Multi-scene adjustable power load management system based on layer visualization

Abstract

The invention discloses a multi-scene adjustable power load management system based on layer visualization. The multi-scene adjustable power load management system comprises a network load interaction engine and a load management service application, the network load interaction engine is used for processing services provided by a power grid graph and providing data and calculation services for upper-layer load management business application, and the map display and interaction assembly and the network load interaction engine are in butt joint with a marketing data center. User archive information, a hitching relation between a 10kV feeder line and a public and special transformer, a corresponding relation between the public and special transformer and a user, an industry category and power utilization category information are acquired, and a power utilization information acquisition system and a power sale market intelligent analysis and prediction system are connected; a visual page is provided for load management business application, and observable and measurable adjustable resources, adjustable and controllable load management, resource capability analysis and management, virtual power plant high-order application and photovoltaic resource high-order application are realized. According to the scheme of the invention, a novel power system with network-load interaction and a source moving along with the load is constructed.

Description

Technical Field

[0001] The present invention belongs to the field of power grid resource management, and in particular relates to a multi-scenario adjustable power load management system based on layer visualization. Background Art

[0002] The new power load management system played an important role in the battle to ensure power supply during the peak summer season. In terms of business, the new power load management system needs to integrate the main grid, distribution grid and demand-side resources to form an integrated management capability to support the construction of a new power system with grid-load interaction and source-load dynamics. The in-depth application of the new load management system requires the construction of load layering and zoning management capabilities, extensive access to various resources, strengthening classification and grading dynamic management and monitoring, promoting the connection between the new power load management system and the power grid, and continuously improving the precise regulation and control capabilities.

[0003] However, the current load management system uses the reuse of metering and pre-use resources as a transition, which is still far from the relevant management regulations and work requirements. There are still deficiencies in the construction of the load management system, the level of load resource management, the depth of hierarchical and zoning management, the effect of customer branch monitoring, the interactive ability of load management, and the system security protection ability. At present, the accurate resource topology relationship of "station-line-transformer-user" has not been established, and the potential of adjustable resources is not deeply explored, which cannot meet the local supply and demand adjustment and needs. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the present invention provides a multi-scenario adjustable power load management system based on layer visualization to solve the technical problems of hierarchical and regional management of load resources, mining the potential of adjustable resources, and local supply and demand regulation.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions.

[0006] The present invention first discloses a multi-scenario adjustable power load management system based on layer visualization, comprising: Grid-load interaction engine and load management business applications; The grid-load interaction engine is used to process the services provided by a grid map, provide data and computing services, map display and interactive public components for upper-level load management business applications, including grid-load data interaction, load resource aggregation, load analysis and calculation, and capability exposure services. The grid-load interaction engine serves as a unified data entry and exit, connects to the marketing data middle platform, obtains user profile information, the connection relationship between 10kV feeder and public-private transformer, the corresponding relationship between public-private transformer and user, industry category, electricity category information, and obtains metering point load data by connecting to the electricity information collection system, and connects to the power sales market intelligent analysis and prediction system to obtain distributed photovoltaic historical load and predicted load data; The load management business application is used to provide a visualization page to achieve observable and measurable adjustable resources, adjustable and controllable load management, resource capacity analysis management, high-level applications of virtual power plants, and high-level applications of photovoltaic resources. According to business functions and data requirements, the map service is called from the grid-load interaction engine to obtain corresponding data, or data is directly obtained from the system database.

[0007] The present invention further includes the following preferred embodiments: The power grid map is used to provide map and grid services, power grid topology analysis services, and spatial analysis services, connect to the power grid resource business middle platform, obtain device coordinate information or homologous system device location information, and grid topology relationships from the energy management system; connect to the enterprise-level measurement center to obtain switch measurement data; and connect to the meteorological center to obtain meteorological data.

[0008] The power load management system is further used to associate user information with substations, lines and transformer equipment in the power grid in different connection modes for users of different voltage levels.

[0009] For dedicated transformer users, the user access point should be at the end of the feeder, and a distribution transformer should be established inside the user; for dedicated line users, the user access point is located at the end of the feeder, and a distribution transformer is established inside the user; for low-voltage users, the user access point is located after the distribution transformer.

[0010] The system is further used for checking and managing the relationship between dedicated transformer users and dedicated line users; For the relationship between dedicated transformer users, check whether the relationship between the dedicated transformer on the marketing side and the middle station feeder is correct, and then re-maintain the relationship between the dedicated transformer and the feeder through the same-source maintenance tool. For the verification and management of dedicated line user relationships, check whether the relationship between the dedicated transformer and dedicated line on the marketing side and the substation on the middle station side is correct, and then re-maintain the relationship between the dedicated transformer and dedicated line and the substation through the same-source maintenance tool.

[0011] The network-load interaction engine further comprises: The archive management service module is used to provide storage, management and query services for archives related to grid-load interaction, covering demand response, orderly power consumption, and power conservation resource management libraries; The computing task management service module is used to receive, split, schedule and manage computing tasks, including user baseline calculation, actual load calculation and aggregation calculation.

