Power grid dispatching method and system based on mass power data mining

Abstract

The invention relates to the technical field of power grids, and provides a power grid dispatching method and system based on mass power data mining, and the method comprises the steps: carrying out the clustering of the power consumption data of household users, and dividing the data into different power consumption categories; clustering the electricity consumption data of the resident electric vehicle users, and dividing the electricity consumption data into different electricity consumption categories; the method comprises the following steps: performing cross classification on electricity consumption categories of resident users, electricity price periods and peak and valley periods to obtain classification sets, grouping and aggregating the classification sets into a transaction data set based on the resident users, and screening a frequent item set based on a formulated screening rule and an optimized association searching thought of a multi-minimum support strategy to obtain a frequent item set; establishing an association rule according to items contained in the frequent item set; based on the association rule, the power grid scheduling strategy is formulated, and the power grid is scheduled, so that the power utilization behavior adjustment of the user can be guided more actively, the power utilization load stability and the energy utilization efficiency are improved, the clean energy consumption is promoted, and finally the optimization and upgrading of the power system are promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grids, and in particular relates to a power grid dispatching method and system based on massive power data mining. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The adjustment and optimization of the time-of-use electricity price mechanism has become one of the key measures. The policy aims to encourage users to change their electricity usage patterns and alleviate the contradiction between electricity supply and demand by formulating electricity price levels for different time periods based on changes in grid load.

[0004] The proportion of new energy on the power generation side has increased, and the time-of-use electricity price that adapts to the interaction between sources and loads and market-driven needs to be improved. With the construction of new power systems, the demand for fully mobilizing flexible resources to participate in the overall optimization of power system resources has increased. However, currently, due to the shortage of flexible resources and the relatively rigid power operation mechanism, my country's flexible resources have obviously not been effectively released. At the same time, with the construction of power marketization, the market transaction price curve is affected by factors such as new energy output, and there is a misalignment with the traditional peak and valley time-of-use electricity price period division. Therefore, time-of-use electricity prices need to further optimize peak and valley periods, promote the orderly connection between plans and market price systems, and form a market-driven promotion of renewable energy consumption.

[0005] The peak-valley characteristics of the load on the power grid side change, and the time-of-use electricity price mechanism that is conducive to reducing peak loads has not yet been fully established. In recent years, the corresponding peak power demand in various places has surged compared with normal periods, posing greater challenges to the stable operation of the power system. However, the power capacity has been idle for a long time during most periods of the year. Although the overall goal of reducing the peak-valley difference was considered at the beginning of the establishment of the time-of-use electricity price system, it is expected to effectively reduce the large amount of investment caused by a very small part of the peak load. Therefore, in order to better adapt to new situations such as changes in the peak-valley characteristics of the power system, it is necessary to master the calculation method of the execution effect of the time-of-use electricity price mechanism, so as to continuously adjust and optimize the time-of-use electricity price policy.

[0006] The user-side market mechanism needs to be improved, and the role of time-of-use electricity prices in stimulating the development of new market players needs to be enhanced. When the time-of-use electricity price system was first formulated, it did not take into account the emergence and development of new elements in the energy field such as distributed energy and electric vehicles. Therefore, the peak-valley price difference incentives are insufficient, and it is impossible to provide more room for new energy and new energy storage systems to adjust the electricity price system. At the same time, the current time-of-use electricity price policy still has problems such as inaccurate time division, unclear peak-valley electricity price difference, incomplete peak electricity price mechanism, and insufficient connection with the construction and development of the power market.

[0007] Therefore, it is necessary to apply professional analysis methods, establish a correlation analysis system between time-of-use electricity price adjustments and users' electricity consumption behavior, conduct analysis on the correlation between time-of-use electricity price adjustments and users' electricity consumption behavior, quantify and measure the impact of time-of-use electricity prices on residential users' electricity consumption behavior, understand the implementation effects and shortcomings of time-of-use electricity price policies, and provide data support for the adjustment of time-of-use electricity price policies.

