Low-voltage user demand response control method, device and system based on knowledge graph

Abstract

The present invention discloses a method, device and system for regulating and controlling demand response of low-voltage users based on knowledge graph, wherein the method for constructing the knowledge graph includes: obtaining historical power data of multiple power devices for each user in a predetermined area, calculating the historical load power of each power device according to the historical power data, and then obtaining historical frequency data of different power adjustable margin levels; obtaining current power data of the power device, calculating the real-time load power and real-time power sequence fluctuation level of the power device according to the current power data, and calculating the real-time power adjustable margin level in combination with the historical load power, so as to determine the power adjustable margin level frequency corresponding to the real-time power adjustable margin level based on the historical frequency data; constructing a triple relationship table, and using a graphic database to generate a low-voltage user demand response knowledge graph based on the triple relationship table. The present invention can realize the intelligent recommendation function of the demand-side response strategy.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid optimization and dispatching, and relates to a demand-side response (Demand Response, DR) of low-voltage power users, and specifically to a low-voltage user demand response regulation method, device and system based on a knowledge graph. Background Art

[0002] With the rapid development of my country's economy, on the one hand, the demand for electricity is still gradually increasing, and the peak-to-valley difference in electricity consumption in urban or remote areas is becoming increasingly significant; on the other hand, residential power equipment is becoming increasingly intelligent, flexible and controllable. Therefore, electricity users take the approach of reducing loads in accordance with the power supply party's stimulus policies, change their inherent habitual electricity consumption patterns, and actively respond to demand, which can effectively improve the voltage quality of weak power grids and help the power grid operate safely and reliably.

[0003] At present, there have been a large number of related studies in the field of smart home appliance control for low-voltage users. These studies have optimized the load power from different angles by analyzing the energy consumption of household equipment, reducing the peak-to-valley difference of the load, and realizing active and effective response of the demand-side load. However, since the user's electricity consumption behavior is not considered in depth, and the optimization objects are mostly single household power devices, it is difficult to explore the potential relationship between different users and different devices. In addition, some studies have attempted to use knowledge graphs to solve the problem of demand-side response for low-voltage users, but existing studies select candidate users based on similar time periods or similar power, without considering the differences in user electricity consumption behavior, and have not quantified comprehensive recommendation indicators, making it difficult to achieve actual engineering applications. Summary of the invention

[0004] According to a first aspect of the present invention, a method for constructing a low-voltage user demand response knowledge graph is disclosed, comprising:

[0005] For each user in a predetermined area, historical power data of multiple power devices are obtained, and historical load power of each power device is calculated according to the historical power data, so as to obtain historical frequency data of different power adjustable margin levels;

[0006] Acquire current power data of the power device, calculate the real-time load power and real-time power sequence fluctuation level of the power device according to the current power data, and calculate the real-time power adjustable margin level in combination with the historical load power, so as to determine the power adjustable margin level frequency corresponding to the real-time power adjustable margin level based on the historical frequency data;

[0007] Construct a triple relationship table, and use a graph database to generate a low-voltage user demand response knowledge graph based on the triple relationship table; wherein the triple relationship table includes the belonging relationship between each user and the power equipment and the corresponding relationship between each power equipment and its real-time load power, real-time power sequence fluctuation level, real-time power adjustable margin level and power adjustable margin level frequency.

[0008] In some examples, the nodes of the low-voltage user demand response knowledge graph include user nodes, power device nodes, real-time load power nodes, real-time power sequence fluctuation level nodes, real-time power adjustable margin level nodes, and power adjustable margin level frequency nodes; the relationships of the low-voltage user demand response knowledge graph include: the belonging relationship between user nodes and power device nodes, the real-time load relationship between user device nodes and real-time load power nodes, the power quality relationship between user device nodes and real-time power sequence fluctuation level nodes, the schedulability relationship between user device nodes and real-time power adjustable margin nodes, and the electricity consumption behavior relationship between user device nodes and real-time power adjustable margin level frequency nodes.

[0009] In some examples, the current power data is power data collected by the smart meter at a predetermined sampling period in a latest demand response cycle; the historical power data is power data collected by the smart meter at a predetermined sampling interval in multiple demand response cycles.