[0012] The load resource aggregation includes the following processes: 1: Baseline computing task scheduling: 1.1: Split multi-day tasks into single-day tasks and generate single-day tasks in advance before the execution date; 1.2: Determine the execution order based on the date, priority, complexity, and number of users of the daily tasks; 1.3: Write the users and parameters required for the calculation of a single-day task into the Kafka message queue to be calculated. The calculation execution program only needs to query the load data. The number of tasks written into the message queue for a single day shall not exceed 3. 1.4: Scheduling tasks changes the task execution status according to the daily task execution progress; 1.5: Supports task calculation termination midway and supports single-day task recalculation; 2: Perform baseline calculations 2.1: The baseline calculation program supports multiple baseline algorithms, each of which has multiple parameters for adjusting the calculation rules; 2.2: Deploy the baseline calculation program in a multi-process cluster, and automatically search for the current task to be executed in conjunction with the scheduling task. Each progress multi-threads the data to be calculated to achieve high-speed baseline calculation; 2.3: Update the calculation status to redis in real time and update it to the database regularly to achieve real-time monitoring of the calculation process and status; 2.4: Provide a complete classification of the reasons for baseline calculation failure; 3: Computational task scheduling 3.1: Split multi-day tasks into single-day tasks; 3.2: In batch processing mode, the execution order is determined according to the date, priority, complexity, and number of users of historical tasks; 3.3: Scheduling tasks changes the task execution status according to the daily task execution progress; 3.4: Supports task calculation termination midway and supports single-day task recalculation; 4: Perform aggregate calculations: 4.1: Execute historical load and baseline load task by task in batch mode; 4.2: Real-time aggregation tasks are processed in parallel in groups. Tasks in the same group are processed in the same computing cluster. Small and medium-sized tasks use the 1×n method, where n is high availability and a single server can perform aggregation calculations of all tasks. Large-scale tasks use the (1×m+1)×n method, where m is multi-task group aggregation and 1 is summary aggregation. 4.3: The load of the day and the previous day at the quarter hour, and the load of the last two hours in the minute are initially obtained from HBase and cached in memory. Subsequently, the cache is updated from the Kafka message queue in a streaming manner and cached in memory. The archive data is initialized from the database and updated regularly. 4.4: In aggregate computing, streaming incremental updates or batch real-time computing are used according to the complexity; 4.5: Use multi-threading to perform aggregate calculations; 5: Execution Assessment 5.1: Cache the user's real-time load in memory, obtain it from HBase during initialization, and update it from the message queue later for sharing by all evaluation tasks; 5.2: Standardize the input and return parameters of the evaluation function, and use the jar upload and dynamic loading method for the evaluation implementation class; 5.3: Store the evaluation results into the database according to the configuration; 6: Perform data extraction: 6.1: Multiple nodes perform data extraction step by step. According to the dependency relationship between nodes, the subsequent nodes are executed after the previous nodes are executed; 6.2: The node executes the preceding and subsequent database operations. The main processing function implements cross-database data query and storage. The main processing function supports multi-threading, and the query results support js function processing; 6.3: Monitor the process and restore the failed nodes to the pending state after processing; 7: General interface service: 7.1: Generate data query interface through configuration; 7.2: By executing the chain, the results of the previous query are used as the parameters of the subsequent query, thus achieving flexible data query; 7.3: Assemble the query results into the expected JSON format through configuration.

[0013] The multiple baseline algorithms include: Baseline calculation algorithm: Get the user to be processed; traverse the rule array with the date of the daily baseline task and the user's working day attribute as parameters, and use the judgment function of the algorithm class corresponding to the rule to determine whether the current date to be processed meets the rule. If it meets the rule, use the calculation function of the algorithm class corresponding to the rule to calculate the baseline of the day, and the loop ends; Working day algorithm: determine normal days based on the number of days in advance, the maximum recursive number of days, and the user's abnormal day information; start a loop to try to calculate the baseline with the maximum number of samples as the initial value, and end the loop if successful; pull data for the number of days of samples from hbase, first make rational points for each device in the time period, if there is no data or too many defects, the data of this device on that day is invalid, and the data of this user on that day is invalid; use the data of each device to calculate the data of each sampling point of the user on that day; calculate the mean value of each day in the time period; fix the left window for the calculated mean, slide the right window, and evaluate whether there are samples to be removed. If there are samples to be removed, slide the right window to add new samples and evaluate again. If the sample set boundary is reached, query hbase again to pull a batch of samples. If all samples are traversed and there is no result, the round of calculation ends; if the required number of samples is reached, use qualified samples to calculate the baseline value. When calculating the sample baseline mean, the maximum or minimum sample needs to be removed according to the configuration; Holiday algorithm: Find the same holiday as last year and remove the response day. If it is empty, find the same holiday as the previous year and remove the response day; find the best matching sample day among the remaining sample days; determine whether it is necessary to correct according to capacity, and correct it if necessary; if no sample day is found, use the non-working day baseline calculation method; if the baseline calculation does not find a sample set of load conditions, there is no baseline value, and the application needs to be removed from the candidate list, and the aggregate statistics do not count the value of -1; Equipment missing point interpolation algorithm: Determine whether the missing point ratio exceeds the standard within the time period. If it exceeds the standard, the device’s daily sample will be invalidated, and the user’s daily sample will be invalidated; find all sampling segments with continuous defects, and for each sampling segment, find the two points closest to this sampling segment and record them as (x1, y1), (x2, y2). The defect is (x, y), y=y1+(y2-y1) / (x2-x1)×(x-x1).

[0014] Accordingly, the present application also discloses a terminal, including a processor and a storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the multi-scenario adjustable power load management system based on the aforementioned layer visualization.

[0015] Correspondingly, the present application also discloses a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the aforementioned multi-scenario adjustable power load management system based on layer visualization is implemented.

[0016] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a multi-scenario adjustable power load management system based on layer visualization, based on the power grid resource business middle platform and the basic capabilities of the power grid one map, through the attachment of load resources and data quality management, to realize the station-line-transformer-household-load model construction, based on the efficient multi-scenario load aggregation computing capability, to realize the on-map management and analysis of resources such as adjustable loads, distributed photovoltaics, charging stations, user-side energy storage, etc., to support the construction of a new power system with grid-load interaction and source following load. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the marketing archive management process in the present invention.

[0018] Figure 2 It is a schematic diagram of the user-feeder-dedicated public transformer management process in the present invention.

