates, the current dispatch sample data is excluded; when the load change features of a sliding window are not matched with the load change feature template, or the power generation change features of a sliding window are not matched with the power generation change feature template, the sliding window with the earliest time and the matching result of not matching is taken as the target window, and the load change feature with the matching result of not matching in the target window is taken as the first target feature, or the power generation change feature with the matching result of not matching in the target window is taken as the second target feature;

[0092] If the load and generation characteristics in the window match the template, it is considered that the window has no precursor characteristics, and the corresponding sample is removed.

[0093] If there are unmatched characteristics, record the window where the earliest unmatched characteristics appear, and extract the unmatched load or generation characteristics as target characteristics. Through time priority screening, lock the initial abnormal point of the burst scheduling. As a precursor characteristic sample.

[0094] (5) Cluster the first target characteristics of multiple scheduling data samples to obtain multiple first characteristic clusters, and cluster the second target characteristics of multiple scheduling data samples to obtain multiple second characteristic clusters;

[0095] K-means or DBSCAN clustering is used for the first target characteristics (load mismatch characteristics) and the second target characteristics (generation mismatch characteristics) respectively, to find multiple typical precursor characteristic sample clusters (such as "load steep rise + photovoltaic sudden drop").

[0096] (6) Retain the first characteristic clusters and the second characteristic clusters whose data amount in the cluster is greater than a preset data amount threshold, to obtain the first target characteristic clusters and the second target characteristic clusters;

[0097] Only clusters whose sample number in the cluster is greater than a preset value (such as 50) are retained, to ensure statistical significance. In addition, low-quality clusters can also be removed in combination with indicators such as silhouette coefficient and intra-cluster distance.

[0098] (7) Calculate the average value and standard deviation of multiple dimension parameters of all load change characteristics in the first target characteristic clusters, construct the value range of multiple dimension parameters based on the average value and standard deviation of multiple dimension parameters of the load change characteristics, and construct the load change characteristic template based on the value range of multiple dimension parameters of the load change characteristics;

[0099] Based on the principle of three standard deviations, the characteristic template is counted to obtain multiple typical load change characteristic templates. For example, the load change characteristic template can be , wherein, is a flag bit, -1 represents a decrease, 1 represents an increase, and 0 represents a fluctuation; represents a slope range; represents a maximum difference range; represents a standard deviation range.

[0100] (8) Calculate the average value and standard deviation of multiple dimension parameters of all generation change characteristics in the second target characteristic clusters, construct the value range of multiple dimension parameters based on the average value and standard deviation of multiple dimension parameters of the generation change characteristics, and construct the generation change characteristic template based on the value range of multiple dimension parameters of the generation change characteristics;

[0101] The principle of the power generation change feature template is consistent with that of the load change feature template, and will not be described herein again.

[0102] (9) Constructing a dispatching precursor feature template library based on the plurality of load change feature templates and the plurality of power generation change feature templates.

[0103] Figure 4 For a change feature matching flowchart in an embodiment of the present application, as shown in FIG. 1, in an embodiment of the present application, the load change feature and the power generation change feature are matched with a pre-constructed dispatching precursor feature template library, which includes the following steps. Figure 4

[0104] S410, comparing the load change feature with a load change feature template and comparing the power generation change feature with a power generation change feature template;

[0105] S420, when the value of each dimension parameter in the load change feature falls within the value range of the corresponding dimension parameter in the load change feature template, it is determined that the load change feature matches the load change feature template, otherwise it is determined that the load change feature does not match the load change feature template; when the value of each dimension parameter in the power generation change feature falls within the value range of the corresponding dimension parameter in the power generation change feature template, it is determined that the power generation change feature matches the power generation change feature template, otherwise it is determined that the power generation change feature does not match the power generation change feature template.

[0106] Since the feature template constructed in the foregoing includes the value range of the plurality of dimension parameters, the value of each dimension in the change feature is compared with the value range of the corresponding dimension, so as to determine whether the change feature matches the feature template.