[0010] In some examples, before calculating the historical frequency data of each power device operating at different power adjustable margin levels based on the historical power data, the historical power data is also repaired, and the repair includes: using one of adjacent time period completion, similar day replacement, interpolation fitting and similar device replacement to complete the data according to the data missing situation, and using adjacent time period replacement to process peak power.

[0011] In some examples, the graph database is Neo4j, and the triple relationship table is generated into a csv text and input into Neo4j to generate the low-voltage user demand response knowledge graph.

[0012] According to a second aspect of the present invention, a low-voltage user demand response regulation method is disclosed, comprising:

[0013] Generate a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of the above schemes;

[0014] The low-voltage user demand response knowledge graph is searched using a graphic query instruction to generate a power device regulation strategy; wherein the power device regulation strategy generated by the search includes the selected power devices and their regulation priority.

[0015] In some examples, the method further includes obtaining user search conditions, wherein the search conditions include the user's control priority information for the power device, and adjusting the power device control strategy generated by the search based on the control priority information.

[0016] According to a third aspect of the present invention, a low-voltage user demand response control device is disclosed, comprising:

[0017] A graph generation unit is configured to generate a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of the aforementioned solutions;

[0018] The search unit is configured to use graphic query instructions to search the low-voltage user demand response knowledge graph to generate a power device regulation strategy; wherein the power device regulation strategy generated by the search unit includes the selected power devices and their regulation priority.

[0019] In some examples, it also includes: an input interface configured to obtain user-defined search conditions, wherein the customized search conditions include the user's control priority information for the power device; and an adjustment unit configured to adjust the power device control strategy generated by the search unit based on the control priority information.

[0020] According to a fourth aspect of the present invention, a low-voltage user demand response control system is disclosed, comprising: a control server configured at a regional power supply end, and smart meters, smart sockets and a user demand response server configured at each user end in the region;

[0021] The smart meter is used to collect power data of each power device of the user in real time, and send it to the control server through the user demand response server;

[0022] The control server generates a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of the above schemes, and searches the low-voltage user demand response knowledge graph using a graphic query instruction to generate a power device control strategy, and then sends a demand response task instruction to the user demand response server;

[0023] The user demand response server receives the demand response task instruction and controls the switching of the power device through the smart socket.

[0024] The present invention establishes a comprehensive demand-side indicator system that takes into account power adjustable margin, user energy consumption behavior, power quality assessment, etc., and grades the indicator values ​​to facilitate the establishment, display and search of knowledge graphs. In addition, the knowledge graph established based on the triple relationship table associates various indicators. On the one hand, the internal relationship of the indicators can be visualized. On the other hand, the knowledge graph can input search indicators according to decision preferences, quickly and intuitively display search results, and realize the intelligent recommendation function of demand-side response. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0026] Figure 1 A schematic diagram of the composition of a low-voltage user demand response control system according to an embodiment of the present invention;

[0027] Figure 2 A schematic diagram of a method for constructing a low-voltage user demand response knowledge graph according to an embodiment of the present invention;

[0028] Figure 3 An example of a low-voltage user demand response knowledge graph (part) constructed using the present invention;

[0029] Figure 4 A schematic diagram of a low-voltage user demand response control method according to an embodiment of the present invention;

[0030] Figure 5 It is a schematic diagram of the composition of a low-voltage user demand response control device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0032] Terminology explanation:

[0033] Neo4j database: is a high-performance NoSQL graph database. Its storage structure is very similar to a graph in mathematics, with the relationship between nodes as the basic content of storage. The graph in Neo4j refers to a directed graph, and all relationships must be directed. Neo4j does not have the concept of sub-library and sub-table, and all nodes and relationships are stored in one library. Since there is no table structure, it is very suitable for storing semi-structured data.

[0034] Cypher language: It is Neo4j's declarative graph query language, which allows users to efficiently query graph data without having to write graph structure traversal code. Cypher is designed to be similar to SQL and is suitable for developers and professional operators who do ad-hoc queries on databases. Its capabilities include: creating, updating, and deleting nodes and relationships; querying and modifying nodes and relationships through pattern matching; managing indexes and constraints, etc.