[0019] Figure 3 It is a schematic diagram of data storage of the network-load interaction engine in the present invention.

[0020] Figure 4 It is a schematic diagram of the interaction between the network-load interaction engine and the application layer in the present invention.

[0021] Figure 5 It is a schematic diagram of data query configuration in the present invention.

[0022] Figure 6 It is a schematic diagram of the execution chain configuration in the present invention.

[0023] Figure 7 It is a schematic diagram of the execution environment configuration in the present invention.

[0024] Figure 8 It is a schematic diagram of the return parameter configuration in the present invention.

[0025] Fig. 9 It is a logical diagram of the adjustable load resource calculation in the present invention.

[0026] Fig.10 It is a schematic diagram of the business logic of the aggregation computing interface in the present invention.

[0027] Fig.11 It is a schematic diagram of the logical interaction of the aggregate computing interface in the present invention.

[0028] Fig.12 It is the overall architecture diagram of the present invention.

[0029] Fig.13 It is the business architecture diagram in the present invention.

[0030] Fig.14 It is a data architecture diagram in the present invention.

[0031] Fig.15 It is a schematic diagram of the load layer home page in the present invention.

[0032] Fig.16 It is a schematic diagram of user resource management in the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present invention more clear, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0034] The embodiments described in this application are only some embodiments of the present invention, not all embodiments. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the protection scope of the present invention.

[0035] In view of the deficiencies in the prior art, the present invention proposes a multi-scenario adjustable power load management system based on layer visualization, including a grid-load interaction engine and a load management business application.

[0036] The grid-load interaction engine is used to process the services provided by the grid map, provide data and computing services, map display and interaction and other public components for upper-layer applications, and provide data and computing services, map display and interaction and other public components for load management business applications, including grid-load data interaction, load resource aggregation, load analysis and calculation, capability exposure services and other functions; the grid-load interaction engine serves as a unified data entry and exit, connects to the marketing data middle platform, obtains user profile information, the connection relationship between 10kV feeder and public-private transformer, the corresponding relationship between public-private transformer and user, industry category, electricity category and other information. Connect to the electricity information collection system to obtain metering point load data; connect to the power sales market intelligent analysis and prediction system to obtain distributed photovoltaic historical load and forecast load data.

[0037] The load management business application is used to realize the construction of diversified business scenarios for load management, create personalized and characteristic applications, and provide visualization pages; realize the functions of observable and measurable adjustable resources, adjustable and controllable load management, resource capacity analysis and management, high-level applications of virtual power plants, and high-level applications of photovoltaic resources. According to business functions and data requirements, the map service is called from the grid-load interaction engine to obtain corresponding data, or data is directly obtained from the system database.

[0038] Based on the current situation of station-line-transformer-user resource connection in a power grid diagram, the present invention connects the load resources to the power grid frame, providing a basis for subsequent aggregate calculation of load resources based on the grid structure and visual display of superimposed layers.

[0039] The grid map integrates data from various business systems of the grid, including equipment parameters, asset ledgers, topological relationships, real-time measurements, etc., to digitally model the physical grid of all voltage levels and all links of "generation, transmission, transformation, distribution, and use". It forms a "one map, multiple layers, and multiple states" homologous maintenance and comprehensive application model with the grid topology as the core, providing graphical services for the production and operation of the grid. The connection between users and station-line transformers refers to the association of user information with equipment such as substations, lines, and transformers in the grid. This association enables the grid system to grasp the user's power consumption in real time, including power consumption, power load, etc., and can accurately determine the scope of the fault based on the topological structure and measurement information of the grid, thereby improving the efficiency of fault handling. Currently, power employees use homologous maintenance tools to connect users to access points.

[0040] According to the actual power supply plan, different connection methods are used for users of different voltage levels. For dedicated transformer users, the user access point should be at the end of the feeder (public line), and a distribution transformer should be built inside the user; for dedicated line users, the user access point should be at the end of the feeder (dedicated line), and a distribution transformer should be built inside the user; for low-voltage users, the user access point should be after the distribution transformer.

[0041] In view of the accuracy of marketing archive data, the present invention conducts verification and management on marketing archive data, mainly including user-special transformer relationship, special transformer-dedicated line relationship, such as Figure 1 As shown. The user topology verification tool is used to export the missing list of "user-dedicated transformer" relationships (see relationship ① in the figure) and the list of "dedicated transformer-dedicated line" relationships (see relationship ② in the figure). The district and county company power inspection personnel will verify and confirm the data, and the same source maintenance tool will be used to correct the data.

[0042] In view of the accuracy of user-feeder-dedicated public transformer, the "dedicated transformer-public line" and "dedicated line-substation" data are verified and managed. Figure 2 The user-feeder-dedicated public transformer management process is shown. The user topology verification tool is used to export the mismatch list between the dedicated transformer on the marketing side and the distribution transformer on the middle platform side, and the missing list of the "dedicated line-substation" relationship. The district and county company power inspection personnel will verify and confirm, and use the same source maintenance tool to attach the "user-dedicated transformer" relationship to the public line of the power grid resource business middle platform, and attach the dedicated line resources to the substation of the power grid resource business middle platform.

[0043] The specific governance solutions for marketing data and homologous data are as follows: (1) Verification and management of the relationship between the dedicated transformer and the user. Check whether the relationship between the dedicated transformer on the marketing side and the feeder on the middle station is correct, and then re-maintain the relationship between the dedicated transformer and the feeder through the same source maintenance tool. (See Figure 2 Relationship ①) 1) If the information of the special transformer on the marketing side is missing on the middle station side, the power inspection personnel of the marketing district and county will use the same source maintenance tool to add a new special transformer file and maintain the relationship between the special transformer and the feeder.