[0107] Figure 5 For a change feature extraction flowchart in an embodiment of the present application, as shown in FIG. 2, in an embodiment of the present application, the extraction process of the load change feature or the power generation change feature includes the following steps. Figure 5

[0108] S510, dividing a plurality of time points in a current time period or a future time period into a plurality of time windows based on a sliding average method, and calculating the average value of a plurality of data points in each time window. wherein, i represents the time window number, and the data point is the load or the power;

[0109] The sliding average method is used to divide the plurality of time points into the plurality of time windows, for example, defining a sliding window width w (such as 5 minutes, 15 minutes) and a sliding step s.

[0110] S520, comparing the average values corresponding to any two adjacent time windows, and when the following condition is met, ​​​​when the trend of the data points at the plurality of time points is increasing; or when the trend of the data points at the plurality of time points is decreasing; or otherwise, when the trend of the data points at the plurality of time points is none, wherein, a ratio is set;

[0111] In this embodiment, the overall value trend is extracted by the moving average method, and then the value comparison of adjacent windows is performed to obtain the change trend.

[0112] S530, when the trend of the data points at the plurality of time points is increasing or decreasing, calculating the slope of the data points at the plurality of time points , wherein the slope is mathematically expressed as:

[0113]

[0114] In the formula, is the average value of the last time window, is the average value of the first time window, is the middle time point of the last time window, is the middle time point of the first time window;

[0115] If there is a one-way change trend, for example, rising or falling, the corresponding change amplitude feature, i.e., the slope , is extracted. The slope is used to reflect the change rate of the one-way change. If there is no one-way change trend, the slope is 0. In the case where there is no one-way change trend, the fluctuation feature is more important.

[0116] S540, extracting the difference between the maximum value and the minimum value of the data points at the plurality of time points , and extracting the variance after normalizing the data points at the plurality of time points;

[0117] S550, determining a trend flag based on the trend of the plurality of time points, and constructing a change feature based on the trend flag, the slope , the difference , and the variance , wherein the change feature is a load change feature or a power generation change feature, and the slope is set to 0 when the trend of the data points at the plurality of time points is none.

[0118] The fluctuation feature includes the difference between the maximum value and the minimum value and the variance . Therefore, the final change feature is a four-dimensional vector, i.e., .

[0119] a pre-response module 340, configured to control the charging and discharging and temperature adjustment of the energy storage unit based on the dispatching precursor feature, so that the energy storage unit is in a response ready state, wherein the charging and discharging range of the energy storage unit in the response ready state is further limited.

[0120] Finally, if there is a dispatching precursor feature matching the template, the storage unit is in a response ready state by using the BMS system. The charging and discharging range of the energy storage unit in the response ready state is further limited, that is, a certain charging / discharging redundancy is reserved. Specifically as follows:

[0121] (1) When the dispatching precursor feature represents a decrease in power generation or an increase in grid load, the charging and discharging range of the energy storage unit is adjusted to , wherein is the lower limit of the power, is the upper limit of the power, is the power adjustment value;

[0122] Generally, in order to prolong the battery life, the SOC upper and lower limit values are usually limited when the integrated step-up substation performs arbitrage of charging in the low valley period and discharging in the peak period. For example, , that is , .

[0123] If the dispatching precursor feature represents a decrease in power generation or an increase in grid load, it may receive a dispatching instruction for grid-connected discharging. In this case, a certain discharging redundancy is required, so the SOC lower limit value needs to be adjusted higher to ensure that the battery has sufficient power during the arbitrage of charging and discharging. For example, , the charging and discharging range of the adjusted energy storage unit is adjusted to At this time, the normal process of charging in the low valley period and discharging in the peak period can still be continued, while ensuring the lower limit of the power.

[0124] (2) When the dispatching precursor feature represents a decrease in grid load, the charging and discharging range of the energy storage unit is adjusted to .

[0125] Similarly, if the dispatching precursor feature represents a decrease in grid load, the energy storage unit needs to be charged to absorb excess power. At this time, the charging and discharging range adjustment value is reserved to reserve a certain chargeable capacity.