[0008] More importantly, the current power grid dispatching is unable to grasp the behavioral rules of residents' electricity use, resulting in energy waste. Summary of the invention

[0009] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a power grid dispatching method and system based on massive power data mining. Based on massive power data mining, the correlation analysis between residential users' power consumption behavior and the execution of time-of-use electricity price mechanism is carried out, and on this basis, a scientific power grid dispatching strategy is formulated, which can more actively guide users to adjust their power consumption behavior, improve the stability of power load and energy utilization efficiency, promote the consumption of clean energy, and ultimately promote the optimization and upgrading of the power system.

[0010] In order to achieve the above object, the present invention adopts the following technical solution:

[0011] A first aspect of the present invention provides a power grid dispatching method based on massive power data mining, which comprises:

[0012] Obtain the electricity consumption data of residential users, and classify residential users into two categories: residential household users and residential electric vehicle users;

[0013] Cluster the electricity consumption data of household users and divide them into different electricity consumption categories; cluster the electricity consumption data of electric vehicle users and divide them into different electricity consumption categories;

[0014] The electricity consumption categories of residential users are cross-classified with electricity price periods and peak and valley periods to obtain classification sets. After the classification sets are grouped and aggregated into transaction data sets based on residential users, frequent item sets are screened based on the idea of ​​optimizing Apriori association relationship search with the formulated screening rules and multi-minimum support strategy, and association rules are formed according to the items contained in the frequent item sets.

[0015] Based on the association rules, the power grid dispatching strategy is formulated and the power grid is dispatched.

[0016] Furthermore, if the association rule is that the peak period associated with the stage where the time-of-use electricity price is not implemented is a high-energy consumption user, the grid dispatching strategy is to dispatch backup power sources or energy storage equipment in advance to ensure stable power supply during the peak period;

[0017] If the association rule is that the second associated peak period of the time-of-use electricity price stage is a medium energy consumption user, then the grid dispatching strategy is to reduce power generation and transmission losses;

[0018] If the association rule is that the deep valley section in the second stage of the time-of-use electricity price is associated with medium energy consumption users, the grid dispatching strategy is to dispatch renewable energy generation or energy storage equipment to generate or discharge in the deep valley section to balance the grid load.

[0019] Furthermore, the field categories of the power consumption data include user type, power consumption period and power consumption.

[0020] Furthermore, the clustering step includes:

[0021] Construct daily or monthly electricity consumption data into time series;

[0022] After extracting the statistical and shape features of the time series, the electricity consumption categories are obtained through K-Shape clustering based on the restricted learning machine.

[0023] Furthermore, the electricity consumption categories include high, medium and low.

[0024] Furthermore, the step of K-Shape clustering based on restricted learning machine includes:

[0025] Determine the number of hidden layer nodes, activation function and cluster number of the unsupervised extreme learning machine, and then randomly select a time series as the initial centroid;

[0026] Input the time series into the unsupervised extreme learning machine to obtain the time series embedding;

[0027] Based on time series embedding, the shape distance between each time series and the initial centroid is calculated, the probability of each time series becoming the centroid is calculated, and the second centroid is determined; based on time series embedding, the shape distance between each time series and the two centroids is calculated, and the third centroid is determined;

[0028] The three centroids are input into the k-shape algorithm, and the time series are continuously redistributed to the nearest centroid through iteration, and the centroid position is updated until the convergence condition is reached.

[0029] A second aspect of the present invention provides a power grid dispatching system based on massive power data mining, comprising:

[0030] A data acquisition module is configured to: acquire electricity consumption data of residential users, and classify the residential users into two categories: residential household users and residential electric vehicle users;

[0031] A clustering module is configured to: cluster the electricity consumption data of household users and classify them into different electricity consumption categories; cluster the electricity consumption data of electric vehicle users and classify them into different electricity consumption categories;

[0032] The data mining module is configured to: cross-classify the electricity consumption categories of residential users with electricity price periods and peak and valley periods to obtain classification sets, and after grouping and aggregating the classification sets into transaction data sets based on the residential users, filter the frequent item sets based on the idea of ​​optimizing Apriori association relationship search based on the formulated screening rules and multi-minimum support strategy, and form association rules based on the items contained in the frequent item sets;

[0033] The dispatching module is configured to formulate a power grid dispatching strategy based on association rules and dispatch the power grid.