[0044] 2) If the information of the dedicated transformer on the marketing side matches that on the middle station side, the power inspection personnel in the marketing district and county will confirm whether the relationship between the dedicated transformer and the feeder on the middle station is correct. If not, the relationship between the dedicated transformer and the feeder will be re-maintained through the same-source maintenance tool.

[0045] 3) If the dedicated transformer on the central station side is missing on the marketing side, the power inspection personnel of the marketing district and county will confirm whether the dedicated transformer has been removed. After confirmation, the dedicated transformer file will be deleted through the same source maintenance tool. If the dedicated transformer is operating normally, the new dedicated transformer creation process will be carried out through the marketing system, and the marketing side dedicated transformer ID will be re-associated in the same source system.

[0046] (2) Verification and management of the relationship between dedicated lines and users. Verify whether the relationship between the dedicated line on the marketing side and the substation on the middle platform side is correct, and then re-maintain the relationship between the dedicated line and the substation through the same source maintenance tool. (See Figure 2 Relationship②③) 1) If the dedicated line information on the marketing side is missing on the middle office side, the marketing district and county electricity inspection personnel will use the same source maintenance tool to add a new dedicated line file and maintain the relationship between the dedicated line and the substation.

[0047] 2) If the dedicated line information on the marketing side matches the middle office side, the marketing district and county electricity inspection personnel will confirm whether the relationship between the middle office dedicated transformer dedicated line and the substation is correct. If not, the relationship between the dedicated transformer dedicated line and the substation will be re-maintained through the same source maintenance tool.

[0048] 3) If the dedicated line on the central office side is missing on the marketing side, the marketing district and county power inspection personnel will confirm whether the dedicated line has been removed. After confirmation, the dedicated line file will be deleted through the same source maintenance tool. If the dedicated line is running normally, the dedicated line dedicated change new process will be carried out through the marketing system, and the marketing side line ID will be re-associated in the same source system.

[0049] The purpose of load aggregation calculation based on grid-load interaction is to improve the operating efficiency and stability of the power grid by integrating and optimizing load resources, and to achieve optimal configuration and scheduling of load resources.

[0050] The network-load interaction engine adopts a master-slave synchronization architecture to ensure data consistency and system stability. The archive management service module is used to provide storage, management and query services for network-load interaction related archives, covering resource management libraries such as demand response, orderly power consumption, and power conservation. The computing task management service module is responsible for receiving, splitting, scheduling and managing computing tasks, including user baseline calculations, actual load calculations, and aggregate calculations.

[0051] The database design uses Hbase database and MySQL database. Hbase database is used to store real-time aggregation results and provide efficient read and write performance. MySQL database is divided into load data independent database and application database. The load data independent database stores load-related data such as user quarter-hour data, while the application database stores other application-related data. Figure 3 It is used to store data for the network interaction engine. The service process is as follows: The data receiving service module receives data from different sources, such as load data, user information, etc. The general data query service module provides a general query interface for data, supporting data query requirements in different scenarios. The data extraction and orchestration execution service module is responsible for the data extraction, conversion and loading process to ensure the accuracy and consistency of the data. The streaming computing service module uses streaming computing technology to calculate and aggregate the user's real-time load and provide real-time load data. The calculation result query service module provides a query interface for the calculation results, supporting the query of historical aggregation results and real-time aggregation results. The computing task receiving service module receives computing task requests from the outside. The back-end computing task delivery service splits the task according to the task requirements and delivers it to the corresponding computing nodes for execution.

[0052] The grid-load graph provides a visual display of grid-load interaction, helping users to intuitively understand the situation of grid-load interaction. The unified statistical program is used for the planned development of a unified statistical program to collect statistics and analyze relevant data on grid-load interaction. The virtual power plant library is used to store and manage relevant data on virtual power plants, including the composition, operating status, and scheduling strategies of virtual power plants. The grid-load interaction engine architecture design takes into account the scalability of the system and can support the needs of future new functions and services. Figure 4 It shows the interaction between the network load interaction engine and the application layer.

[0053] Load resource aggregation based on grid-load interaction includes the following processes: 1: Baseline computing task scheduling: 1.1: Split multi-day tasks into single-day tasks and generate single-day tasks in advance before the execution date; 1.2: The execution order is determined based on the date of the single-day task (the switch determines whether the earlier task or the latest task is executed first), priority, complexity, and number of users. For example, the default is to execute in reverse order of date, with the same date and higher priority first, and the result value order of complexity × number of users (smaller value is executed first), and the execution order can be controlled by the user; Preferably, the execution order calculation formula is: Date serial number × 10000 + priority serial number (priority serial number on the same day, if it exceeds 100, it will be normalized to a serial number between 1-100) × 100 + calculation amount serial number (calculation amount serial number on the same day with the same priority, if it exceeds 100, it will be normalized to a serial number between 0-100) 1.3: Write the users and parameters required for the calculation of the daily task into the message queue to be calculated (kafka). The calculation execution program only needs to query the load data. The number of daily tasks written into the message queue shall not exceed 3 to avoid abnormal count values ​​of individual task execution progress causing the calculation process to be stuck, and to avoid too many tasks being piled up in the message queue. 1.4: Scheduling tasks changes the task execution status according to the daily task execution progress; 1.5: Supports midway termination of task calculation and recalculation of single-day tasks.

[0054] 2: Perform baseline calculations 2.1: The baseline calculation program supports multiple baseline algorithms, and each baseline algorithm has rich parameters for adjusting the calculation rules to meet changing business needs.