[0126] (3) When the dispatching precursor feature represents power generation fluctuation or grid load fluctuation, the batteries in the energy storage unit are divided into two groups, one group is adjusted to , and the other group is adjusted to ;

[0127] If it is an abnormal fluctuation, the battery pack is divided, one group is used to respond to the discharge scheduling instruction, and one group is used to respond to the charging scheduling instruction.

[0128] (4) Obtain the real-time temperature of the energy storage unit, and perform closed-loop control on the air conditioning system and the battery heating system based on the real-time temperature of the energy storage unit, so that the real-time temperature of the energy storage unit is in a preset temperature interval.

[0129] The performance of the battery pack is best at 20-40 DEG C. In order to ensure timely response, the battery temperature needs to be maintained at about 20 DEG C, and the scheduling instruction is waited. Once the scheduling instruction arrives, it can be responded immediately, without waiting for the charging and discharging of the battery, and without waiting for the preheating / cooling of the battery.

[0130] When no scheduling instruction is received within a set time (such as 6 hours) after the current time point, the response readiness state of the energy storage unit is released. The charging and discharging arbitrage can be normally performed.

[0131] The power prediction model is used for predicting the power generation of the renewable energy power station, and the predicted power generation is obtained. Then, the predicted power generation and the actual load of the power grid are subjected to feature extraction, and the change features are extracted, so as to reflect the fluctuation or change trend of the data. Then, the change features are matched with the pre-constructed scheduling precursor feature template library, and if the match is matched, the match is regarded as the precursor feature of the scheduling instruction. Once the precursor feature is obtained, the charging and discharging and temperature regulation control of the energy storage unit are performed, so that the energy storage unit is in a response readiness state, and once the scheduling instruction arrives, a fast response can be performed. In addition, the energy storage unit in the response readiness state only further limits the charging and discharging range, and leaves sufficient charging capacity / discharge capacity. The valley electricity storage and peak electricity grid arbitrage can be continued, so that the fast response is achieved, and the peak-valley electricity arbitrage is not affected.

[0132] As shown in Figure 6 The application also provides a power boost integrated power transformation control method, comprising:

[0133] S610, obtain meteorological prediction data at a plurality of time points in a current time period, and obtain actual load of the power grid in the current time period after authorization, wherein the current time period is a time period of a target time length before the current time point;

[0134] S620, predict the predicted power generation of a future time period based on the meteorological prediction data and a pre-constructed output prediction model, wherein the future time period is a time period of a target time length after the current time point;

[0135] S630, extracting a load change feature of the actual load of the current time period and a power generation change feature of the predicted power generation of the future time period; matching the load change feature and the power generation change feature with a pre-constructed scheduling precursor feature template library, and determining that the load change feature or the power generation change feature is a scheduling precursor feature when the load change feature or the power generation change feature matches any one feature template in the pre-constructed scheduling precursor feature template library.

[0136] S640, performing charging and discharging and temperature adjustment control on the energy storage unit based on the scheduling precursor feature, so that the energy storage unit is in a response ready state, wherein the charging and discharging range of the energy storage unit in the response ready state is further limited.

[0137] The power boosting integrated substation control method of the present application predicts the power generation of a renewable energy power station such as a wind power station or a photovoltaic power station through a pre-constructed power prediction model to obtain predicted power generation. Then, the predicted power generation and the actual load of the power grid are subjected to feature extraction to extract change features, thereby reflecting the fluctuation or change trend of the data. Then, the change features are matched with a pre-constructed scheduling precursor feature template library, and if a match is found, the change features are regarded as precursor features of scheduling instructions. Once the precursor features are obtained, charging and discharging and temperature adjustment control are performed on the energy storage unit, so that the energy storage unit is in a response ready state, and once the scheduling instructions are received, a fast response can be performed. In addition, the energy storage unit in the response ready state is only further limited in the charging and discharging range, and the capacity for charging and the amount of electricity for discharging are reserved. The valley electricity storage and peak electricity grid connection arbitrage can be continued, so that the fast response is achieved and the peak-valley electricity arbitrage is not affected.

[0138] The present embodiment also provides a computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement any one of the control methods in the present embodiment, wherein the control method is the execution logic of the present system.