[0035] Technical indicators of user demand side response

[0036] (1) Power adjustable margin

[0037] The power equipment on the demand side of the household has a large regulation potential. Through the active switching of the demand side equipment, it responds to the operation instructions of the power grid, reduces the peak and valley load, and improves the power quality of the power grid. To this end, it is necessary to first evaluate the user's adjustable potential. Taking multiple users in the same community as an example, the data collection cycle of the smart meter is 1 minute, and the demand response is generally based on 1 hour (a demand response cycle). Therefore, it is necessary to sum the power data every minute. The formula is as follows:

[0038]

[0039] In the formula, u represents the user number; n is the number of the power device; represents the hourly load power of the nth power device of the uth user; Indicates the real-time power data of the nth device of the uth user every 1 minute; Indicates the real-time switching status of the nth power device of the uth user, A value of 1 indicates that the switch is on, and 0 indicates that the switch is off.

[0040] In the actual use of power equipment, the rated power is the maximum power value allowed by the equipment, but due to the user's usage habits or the equipment's operating conditions, it is generally not operated at the rated power. To evaluate the power adjustable margin of the user's power equipment, it is necessary to evaluate it based on the user's historical power data, that is, to calculate it based on the maximum and minimum values ​​of the historical load power. The formula is as follows:

[0041]

[0042] In the formula, max and min are the historical load power respectively. The maximum and minimum values ​​of ; is the power adjustable margin of the nth device of the uth household, and its value range is [0,1].

[0043]

[0044] is the adjustable margin level of the current power. Generally, the higher the level, the larger the adjustable power value. ceil is the upward rounding function; R Q is the power margin level number.

[0045] (2) User energy consumption behavior indicators

[0046] Power adjustable margin level of power equipment It reflects the current power level of the equipment. The higher the value, the greater the power. The equipment is closer to full load operation, and the greater the power value that can be switched. Based on historical power data, the present invention counts the frequency of the equipment power adjustable margin level at each moment to indicate the user's energy usage habits for each power device. The formula is as follows:

[0047]

[0048] r q ∈[0,R Q ] and r q ∈N(5)

[0049] In the formula, For user u's device n in the rth q The frequency of the power adjustable margin level; r q is the current power adjustable margin level variable; T end is the total number of data moments; tabulate{} is the statistical moment function of the sequence.

[0050] (3) Power quality indicators

[0051] The equipment power of low-voltage users has a short startup characteristic, that is, a periodic pulse characteristic appears, with a peak value in a certain period, then a sudden drop, and then a flattening. The simultaneous start and stop of large-scale power equipment will have an adverse impact on the power quality of the low-voltage distribution network, and will also have an adverse effect on the operation of other low-voltage equipment. The present invention uses 1h as the statistical period, each minute as the sampling period, and a total of 60 sampling moments to calculate the power peak standard deviation of the power equipment:

[0052]

[0053] In the formula, is the standard deviation of the device's power peak value. The larger the value, the greater the power fluctuation of the device. mean() is the function for finding the average value of the sequence, that is, calculating the mean of the 60 power sampling values.

[0054] Then the calculation example of the power sequence fluctuation level used to represent the power quality is:

[0055]

[0056] It is the fluctuation level of the power sequence. The larger the value, the greater the fluctuation load of the power sequence and the worse the power quality. The smaller the value, the more stable the power and the better the power quality.

[0057] Figure 1 The figure is a schematic diagram of the composition of the low-voltage user demand response control system according to an embodiment of the present invention. As shown in the figure, the system 100 includes: a control server 101 configured at the regional power supply end (grid side), and smart meters 102, smart sockets 103, user demand response servers 104 and multiple power devices 105 configured at each user end in the region, such as the power devices 105, such as air conditioners, refrigerators, washing machines, water heaters and other household appliances.

[0058] The smart meter 102 is configured on the main line, and is responsible for collecting, measuring and transmitting the original electric energy data. It has a two-way data communication function, and actually collects the power data of each power device in real time and sends it to the user demand response server 104. In actual operation, smart meters mostly use 4G wireless communication, which is easily affected by the installation environment, resulting in poor communication and affecting data quality. Therefore, the present invention repairs the power data collected by the smart meter, and uses one of the adjacent time period completion, similar day replacement, interpolation fitting and similar device replacement to complete the data according to the data missing situation, and uses the adjacent time period replacement to process the peak power. Specifically including:

[0059] (1) When the collected power data is missing at a few moments (1 to 3 moments), the adjacent time period filling method is used to directly fill the power data of the missing time period with the power data of the adjacent time period.