[0034] Furthermore, if the association rule is that the peak period associated with the stage where the time-of-use electricity price is not implemented is a high-energy consumption user, the grid dispatching strategy is to dispatch backup power sources or energy storage equipment in advance to ensure stable power supply during the peak period;

[0035] If the association rule is that the second associated peak period of the time-of-use electricity price stage is a medium energy consumption user, then the grid dispatching strategy is to reduce power generation and transmission losses;

[0036] If the association rule is that the deep valley section in the second stage of the time-of-use electricity price is associated with medium energy consumption users, the grid dispatching strategy is to dispatch renewable energy generation or energy storage equipment to generate or discharge in the deep valley section to balance the grid load.

[0037] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in a power grid dispatching method based on massive power data mining as described above.

[0038] The fourth aspect of the present invention provides a computer device, including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, wherein when the processor executes the program, the steps in a power grid dispatching method based on massive power data mining as described above are implemented.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention conducts a correlation analysis between residential users' electricity consumption behavior and the execution of the time-of-use electricity price mechanism based on massive power data mining, and scientifically formulates power grid dispatching strategies on this basis, which can more actively guide users to adjust their electricity consumption behavior, improve the stability of power load and energy utilization efficiency, promote the consumption of clean energy, and ultimately promote the optimization and upgrading of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 It is a flow chart of a power grid dispatching method based on massive power data mining according to the first embodiment of the present invention;

[0043] Figure 2 This is a flowchart of correlation analysis between residential users' electricity consumption behavior and time-of-use electricity price execution according to the first embodiment of the present invention;

[0044] Figure 3 It is a structural diagram of a computer device according to a fourth embodiment of the present invention. DETAILED DESCRIPTION

[0045] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0046] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0047] Embodiment 1

[0048] This embodiment provides a power grid scheduling method based on massive power data mining.

[0049] This embodiment provides a power grid dispatching method based on massive power data mining, which classifies the electricity consumption behavior of residential users through the K-Shape clustering algorithm, grasps the behavioral rules of residential electricity consumption through the association rule algorithm to verify the implementation effect of the time-of-use electricity price mechanism, and conducts power grid dispatching on this basis to make it more in line with the behavioral rules of residential electricity consumption and avoid energy waste.

[0050] Insightful analysis of the implementation of residential time-of-use electricity prices can help us understand the rules and characteristics of the implementation of time-of-use electricity prices, provide decision-making support for relevant departments, optimize the allocation of power resources, and help improve the utilization efficiency and service quality of power resources. Establishing a correlation analysis system between time-of-use electricity price adjustments and user electricity consumption behavior, quantitatively studying the impact of time-of-use electricity prices on residential users' electricity consumption behavior, and providing auxiliary support for the optimization of time-of-use electricity price implementation strategies will help the power grid further encourage residential users to participate in power peak shifting and valley filling, help ensure power supply, and ensure power grid security.

[0051] In order to improve the efficiency and comprehensiveness of the company's analysis of the current status of the implementation of time-of-use electricity prices for residents, the present invention conducts an in-depth study on the characteristics of residents' electricity consumption behavior and the changes in residents' electricity consumption behavior before and after the implementation of the time-of-use electricity price mechanism.

[0052] The present invention mainly solves the problem of verifying the effectiveness of the implementation of the time-of-use electricity price mechanism. Aiming at the potential rules existing in the electricity consumption behavior of residential users, a research method for the implementation and impact of time-of-use electricity prices for residential users based on massive power data mining is designed, and power grid scheduling is carried out on this basis.

[0053] This embodiment provides a power grid dispatching method based on massive power data mining, including:

[0054] (1) Cluster analysis method of residential electricity consumption behavior: Residential electricity consumption is divided into two subcategories: household electricity consumption and electric vehicle electricity consumption. The electricity consumption time series data of the two types of residential users are collected and preprocessed, and the cluster centers are initialized. The improved unsupervised extreme learning machine K-Shape algorithm is used to measure the similarity between each user's electricity consumption behavior and the cluster center using shape distance (such as SBD distance). Users are assigned to the most similar cluster center and the cluster center is iteratively updated until the clustering result is stable. Finally, the classification result of residential electricity consumption behavior is obtained.