[0139] The present embodiment also provides an electronic terminal, comprising a processor and a memory.

[0140] The memory is configured to store a computer program, and the processor is configured to execute the computer program stored in the memory to enable the terminal to execute any one of the control methods in the present embodiment.

[0141] The computer readable storage medium in the embodiment can be understood by those skilled in the art that all or part of the steps of the above-mentioned control method embodiments can be completed by a computer program related hardware. The computer program can be stored in a computer readable storage medium. The program executes the steps of the above-mentioned control method embodiments when executed; and the storage medium includes ROM, RAM, magnetic disc or optical disc and other storage medium that can store program codes.

[0142] The electronic terminal provided in the embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected with the processor and the transceiver and complete communication between each other. The memory is used for storing a computer program. The communication interface is used for communication. The processor and the transceiver are used for running the computer program, so that the electronic terminal executes each step of the above control method.

[0143] In the embodiment, the memory can include a random access memory (RAM) and can also include a non-volatile memory, for example, at least one disk memory.

[0144] The processor can be a general processor, including a central processing unit (CPU), a network processor (NP) and the like; and can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0145] In the above embodiments, although the present application has been described in conjunction with specific embodiments thereof, it will be readily apparent to those skilled in the art that many substitution, modification and / or simplifications can be made in the specific embodiments without departing from the scope of the present application, which is defined by the following claims. The embodiments of the present application are intended to embrace all such alternatives, modifications and variations as falling within the scope of the appended claims.

[0146] The above embodiments are only illustrative of the principles of the present application and its efficacy, and are not intended to limit the present application. Any modification or change made by any person skilled in the art without departing from the spirit and scope of the present application shall be covered by the claims of the present application.

Claims

1. A step-up integrated substation system, characterized by comprising: The method comprises the following steps: an acquisition module is configured to acquire meteorological prediction data at multiple time points in a current time period and to acquire actual load of a power grid in the current time period after authorization, wherein the current time period is a time period of a target length before a current time point; a prediction module is configured to predict predicted power generation in a future time period based on the meteorological prediction data and a pre-constructed power output prediction model, wherein the future time period is a time period of the target length after the current time point; a feature matching module is configured to extract load change features of the actual load in the current time period and to extract power generation change features of the predicted power generation in the future time period; the load change features and the power generation change features are matched with a pre-constructed dispatching precursor feature template library, and when the load change features or the power generation change features match any one of the feature templates in the pre-constructed dispatching precursor feature template library, the load change features or the power generation change features are determined to be dispatching precursor features; a pre-response module is configured to control charging and discharging and temperature adjustment of an energy storage unit based on the dispatching precursor features, so that the energy storage unit is in a response-ready state, wherein the charging and discharging range of the energy storage unit in the response-ready state is further limited.

2. The integrated booster substation system of claim 1, wherein Further comprising: a low-voltage end input module configured to access a low-voltage end power grid; an energy storage unit connected to the low-voltage end input module and configured to store or release direct current; a bidirectional energy storage converter connected to the energy storage unit and the low-voltage end input module and configured to convert direct current input by the energy storage unit or the low-voltage end input module into low-voltage alternating current or to convert low-voltage alternating current into direct current; a transformer with a primary side connected to the bidirectional energy storage converter and a secondary side connected to the power grid through a grid-connected circuit breaker and configured to step up the alternating current to obtain medium-voltage alternating current and connect the medium-voltage alternating current to the power grid or to step down medium-voltage alternating current in the power grid to obtain low-voltage alternating current.