[0060] (2) When data for a large number of consecutive moments (e.g., 4 to 10 moments) are missing, the method of similar day substitution is used to replace the load data of the same period of adjacent days. If data of the same period of adjacent days are still missing, the cubic spline interpolation function (e.g., the built-in function of Excel or Matlab) is used to fit and complete the missing data.

[0061] (3) When a large number of consecutive moments (for example, more than 20 moments or even several days) are missing, the method of using adjacent similar devices to complete the data is used. The power per unit value of adjacent users and the same type of equipment in the same time period is taken, and then multiplied by the load average value of the closest data day of this device to approximate the load data of the missing date.

[0062] (4) When an abnormal peak power occurs that exceeds the rated power of the equipment, the power data of the adjacent time period is also used as a substitute.

[0063] The present invention repairs and completes the data by adopting similar time period repair, similar day repair, adjacent load approximate replacement and cubic spline interpolation method based on data statistics, so as to ensure that the data is as close as possible to the original data in terms of change trend and value size.

[0064] Each of the power devices 105 is connected to the main line via the smart socket 103. The user demand response server 104 is connected to the smart meter 102 and the smart socket 103 for data on one hand, and is connected to the control server 101 for data on the other hand.

[0065] Figure 1 The data connection in is represented by a dotted line, which means that both wired connection and wireless connection can be used.

[0066] In the present invention, the control server 101 obtains the power data of the power equipment of each user in the area collected by the smart meter 102, and constructs a low-voltage user demand response knowledge graph based on the power data, and then searches the low-voltage user demand response knowledge graph to generate a power equipment control strategy, and sends a demand response task instruction to the user demand response server 104. The user demand response server 104 receives the demand response task instruction, and controls the switching of the power device 105 through the smart socket 103 to achieve low-voltage user demand response control.

[0067] Figure 2 A schematic flow chart of a method for constructing a low-voltage user demand response knowledge graph based on power data of the control server. As shown in the figure, the method includes the following steps:

[0068] Step 201: For each user in a predetermined area, historical power data of multiple power devices are obtained, and historical load power of each power device is calculated based on the historical power data, so as to obtain historical frequency data of different power adjustable margin levels.

[0069] For example, for power device n of user u, obtain continuous T endThe historical power data collected by the smart meter at a predetermined sampling period (e.g., 1 minute) within a demand response period T (e.g., 1 hour) According to the above formula (1), the load power of each demand response period T is calculated That is, we get T end Historical load power

[0070] For the T end Sort the historical load powers and get their maximum value and minimum value Then, according to the above formula (2), the power adjustable margin corresponding to each historical load power is calculated:

[0071] Based on the preset power adjustable margin level number R Q , calculate the power adjustable margin according to the above formula (3) Corresponding power adjustable margin level

[0072] Calculate R using the above formula (4) Q The proportion corresponding to each power adjustable margin level is obtained to obtain the frequency corresponding to each power adjustable margin level.

[0073] For other power devices of user u and various power devices of other users, calculations are performed according to the above method to obtain relevant historical data.

[0074] The above historical data is stored as a queryable historical data table for subsequent use. The historical data table contains the data in a predetermined historical time period (e.g., T end The load power of each power device of each user within a demand response cycle and its maximum value and minimum value Frequency and other data corresponding to each power adjustable margin level.

[0075] Step 202: obtain the current power data of the power device, calculate the real-time load power and real-time power sequence fluctuation level of the power device according to the current power data, and calculate the real-time power adjustable margin level in combination with the historical load power, so as to determine the power adjustable margin level frequency corresponding to the real-time power adjustable margin level based on the historical frequency data.

[0076] For example, the current power data collected by the smart meter at a predetermined sampling period (for example, 1 minute) within the most recent demand response period T (for example, 1 hour) is obtained. According to the above formula (1), the real-time load power of the power equipment is calculated And calculate the power peak standard deviation according to the above formula (6) (first calculate 60 current power data sampling values The mean Then, the power peak standard deviation is calculated according to the aforementioned formula (6), and the real-time power sequence fluctuation level is obtained according to the preset power sequence fluctuation level (such as the aforementioned formula (7)).