[0055] Baseline calculation algorithm: Get the user to be processed; traverse the rule array with the date of the daily baseline task and the user's working day attribute as parameters, and use the judgment function of the algorithm class corresponding to the rule to determine whether the current date to be processed meets the rule. If it meets the rule, use the calculation function of the algorithm class corresponding to the rule to calculate the baseline of the day, and the loop ends; Working day algorithm: determine normal days based on the number of days in advance, the maximum recursive number of days and the user's abnormal day information; start the loop with the maximum number of samples as the initial value (to 1) to try to calculate the baseline, and end the loop if successful; pull the data of the sample number (or more) days from hbase (according to the sample set selection method), first make reasonable points for each device in the time period, if there is no data or too many defects, the data of this device on that day is invalid, and the data of this user on that day is invalid; use the data of each device to calculate the data of each sampling point of the user on that day; calculate the mean value of each day in the time period; fix the left window for the calculated mean, slide the right window (the initial window range is the number of samples), and evaluate whether there are samples to be removed. If there are samples to be removed, slide the right window to add new samples after removal and then evaluate again. If the sample set boundary is reached, query hbase again to pull a batch of samples (not exceeding the normal day set). If all samples are traversed and there is no result, the round of calculation ends. If the required number of samples is reached, the baseline value (maximum and mean) is calculated using qualified samples. When calculating the sample baseline mean, the maximum or minimum sample needs to be removed according to the configuration; Holiday algorithm: Find the same holiday as last year and remove the response day. If it is empty, find the same holiday in the previous year (up to the maximum number of years) and remove the response day; find the most matching sample day (proportionally) among the remaining sample days; determine whether it is necessary to correct according to capacity, and if necessary, correct it (m to n days before the calculation date / m to n days before the sample year); if no sample day is found, use the non-working day baseline calculation method; if the baseline calculation does not find a sample set of load conditions, there is no baseline value (set to -1), and the application needs to be removed from the candidate list, and the aggregate statistics do not count the value of -1.

[0056] Equipment missing point interpolation algorithm: Determine whether the missing point ratio exceeds the standard within the time period. If it exceeds the standard, the equipment sample of the day will be invalidated, and the user sample of the day will be invalidated; find all sampling segments with continuous defects, and find the two points closest to each sampling segment, recorded as (x1, y1), (x2, y2), the defect is (x, y), y=y1+(y2-y1) / (x2-x1)×(x-x1) 2.2: Deploy the baseline calculation program in a multi-process cluster, and automatically search for the current task to be executed in conjunction with the scheduling task. Each progress multi-threads the data to be calculated to achieve high-speed baseline calculation; 2.3: Update the calculation status to redis in real time and update it to the database regularly to achieve real-time monitoring of the calculation process and status; 2.4: Provide a complete classification description of the causes of baseline calculation failure to improve maintainability; 3: Computational task scheduling 3.1: Split multi-day tasks into single-day tasks; 3.2: In batch mode, the execution order is determined based on the date of historical tasks (the switch determines whether to execute the earlier task or the most recent task first), priority, complexity, and number of users; 3.3: Scheduling tasks changes the task execution status according to the daily task execution progress; 3.4: Supports task calculation termination midway and supports single-day task recalculation; 4: Perform aggregate calculations 4.1: Execute historical load and baseline load task by task in batch mode; 4.2: Real-time aggregation tasks are processed in parallel in groups. Tasks in the same group (the user set includes users of tasks in the same group) are performed in the same computing cluster. Small and medium-sized tasks (less than 2 million users) use the 1×n method, where n is the number of highly available services. Aggregate calculations of all tasks can be performed in a single server. Large-scale tasks (more than 2 million users) use the (1×m+1)×n method (m is multi-task group aggregation, and 1 is summary aggregation).

[0057] 4.3: The load of the day and the previous day at the quarter hour, and the load of the last two hours in the minute are initialized and cached from HBase (OTS) to the memory. Subsequently, the cache is updated from the message queue (Kafka) in a streaming manner to the memory. The archive data is initialized from the database and updated regularly. 4.4: In aggregate computing, streaming incremental updates or batch real-time computing are used according to the complexity; 4.5: Use multi-threading to perform aggregation calculations.

[0058] 5: Execution Assessment 5.1: Cache the user's real-time load in memory, obtain it from HBase during initialization, and update it from the message queue later for sharing by all evaluation tasks; 5.2: Standardize the input and return parameters of the evaluation function, and use the jar upload and dynamic loading method for the evaluation implementation class; 5.3: Store the evaluation results into the database according to the configuration; 6: Perform data extraction 6.1: Multiple nodes perform data extraction step by step. According to the dependency relationship between nodes, the subsequent nodes are executed after the previous nodes are executed; 6.2: The node executes the preceding and subsequent database operations. The main processing function implements cross-database data query and storage. The main processing function supports multi-threading, and the query results support js function processing; 6.3: Monitor the process and restore the failed nodes to the pending state after processing; 7: General interface service 7.1: Generate data query interface through configuration. Figure 5 A data query configuration is shown.

[0059] 7.2: By executing the chain, the results of the previous query are used as parameters of the subsequent query, thereby achieving flexible data query. Figure 6 and Figure 7 The execution chain configuration and the execution environment configuration are shown respectively.

[0060] 7.3: Assemble the query results into the expected JSON format through configuration; Figure 8 The return parameter configuration is shown.

[0061] The present invention covers the key elements such as the reverse index mechanism of grid data, aggregate calculation mode, preprocessing of calculation data, data aggregate calculation process, hierarchical streaming aggregation technology, data loading strategy and processing of user sets in different scenarios for the connection and aggregation calculation method of adjustable load resources in multiple scenarios. These elements constitute the core architecture of the load aggregation calculation algorithm based on grid-load interaction, such as Fig. 9 shown.