[0055] The process of clustering residential users' electricity consumption behavior is as follows: first, according to the differences in the types of residential users, the electricity consumption data of residential users are divided into household users' electricity consumption data and electric vehicle users' electricity consumption data; then, combining the time series characteristics and electricity consumption characteristics of the two types of residential users' electricity consumption behaviors, the improved unsupervised extreme learning machine K-Shape clustering algorithm based on time series is adopted. The specific steps are as follows:

[0056] 1) Collect and preprocess the time series data of electricity consumption of residential users, perform data cleaning, including data operations such as removing erroneous data and supplementary control, initialize the number of hidden layer nodes and activation function of the unsupervised extreme learning machine, and input the cleaned data into the model to obtain data embedding;

[0057] 2) Initialize a centroid, calculate the similarity distance between each user's electricity consumption behavior and the cluster center based on shape distance metrics such as SBD, calculate the probability of each sample becoming the centroid, obtain the second centroid, and obtain the third centroid in the same way as above;

[0058] 3) The centroid obtained in the above way is used as the initial cluster center and input into the standard K-shape algorithm until the clustering result is stable or the preset number of iterations is reached. Finally, the classification results of residential users' electricity consumption behavior are obtained, including the cluster category and cluster center to which the user belongs, and then the characteristics of users' electricity consumption behavior in different clusters are analyzed.

[0059] (2) Association analysis method between residential users' electricity consumption behavior and time-of-use electricity price implementation: The classification results of residential users' electricity consumption behavior (electricity consumption category) are cross-classified with electricity price period information and peak and valley period information to obtain a classification set, which is then grouped and aggregated into a transaction data set. Then, based on the established screening rules and the idea of ​​optimizing Apriori association relationship search with multiple minimum support strategies, frequent item sets are screened, and on this basis, association rules are established and the lift is calculated to obtain the conclusion of the association analysis between residential users' electricity consumption behavior and time-of-use electricity price implementation.

[0060] Among them, the process of association analysis between residential users' electricity consumption behavior and time-of-use electricity price implementation is as follows: based on the results of clustering residential users' electricity consumption behavior, according to the different electricity price periods to which the electricity consumption data belongs, and the predetermined daily electricity consumption peak and valley periods, the electricity consumption categories of residential users are further cross-classified with different electricity price periods and electricity consumption peak and valley periods, and the classification sets of residential household users and residential electric vehicle users in different electricity price periods, different peak and valley periods, and different electricity consumption categories are obtained; the classification sets are grouped and aggregated according to users, and processed into the transaction data set required for Apriori association analysis; according to the purpose of studying the impact of time-of-use electricity price implementation on residential users' electricity consumption behavior, the required item sets are determined, the corresponding support is calculated, and the frequent item sets are screened; in the acquired frequent item sets, the required association rules are formed; the lift index is calculated based on the corresponding support and confidence, and the required conclusions are drawn.

[0061] Among them, the transaction data set, item set, support, frequent item set, association rule, confidence and lift are all special terms in Apriori association analysis. The transaction data set refers to a set of attribute subsets and other sample identifiers of each sample data (in the present invention, a residential user electricity consumption data containing multiple category labels), and each attribute subset contains attributes that can describe the classification characteristics of the sample, such as the electricity price period, the electricity consumption category of a specific peak and valley period. The item set refers to a subset of the full attribute set. For example, in an association analysis, the full attribute set used to describe the classification characteristics of the sample is {1,2,...,n} (n is an integer greater than 2), then {1}, {2}, {1,2}, etc. are all an item set. The support refers to the ratio of the frequency of an item set appearing in a transaction data set to the length of the transaction data set. For example, if an item set appears in m (m is a positive integer) attribute subsets of the transaction data set, and the total length of the transaction data set is l (l is an integer not less than m), then the support of the item set is m / l. Frequent item sets refer to item sets whose support exceeds a preset threshold. Association rules refer to the way of splitting frequent item sets according to a certain pre-set causal relationship. For example, for the frequent item set {1,2,3}, if the set of item 1 and item 2 {1,2} is set as the cause item set, and the set of item 3 {3} is set as the result item set, then {1,2->3} is its corresponding association rule. Confidence refers to the ratio of the support of the corresponding frequent item set (i.e., the union of the cause item set and the result item set) to the support of the cause item set in an association rule. Lift refers to an evaluation index of association rules, which is the ratio of the corresponding confidence of the association rule to the support of the result item set. This value can characterize the effect of the association rule, with 1 as the boundary. The larger the lift, the stronger the positive influence of the cause item set on the result item set, and vice versa. The lift of 1 means that the cause item set and the result item set are independent of each other and have no causal relationship.