3. The integrated booster substation system of claim 1, wherein The method for constructing the power output prediction model comprises: acquiring power output sample data at multiple historical time points, wherein the power output sample data comprises values of multiple meteorological factors and power generation; Extract multiple meteorological factor value sequences from the power output sample data. and power generation value sequence And calculate the value sequences of various meteorological factors. and power generation value sequence Relevance The relevance The mathematical expression is: In the formula, Represents the sequence of meteorological factor values The values ​​in For the power generation value sequence The values ​​in Represents the sequence of meteorological factor values and power generation value sequence The joint probability distribution, Represents the sequence of meteorological factor values Marginal probability distribution, Represents the sequence of power generation values The marginal probability distribution; correlation degree a meteorological factor greater than the preset correlation degree threshold as a target meteorological factor ; The target meteorological factor is removed from the output sample data. The value and corresponding target sample data of power generation; clustering the target sample data to obtain multiple clusters; comparing the average profile coefficient with a preset profile coefficient threshold value; when the average profile coefficient is greater than or equal to the preset profile coefficient threshold value, performing multiple linear regressions on multiple target meteorological factors and power generation based on the target sample data in each cluster to obtain multiple linear regression models of the multiple clusters, and constructing a power output prediction model based on the multiple linear regression models; When the average profile coefficient is less than a preset profile coefficient threshold, performing monadic multiple regression on each target meteorological factor and power generation based on the plurality of target sample data to obtain a monadic multiple regression model of the plurality of target meteorological factors; and determining the correlation degrees of the plurality of target meteorological factors performing normalization to obtain weights of the plurality of target meteorological factors; and constructing an output prediction model based on the weights of the plurality of target meteorological factors and the monadic multiple regression model.

4. The integrated booster substation system of claim 3, wherein The method for constructing the dispatching precursor feature template library comprises: For any one target sample data , the Euclidean distance between the target sample data and other target sample data in the cluster to which it belongs is calculated , and the average of the plurality of Euclidean distances is calculated to obtain the intra-cluster similarity ; For any one target sample data , the Euclidean distance between the target sample data and the target sample data in each non-belonging cluster is calculated , and the average of the plurality of Euclidean distances is calculated to obtain the average distance between the target sample data and each non-belonging cluster , and the smallest average distance is screened out to obtain the inter-cluster separation degree ; based on the intra-cluster similarity and the inter-cluster separation calculating target sample data of the profile coefficient wherein the profile coefficient is mathematically expressed as The average value of the profile coefficients of all target sample data is calculated to obtain the average profile coefficient.

5. The integrated booster substation system of claim 1, wherein acquiring dispatching sample data of a dispatching type of burst dispatching, wherein the dispatching sample data comprises power generation and power grid load values at multiple historical time points within a preset length of time before a dispatching instruction arrival time point; ​ A sliding window with a target length is constructed, and the sliding window slides forward from a time point of arrival of a scheduling instruction as a starting point. Each time the sliding window slides, a load change feature and a power generation change feature of scheduling sample data in the sliding window are extracted; The load change features and the power generation change features of the multiple sliding windows are matched with preset load change feature templates and preset power generation change feature templates, respectively, to obtain matching results of the multiple sliding windows; When the load change features and the power generation change features of the multiple sliding windows are matched with the corresponding feature templates, the current scheduling sample data is excluded. When there is a load change feature of a sliding window that is not matched with a load change feature template, or there is a power generation change feature of a sliding window that is not matched with a power generation change feature template, a sliding window with the earliest time and a matching result of not matching is taken as a target window, and a load change feature with a matching result of not matching in the target window is taken as a first target feature, or a power generation change feature with a matching result of not matching in the target window is taken as a second target feature; The first target features of the multiple scheduling data samples are clustered to obtain multiple first feature clusters, and the second target features of the multiple scheduling data samples are clustered to obtain multiple second feature clusters; First feature clusters and second feature clusters with data amounts greater than a preset data amount threshold are retained to obtain first target feature clusters and second target feature clusters; The average values and standard deviations of multiple dimension parameters of all the load change features in the first target feature clusters are calculated, the value ranges of the multiple dimension parameters are constructed based on the average values and the standard deviations of the multiple dimension parameters of the load change features, and a load change feature template is constructed based on the value ranges of the multiple dimension parameters of the load change features; The average values and standard deviations of multiple dimension parameters of all the power generation change features in the second target feature clusters are calculated, the value ranges of the multiple dimension parameters are constructed based on the average values and the standard deviations of the multiple dimension parameters of the power generation change features, and a power generation change feature template is constructed based on the value ranges of the multiple dimension parameters of the power generation change features; A scheduling precursor feature template library is constructed based on the multiple load change feature templates and the multiple power generation change feature templates.