[0062] The business logic of the aggregate calculation query interface is as follows Fig.10 . Determine the query method to be used based on the event number, date, primary dimension level, attribute, and label. In the static_job table (associated with static_task and calcu_task), check whether there is a statistical task for the corresponding event number and date. If there is no statistical task, an error message is returned. If there is a statistical task, you need to determine which query method to use: 1. If the date is historical and the attribute is in the statistical attributes (attributes currently only support industries of different levels), it means large granularity, then it will be directly returned after querying from the statistical results of hbase; 2. If the date is historical and the attribute is not included in the statistical attributes, indicating small-granularity statistics, first use the industry (level 2 and below) and label to filter out the user set, use the metering points in the user set to query the load data from HBase, and then aggregate the required main dimensions according to the main dimensions to which the users belong; 3. If the date is real-time and the attribute is in the statistical attributes (attributes currently only support industries of different levels), indicating a large granularity, it will be directly returned after querying from the statistical results of real-time aggregate calculation; 4. If the date is real-time and the attribute is not included in the statistical attributes, it indicates small-granularity statistics. In this case, the user set is first filtered out by industry (level 2 and below) and tags, and the metering points in the user set are used to obtain the real-time load data in the cache. Then, the required primary dimensions are aggregated and counted based on the primary dimensions to which the users belong.

[0063] Fig.11The aggregate computing interface logical interaction is shown.

[0064] The demand-side resource visualization layer display is based on a map of the power grid, superimposing various types of load resource layers, constructing a map of demand-side adjustable load resources and a network of wind, solar, storage, charging and discharge, so as to realize the management of load resources on the map.

[0065] Based on the construction of a new type of power load management system, it interacts with business and data internally with functional modules such as virtual power plant operation management, demand response organization and implementation, and externally with the marketing data middle platform, electricity consumption information collection system, and power sales market intelligent analysis and prediction system.

[0066] Relying on the basic technical service capabilities of the power grid map, two functional modules are constructed: the grid-load interaction engine and the load management business application. Among them, the power grid map is used to provide technical service capabilities such as map and grid services, power grid topology analysis services, and spatial analysis services. The power grid map connects to the power grid resource business middle platform to obtain equipment coordinate information or homologous system equipment location information and grid topology relationships from the energy management system; connects to the enterprise-level measurement center to obtain switch measurement data; and connects to the meteorological center to obtain meteorological data. The overall architecture diagram is as follows Fig.12 . The grid-load interaction engine is used to process the services provided by a grid map, provide data and computing services, map display and interaction and other public components for upper-level applications, and provide data and computing services, map display and interaction and other public components for load management business applications, including grid-load data interaction, load resource aggregation, load analysis and calculation, capability exposure services and other functions; load management business applications are used to realize the construction of diversified business scenarios for load management, create personalized and characteristic applications, and provide visualization pages; realize the observable and measurable adjustable resources, adjustable and controllable load management, resource capacity analysis and management, virtual power plant high-level applications, photovoltaic resource high-level applications and other functions. The application architecture diagram is as follows Fig.13 .

[0067] In addition, the grid-load interaction engine serves as a unified data entry and exit, connecting to the marketing data middle platform to obtain user profile information, the connection relationship between 10kV feeders and public and private transformers, the correspondence between public and private transformers and users, industry categories, electricity consumption categories and other information; connecting to the electricity consumption information collection system to obtain metering point load data; connecting to the power sales market intelligent analysis and prediction system to obtain distributed photovoltaic historical load and predicted load data.

[0068] Load management business applications, based on business functions and data requirements, call map services from the network-load interaction engine and obtain corresponding data, or they may directly obtain data from the system database. Fig.14 .

[0069] As an application module, the new power load management system is integrated with the business middle platform to push the load resource files that need to be attached, obtain the grid topology relationship, meteorological data, and the power grid one-map layer display components; integrate with the marketing professional external system, the marketing data middle platform, the electricity consumption information collection system, and the power sales market intelligent analysis and prediction system; integrate with the new power load management system internally to obtain the business data that needs to be calculated and displayed. The integration content is as follows:

[0070] For different types of load resources, we build multiple layers display capabilities such as adjustable load resources, distributed photovoltaics, charging and swapping stations, user-side energy storage, data center diesel generation, and adjustable loads. Fig.15 For one Figure 1 Home page.

[0071] For the dot positioning of demand-side resources on the map, the coordinates are obtained according to the device associated with the user, and different icons are added to distinguish different resource types. Fig.16 Manage user resources.

[0072] Typical business scenarios include response resource mining under key sections, layered and zoned load resource monitoring, demand response organization under section gaps, load heat maps, and electricity saving for residents in heavily overloaded areas.

[0073] In the response resource mining under key sections, it is necessary to conduct fast forward and reverse retrieval based on dynamic topology and user profile information in a grid-load holographic map according to the grid topology data and the connection relationship between the grid topology and high-voltage electricity users.

[0074] In the layered and zoned load resource monitoring, the matching relationship between the power grid and power users is used to aggregate and calculate and analyze the user load and regulation capacity under the section, line, and site to form the total adjustable capacity of the section, line, and site. Specifically, based on the power grid topology data and the connection relationship between the power grid topology and high-voltage power users, the dynamic topology and user profile information are quickly forwarded and reversely retrieved in the grid-load hologram, and the real-time load and adjustable capacity of high-voltage users and their adjustable resources are aggregated and calculated.

[0075] In the demand response organization under the section gap, the demand response includes global gap response and local section gap response. The demand response organization under the section gap uses the "station-line-transformer-user" topological relationship of the grid load holographic one map to accurately screen out the demand response users under the section and site, and quickly organize these users to participate in the market-oriented response on the demand side, guide users to change their electricity consumption behavior, and then reduce the electricity load under the peak section to alleviate the gap pressure. Specifically, according to the grid topology data, the connection relationship between the grid topology and high-voltage electricity users, the grid load holographic one map is used to perform fast forward and reverse retrieval based on dynamic topology and user profile information.