[0062] This embodiment provides a power grid dispatching method based on massive power data mining, whose input is: original data on residential users' electricity consumption including labels such as time and user category, and whose output is: analysis conclusions and dispatching strategies based on Apriori association analysis on the impact of time-of-use electricity prices on residential users' electricity consumption behavior.

[0063] This embodiment provides a power grid dispatching method based on massive power data mining, such as Figure 1 As shown, the specific steps are as follows:

[0064] Step 1: Data preparation.

[0065] The electricity consumption data of residential users in the study area within a certain time span covering multiple electricity price scheme stages are obtained. The field categories of each data include user type, electricity consumption period, electricity consumption, etc., as shown in Table 1.

[0066] Table 1: Field categories that the original data should include

[0067]

[0068] Step 2: Clustering of residential users’ electricity consumption behaviors.

[0069] This embodiment divides the electricity consumption data of residential users into electricity consumption data of residential household users and electricity consumption data of residential electric vehicle users according to the difference of the residential user type field in the original data.

[0070] Afterwards, combined with the time series characteristics and power consumption characteristics of residential users' electricity consumption behavior, the improved unsupervised extreme learning machine K-Shape clustering algorithm is used, combined with outlier processing technology, to more accurately capture the pattern differences of residential users' electricity consumption behavior, effectively process outliers, and improve the accuracy and robustness of clustering. The specific methods are:

[0071] Step 201: Data preprocessing.

[0072] Time series construction: First, the daily or monthly electricity consumption data of the two types of residential users are constructed into time series to reflect the temporal characteristics of users' electricity consumption behavior.

[0073] Outlier detection and processing: A machine learning-based approach is used to detect outliers in time series. Instead of directly classifying into high or low power consumption categories, this implementation uses smoothing or interpolation to replace outliers to reduce information loss and maintain the continuity of the time series.

[0074] Step 202: Feature extraction.

[0075] Extract statistical characteristics (such as mean, variance, kurtosis, skewness) and shape characteristics (such as trend, seasonality, cyclicality) of time series.

[0076] Step 203: Clustering of improved K-Shape based on restricted learning machine.

[0077] Initialization: First determine the number of hidden layer nodes, activation function and cluster number K of the unsupervised extreme learning machine. In this implementation, K is set to 3, corresponding to high, medium and low electricity consumption categories. Then randomly select one time series as the initial centroid.

[0078] Data encoding: Input data into an unsupervised extreme learning machine to obtain data embedding.

[0079] Distance calculation: Use the K-Shape algorithm to calculate the embedded shape distance, such as the cross-correlation-based distance metric SBD to calculate the distance between time series.

[0080] Initial centroid determination: Determine the second centroid based on the calculated shape distance of each sample to become the centroid, and calculate the shape distance of each sample to the two cluster centers again to obtain the third initial centroid.

[0081] Iterative update: The three centroids obtained above are input into the standard k-shape algorithm. Through an iterative process, the time series are continuously redistributed to the nearest centroid and the centroid position is updated until the convergence condition is reached, which is generally when the centroid no longer changes or the maximum number of iterations is reached.

[0082] Output the clustering results of electricity consumption behaviors of two types of residential users, including the cluster category and cluster centroid to which the users belong.

[0083] Step 4: Correlation analysis between residential users’ electricity consumption behavior and time-of-use electricity price implementation, such as Figure 2 shown.

[0084] Step 401: Obtain a classification set.

[0085] Based on the results of the clustering of residential users' electricity consumption behaviors, this embodiment further cross-classifies the electricity consumption categories of residential users with different electricity price periods and peak and valley periods according to relevant data to obtain a classification set. The specific method is:

[0086] (1) According to the implementation of different electricity price strategies during the study period, different electricity price period labels are assigned to the clustering results to characterize the electricity price background of residential users' electricity consumption behavior;

[0087] (2) According to the predetermined peak and valley periods of electricity consumption within the day (for household users, it is divided into peak period, flat period and valley period; for electric vehicle users, it is divided into peak period, peak period, flat period, valley period and deep valley period), different peak and valley period labels are assigned to the clustering results to characterize the grid load background of the residential users' electricity consumption behavior;

[0088] (3) Add aggregation records to aggregate the electricity consumption of individual users and obtain the total electricity consumption. The electricity price period label is consistent with the original record, and the peak and valley period label is recorded as the total electricity consumption.