[0077] At the same time, the real-time power adjustable margin level is calculated according to the above formula (2) in combination with the historical load power. For example, the real-time load power is combined with the historical load power data, and the combined load power is sorted to obtain its maximum value. and minimum value Then, according to the above formula (2), the power adjustable margin corresponding to the real-time load power is calculated: Then, based on the preset power adjustable margin level number R Q , calculate the real-time power adjustable margin according to the above formula (3) Corresponding real-time power adjustable margin level Finally, the power adjustable margin level frequency corresponding to the real-time power adjustable margin level is determined based on the historical frequency data.

[0078] Step 203: construct a triple relationship table, and use a graph database to generate a low-voltage user demand response knowledge graph based on the triple relationship table.

[0079] The present invention obtains historical electricity load structured data and extracts demand-side response technical indicators as the basic elements of the knowledge graph, so there is no disambiguation and fusion process. When constructing various entity nodes and relationships, the present invention adopts the method of constructing user-side demand response triples, which can explicitly express the relationship between various entities. For example, the relationship structure between user number u and its power device number n is as follows:

[0080] Functional device n (head entity) relation User number u (last entity) Device 1, H11 device belong User 1, Household 01 in Community A Device 2, H12 device belong User 1, Household 01 in Community A Device 3, H13 device belong User 1, Household 01 in Community A Device 4, H14 device belong User 1, Household 01 in Community A

[0081] According to the structural relationship in the above table, the entity nodes and subordinate relationships of users and power devices are established in the Neo4j graph database. Then the frequency of the power adjustable margin level is calculated based on the historical power data. According to the current power data Calculate the power series fluctuation level To establish a triple relationship among users, power devices, and technical index results. Among them, the triple relationship table includes the ownership relationship between each user and the power device, and the corresponding relationship between each power device and its real-time load power, real-time power sequence fluctuation level, real-time power adjustable margin level, and power adjustable margin level frequency. For example, construct <device power >-(real-time load)-><H11 device>, indicating that the device power is the real-time load of the H11 device; <H11 device>-(load adjustability)-><adjustable margin level >, indicating the power margin of the current load adjustability of the H11 device

[0082] Exemplarily, generate a csv text of the triple relationship table and input it into Neo4j to generate the low-voltage user demand response knowledge graph. The nodes of the low-voltage user demand response knowledge graph include user nodes, power device nodes, real-time load power nodes, real-time power sequence fluctuation level nodes, real-time power adjustable margin level nodes, and power adjustable margin level frequency nodes; the relationships of the low-voltage user demand response knowledge graph include: the ownership relationship between user nodes and power device nodes, the real-time load relationship between user device nodes and real-time load power nodes, the power quality relationship between user device nodes and real-time power sequence fluctuation level nodes, the schedulability relationship between user device nodes and real-time power adjustable margin nodes, and the electricity consumption behavior relationship between user device nodes and real-time power adjustable margin level frequency nodes. Figure 3 It is an example of the low-voltage user demand response knowledge graph part.

[0083] Figure 4 It is a schematic flow chart of the low-voltage user demand response regulation method according to an embodiment of the present invention. As shown in the figure, the method includes:

[0084] Step 401, generate a low-voltage user demand response knowledge graph by using the above-mentioned low-voltage user demand response knowledge graph construction method.

[0085] Step 402, search the low-voltage user demand response knowledge graph by using a graphical query instruction to generate a power device regulation strategy.

[0086] Among them, the generated power device regulation strategy by the search includes the selected power devices and their regulation priority orders.

[0087] Construct the nodes and relationships of the knowledge graph in sequence according to the above method to generate a low-voltage user demand response knowledge graph. Then, use the query statement Cypher in the Neo4j graph database to search for the best electricity-consuming users and electricity-consuming devices according to the nodes.

[0088] For example, search according to the following two conditions: power adjustable margin is 5, use frequency The lowest node can search out users and their power equipment that indicate that the current load power value is sufficient and is not frequently used by the users, and is the demand-side response control strategy recommended by the current search conditions.

[0089] In some examples, the step further includes obtaining custom search conditions, wherein the custom search conditions include user regulation priority information for the power device, and adjusting the power device regulation strategy generated by the search based on the regulation priority information.

[0090] When selecting control equipment, in addition to the above technical indicators, other factors need to be considered, such as the importance of time periods. For example, the air conditioning load can be reduced or even shut down at night, which is relatively less important. However, the time-of-day electricity price is cheap at this time, and the energy consumption of charging piles or other charging equipment is more important. Therefore, users are allowed to set the control priority of various power equipment, that is, set custom search conditions:

[0091]

[0092] In the formula, represents the current priority ranking variable of n devices of user u; k is the ranking number of the device; N is the total number of power devices of user u.