[0076] In the load heat map and electricity conservation for residents in heavily overloaded areas, the load heat map is drawn using the "station-line-transformer-household" geographic data and load data, which will intuitively display the load distribution and regional load center, and help formulate strategies for electricity conservation; at the same time, the network load holographic one map is used to monitor and display the load rate of the area in real time, which will help screen users in heavily overloaded areas and accurately send users warm reminder text messages to save electricity. Specifically, the load data is superimposed on the network load holographic one map, rendered in the form of a heat map, and the network load holographic one map is used to monitor heavily overloaded areas and search for users in the area, while displaying the coordinates of the area, the relationship between the area and the user, the load rate of the area, and the heavy overload information.

[0077] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a multi-scenario adjustable power load management system based on layer visualization, based on the power grid resource business middle platform and the basic capabilities of the power grid one map, through the attachment of load resources and data quality management, to realize the station-line-transformer-household-load model construction, based on the efficient multi-scenario load aggregation computing capability, to realize the on-map management and analysis of resources such as adjustable loads, distributed photovoltaics, charging stations, user-side energy storage, etc., to support the construction of a new power system with grid-load interaction and source following load.

[0078] Based on the spirit of the present invention, those skilled in the art can easily think that a computer program product can be obtained based on the aforementioned multi-scenario adjustable power load management system based on layer visualization. The computer program product may include a computer-readable storage medium, which carries computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure. That is, the present application also includes a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the aforementioned multi-scenario adjustable power load management system based on layer visualization.

[0079] Computer readable storage medium can be a tangible device that can keep and store the instructions used by the instruction execution device. Computer readable storage medium can be, for example, - but not limited to - electrical storage device, magnetic storage device, optical storage device, electromagnetic storage device, semiconductor storage device or any suitable combination of the above. More specific examples (non-exhaustive list) of computer readable storage medium include: portable computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as punch card or groove protrusion structure with instructions stored thereon and any suitable combination of the above. Computer readable storage medium used here is not interpreted as instantaneous signal itself, such as radio wave or other free propagating electromagnetic wave, electromagnetic wave propagated by waveguide or other transmission medium (for example, light pulse by optical fiber cable) or electrical signal transmitted by wire.

[0080] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0081] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet). In some embodiments, by using the state information of the computer-readable program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA), the electronic circuit may execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-scenario adjustable power load management system based on layer visualization, characterized in that: Includes grid-load interaction engine and load management business applications; The grid-load interaction engine is used to process the services provided by a grid map, provide data and computing services, map display and interactive public components for upper-level load management business applications, including grid-load data interaction, load resource aggregation, load analysis and calculation, and capability exposure services. The grid-load interaction engine serves as a unified data entry and exit, connects to the marketing data middle platform, obtains user profile information, the connection relationship between 10kV feeder and public-private transformer, the corresponding relationship between public-private transformer and user, industry category, electricity category information, and obtains metering point load data by connecting to the electricity information collection system, and connects to the power sales market intelligent analysis and prediction system to obtain distributed photovoltaic historical load and predicted load data; The load management business application is used to provide a visualization page to achieve observable and measurable adjustable resources, adjustable and controllable load management, resource capacity analysis management, high-level applications of virtual power plants, and high-level applications of photovoltaic resources. According to business functions and data requirements, the map service is called from the grid-load interaction engine to obtain corresponding data, or data is directly obtained from the system database.

2. The multi-scenario adjustable power load management system based on layer visualization according to claim 1 is characterized in that: The power grid map is used to provide map and grid services, power grid topology analysis services, and spatial analysis services, connect to the power grid resource business middle platform, obtain device coordinate information or homologous system device location information, and grid topology relationships from the energy management system; connect to the enterprise-level measurement center to obtain switch measurement data; and connect to the meteorological center to obtain meteorological data.

3. The multi-scenario adjustable power load management system based on layer visualization according to claim 2 is characterized in that: The power load management system is further used to associate user information with substations, lines and transformer equipment in the power grid in different connection modes for users of different voltage levels.

4. The multi-scenario adjustable power load management system based on layer visualization according to claim 3 is characterized in that: For dedicated transformer users, the user access point should be at the end of the feeder, and a distribution transformer should be established inside the user; for dedicated line users, the user access point is located at the end of the feeder, and a distribution transformer is established inside the user; for low-voltage users, the user access point is located after the distribution transformer.

5. The multi-scenario adjustable power load management system based on layer visualization according to claim 4 is characterized in that: The system is further used for checking and managing the relationship between dedicated transformer users and dedicated line users; For the relationship between dedicated transformer users, check whether the relationship between the dedicated transformer on the marketing side and the middle station feeder is correct, and then re-maintain the relationship between the dedicated transformer and the feeder through the same-source maintenance tool. For the verification and management of dedicated line user relationships, check whether the relationship between the dedicated transformer and dedicated line on the marketing side and the substation on the middle station side is correct, and then re-maintain the relationship between the dedicated transformer and dedicated line and the substation through the same-source maintenance tool.

6. The multi-scenario adjustable power load management system based on layer visualization according to claim 5 is characterized in that: The network-load interaction engine further comprises: The archive management service module is used to provide storage, management and query services for archives related to grid-load interaction, covering demand response, orderly power consumption, and power conservation resource management libraries; The computing task management service module is used to receive, split, schedule and manage computing tasks, including user baseline calculation, actual load calculation and aggregation calculation.