[0089] (4) According to the above labels, the electricity consumption behavior records of residential users are cross-classified (residential household users and residential electric vehicle users are treated separately). Finally, each record contains the user's residence type, electricity price period, peak and valley time period, and electricity consumption category classification.

[0090] Step 402: Transaction data set processing.

[0091] In this embodiment, the classification set is grouped and aggregated according to the users, and processed into a transaction data set required for Apriori association analysis, and the items included are shown in Table 2.

[0092] Table 2. Description of items included in the transaction data set

[0093]

[0094]

[0095] Combined with the set elements defined in Table 2, an example of a transaction data set is shown in Table 3.

[0096] Table 3. Example table of residential user transaction data set

[0097]

[0098] Step 403: Item set processing and determination of minimum support.

[0099] Each item set in the data item set is given the minimum support it needs to satisfy, and each item set is sorted from large to small according to the minimum support. The database is scanned to generate frequent 1-item sets, and the structure of the item set is reconstructed into a key-value pair structure of frequent item-support.

[0100] Step 404: Filter k-frequent itemsets.

[0101] This embodiment determines the itemsets required for the study, calculates the corresponding support, and screens the frequent itemsets. The specific method is:

[0102] (1) In the process of generating k-item sets, connect the k-1 item set and the 1 item set, and compare the support between the last element in the subset of the k-1 item set and the element in the frequent 1 item set. If the last element is smaller than the element in the 1 item set, take the transaction list of the k-1 transaction list and the 1 transaction list as the transaction list of the k-item candidate set, and determine the support. The support at this time is the minimum value of the minimum support of the included elements.

[0103] (2) Repeat step (1) until no new frequent itemsets are generated.

[0104] (3) Determine the type of required item sets, that is, the item sets that are helpful in testing the “peak shaving and valley filling” capability of the TOU electricity pricing strategy, that is, the electricity consumption levels under the same grid load background in different electricity price periods, such as {1,7}, {3,9}, etc., and formulate the screening rules for the required item sets to ensure that all required item sets can be screened out;

[0105] (4) Based on the rules, the frequent itemsets obtained by the algorithm are screened to obtain the final association rules. In this embodiment, the required association rules are constructed from the obtained frequent itemsets, the lift index is calculated based on the corresponding support and confidence, and the required conclusion is obtained.

[0106] Step 405: Establish association rules.

[0107] (1) According to the items in the frequent itemset, meaningful association rules are formed for subsequent calculation of lift index, some of which are shown in Table 4;

[0108] Table 4. Some association rules

[0109]

[0110]

[0111] (2) Calculate the lift index corresponding to all selected association rules and draw corresponding conclusions. The conclusions can reflect the impact of implementing time-of-use electricity prices on the electricity consumption behavior of residential users from a scientific perspective.

[0112] Step 5: Formulate grid dispatching strategy.

[0113] Based on the analysis results of the above association rules, this embodiment formulates some power grid dispatching strategies, for example:

[0114] (1) For the {1->7} rule:

[0115] Strategy formulation: Before the implementation of time-of-use electricity prices, the peak load is predicted to be large. Therefore, backup power supplies or energy storage equipment can be dispatched in advance to ensure stable power supply during peak periods. Through smart meters and data analysis, users' electricity consumption behavior can be monitored in real time, peak loads can be predicted, and grid operation strategies can be adjusted in advance.

[0116] (2) Regarding the {3->8} rule:

[0117] Strategy formulation: When the time-of-use electricity price is implemented to the second stage, the peak load is reduced. At this time, the grid operation can be optimized to reduce unnecessary power generation and transmission losses. According to the implementation of the time-of-use electricity price and the user's electricity consumption behavior, the grid load distribution is dynamically adjusted to improve energy utilization efficiency.