[0093] The power device control strategy generated by the search is adjusted based on the control priority information, and finally a control scheme that meets the conditions is selected.

[0094] The technical solution of the present invention is further described below through an example.

[0095] Ten users in a low-voltage community A in a certain area are selected, named A community 01 to A community 010, and represented by u1 to u10 respectively; smart meters are used in each household to monitor the power data of 6 power devices, represented by device 1 to device 6 respectively. The data collection cycle of the smart meter is 1 minute, the collection time for each household is 1320 minutes, and each household has 6 power devices, so the data size is 10×1320×6=79200 data in total, and the cycle of demand-side response is 1 hour.

[0096] First, Matlab is used for programming and calculation. The data is read into Matlab in the form of csv text, and then stored and calculated in Matlab in the form of mat matrix. Then, data repair and completion are completed in Matlab, and blank data and outlier data are repaired. The power adjustable margin is calculated based on the repaired data. and power adjustable margin level The frequency of the power adjustable margin level is calculated using Matlab's statistical function tabulate The obtained power adjustable margin level frequency is shown in the following table:

[0097] u1 device 1 u1 device 2 u1 device 3 u1 device 4 u1 device 5 u1 device 6 Margin Level 1 0.788 0.205 0.078 0.969 0.303 0.848 Margin Level 2 0.068 0.295 0.068 0 0.129 0.045 Margin Level 3 0.045 0.356 0.045 0 0.098 0.015 Margin Level 4 0.053 0.128 0.061 0.007 0.114 0.054 Margin Level 5 0.045 0.015 0.038 0.024 0.356 0.038

[0098] In the above table, the higher the power adjustable margin level, the higher the load level and the larger the adjustable power margin. u1 represents user 1. From the table, it can be seen that the different types of power equipment of user 1 have different electricity usage habits. The margin level 1 of device 1 accounts for 0.788, indicating that it is light-loaded most of the time and the power margin available for switching is relatively insufficient. The margin level distribution of device 2 is relatively even, indicating that it can be dispatched at most times. The margin level 5 of device 5 accounts for 0.356, indicating that the device is running at a high load level for many times and the load capacity available for dispatch is relatively strong.

[0099] Real-time load power of each user and each device As shown in the following table (unit: W):

[0100] u1 u2 u3 u4 u5 u6 u7 u8 u9 u10 Device 1 98.738 15.864 16.278 45.236 22.784 20.694 41.464 24.614 54.298 34.006 Device 2 7.658 4.38 5.35 5.862 4.38 8.654 9.296 8.886 4.302 8.186 Device 3 406.2 69.6 72.2 189.2 99.2 95.8 177.6 114.2 226.8 145.4 Device 4 0 0 0 0 0 0 0 0 0 0 Device 5 20 0 0 3 11 22 42 41 40 59 Device 6 96 101 92 563 89 107 100 99 44 39

[0101] According to the current real-time load power Calculating the power adjustable margin of the load at this time and the power adjustable margin level to which it belongs can reflect the adjustable capacity of each load at this time; next, combined with the historical frequency of the power adjustable margin (the power adjustable margin level frequency table in the previous article), it can be determined whether the current real-time load is the power frequency commonly used by the user, that is, the frequency of the power adjustable margin level. If it is a commonly used device, when it is recommended as a cut-off load, it may violate the user's usage habits and reduce the user's energy satisfaction. Therefore, it is best to choose an uncommon frequency as much as possible.

[0102] Combined with the 60 power data collected within the current hour Calculate the peak power standard deviation for that hour and power series fluctuation level

[0103] After calculating various indicators in Matlab, a csv file is generated and used as the input text of the Neo4j graph database. The Neo4j graph database is used to write various triple relationship tables and construct various nodes and their relationships. The specific nodes and relationships are as follows: <device type>-(belongs to)-><user u>, indicating that various power devices belong to 10 different users; construct <device power> >-(real-time load)-><device type>, indicating the device power It is the real-time load of different power equipment; <Equipment Type>-(Load Adjustability)-><Power Adjustability Margin Level >, representing the adjustable power margin of the current load of devices with different powers Construct <H11 device>-(Power quality)-><Power sequence fluctuation level >, representing the power fluctuation quality of the current load of devices with different powers; construct <Adjustable power margin level frequency >-(Power consumption behavior)-><Device type>, representing the margin level frequency of different devices, which reflects the user's power consumption behavior.