6. The integrated booster substation system of claim 5, wherein, The extraction process of the load change feature or the power generation change feature includes: The plurality of time points in the current time period or the future time period are divided into a plurality of time windows based on a sliding average method, and an average value of the plurality of data points in each time window is calculated wherein, denotes a time window number, and the data point is a load or power;​ comparing the average values corresponding to any two adjacent time windows, determining that the trend of the data points at the plurality of time points is increasing when satisfied, determining that the trend of the data points at the plurality of time points is decreasing when satisfied, and otherwise determining that the plurality of time points has no trend, wherein is a set ratio; a trend of the data points at the plurality of time points is increasing or decreasing, calculating a slope of the data points at the plurality of time points wherein the slope is mathematically expressed as: wherein is the average value of the last time window, is the average value of the first time window, is the middle time point of the last time window, is the middle time point of the first time window; extracting a difference between a maximum value and a minimum value of the data points at the plurality of time points extracting a variance after normalizing the data points at the plurality of time points ; Determining trend markers based on trends at multiple time points, and determining trend markers and slopes. Difference and variance Construct change characteristics, wherein the change characteristics are load change characteristics or power generation change characteristics, and when there is no trend in the data points at multiple time points, the slope is... Set to zero.

7. The integrated booster substation system of claim 6, wherein, The load change feature and the power generation change feature are matched with a scheduling precursor feature template library constructed in advance, including: The load change feature is compared with a load change feature template, and the power generation change feature is compared with a power generation change feature template; When the values of each dimension parameter in the load change feature fall within the value range of the corresponding dimension parameter in the load change feature template, it is determined that the load change feature is matched with the load change feature template, otherwise it is determined that the load change feature is not matched. When the values of each dimension parameter in the power generation change feature fall within the value range of the corresponding dimension parameter in the power generation change feature template, it is determined that the power generation change feature is matched with the power generation change feature template, otherwise it is determined that the power generation change feature is not matched.

8. The integrated booster substation system of claim 1, wherein, The scheduling precursor feature is used to control charging and discharging and temperature regulation of an energy storage unit, including: adjust the charge-discharge range of the energy storage unit to wherein, is a lower limit value of the electric quantity, is an upper limit value of the electric quantity, is an electric quantity adjustment value; In the event that the dispatching precursor feature characterizes a drop in grid load, the charge-discharge range of the energy storage unit is adjusted to ; when the dispatching precursor feature represents a fluctuation of power generation or a fluctuation of power grid load, the batteries in the energy storage unit are divided into two groups, the charge-discharge range of one group is adjusted to , and the charge-discharge range of the other group is adjusted to ; Obtaining the real-time temperature of the energy storage unit, and performing closed-loop control on the air conditioning system and the battery heating system based on the real-time temperature of the energy storage unit, so that the real-time temperature of the energy storage unit is in a preset temperature interval.

9. The integrated booster substation system of claim 1, wherein, Also includes: When no scheduling instruction is received within a time period after the current time point, the response readiness state of the energy storage unit is released.

10. A boost integrated substation control method, characterized by, Includes: Obtaining meteorological prediction data at multiple time points in a current time period, and obtaining actual load of the power grid in the current time period after authorization, wherein the current time period is a time period of a target time length before the current time point; Based on the meteorological prediction data and the pre-constructed output prediction model, predicting the predicted power generation in a future time period, wherein the future time period is a time period of a target time length after the current time point; Extracting the load change characteristics of the actual load in the current time period, and extracting the power generation change characteristics of the predicted power generation in the future time period; matching the load change characteristics and the power generation change characteristics with a pre-constructed scheduling precursor feature template library, and determining that the load change characteristics or the power generation change characteristics are scheduling precursor features when the load change characteristics or the power generation change characteristics match any one feature template in the pre-constructed scheduling precursor feature template library; Based on the scheduling precursor feature, the energy storage unit is controlled to charge and discharge and adjust the temperature, so that the energy storage unit is in a response readiness state, and the charge and discharge range of the energy storage unit in the response readiness state is further limited.

Images