[0118] (3) For the {3->20} rule:

[0119] Strategy formulation: When the time-of-use electricity price is implemented in the second stage, the load in the deep valley section increases. At this time, renewable energy generation or energy storage equipment can be dispatched to generate or discharge in the deep valley section to balance the grid load. Through smart grid technology, the load changes in the deep valley section are monitored in real time, and renewable energy generation or energy storage equipment is dispatched to ensure stable operation of the grid.

[0120] This embodiment provides a grid dispatching method based on massive power data mining. Based on massive power data mining, the correlation analysis between the electricity consumption behavior of residential users and the execution of the time-of-use electricity price mechanism is carried out, which helps to prove the impact of the implementation of the time-of-use electricity price mechanism on the electricity consumption behavior of residential users from a quantitative and more scientific perspective. This is of great significance for the formulation method and effect analysis of local time-of-use electricity price mechanisms, and helps to improve the availability of time-of-use electricity price strategies, so that power companies can scientifically formulate grid dispatching strategies, which can more actively guide users to adjust their electricity consumption behaviors, improve the stability of electricity load and energy utilization efficiency, promote the consumption of clean energy, and ultimately promote the optimization and upgrading of the power system.

[0121] Embodiment 2

[0122] This embodiment provides a power grid dispatching system based on massive power data mining.

[0123] This embodiment provides a power grid dispatching system based on massive power data mining, such as Figure 2 As shown, it specifically includes:

[0124] A data acquisition module is configured to: acquire electricity consumption data of residential users, and classify the residential users into two categories: residential household users and residential electric vehicle users;

[0125] A clustering module is configured to: cluster the electricity consumption data of household users and classify them into different electricity consumption categories; cluster the electricity consumption data of electric vehicle users and classify them into different electricity consumption categories;

[0126] The data mining module is configured to: cross-classify the electricity consumption categories of residential users with electricity price periods and peak and valley periods to obtain classification sets, and after grouping and aggregating the classification sets into transaction data sets based on the residential users, filter the frequent item sets based on the idea of ​​optimizing Apriori association relationship search based on the formulated screening rules and multi-minimum support strategy, and form association rules based on the items contained in the frequent item sets;

[0127] The dispatching module is configured to formulate a power grid dispatching strategy based on association rules and dispatch the power grid.

[0128] Furthermore, if the association rule is that the peak period associated with the stage where the time-of-use electricity price is not implemented is a high-energy consumption user, the grid dispatching strategy is to dispatch backup power sources or energy storage equipment in advance to ensure stable power supply during the peak period;

[0129] If the association rule is that the second associated peak period of the time-of-use electricity price stage is a medium energy consumption user, then the grid dispatching strategy is to reduce power generation and transmission losses;

[0130] If the association rule is that the deep valley section in the second stage of the time-of-use electricity price is associated with medium energy consumption users, the grid dispatching strategy is to dispatch renewable energy generation or energy storage equipment to generate or discharge in the deep valley section to balance the grid load.

[0131] It should be noted here that each module in this embodiment corresponds to each step in Example 1 one by one, and the specific implementation process is the same, which will not be repeated here.

[0132] Embodiment 3

[0133] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the power grid scheduling method based on massive power data mining as described in the first embodiment above are implemented.

[0134] Embodiment 4

[0135] This embodiment provides a computer device, such as Figure 3 As shown, it includes a computer-readable storage medium 1003, a processor 1001, a communication interface 1002, and a computer program stored on the computer-readable storage medium 1003 and executable on the processor 1001, wherein the processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 may be connected via a bus or other means. The communication interface 1002 is used to receive and send data, and when the processor 1001 executes the program, the steps in the power grid dispatching method based on massive power data mining as described in the first embodiment above are implemented.

[0136] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), 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 processor, 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.

[0137] 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 1A function specified in one or more boxes.

[0138] 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.

[0139] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A power grid dispatching method based on massive power data mining, characterized in that: include: Obtain the electricity consumption data of residential users, and classify residential users into two categories: residential household users and residential electric vehicle users; Cluster the electricity consumption data of household users and divide them into different electricity consumption categories; cluster the electricity consumption data of electric vehicle users and divide them into different electricity consumption categories; The electricity consumption categories of residential users are cross-classified with electricity price periods and peak and valley periods to obtain classification sets. After the classification sets are grouped and aggregated into transaction data sets based on residential users, frequent item sets are screened based on the idea of ​​optimizing Apriori association relationship search with the formulated screening rules and multi-minimum support strategy, and association rules are formed according to the items contained in the frequent item sets. Based on the association rules, the power grid dispatching strategy is formulated and the power grid is dispatched.