[0104] There are many nodes created by the Neo4j graph database. For the convenience of display, taking the user 1 database image of some nodes (i.e., the low-voltage user demand response knowledge graph) as an example, the results are as Figure 3 shown.

[0105] In the constructed low-voltage user demand response knowledge graph, the Cypher query statement is used to search for power device nodes that meet the conditions. For example, first, filter out the nodes with a power margin level of 5, which can reflect the schedulability of the current power devices. The users and their power devices that meet the conditions are: Device 1 and Device 3 of User 1, Device 2 of User 2, Device 3 of User 3, and Device 6 of User 8.

[0106] Then, further filter the power margin level frequency that reflects the user's power consumption behavior The power sequence fluctuation level that reflects the power quality, and the filtering results are shown in the following table:

[0107] Power margin level Power margin level frequency Power sequence fluctuation level User 1 Device 1 5 <![CDATA[ 0.045 ]]> <![CDATA[ 2 ]]> User 1 Device 3 5 <![CDATA[ 0.038 ]]> <![CDATA[ 4 ]]> User 2 Device 2 5 0.242 1 User 2 Device 5 5 0.333 3 User 3 Device 3 5 <![CDATA[ 0.053 ]]> <![CDATA[ 3 ]]> User 8 Device 6 5 0.201 5

[0108] As can be seen from the above table, the adjustable power margin levels of the above 6 power devices are all 5, that is, the load level is running high and the schedulable capacity is strong. Further considering the power margin level frequency, the larger the frequency, the higher the frequency and dependence of the user on using this load power, and it is not suitable to cut off. Therefore, try to select the device with a small frequency. The power sequence fluctuation level reflects the power quality of the power device. The larger this value, the greater the load fluctuation and the worse the power quality. Therefore, try to select the power device with a large fluctuation level for cutting off to ensure the overall power quality.

[0109] According to the above principle, combining the power margin level frequency and the power sequence fluctuation, the final selected results and their control order are as follows: User 1 Device 3->User 3 Device 3->User 1 Device 1. Since the power margin level frequencies of the three are similar, they are sorted according to the power sequence fluctuation level for priority.

[0110] According to another aspect of the present invention, a low-voltage user demand response control device is disclosed. As Figure 5 shown, the device 500 includes:

[0111] The graph generation unit 501 is configured to generate a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any of the aforementioned schemes.

[0112] The search unit 502 is configured to use a graphic query instruction to search the low-voltage user demand response knowledge graph to generate a power device regulation strategy; wherein the power device regulation strategy generated by the search unit includes the selected power devices and their regulation priority.

[0113] In some examples, the apparatus 500 further includes:

[0114] The input interface 503 is configured to obtain a user-defined search condition, wherein the user-defined search condition includes the user's control priority information for the power device;

[0115] The adjustment unit 504 is configured to adjust the power device regulation strategy generated by the search unit based on the regulation priority information.

[0116] The present invention constructs a technical indicator system for demand-side response of low-voltage users, fully mines historical data information, reflects the controllable ability of user equipment through adjustable power margin and adjustable margin level, reflects the user's electricity consumption behavior through margin level frequency, reflects the user's power quality through equipment power peak standard deviation and power sequence fluctuation level, and defines the priority of power equipment according to the user's objective choice, and finally realizes the combination of multiple evaluation indicators, so as to more comprehensively evaluate the energy consumption of low-voltage users.

[0117] In addition, by establishing a triple relationship table of demand-side response, the present invention can relate the attributes of various indicators to each other, which is convenient for in-depth correlation analysis between indicators. At the same time, the joint programming method of Matlab and Neo4j graphic database is adopted to construct a knowledge graph of demand response of low-voltage users, forming a graph structure of user-load-value-evaluation index value-evaluation index level, and then the query statement Cypher can be used to provide selection search conditions according to decision preferences, which can display search results more intuitively and efficiently, and visualize various internal connections to quickly select demand-side response strategies. In addition, the strategy contains detailed users and their power equipment, as well as the priority order of regulation, which is convenient for practical engineering applications.