7. The multi-scenario adjustable power load management system based on layer visualization according to claim 6 is characterized in that: The load resource aggregation includes the following processes: (1): Baseline computing task scheduling: (1.1): Split multi-day tasks into single-day tasks and generate single-day tasks in advance before the execution date; (1.2): Determine the execution order based on the date, priority, complexity, and number of users of the daily tasks; (1.3): Write the users and parameters required for the calculation of a single-day task into the Kafka message queue to be calculated. The calculation execution program only needs to query the load data. The number of tasks written into the message queue for a single day shall not exceed 3. (1.4): The scheduling task changes the task execution status according to the daily task execution progress; (1.5): Supports task calculation termination midway and supports single-day task recalculation; (2): Perform baseline calculation (2.1): The baseline calculation program supports multiple baseline algorithms, each of which has multiple parameters for adjusting the calculation rules; (2.2): Deploy the baseline calculation program in a multi-process cluster, and automatically search for the current task to be executed in conjunction with the scheduling task. Each progress multi-threaded process processes the data to be calculated to achieve high-speed baseline calculation; (2.3): Update the computing status to redis in real time and update it to the database regularly to achieve real-time monitoring of the computing process and status; (2.4): Provide a complete classification of the causes of baseline calculation failures; (3): Computing task scheduling (3.1): Split multi-day tasks into single-day tasks; (3.2): In batch processing mode, the execution order is determined according to the date, priority, complexity, and number of users of historical tasks; (3.3): The scheduling task changes the task execution status according to the daily task execution progress; (3.4): Supports task calculation termination midway and supports single-day task recalculation; (4): Perform aggregate calculations: (4.1): Execute historical load and baseline load task by task in batch mode; (4.2): Real-time aggregation tasks are processed in parallel in groups. Tasks in the same group are processed in the same computing cluster. Small and medium-sized tasks use the 1×n method, where n is high availability and a single server can perform the aggregation calculation of all tasks. Large-scale tasks use the (1×m+1)×n method, where m is multi-task group aggregation and 1 is summary aggregation. (4.3): The load of the day and the previous day at the quarter hour, and the load of the last two hours in the minute are initialized and cached from HBase to the memory. Subsequently, the cache is updated from the Kafka message queue in a streaming manner to the memory. The archive data is initialized from the database and updated regularly. (4.4): In aggregate calculations, streaming incremental updates or batch real-time calculations are used according to the complexity; (4.5): Use multi-threading to perform aggregate calculations; (5): Execution evaluation (5.1): Cache the user's real-time load in memory, obtain it from HBase during initialization, and update it from the message queue later for sharing by all evaluation tasks; (5.2): Standardize the input and return parameters of the evaluation function, and use the jar upload and dynamic loading method for the evaluation implementation class; (5.3): Store the evaluation results in the database according to the configuration; (6): Execute data extraction: (6.1): Multiple nodes perform data extraction in steps. According to the dependencies between nodes, the subsequent nodes are executed after the previous nodes are completed. (6.2): ​​The node executes the preceding and subsequent database operations. The main processing function implements cross-database data query and storage. The main processing function supports multi-threading, and the query results support js function processing; (6.3): Monitor the process and restore the failed nodes to the pending state after processing; (7): General interface service: (7.1): Generate data query interface through configuration; (7.2): By executing the chain, the results of the previous query are used as the parameters of the subsequent query, thus achieving flexible data query; (7.3): Assemble the query results into the expected JSON format through configuration.

8. The multi-scenario adjustable power load management system based on layer visualization according to claim 7 is characterized in that: The multiple baseline algorithms include: Baseline calculation algorithm: Get the user to be processed; traverse the rule array with the date of the daily baseline task and the user's working day attribute as parameters, and use the judgment function of the algorithm class corresponding to the rule to determine whether the current date to be processed meets the rule. If it meets the rule, use the calculation function of the algorithm class corresponding to the rule to calculate the baseline of the day, and the loop ends; Working day algorithm: determine normal days based on the number of days in advance, the maximum recursive number of days, and the user's abnormal day information; start a loop to try to calculate the baseline with the maximum number of samples as the initial value, and end the loop if successful; pull data for the number of days of samples from hbase, first make rational points for each device in the time period, if there is no data or too many defects, the data of this device on that day is invalid, and the data of this user on that day is invalid; use the data of each device to calculate the data of each sampling point of the user on that day; calculate the mean value of each day in the time period; fix the left window for the calculated mean, slide the right window, and evaluate whether there are samples to be removed. If there are samples to be removed, slide the right window to add new samples and evaluate again. If the sample set boundary is reached, query hbase again to pull a batch of samples. If all samples are traversed and there is no result, the round of calculation ends; if the required number of samples is reached, use qualified samples to calculate the baseline value. When calculating the sample baseline mean, the maximum or minimum sample needs to be removed according to the configuration; Holiday algorithm: Find the same holiday as last year and remove the response day. If it is empty, find the same holiday as the previous year and remove the response day; find the best matching sample day among the remaining sample days; determine whether it is necessary to correct according to capacity, and correct it if necessary; if no sample day is found, use the non-working day baseline calculation method; if the baseline calculation does not find a sample set of load conditions, there is no baseline value, and the application needs to be removed from the candidate list, and the aggregate statistics do not count the value of -1; Equipment missing point interpolation algorithm: Determine whether the missing point ratio exceeds the standard within the time period. If it exceeds the standard, the device’s daily sample will be invalidated, and the user’s daily sample will be invalidated; find all sampling segments with continuous defects, and for each sampling segment, find the two points closest to this sampling segment and record them as (x1, y1), (x2, y2). The defect is (x, y), y=y1+(y2-y1) / (x2-x1)×(x-x1).

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to implement a multi-scenario adjustable power load management system based on layer visualization according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it implements the multi-scenario adjustable power load management system based on layer visualization as described in any one of claims 1-8.

Images