2. A power grid dispatching method based on massive power data mining as claimed in claim 1, characterized in that: If the association rule is that the peak period associated with the stage where the time-of-use electricity price is not implemented is a high-energy consumption user, the grid dispatching strategy is to dispatch backup power sources or energy storage equipment in advance to ensure stable power supply during the peak period; If the association rule is that the second associated peak period of the time-of-use electricity price stage is a medium energy consumption user, then the grid dispatching strategy is to reduce power generation and transmission losses; If the association rule is that the deep valley section in the second stage of the time-of-use electricity price is associated with medium energy consumption users, the grid dispatching strategy is to dispatch renewable energy generation or energy storage equipment to generate or discharge in the deep valley section to balance the grid load.

3. A power grid dispatching method based on massive power data mining as claimed in claim 1, characterized in that: The field categories of the power consumption data include user type, power consumption period and power consumption.

4. A power grid dispatching method based on massive power data mining as claimed in claim 1, characterized in that: The clustering step comprises: Construct daily or monthly electricity consumption data into time series; After extracting the statistical and shape features of the time series, the electricity consumption categories are obtained through K-Shape clustering based on the restricted learning machine.

5. A power grid dispatching method based on massive power data mining as claimed in claim 4, characterized in that: The electricity consumption categories include high, medium and low.

6. A power grid dispatching method based on massive power data mining as claimed in claim 4, characterized in that: The steps of K-Shape clustering based on restricted learning machine include: Determine the number of hidden layer nodes, activation function and cluster number of the unsupervised extreme learning machine, and then randomly select a time series as the initial centroid; Input the time series into the unsupervised extreme learning machine to obtain the time series embedding; Based on time series embedding, the shape distance between each time series and the initial centroid is calculated, the probability of each time series becoming the centroid is calculated, and the second centroid is determined; based on time series embedding, the shape distance between each time series and the two centroids is calculated, and the third centroid is determined; The three centroids are input into the k-shape algorithm, and the time series are continuously redistributed to the nearest centroid through iteration, and the centroid position is updated until the convergence condition is reached.

7. A power grid dispatching system based on massive power data mining, characterized in that: include: A data acquisition module is configured to: acquire electricity consumption data of residential users, and classify the residential users into two categories: residential household users and residential electric vehicle users; A clustering module is configured to: cluster the electricity consumption data of household users and classify them into different electricity consumption categories; cluster the electricity consumption data of electric vehicle users and classify them into different electricity consumption categories; The data mining module is configured to: cross-classify the electricity consumption categories of residential users with electricity price periods and peak and valley periods to obtain classification sets, and after grouping and aggregating the classification sets into transaction data sets based on the residential users, filter the frequent item sets based on the idea of ​​optimizing Apriori association relationship search based on the formulated screening rules and multi-minimum support strategy, and form association rules based on the items contained in the frequent item sets; The dispatching module is configured to formulate a power grid dispatching strategy based on association rules and dispatch the power grid.

8. A power grid dispatching system based on massive power data mining as claimed in claim 7, characterized in that: If the association rule is that the peak period associated with the stage where the time-of-use electricity price is not implemented is a high-energy consumption user, the grid dispatching strategy is to dispatch backup power sources or energy storage equipment in advance to ensure stable power supply during the peak period; If the association rule is that the second associated peak period of the time-of-use electricity price stage is a medium energy consumption user, then the grid dispatching strategy is to reduce power generation and transmission losses; If the association rule is that the deep valley section in the second stage of the time-of-use electricity price is associated with medium energy consumption users, the grid dispatching strategy is to dispatch renewable energy generation or energy storage equipment to generate or discharge in the deep valley section to balance the grid load.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a power grid dispatching method based on massive power data mining as described in any one of claims 1 to 6 are implemented.

10. A computer device comprising a computer-readable storage medium, a processor, and a computer program stored in the computer-readable storage medium and executable on the processor, characterized in that: When the processor executes the program, the steps in the power grid scheduling method based on massive power data mining as described in any one of claims 1-6 are implemented.

Images