[0118] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing a knowledge graph of low-voltage user demand response, characterized in that: include: For each user in a predetermined area, historical power data of multiple power devices are obtained, and historical load power of each power device is calculated according to the historical power data, so as to obtain historical frequency data of different power adjustable margin levels; Acquire current power data of the power device, calculate the real-time load power and real-time power sequence fluctuation level of the power device according to the current power data, and calculate the real-time power adjustable margin level in combination with the historical load power, so as to determine the power adjustable margin level frequency corresponding to the real-time power adjustable margin level based on the historical frequency data; Construct a triple relationship table, and use a graph database to generate a low-voltage user demand response knowledge graph based on the triple relationship table; wherein the triple relationship table includes the belonging relationship between each user and the power device, and the corresponding relationship between each power device and its real-time load power, real-time power sequence fluctuation level, real-time power adjustable margin level and power adjustable margin level frequency; The nodes of the low-voltage user demand response knowledge graph include user nodes, power equipment nodes, real-time load power nodes, real-time power sequence fluctuation level nodes, real-time power adjustable margin level nodes and power adjustable margin level frequency nodes; the relationships of the low-voltage user demand response knowledge graph include: the belonging relationship between user nodes and power equipment nodes, the real-time load relationship between user equipment nodes and real-time load power nodes, the power quality relationship between user equipment nodes and real-time power sequence fluctuation level nodes, the schedulability relationship between user equipment nodes and real-time power adjustable margin nodes, and the electricity consumption behavior relationship between user equipment nodes and real-time power adjustable margin level frequency nodes.

2. The method for constructing a low-voltage user demand response knowledge graph according to claim 1, characterized in that: The current power data is the power data collected by the smart meter at a predetermined sampling period in the latest demand response cycle; the historical power data is the power data collected by the smart meter at a predetermined sampling interval in multiple demand response cycles.

3. The method for constructing a low-voltage user demand response knowledge graph according to claim 2, characterized in that: Before calculating the historical frequency data of each power device operating under different power adjustable margin levels based on the historical power data, the historical power data is also repaired; the repair includes: according to the data missing situation, one of adjacent time period completion, similar day replacement, interpolation fitting and similar device replacement is used to complete the data, and adjacent time period replacement is used to process peak power.

4. The method for constructing a low-voltage user demand response knowledge graph according to any one of claims 1 to 3, characterized in that: The graph database is Neo4j, and the triple relationship table is generated into a csv text and input into Neo4j to generate the low-voltage user demand response knowledge graph.

5. A low-voltage user demand response control method, characterized in that: include: Generate a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of claims 1 to 4; The low-voltage user demand response knowledge graph is searched using a graphic query instruction to generate a power device regulation strategy; wherein the power device regulation strategy generated by the search includes the selected power devices and their regulation priority.

6. The low-voltage user demand response control method according to claim 5 is characterized in that: The method also includes obtaining user search conditions, wherein the search conditions include the user's control priority information for the power device, and adjusting the power device control strategy generated by the search based on the control priority information.

7. A low-voltage user demand response control device, characterized in that: include: A graph generation unit is configured to generate a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of claims 1 to 4; The search unit is configured to use graphic query instructions to search the low-voltage user demand response knowledge graph to generate a power device regulation strategy; wherein the power device regulation strategy generated by the search unit includes the selected power devices and their regulation priority.

8. The low-voltage user demand response control device according to claim 7, characterized in that: Also includes: An input interface is configured to obtain a user-defined search condition, wherein the user-defined search condition includes the user's control priority information for the power device; An adjustment unit is configured to adjust the power device regulation strategy generated by the search unit based on the regulation priority information.

9. A low-voltage user demand response control system, characterized in that: include: The control server is configured at the regional power supply end, and the smart meters, smart sockets and user demand response servers are configured at each user end in the region; The smart meter is used to collect power data of each power device of the user in real time, and send it to the control server through the user demand response server; The control server generates a low-voltage user demand response knowledge graph using the low-voltage user demand response knowledge graph construction method according to any one of claims 1 to 4, and searches the low-voltage user demand response knowledge graph using a graphic query instruction to generate a power device control strategy, and then sends a demand response task instruction to the user demand response server; The user demand response server receives the demand response task instruction and controls the switching of the power device through the smart socket.

Images